Center for Bioethics N504 Boynton 410 Church Street Minneapolis MN 55455 612-624-9440 Fax: 612

-624-9108 www.bioethics.umn.edu Ms. Laure Campbell Research Subject Advocate University of Minnesota MMC 504 Mayo 8504A (Campus Delivery Code) 420 Delaware St SE Minneapolis, MN 55455 November 7, 2012 Dear Ms. Campbell: I would like to make a two-fold request. First, I would like to request an investigation into the circumstances surrounding the suicide of Dan Markingson in an AstraZeneca-sponsored research study of Seroquel in the Department of Psychiatry. As you probably know, a group of faculty members and I made a similar request to the Board of Regents in late 2010, but they refused to look into the case. Second, I am concerned that the absence of adequate human subject protections demonstrated in that case may still be placing research subjects in danger, especially in the Department of Psychiatry. Charles Schulz in the Department of Psychiatry is currently recruiting subjects for an AstraZenecafunded study of Seroquel, despite the fact that in 2010 the company settled a $520 million lawsuit for fraudulent marketing of Seroquel in which Schulz was implicated. The Clinical and Translational Science Institute is listed as the sponsor of that trial on Clinicaltrials.gov. For that reason, I would like to request an inquiry into the adequacy of human subject protections at the university, with special attention to the Department of Psychiatry. I should also mention that the Markingson case raises issues of concern to the federal government – in particular, to the Senate Finance Committee, which investigated the University of Minnesota for conflict-of-interest violations in 2009. The study in which Markingson died was published with the aid of federal funding from the National Institutes of Mental Health (Attachment A.) The medical care of Markingson was paid for by a health insurance company through an arrangement with Minnesota Care, a public assistance program which also receives federal funds (Attachment B.) In order to make the case for action as forcefully as possible, I am including a summary of the problems that I believe need to be investigated. Summary Dan Markingson was a 26 year-old man from St. Paul and a recent graduate of the University of Michigan who began showing signs of mental illness in the summer of 2003. His thoughts became paranoid and delusional, and he became convinced that he was part of a vast cult, which was calling on him to murder people, including his mother. On November 12, 2003, Markingson was taken to Fairview Hospital in Minneapolis, where he was seen by Dr. Stephen Olson, a psychiatrist at the University of Minnesota. Olson recommended involuntary commitment, and a court agreed. Later,

despite objections by his mother, Mary Weiss, Olson recruited Markingson into a clinical trial of antipsychotic drugs. The clinical trial, which was known as the CAFÉ study (an acronym for Comparison of Atypicals in First-Episode Schizophrenia), was sponsored by AstraZeneca, the manufacturer of Seroquel (quetiapine), and managed by Quintiles, a Contract Research Organization (Attachment C.) The CAFÉ study was a randomized, double-blind trial comparing the effectiveness of three different atypical antipsychotic drugs: Zyprexa (olanzapine), Risperdal (risperidone) and Seroquel (quetiapine.) The University of Minnesota was one of 26 sites for the trial, which lasted a full year. After Markingson was enrolled in the CAFÉ study, he spent about two weeks in Fairview Hospital before being discharged against his mother’s wishes to a halfway house. Over the next five months Ms. Weiss repeatedly expressed her concerns about her son’s medical condition, especially his increasing agitation and rage. Her warnings were largely ignored. Finally, in desperation, she warned the study coordinator that her son might kill himself. On May 7, 2004, Markingson mutilated himself with a box cutter so violently that he nearly decapitated himself. His body was found by halfway house workers in the shower, along with a suicide note that said, “I went through this experience smiling.” Blood tests later showed that Markingson had been taking Seroquel (Attachment D.) In the two years since Markingson’s suicide became public, officials at the University of Minnesota have consistently maintained that the case has been thoroughly investigated, that the university has broken no laws, and that neither the university nor the CAFÉ study investigators bear any responsibility for Markingson’s death. However, the CAFÉ study investigators and the University of Minnesota IRB violated an alarming number of widely accepted ethical guidelines governing the conduct of medical research, which, if uncorrected, endanger the welfare of human subjects at the university. Those violations include the following:  Markingson was coerced into the CAFÉ study by the threat of involuntary commitment.  Markingson was incompetent to consent to the study.  The CAFE study improperly enrolled psychotic subjects at risk of homicide or violence.  University of Minnesota investigators ignored warnings that Markingson was in danger of committing suicide.  University of Minnesota investigators illegally released private health information to CAFÉ study sponsors.  The CAFÉ study consent form failed to disclose serious risks to subjects.  The CAFÉ study targeted vulnerable subjects for recruitment.  AstraZeneca and a University of Minnesota investigator manipulated research results to promote Seroquel.  University of Minnesota investigators failed to disclose important financial conflicts of interest to subjects. In the following pages, I have provided further information for your review. I have also included

supporting documentation in an appendix. Markingson was coerced into the CAFÉ study by the threat of involuntary commitment. Because of the fact that Markingson was under a commitment order, I do not believe that he was in a position to give his voluntary informed consent for the CAFÉ study. On November 12, 2003, when Markingson was admitted to Fairview Hospital, Dr. Olson recommended that he be involuntarily committed to a state mental institution, on the grounds that he was mentally ill and dangerous. The Dakota County court agreed. A week later, Olson asked the court to give Markingson a “stay of commitment,” which allowed Markingson to avoid involuntary confinement as long as he agreed to follow the recommendations of his treatment team. Again, the court agreed. On November, 20, the court issued a legal order requiring Markingson to “remain hospitalized, cooperate with the treatment plan at Fairview University Medical Center until medical discharged, and follow all of the aftercare recommendations of his treatment team.” The day after the court issued an order requiring Markingson to comply with the recommendations of his psychiatrist, Dr. Olson enrolled him in the CAFÉ study. The CAFÉ study coordinator had Markingson sign a consent form for the study when his mother was not present, and kept him in the study despite her objections. At no time did Dr. Olson inform the court that he had enrolled Markingson in a research study (Attachment E.) Markingson was incompetent to consent to the study. It is very unlikely that Markingson, who was acutely psychotic, was mentally capable of consenting to the CAFÉ study. During the period leading up to his enrollment, Markingson had been repeatedly judged incapable of consenting to neuroleptic (antipsychotic) drugs. On November 14, 2003 Dr. Olson signed a commitment document stating that Markingson “lacks the ability to make decisions regarding such treatment.” (Attachment E). On November 17, a pre-petition screening team recommended commitment, noting Markingson’s bizarre beliefs and his refusal to acknowledge his mental illness. On November 19, a clinical psychologist confirmed those assessments, writing that Markingson “is believed not to have the capacity to make decisions regarding neuroleptic medication.” Yet on November 21, when Markingson was asked to consent to the CAFÉ study, this assessment of his mental state was reversed and the CAFÉ study team judged him competent (Attachment F). While it is possible that a psychotic patient’s mental capacity could have improved in two days, there are good reasons to doubt this happened. First, the final competence assessment was made not by an independent party, but by the study coordinator for the CAFÉ study, Jean Kenney, whose job it was to recruit subjects for the study. This is hardly an impartial, disinterested assessment. Second, Kenney was a social worker, not a psychiatrist or psychologist trained to make competence assessments. Third, even after Mr. Markingson had been judged competent to consent to the CAFÉ study, his involuntary commitment order was not lifted. This suggests that his mental state had not changed dramatically. Finally, one of the most persistent features of Markingson’s psychosis was a lack of insight into his condition. Markingson did not believe he had a mental illness (Attachment H). It is unlikely that he could be competent to consent to a study comparing treatments for his mental illness when he would not even acknowledge that he was mentally ill. The CAFE study improperly enrolled psychotic subjects at risk of homicide or violence. Most clinical trials of antipsychotic drugs prohibit researchers from enrolling subjects who are at risk of suicide or violence, to minimize the possibility that these subjects will harm themselves or others in the trial. While the CAFÉ study did not allow researchers to recruit subjects at risk of suicide, it permitted the recruitment of subjects at risk of violence (Attachment G). The University of

Minnesota Institutional Review Board approved the protocol without requiring that potentially violent subjects be excluded. This allowed the CAFÉ study team to enroll Markingson, who had been involuntarily committed precisely because he was threatening homicide. In fact, at the time of his enrollment, there seems to have been broad agreement that he was at high risk of acting out his delusions. University of Minnesota investigators ignored warnings that Markingson was in danger of committing suicide. Records indicate that Markingson experienced little if any improvement during the five and a half months he was in the CAFÉ study. The documents show a subject who, by the most generous interpretation possible, failed to improve over a period of five and a half months; whose own psychiatrist acknowledged so little improvement that his stay of commitment could not be lifted; whose mother was convinced that his condition was spiraling dangerously downward; and whose life ended in a grisly suicide. As Markingson grew increasingly agitated, Mary Weiss repeatedly warned the CAFÉ study investigators that he was in danger of killing himself. On April 15, 2004, Ms. Weiss called the CAFÉ study coordinator and asked, “Do we have to wait until he kills himself or someone else before anyone does anything?” After two letters to Dr. Schulz went unanswered, Ms. Weiss wrote a third letter on April 26, warning him, “Dan has a rage within him, just below the surface, and desperately needs help in dealing with it. Please don’t wait until it comes boiling out!” (Attachment H). Nonetheless, the CAFÉ study investigators dismissed the concerns of Ms. Weiss and left Markingson in the CAFÉ study until he stabbed himself to death. CAFÉ study investigators illegally released private health information to the study sponsors. According to the Health Insurance Portability and Accountability Act (HIPAA), health care providers cannot release a patient’s private health information for use in research without the express written consent of the patient. Violation of the law is a felony. However, Markingson’s private health information was given to AstraZeneca and Quintiles without his authorization. Included in Markingson’s medical records from Fairview Hospital is an unsigned authorization form. When Dr. Olson was questioned about possible HIPAA violations in his deposition, he replied that he did not know whether Mr. Markingson had authorized the release of his health information and that he did not know the specifics of HIPAA requirements (Attachment I.) The CAFÉ study consent form failed to disclose serious risks to subjects. According to the informed consent document for the CAFÉ study, the main risks to study subjects were the side-effects of the drugs being tested and the possibility that the drugs would not be effective (Attachment E.) However, the consent document did not mention that subjects in the CAFÉ were also required to forgo important therapeutic advantages of being treated outside the study. Ordinarily, if a treatment is not effective or is causing serious side-effects, the treating psychiatrist can try another treatment or add adjunct treatments. However, the CAFÉ study placed limits on the kinds of adjunct treatments that could be used, and subjects could not be changed to another treatment without being dropped from the study. On May 12, 2004, Dr. Olson asked the University of Minnesota IRB to approve a new consent form, which disclosed to subjects that all three study drugs carried the risk of diabetes. This change was prompted by a warning issued by the FDA. However, litigation against AstraZeneca has subsequently revealed that the company knew about the risk of diabetes and other metabolic side-effects as early as 2000, three years before Mr. Markingson was recruited into the CAFÉ study, and possibly as early as

1997 (Attachment J.) AstraZeneca did not inform subjects of these risks until 2004. The CAFÉ study targeted vulnerable subjects for recruitment. Before Markingson was enrolled into the CAFÉ study, Quintiles had placed the University of Minnesota trial site on probation for being slow to recruit subjects. The Minnesota site was “struggling to get patients,” according to internal email correspondence, and Quintiles and was pressuring it to step up its efforts. However, as the St. Paul Pioneer Press has reported, in April 2003 the University of Minnesota’s Department of Psychiatry established a locked inpatient unit designed for severely psychotic patients. On this psychosis unit, called Station 12, every patient could be evaluated for recruitment as a potential research subject (Attachment K.) While the establishment of Station 12 increased recruitment numbers dramatically, I am concerned that severely psychotic patients on a locked unit are especially vulnerable to coercion or undue influence. A University of Minnesota investigator helped AstraZeneca manipulate research data. In April 2010, AstraZeneca agreed to pay $520 million to settle two federal investigations and two whistleblower lawsuits alleging that it had marketed Seroquel illegally and concealed its health risks. Documents unsealed in that litigation indicate that Dr. Charles Schulz, the co-investigator on the CAFÉ study, was involved in the misleading presentation of least two AstraZeneca studies. As the St. Paul Pioneer Press has reported, an AstraZeneca trial called Study 15 found that Seroquel performed worse than Haldol, a older, generic antipsychotic drug, and also that Seroquel increased the risk weight gain and diabetes. Yet Schulz claimed publicly and in scientific presentations that Seroquel had been shown superior to Haldol (Attachment L). Later, as reported in City Pages, Schulz helped AstraZeneca conduct a trial called Study 41, which found that extended-release Seroquel was no more effective than a placebo. AstraZeneca then repeated the study overseas in the developing world, where the results proved much better. Subsequently, AstraZeneca enlisted Schulz to publicize the positive study in press releases, while keeping the negative study quiet (Attachment L). I am also concerned that the design of the CAFÉ study was biased to produce a favorable result for Seroquel. As a meta-analysis in the American Journal of Psychiatry has shown, 90% of similar head-to-head comparisons of atypical antipsychotics come out positively for whichever company has designed and funded the trial. Several experts, including the editor of the British Journal of Psychiatry, have argued that the CAFÉ study was similarly biased (Attachment M.) University of Minnesota investigators failed to disclose crucial financial conflicts of interest to subjects. Research subjects have a right to know if study investigators have financial conflicts of interest that might influence their care. However, subjects in the CAFÉ study were not informed that the study investigators had significant financial relationships with AstraZeneca, the CAFÉ study sponsor, as well as the manufacturers of the other atypical antipsychotics being studied. According to the St. Paul Pioneer Press, Dr. Charles Schulz, a co-investigator on the CAFÉ study, received over $570, 000 from the pharmaceutical industry from 2002 to 2008, with $112,000 coming from AstraZeneca. Also, subjects were not informed that the Department of Psychiatry was paid $15,648 for each subject who completed the CAFÉ study. This payment was arranged so that the longer a subject stayed in the study, the more the university would be paid. In total, the CAFÉ study generated $327,000 for the University of Minnesota’s Department of Psychiatry (Attachment N.) Almost as alarming as these ethical violations is the refusal of University of Minnesota officials

to address them. Many of the violations reflect larger structural problems with the oversight of clinical research at this institution which continue to place subjects at risk. Ordinarily, the primary responsibility for protecting human subjects rests with the Institutional Review Board (IRB). In this case, however, the Institutional Review Board compounded the problem by failing to address critical problems with the protocol, such as the inclusion of potentially violent subjects. In a deposition, the director of the IRB, Moira Keane, claimed that protecting subjects was not the responsibility of the IRB (Attachment O.) The suicide of Markingson was first made public through a series of articles in the St. Paul Pioneer Press in May 2008. Since that time, officials at the University of Minnesota have repeatedly asserted that the case has already been reviewed and that the university has been cleared of any blame. This claim is misleading at best. To the best of my knowledge, the University of Minnesota has conducted no internal investigation apart from the routine “adverse event” review by the IRB, which was itself implicated in the case. The only external investigations have resulted from complaints filed by Mary Weiss and Mike Howard (a family friend) to agencies which have little if any responsibility for the protection of human subjects, such as the Minnesota Board of Medical Practice. While it is true that the FDA investigated the case in 2005, in response to a complaint by Mary Weiss and Mike Howard, the FDA inspector restricted its investigation to very narrow grounds and produced a deeply flawed inspection report (Appendix P.) I have been given no reason to believe that officials at the University of Minnesota are willing to address the problems raised by this case. Two months after a lawsuit filed by Mary Weiss was dismissed on technical grounds, the University of Minnesota filed a legal action – a notice to assess costs -- against Ms. Weiss, demanding that she pay the university $57,000 (Attachment Q). In December 2010, a group of university faculty members sent a public letter to the Board of Regents asking for an external investigation. That request was followed by a supporting letter from a campus organization, Faculty for the Renewal of Public Education. In February, the Board of Regents refused that request, claiming that the case had already been investigated (Attachment R.) Since that time, similar communications have come from the Mark Rotenberg in the Office of the General Counsel; Aaron Friedman, the Vice-President for Health Sciences and Dean of Medicine; and Tim Mulcahy, the Vice-President for Research (Appendix S). The continued failure of University of Minnesota officials to take these troubling issues seriously has left me deeply concerned about the protection of research subjects at our institution. I hope you will agree to look into my concerns. If you have any further questions, I would be happy to talk to you further. Yours sincerely,

Carl Elliott MD PhD Professor, Center for Bioethics

ATTACHMENT A

Article

Efficacy and Tolerability of Olanzapine, Quetiapine, and Risperidone in the Treatment of Early Psychosis: A Randomized, Double-Blind 52-Week Comparison
Joseph P. McEvoy, M.D. Jeffrey A. Lieberman, M.D. Diana O. Perkins, M.D., M.P.H. Robert M. Hamer, Ph.D. Hongbin Gu, Ph.D. Arthur Lazarus, M.D., M.B.A. Dennis Sweitzer, Ph.D. Christina Olexy Peter Weiden, M.D. Stephen D. Strakowski, M.D.
Objective: This 52-week randomized, double-blind, flexible-dose, multicenter study evaluated the overall effectiveness (as measured by treatment discontinuation rates) of olanzapine, quetiapine, and risperidone in patients early in the course of psychotic illness. Method: Patients were randomly assigned to treatment with olanzapine (2.5– 20 mg/day), quetiapine (100–800 mg/ day), or risperidone (0.5–4 mg/day) administered in twice-daily doses. Statistical analyses tested for noninferiority in allcause treatment discontinuation rates up to 52 weeks (primary outcome measure) based on a prespecified noninferiority margin of 20%. Results: A total of 400 patients were randomly assigned to treatment with olanzapine (N=133), quetiapine (N=134), or risperidone (N=133). The mean modal prescribed daily doses were 11.7 mg for olanzapine, 506 mg for quetiapine, and 2.4 mg for risperidone. At week 52, allcause treatment discontinuation rates were 68.4%, 70.9%, and 71.4% for olanzapine, quetiapine, and risperidone, respectively. Reductions in total score on the Positive and Negative Syndrome Scale (PANSS) were similar for the three treatment groups, but reductions in PANSS positive subscale scores were greater in the olanzapine group (at 12 weeks and at 52 weeks or withdrawal from study) and the risperidone group (at 12 weeks). The most common elicited adverse events for olanzapine were drowsiness (53%), weight gain (51%), and insomnia (38%); for quetiapine, drowsiness (58%), increased sleep hours (42%), and weight gain (40%); and for risperidone, drowsiness (50%), menstrual irregularities in women (47%), and weight gain (41%). Conclusions: Olanzapine, quetiapine, and risperidone demonstrated comparable effectiveness in early-psychosis patients, as indicated by similar rates of allcause treatment discontinuation. (Am J Psychiatry 2007; 164:1050–1060)

atients experiencing a first episode of psychosis have a better therapeutic response to antipsychotic medications than do chronic, multiepisode patients (1, 2). Despite this good clinical response, however, the majority of first-episode patients discontinue their initial antipsychotic medication, often not continuing treatment with another medication (1–3), which places them at high risk of psychotic relapse and clinical deterioration (4). Comparative studies of atypical versus conventional antipsychotics in patients with first-episode psychosis demonstrate reduced extrapyramidal side effects and equal or slightly superior efficacy for the atypical antipsychotics (5– 10). First-episode patients respond to lower doses and demonstrate a greater sensitivity to antipsychotic treatment-related side effects (6, 7, 11) than do multiepisode patients. Few studies have compared atypical antipsychotics to determine their relative effectiveness in a first-episode population. Studies comparing olanzapine and risperidone in first-episode patients suggest similar efficacy for

P

the two treatments, with few extrapyramidal side effects, although olanzapine is associated with more weight gain (12, 13). Preliminary noncomparative studies suggest that quetiapine is efficacious and well-tolerated in first-episode patients (14–16). The purpose of this study was to determine the overall effectiveness of quetiapine relative to two established standards, olanzapine and risperidone, in patients early in the course of psychotic illness. The primary outcome measure was the percentage of patients discontinuing their assigned antipsychotic (all-cause treatment discontinuation) during the 52 weeks of treatment. This measure integrates the efficacy and tolerability of each drug over time. The primary hypothesis was that quetiapine was not inferior to olanzapine or risperidone in the rate of all-cause treatment discontinuation in earlypsychosis patients. No prior data suggested that superiority for quetiapine was a likely outcome. Previous studies suggested that first-episode patients receive good therapeutic benefit from olanzapine with doses in the range of 10–15 mg/day (6) and that the effec-

This article is featured in this month’s AJP Audio.

1050

ajp.psychiatryonline.org

Am J Psychiatry 164:7, July 2007

MCEVOY, LIEBERMAN, PERKINS, ET AL. TABLE 1. Baseline Characteristics of 400 Early-Psychosis Patients Randomly Assigned to Treatment With Olanzapine, Quetiapine, or Risperidonea Treatment Group Characteristic Female Ethnicity White Black Other DSM-IV diagnosis Schizophrenia Schizophreniform disorder Schizoaffective disorder Antipsychotic naive Illness onset >60 months before baseline Inpatient treatment Age >40 years Previous antipsychotic treatment ≥16 weeks total Duration of previous antipsychotic use (weeks) Mean (SD) Median (range) Duration of illness (months) Mean (SD) Median (range) Age (years) Mean (SD) Median (range) Age at onset (years) Mean (SD) Median (range)
a

Olanzapine (N=133) N % 32 24.1 61 61 11 81 35 17 32 1 29 3 7 Mean/ Median 6.9 4.0 11.0 5.4 24.7 23.1 23.4 21.8 45.9 45.9 8.3 60.9 26.3 12.8 24.2 0.8 21.8 2.3 7.1 SD/ Range 8.81 1.0–52.0 12.86 0.4–62.3 5.8 16.5–42.0 5.3 16.2–41.3

Quetiapine (N=134) N % 42 31.3 66 60 8 75 42 17 36 4 29 2 6 Mean/ Median 6.6 4.0 15.1 7.3 25.0 23.0 23.9 22.2 49.3 44.8 6.0 56.0 31.3 12.7 26.9 3.1 21.6 1.5 6.1 SD/ Range 7.34 1.0–46.3 20.04 0.9–166.4 6.1 16.4–44.4 5.7 15.3–43.3

Risperidone (N=133) N % 34 25.6 78 51 4 75 38 20 28 4 26 2 3 Mean/ Median 5.4 4.0 12.7 6.1 23.9 22.6 23.0 21.4 58.7 38.4 3.0 56.4 28.6 15.0 21.1 3.2 19.7 1.5 2.9 SD/ Range 4.97 0.0–27.0 17.90 0.4–124.0 5.5 16.5–43.9 5.7 13.0–43.6

All Patients (N=400) N % 108 27.0 205 172 23 231 115 54 96 9 84 7 16 Mean/ Median 6.3 4.0 12.9 6.5 24.5 23.0 23.5 21.8 51.3 43.0 5.8 57.8 28.8 13.5 24.1 2.4 21.1 1.8 5.4 SD/ Range 7.20 0.0–52.0 17.29 0.4–166.4 5.8 16.4–44.4 5.6 13.0–43.6

Treatment groups did not differ significantly on any characteristic. been continuously ill for at least 1 month and no more than 5 years. Patients were excluded if a prior psychotic episode had remitted for 3 months or more or if they had prior antipsychotic drug treatment for more than 16 cumulative weeks. All patients had a score ≥4 on at least one Positive and Negative Syndrome Scale (PANSS; 17) psychosis item (delusions, conceptual disorganization, hallucinatory behavior, grandiosity, or suspiciousness/ persecution) and a score ≥4 (moderately ill) on the severity item of the Clinical Global Impression scale (CGI; 19) at the point of maximum severity of illness to date. Female participants of childbearing potential had to be using a medically acceptable form of contraception. We excluded patients who did not speak English; had a history of mental retardation; were pregnant or nursing; had a serious, unstable medical illness; had a known allergy to one of the study medications; were at serious risk of suicide; or had participated in an investigational drug trial within 30 days before the first treatment visit.

tiveness of risperidone may be reduced by extrapyramidal side effects if doses exceed 2–4 mg/day (3, 5, 9). Given quetiapine’s low liability for extrapyramidal side effects, we speculated that doses up to 800 mg/day would be tolerable in our study population. Secondary measures of psychopathology, quality of life, and side effects were obtained to delineate differential effects of each drug on efficacy, tolerability, and safety.

Method
Study Design
This was a 52-week randomized, double-blind, flexible-dose, multicenter study of patients early in the course of schizophrenia, schizoaffective disorder, or schizophreniform disorder assigned to treatment with olanzapine, quetiapine, or risperidone.

Study Population
Participants were recruited from inpatient, outpatient, and emergency department services for the evaluation and treatment of psychosis. The study was approved by the institutional review board at each site, and written informed consent was obtained from the patients or their legally authorized representatives. Patients had to be able to participate in the informed consent process or have a legal guardian available to provide informed consent. Consenting patients 16–40 years of age were eligible for the study if they met DSM-IV criteria for schizophrenia, schizophreniform disorder, or schizoaffective disorder. Patients had to be in the first episode of their psychotic illness and had to have Am J Psychiatry 164:7, July 2007

Study Treatments
Patients were randomly assigned to treatment with olanzapine (2.5–20 mg/day), quetiapine (100–800 mg/day), or risperidone (0.5–4 mg/day). On days 1 and 2, each patient received one capsule of olanzapine (2.5 mg), quetiapine (100 mg), or risperidone (0.5 mg) in the evening. At the treating physician’s discretion, the dose could be increased by one capsule every other day—i.e., on days 3 and 4, one capsule in the morning and one in the evening; on days 5 and 6, one capsule in the morning and two in the evening; and so on, up to a maximum of four capsules twice daily. Any previous antipsychotic therapy was tapered and discontinued during the first 2 weeks of double-blind treatment, and no ajp.psychiatryonline.org

1051

EFFICACY AND TOLERABILITY OF ATYPICAL ANTIPSYCHOTICS TABLE 2. Least Square Mean (LSM) Change From Baseline on Efficacy Measures in Early-Psychosis Patients at Weeks 12 and 52 of Treatment With Olanzapine, Quetiapine, or Risperidonea Treatment Group Olanzapine Baseline (N=134) Variable Positive and Negative Syndrome Scale Positive subscale score Negative subscale score General psychopathology subscale score Total score Clinical Global Impression scale, severity item Calgary Depression Scale for Schizophrenia score Heinrichs-Carpenter Quality of Life Scale, social subscale Heinrichs-Carpenter Quality of Life Scale, vocational subscale
a

Quetiapine Week 52 (N=37) LSM Change SE Baseline (N=133) Mean SD Week 12 (N=96) LSM Change SE Week 52 (N=44) LSM Change SE

Week 12 (N=85) LSM Change SE

Mean

SD

18.8 19.9 35.6 74.3 4.3 12.9 9.1 20.4

5.12 6.27 8.56 16.27 0.75 4.15 6.80 10.18

–5.2 –2.9 –6.3 –14.3 –0.9 –1.1 1.1 1.6

0.36 0.37 0.57 1.12 0.07 0.23 0.58 0.85

–7.1 –3.5 –7.9 –18.4 –1.3 –1.2 3.0 4.7

0.51 0.51 0.81 1.60 0.11 0.33 0.86 1.25

18.6 19.5 36.1 74.2 4.3 13.2 8.7 20.5

4.99 6.18 8.28 15.15 0.69 4.30 7.10 9.64

–4.0b –2.1 –5.5 –11.6 –0.8 –1.5 –0.3 0.2

0.35 0.36 0.56 1.11 0.07 0.23 0.58 0.85

–5.3c –2.8 –7.6 –15.6 –1.2 –2.1 2.2 2.9

0.51 0.52 0.82 1.61 0.11 0.33 0.88 1.29

Analyzed using a mixed random coefficients model with fixed effects for treatment, baseline, and center and with random effects for the intercept and log (time). The listed Ns for weeks 12 and 52 for each group are maximums for the visit. Because of sporadic missing data, the number used in the analysis for each specific variable may be slightly different. b Week 12: quetiapine versus olanzapine, p=0.017; quetiapine versus risperidone, p=0.031. c Week 52: quetiapine versus olanzapine, p=0.013. TABLE 3. Elicited Adverse Events of Moderate Severity or Worsea in 400 Early-Psychosis Patients During Treatment With Olanzapine, Quetiapine, or Risperidone (Intent-to-Treat Population) Treatment Group Olanzapine (N=133) Adverse Event Daytime drowsiness Weight gain Increased sleep hours Insomnia Menstrual irregularitiesb Decreased sex drive Akinesia Dry mouth Akathisia Decreased sexual arousal Decreased orgasm Orthostatic faintness Constipation Sialorrhea Skin rash Gynecomastia Urinary hesitancy Incontinence or nocturia Galactorrhea
a

Quetiapine (N=134) N 77 54 56 39 10 35 33 46 25 22 21 26 16 8 7 3 7 5 0 % 57.5 40.3 41.8 29.1 23.8 26.1 24.6 34.3 18.7 16.4 15.7 19.4 11.9 6.0 5.2 2.2 5.2 3.7 0.0

Risperidone (N=133) N 66 55 36 45 16 36 36 21 30 24 25 17 18 18 9 13 4 4 3 % 49.6 41.4 27.1 33.8 47.1 27.1 27.1 15.8 22.6 18.1 18.8 12.8 13.5 13.5 6.8 9.8 3.0 3.0 2.3

All Patients (N=400) N 214 177 137 135 36 108 101 96 82 75 68 58 45 33 26 25 18 14 6 % 53.5 44.3 34.3 33.8 33.3 27.0 25.3 24.0 20.5 18.8 17.0 14.5 11.3 8.3 6.5 6.3 4.5 3.5 1.5

N 71 68 45 51 10 37 32 29 27 29 22 15 11 7 10 9 7 5 3

% 53.4 51.1 33.8 38.4 31.3 27.8 24.1 21.8 20.3 21.8 16.5 11.3 8.3 5.3 7.5 6.8 5.3 3.8 2.3

The table includes patients for whom adverse events were scored at least “moderate” in severity by the treating clinician at some point during the study. b Percentages are based on the total number of women in the study.

subsequent use of an additional antipsychotic was permitted. Treatment with an adjunctive antidepressant or mood stabilizer during the first 8 weeks of treatment was not allowed unless approved by the project medical officer. Anticholinergic medications for acute extrapyramidal side effects were permitted for up to a total of 2 weeks over the course of the trial. Clinicians were encouraged to lower the dose of antipsychotic to relieve extrapyramidal side effects. Otherwise, adjunctive medications (prescribed to address an aspect of psychopathology inadequately controlled by the assigned antipsychotic) and concomitant medications (pre-

scribed to treat a side effect or a comorbid medical illness) could be used without restriction. When an adjunctive or concomitant medication was prescribed, its name, modal dose, and indication (selected from forced-choice lists) were recorded.

Assessments
The screening evaluation included a diagnostic interview (the Structured Clinical Interview for DSM-IV [18]), medical history, physical examination, measurement of vital signs, and laboratory tests. Confirmation that the illness met clinical severity criteria Am J Psychiatry 164:7, July 2007

1052

ajp.psychiatryonline.org

MCEVOY, LIEBERMAN, PERKINS, ET AL. FIGURE 1. Treatment Discontinuation by 52 Weeks in 400 Early-Psychosis Patients Taking Olanzapine, Quetiapine, or Risperidone (Intent-to-Treat Population)a Risperidone Baseline (N=133) Mean SD Week 12 (N=86) LSM Change SE Week 52 (N=37) LSM Change SE
Continued treatment (N=42, 31.6%) Discontinued treatment, all causes (N=91, 68.4%): Administrative causes (N=5, 3.8%) Clinical causes: Inadequate therapeutic effect (N=15, 11.3%) Unacceptable side effects (N=14, 10.5%) Patient decision (N=57, 42.9%) Continued treatment (N=39, 29.1%) Discontinued treatment, all causes (N=95, 70.9%)b: Administrative causes (N=14, 10.5%)c Clinical causesd: Inadequate therapeutic effect (N=16, 11.9%) Unacceptable side effects (N=13, 9.7%) Patient decision (N=52, 38.8%) Continued treatment (N=38, 28.6%) Discontinued treatment, all causes (N=95, 71.4%): Administrative causes (N=13, 9.8%) Clinical causes: Inadequate therapeutic effect (N=12, 9.0%) Unacceptable side effects (N=13, 9.8%) Patient decision (N=57, 42.9%) Olanzapine (N=133) Quetiapine (N=134) Risperidone (N=133)

18.4 19.4 35.1 73.0 4.2 13.0 9.0 21.7

5.15 6.09 8.73 15.94 0.85 4.01 7.20 11.09

–5.1 –2.6 –6.2 –13.7 –0.9 –1.0 1.2 1.5

0.36 0.37 0.57 1.12 0.07 0.24 0.59 0.86

–6.6 –3.6 –8.4 –18.5 –1.3 –1.3 3.7 5.7

0.52 0.52 0.83 1.63 0.11 0.33 0.91 1.32

was established by a modified, abbreviated version of the PANSS that included items P1–P6 and rated symptom severity at the point of maximum severity of illness. Study visits occurred at baseline, at weekly intervals for the first 6 weeks, every other week for the next 6 weeks, and monthly thereafter. All clinical and laboratory assessments were obtained at baseline, week 12, and week 52 or when the patient terminated the study before week 52. Measures of psychopathology, function, tolerability, and safety were completed at intermediate visits as specified in a schedule of events for the study. The primary outcome measure was the proportion of patients who withdrew from the study prior to 52 weeks of treatment (“allcause pharmacological treatment discontinuation”). The reason for discontinuation was recorded according to a predetermined algorithm: 1) administrative discontinuation due to an independent external event (e.g., moving with family to another state); 2) a clinician decision to discontinue treatment because of inadequate therapeutic effect or intolerable side effects whether or not the patient wanted to discontinue; or 3) a patient decision to discontinue although the clinician believed the treatment to be adequately efficacious, tolerable, and safe. Efficacy was measured in two domains: 1) psychopathology and 2) social and occupational functioning. Psychopathology was assessed by the PANSS, the CGI, and the Calgary Depression Scale for Schizophrenia (20). Social and occupational functioning was assessed with the Heinrichs-Carpenter Quality of Life Scale (21). Clinical response was defined as a score ≤3 on all PANSS items and ≤3 on the CGI severity item at any time during the trial. The number of pills taken was determined by pill counts, and use of concomitant or adjunctive medication was recorded at each study visit. The modal dose for each patient within a treatment group was the dose prescribed for that patient on the maximum number of days during the trial. The mean modal dose for each treatment group was defined as the mean of the modal doses prescribed for the patients assigned to that treatment. Am J Psychiatry 164:7, July 2007

Total all-cause discontinuations (N=281, 70.3%): Administrative causes (N=32, 8.0%) Clinical causes: Inadequate therapeutic effect (N=43, 10.8%) Unacceptable side effects (N=40, 10.0%) Patient decision (N=166, 41.5%)

The Blackwelder noninferiority method (24) was used for comparisons between quetiapine and olanzapine or risperidone using a protocol-defined 20% equivalence margin. b Percentage differences, quetiapine versus olanzapine: 2.5 (95% CI=–8.55 to 13.50); quetiapine versus risperidone: –0.5 (95% CI=–11.4 to 10.33). c Percentage differences, quetiapine versus olanzapine: 6.7 (95% CI=0.58 to 12.79); quetiapine versus risperidone: 0.7 (95% CI=–6.56 to 7.90). d Percentage differences, inadequate therapeutic effect, quetiapine versus olanzapine: 0.7 (95% CI=–7.02 to 8.35); quetiapine versus risperidone: 2.9 (95% CI=–4.42– to .26); unacceptable side effects, quetiapine versus olanzapine: –0.8 (95% CI=–8.06 to 6.41); quetiapine versus risperidone: –0.1 (95% CI=–7.19 to 7.04); patient decision, quetiapine versus olanzapine: –4.1 (95% CI=–15.8 to 7.73); quetiapine versus risperidone: –4.1 (95% CI=–15.8 to 7.73). Clinicians rated the severity of 19 medication-related elicited adverse events on a checklist at each visit. Severity of akathisia was determined with the Barnes Akathisia Rating Scale (23), parkinsonian signs with the Simpson-Angus Scale (22), and dyskinetic movements with the Abnormal Involuntary Movement Scale (19). The most severe scores recorded at any time during the study period are reported. Laboratory tests evaluated glucose, lipids, and prolactin levels. At each blood draw, the patient’s report of the number of hours since consumption of any food or caloric drink was recorded. Fasting was defined as no caloric consumption for 8 or more hours prior to the blood draw. Weight and waist circumference were recorded. ajp.psychiatryonline.org

a

1053

EFFICACY AND TOLERABILITY OF ATYPICAL ANTIPSYCHOTICS TABLE 4. Change From Baseline in Weight and Related Measures in Early-Psychosis Patients at Weeks 12 and 52 of Treatment With Olanzapine, Quetiapine, or Risperidonea Olanzapine Measure Baseline (N=134) Mean 172.0 172.9 168.8 25.8 25.0 28.5 35.2 35.6 SD 43.77 38.90 57.26 6.20 4.66 9.17 5.16 7.58 Week 12 (N=85) LSM SE Change 1.01 15.7b 1.21 16.1b 19.1 3.05 2.4d 0.15 0.18 2.3d 2.4 0.28 1.7 2.2 N 58 43 15 75 53 22 20 14 0.30 0.57 % 59.8f 59.7g 60.0 78.1i 74.6 88.0k 30.3 82.4l Week 52 (N=37) LSM SE Change 24.4b 1.75 24.9b 2.05 14.3 1.68 3.7d 0.26 3.6d 0.30 3.8 0.52 3.5 3.2 N 28 23 5 31 26 5 12 5 0.51 1.30 % 80.0f 79.3 83.3 88.6 89.7j 83.3 46.2 100.0

Weight (lbs) Male Female Body mass index (kg/m2) Male Female Waist circumference (in) Male Female Weight gain ≥7% Male Female Body mass index increase ≥1 unit Male Female Male waist circumference >40 in Female waist circumference >35 in
a

Changes in continuous measures were analyzed using a mixed model similar to that used for the efficacy measures, and changes in categorical measures were analyzed using logistic regression with treatment as the predictor. The listed Ns for weeks 12 and 52 for each group are maximums for the visit. Because of sporadic missing data, the number used in the analysis for each specific variable may be slightly different. b Weight and male weight: quetiapine or risperidone vs. olanzapine (weeks 12 and 52), p<0.01. c Female weight: quetiapine vs. olanzapine (weeks 12 and 52), p<0.001; quetiapine vs. risperidone (weeks 12 and 52), p<0.01. d Body mass index and male body mass index: quetiapine or risperidone vs. olanzapine (weeks 12 and 52), p<0.01. e Female body mass index: quetiapine vs. olanzapine (weeks 12 and 52), p<0.001; quetiapine vs. risperidone (weeks 12 and 52), p<0.01. f Weight gain ≥7%: quetiapine or risperidone vs. olanzapine (weeks 12 and 52), p<0.05. g Male weight gain ≥7%: quetiapine vs. olanzapine (week 12), p<0.01; olanzapine vs. risperidone (weeks 12 and 52), p<0.05. h Female weight gain ≥7%: quetiapine vs. olanzapine (week 12), p<0.01; quetiapine vs. risperidone (week 52), p<0.05. i Body mass index increase ≥1 unit: quetiapine vs. olanzapine (weeks 12 and 52), p<0.05; olanzapine vs. risperidone (week 12), p<0.01. j Male body mass index increase ≥1 unit: quetiapine vs. olanzapine (week 12), p<0.05; olanzapine vs. risperidone (week 52), p<0.05. k Female body mass index increase ≥1 unit: quetiapine vs. olanzapine (week 12), p<0.01; olanzapine vs. risperidone (week 12), p<0.05. l Female waist circumference >35 inches: olanzapine vs. risperidone (week 12), p<0.05. FIGURE 2. Time to All-Cause Treatment Discontinuation in 400 Early-Psychosis Patients Taking Olanzapine, Quetiapine, or Risperidonea
Proportion of Patients Discontinuing From Study 1.0 Patients receiving olanzapine (N=133) 0.8 Patients receiving quetiapine (N=134) Patients receiving risperidone (N=133) 0.6

Statistical Analysis
The protocol-designated primary hypothesis was that quetiapine was not inferior to olanzapine or risperidone in the rate of allcause treatment discontinuation in early-psychosis patients. The primary hypothesis was tested with the protocol-designated statistical analysis of Blackwelder’s (24) noninferiority normal approximation method with a noninferiority margin of 0.20 (20%), using a significance level of 0.025 for each of the two pairwise comparisons. All analyses were specified in a statistical analysis plan that was finalized before the blind was broken. Kaplan-Meier survival curves and a log-rank test were used to assess time to discontinuation. Pairwise comparisons of time to discontinuation between treatments were performed using the log-rank test. Baseline measures of demographic and clinical characteristics were compared using Fisher’s exact test for categorical variables or a Kruskal-Wallis test for continuous variables. Efficacy measures (PANSS, Calgary Depression Scale for Schizophrenia, CGI, and Heinrichs-Carpenter Quality of Life Scale) were tested using a mixed random coefficients model with fixed effects for treatment, baseline, and center and with random effects for the intercept and log (time). Efficacy analyses used a modified intent-to-treat population, defined as patients who were randomly assigned to a treatment and returned for at least one postrandomization assessment. Baseline descriptive statistics are presented using the intent-totreat population, which contained all patients who underwent random assignment to a treatment. Am J Psychiatry 164:7, July 2007

0.4

0.2

0.0 0 10 20 30 40 50 Time in Study (weeks)

a

Several patients came late for their 52-week visit and thus survived longer than the 52 weeks shown in the figure. There was no significant difference between treatment groups in time to all-cause treatment discontinuation. ajp.psychiatryonline.org

1054

MCEVOY, LIEBERMAN, PERKINS, ET AL.

Treatment Group Quetiapine Baseline (N=133) Mean 170.3 175.2 159.4 25.5 25.4 25.9 34.9 34.8 SD 41.12 36.44 48.60 5.22 5.03 5.67 5.08 5.84 Week 12 (N=96) LSM SE Change 1.00 8.12b 1.28 9.38b 1.41 4.65c 0.15 1.2d 0.19 1.4d 0.24 0.8e 1.1 0.5 N 26 20 6 45 31 14 9 15 0.32 0.44 % 29.2f 35.7 18.2h 50.6i 55.4j 42.4k 19.6 55.6 Week 52 (N=44) LSM SE Change 12.49b 1.73 15.2b 2.17 6.47c 2.55 1.9d 0.26 2.2d 0.32 1.1e 0.43 2.1 0.0 N 15 11 4 19 12 7 3 8 0.59 0.90 % 50.0f 64.7 30.8h 63.3i 70.6 53.8 21.4 61.5 Baseline (N=133) Mean 173.1 177.4 161.1 26.1 25.7 27.2 35.4 36.2 SD 42.43 39.16 49.29 5.62 5.01 7.01 5.55 7.69 Risperidone Week 12 (N=86) LSM SE Change 8.87 1.01 8.18 1.21 11.0 1.72 1.4 0.15 1.2 0.18 1.8 0.29 1.2 0.1 N 27 20 7 49 39 10 10 7 0.31 0.59 % 32.5g 31.3 36.8 59.8 61.9 52.6 18.5 46.7 Week 52 (N=37) LSM SE Change 14.5 1.74 13.0 2.06 19.1 3.05 2.3 0.27 2.0 0.30 3.1 0.51 2.4 0.8 N 19 11 8 23 15 8 5 3 0.56 1.21 % 57.6g 45.8 88.9 69.7 62.5 88.9 25.0 50.0

Postbaseline rates of elicited adverse events were compared using Fisher’s exact test. Between-groups differences that met the significance threshold of p≤0.05 are reported without adjustment for multiple comparisons. Because extrapyramidal side effects were minimized by reducing the antipsychotic dose as soon as possible when symptoms appeared, the scores reported for the SimpsonAngus Scale, the Abnormal Involuntary Movement Scale, and the Barnes Akathisia Rating Scale represent worst-case postbaseline values. These were compared using Fisher’s exact test for categorical variables and the Kruskal-Wallis test for continuous variables. Changes in continuous measures of weight, body mass index (BMI), waist circumference, and metabolic parameters were analyzed using a mixed model similar to that used for the efficacy measures, while changes in categorized measures of these parameters were analyzed using logistic regression with treatment as the predictor. Sensitivity analyses for secondary variables were performed using last observation carried forward and observed case analyses to investigate whether the results of the mixed models were similar to those obtained using the observed case and last observation carried forward methods. All analyses of efficacy, weight, and metabolic measures were tested at the nominal significance threshold of p≤0.05, without adjustment for multiple comparisons.

There were no significant differences between treatment groups. Patients showed moderate levels of psychopathology at baseline, with a mean total score of 73.8 (SD=15.8) on the PANSS, a mean score of 4.3 (SD=0.8) on the CGI severity item, and a mean total score of 13.0 (SD=4.2) on the Calgary Depression Scale for Schizophrenia. After caseby-case discussions with site investigators, the project medical officer (J.P .M.) allowed the enrollment of nine patients who had been ill for more than 60 months, seven patients who were over 40 years of age, and 16 patients who had taken antipsychotics for more than 16 weeks.

Pharmacological Treatments
The mean modal number of capsules prescribed per day was 4.7 (SD=2.1) for olanzapine, 5.1 (SD=2.2) for quetiapine, and 4.7 (SD=2.0) for risperidone, which resulted in a mean modal prescribed daily dose of 11.7 mg (SD=5.3) for olanzapine, 506 mg (SD=215) for quetiapine, and 2.4 mg (SD=1.0) for risperidone. Over the course of the trial, 19% of patients in the olanzapine group, 20% of patients in the quetiapine group, and 11% of patients in the risperidone group were brought to the maximum allowed dose of four capsules twice daily.
ajp.psychiatryonline.org

Results
Baseline Characteristics
Table 1 presents demographic and clinical characteristics for the three treatment groups and the whole cohort.
Am J Psychiatry 164:7, July 2007

1055

EFFICACY AND TOLERABILITY OF ATYPICAL ANTIPSYCHOTICS TABLE 5. Change from Baseline in Metabolism-Related Measures and Prolactin Level in Early-Psychosis Patients at Weeks 12 and 52 of Treatment With Olanzapine, Quetiapine, or Risperidonea Olanzapine Variable Baseline (N=134) Mean 99.7 85.3 174.8 48.0 47.5 50.0 27.9 117.3 73.0 N 80 6 4 19 15 6 25 16 SD 58.13 9.83 34.28 12.45 12.25 13.58 27.70 12.79 9.60 % 64.0 7.7 5.13 24.4 25.0 42.9 18.9 12.1 Week 12 (N=85) LSM SE Change 32.3 11.77 1.7 1.39 8.9 4.62 –3.8 1.08d –3.9 1.17 –2.5 2.59 –16.4 2.76f 1.5 1.18g 0.0 0.84 N % 59 62.1 11 25.0 5 11.9 13 29.5 12 32.4 2 28.6 18 18.4 13 13.3 Week 52 (N=37) LSM SE Change 66.4 12.90 b 8.6 1.59 15.7 4.30 –6.5 0.91d –6.7 0.98e –3.9 2.37 –15.9 2.56f 8.5 1.22g 4.8 0.82h N % 78 75.0 22 40.0 14 25.5 23 41.8 23 48.9 4 50.0 41 37.3j 27 24.5

Fasting triglycerides level (mg/dl) Fasting glucose level (mg/dl) Fasting total cholesterol level (mg/dl) Fasting high-density lipoprotein cholesterol level (mg/dl) Male Female Prolactin level Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) Fasting ≥8 hours Fasting triglycerides level >150 mg/dl Fasting glucose level ≥100 mg/dl Fasting total cholesterol level ≥200 mg/dl Male fasting high-density lipoprotein cholesterol level <40 mg/dl Female fasting high-density lipoprotein cholesterol level <50 mg/dl Systolic blood pressure ≥130 mm Hg Diastolic blood pressure ≥85 mm Hg
a

Changes in continuous measures were analyzed using a mixed model similar to that used for the efficacy measures, and changes in categorical measures were analyzed using logistic regression with treatment as the predictor. The listed Ns for weeks 12 and 52 for each group are maximums for the visit. Because of sporadic missing data, the number used in the analysis for each specific variable may be slightly different. b Fasting triglycerides level: quetiapine vs. risperidone (week 52), p<0.05; olanzapine vs. risperidone (week 52), p<0.05. c Fasting total cholesterol level: quetiapine vs. risperidone (week 52), p<0.05. d Fasting high-density lipoprotein cholesterol level: quetiapine vs. olanzapine (week 52), p<0.05; olanzapine vs. risperidone (weeks 12 and 52), p<0.05. e Male fasting high-density lipoprotein cholesterol level: quetiapine vs. olanzapine (week 52), p<0.05; olanzapine vs. risperidone (week 52), p<0.05. f Prolactin level: olanzapine or quetiapine vs. risperidone (weeks 12 and 52), p<0.001. g Systolic blood pressure: quetiapine vs. risperidone (weeks 12 and 52), p<0.01; olanzapine vs. risperidone (weeks 12 and 52), p<0.05. h Diastolic blood pressure: olanzapine vs. risperidone (week 52), p<0.05. i Fasting triglycerides level >150 mg/dl: quetiapine vs. risperidone (weeks 12 and 52), p<0.01; quetiapine vs. olanzapine (week 12), p<0.05. j Systolic blood pressure ≥130 mm Hg: olanzapine vs. risperidone (week 52), p<0.05. k Diastolic blood pressure ≥85 mm Hg: quetiapine vs. risperidone (week 52), p<0.05.

During the study, patients received adjunctive medications mainly for dysphoria/depression (25.7%), anxiety (16.5%), insomnia (15.2%), and agitation/excitement (9.9%). There were no significant differences in postbaseline adjunctive medication use between treatment groups.

Primary Outcome Measure
Figure 1 shows rates of all-cause treatment discontinuation across the three treatment groups. Overall, 70% of patients discontinued treatment before 52 weeks: 68.4% of those assigned to olanzapine, 70.9% of those assigned to quetiapine, and 71.4% of those assigned to risperidone. Based on the prespecified primary outcome measure of a 20% margin for clinically significant inferiority, quetiapine proved noninferior to olanzapine or risperidone. The absolute difference between quetiapine and olanzapine was 2.5%, with an upper-bound one-sided 97.5% confidence interval (CI) of 13.5%, while the absolute difference between quetiapine and risperidone was 0.5%, with an upper-bound one-sided 97.5% CI of 10.3%. Patients receiving olanzapine had fewer administrative discontinuations (3.8%) than those receiving quetiapine (10.5%) or risperidone (9.8%), but there were no other notable differences across the treatment groups in reasons for discontinuation. The most frequent reason for discontinuation across

the entire study population was patient decision despite the recommendations of the treating clinician to continue treatment (41.5%). Only 10.8% discontinued because of inadequate therapeutic effect, and only 10.0% because of intolerable side effects. Figure 2 displays the survival curves to all-cause treatment discontinuation. The median times to all-cause discontinuation for olanzapine (28 weeks), quetiapine (25 weeks), and risperidone (25 weeks) did not differ significantly.

Secondary Outcome Measures
Table 2 presents the least square mean change from baseline scores on efficacy measures at 12 and 52 weeks. All treatment groups showed improvements in symptoms, with no significant differences across groups in PANSS total scores. At 12 weeks, the mean change from baseline in the PANSS positive subscale scores showed greater reductions for olanzapine (–5.2) and risperidone (–5.1) than for quetiapine (–4.0; quetiapine versus olanzapine, p=0.017; quetiapine versus risperidone, p=0.031), but this significant difference persisted only with olanzapine at week 52 (–5.3 for quetiapine versus –7.1 for olanzapine, p=0.013). On all other measures, the three treatment groups did not differ significantly.
Am J Psychiatry 164:7, July 2007

1056

ajp.psychiatryonline.org

MCEVOY, LIEBERMAN, PERKINS, ET AL.

Treatment Group Baseline (N=133) Mean 115.4 85.8 180.5 48.7 47.5 51.5 35.5 119.5 75.7 N 80 15 8 22 15 9 34 19 SD 72.56 9.25 38.50 12.99 13.11 12.51 36.90 12.42 10.56 % 62.5 19.0 10.4 27.9 28.3 40.9 25.8 14.4 Quetiapine Week 12 (N=96) LSM SE Change 52.9 12.16 3.8 1.42 19.3 4.78 –2.2 1.12 –3.0 1.37 –1.2 1.81 –18.5 2.92f 1.9 1.16g 0.5 0.83 N % 56 62.9 22 53.7i 5 12.5 19 46.3 9 33.3 6 42.9 26 25.5 18 17.6 Week 52 (N=44) LSM SE Change 68.1 13.37b 6.2 1.67 25.2 4.46c –3.6 0.95d –3.6 1.12f –4.5 1.73 –18.7 2.66f 7.5 1.21g 4.1 0.82 N % 74 72.5 29 56.9i 10 20.0 23 45.1 15 41.7 9 60.0 40 35.7 32 28.6k Baseline (N=133) Mean 116.1 86.5 176.3 47.4 46.6 49.8 32.7 118.4 73.9 N 80 16 6 17 16 9 23 16 SD 68.38 12.31 34.27 11.81 11.00 14.08 40.28 11.59 9.57 % 63.0 20.5 7.6 21.8 28.1 50.0 18.3 12.7 Risperidone Week 12 (N=86) LSM SE Change 18.2 12.81 1.5 1.50 7.2 5.02 –0.5 1.18 –1.6 1.34 3.2 2.25 13.3 2.90 –2.8 1.25 –1.5 0.89 N % 48 58.5 10 27.0 1 2.8 13 35.1 9 32.1 4 44.4 12 13.8 8 9.2 Week 52 (N=37) LSM SE Change 19.1 13.92 4.8 1.70 11.4 4.65 –2.6 0.99 –2.9 1.13 –1.9 1.93 12.1 2.61 2.7 1.27 1.8 0.86 N % 70 72.2 14 29.8 6 12.5 18 38.3 12 34.3 7 58.3 24 23.8 16 15.8

Sixty-four percent of patients in the olanzapine group, 58% of patients in the quetiapine group, and 65% of patients in the risperidone group met the treatment response criteria (≤3 for all PANSS items and ≤3 for the CGI severity item) at some point during the study. The rates of response were not significantly different between the treatment groups.

the other two groups. Sialorrhea was more common in the risperidone group than in the other two groups. Gynecomastia was more common in the risperidone group than in the quetiapine group. Hypersomnia was more common in the quetiapine group than in the risperidone group. Extrapyramidal Symptoms. Over the course of the trial, only 16% of patients had a rating >1 (mild) on any Simpson-Angus Scale item, only 7% had a rating >2 (mild) on the global severity item of the Barnes Akathisia Rating Scale, and only 1% had a score >2 (mild) on the global severity item of the Abnormal Involuntary Movement Scale. There were no significant differences across treatment groups. The proportion of patients receiving concomitant medications for parkinsonism or akathisia was lower in the quetiapine group (4%) than in the olanzapine group (11%, p=0.021).

Safety and Tolerability
Adverse Events. A total of 18 serious adverse events occurred, four in the olanzapine group and seven each in the quetiapine and risperidone groups. These events included two suicide attempts and one alleged homicide in the olanzapine group, two completed suicides and one case of suicidal ideation in the quetiapine group, and one suicide attempt in the risperidone group. The rates of elicited adverse events that clinicians scored as moderate or severe are presented in Table 3. The most frequent adverse events in the olanzapine group were daytime drowsiness (53%), weight gain (51%), and insomnia (38%); in the quetiapine group, daytime drowsiness (58%), increased sleep hours (42%), and weight gain (40%); in the risperidone group, daytime drowsiness (50%), menstrual irregularities in women (47%), and weight gain (41%). Dry mouth was more common in the quetiapine group than in
Am J Psychiatry 164:7, July 2007

Physical Measures and Laboratory Tests
Weight and BMI. Olanzapine was associated with the greatest increases in body weight and related measures (Table 4). At week 12, the olanzapine group had more weight gain, a greater increase in BMI, and a higher proportion of patients with a BMI increase of at least 1 unit compared with the quetiapine and risperidone groups. Similar differences between olanzapine and quetiapine or
ajp.psychiatryonline.org

1057

EFFICACY AND TOLERABILITY OF ATYPICAL ANTIPSYCHOTICS

risperidone were also observed at week 52 except in the proportion of patients with a BMI increase of at least 1 unit in the risperidone group. Furthermore, 80% of patients in the olanzapine group had gained ≥7% of their baseline weight at week 52, compared with 50% and 58% of the quetiapine and risperidone groups, respectively (observed cases). Risperidone was associated with greater increases than quetiapine in weight and BMI in women (p<0.01). Metabolic Measures. Of the three drugs, risperidone was associated with the smallest elevations in fasting levels of triglycerides and cholesterol and the smallest reduction in high-density lipoprotein cholesterol level (Table 5). Prolactin. Patients in the risperidone group had greater increases in prolactin levels than those in the olanzapine and quetiapine groups at weeks 12 and 52 (Table 5).

fects at higher doses led us to use lower dose ranges of olanzapine and risperidone in this study than those used in the Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE). However, given quetiapine’s low extrapyramidal side effects liability, we expected that the quetiapine doses used in CATIE would be tolerable in our study population. This may have been a factor in the comparable effectiveness demonstrated by quetiapine. In the CATIE phase 1 and phase 2 schizophrenia trials (26–28), higher mean modal doses of olanzapine and risperidone but similar mean modal doses of quetiapine were used; in these trials, olanzapine and, less consistently, risperidone proved to be more effective than quetiapine. It remains an empirical question whether higher doses will improve the relative effectiveness of quetiapine in chronic patients. First-episode patients with psychotic illness tend to be treatment responsive (5, 6, 8, 29). Overall, 62% of our study population met response criteria at some point during the trial (64%, 58%, and 65% in the olanzapine, quetiapine, and risperidone groups, respectively), as indicated by mild or absent positive, negative, or mood symptoms and a global rating of illness severity as mild or less. While clinical improvement was good, these patients were sensitive to medication side effects. Moderate to severe daytime drowsiness, increased sleep hours, weight gain, and menstrual irregularities were commonly experienced. It is possible that these treatments would have been associated with less sedation if olanzapine and risperidone had been given as a single bedtime dose. Over half of patients who remained in treatment with any of these agents at 1 year gained more than 7% of their body weight. Weight gain was more prominent and more likely in the olanzapine group, which is consistent with findings from other studies. Extrapyramidal side effects were uncommon and not severe, probably because low doses of olanzapine and risperidone were used. Patients early in the course of psychotic illness (and their families) who are willing to participate in a trial such as this one may represent an especially insightful and agreeable subgroup of early-psychosis patients. Still, about 70% of patients in this study discontinued prior to 1 year. More than half of the discontinuations were against the preferences of the treating clinicians, indicating a need to improve methods of preemptively identifying and addressing patients’ reasons for leaving treatment (3, 9). Overall, five patients made suicide attempts, of which two resulted in completed suicides. These events occurred despite the close attention provided in clinical research aftercare programs. Since our primary aim was to test the durability of the study drugs over time rather than to measure change from a carefully determined baseline, we allowed patients who had some prior antipsychotic treatment to participate. Thus, our study population may differ in baseline characteristics from populations in some first-episode studies.
Am J Psychiatry 164:7, July 2007

Discussion
This is the first double-blind randomized clinical trial comparing three atypical antipsychotics in patients early in the course of psychotic illness. We evaluated the overall effectiveness of olanzapine, quetiapine, and risperidone in 400 patients over a 52-week period, with the rate of allcause treatment discontinuation as the primary outcome measure. All-cause discontinuation rates were comparable for all three drugs, and quetiapine was noninferior to olanzapine and risperidone. In order to test noninferiority, one must test the null hypothesis that the standard treatment is better than the experimental treatment by at least some specified amount. The specified amount in this study, 20%, was the margin of difference in discontinuation rates that was pragmatically testable within the available budget and sample size parameters; smaller prespecified margins would have required unattainable sample sizes. Our results suggest that the differences in discontinuation rates between these three treatments are much smaller than 20%. There were no significant differences between groups in PANSS total score or percentages of patients meeting predetermined treatment response criteria. However, greater reductions were seen in positive symptom subscale scores at 12 weeks in the olanzapine and risperidone groups than in the quetiapine group, and this advantage persisted with olanzapine at 52 weeks. A similar slight advantage for risperidone on the PANSS positive symptom subscale was seen in another recently reported trial (25). These results suggest that quetiapine, although comparably effective in the treatment of patients early in the course of schizophrenia, schizoaffective disorder, or schizophreniform disorder as assessed on global measures, may be somewhat less potent than olanzapine and risperidone at the doses administered in this study. Previous studies suggesting that first-episode patients receive therapeutic benefit from antipsychotic doses lower than those required for chronic patients and that first-episode patients develop unnecessary extrapyramidal side ef-

1058

ajp.psychiatryonline.org

MCEVOY, LIEBERMAN, PERKINS, ET AL.
5. Emsley RA, Risperidone Working Group: Risperidone in the treatment of first-episode psychotic patients: a double-blind multicenter study. Schizophr Bull 1999; 25:721–729 6. Sanger TM, Lieberman JA, Tohen M, Grundy S, Beasley C Jr, Tollefson GD: Olanzapine versus haloperidol treatment in firstepisode psychosis. Am J Psychiatry 1999; 156:79–87 7. Malla AK, Norman RM, Scholten DJ, Zirul S, Kotteda V: A comparison of long-term outcome in first-episode schizophrenia following treatment with risperidone or a typical antipsychotic. J Clin Psychiatry 2001; 62:179–184 8. Lieberman JA, Tollefson G, Tohen M, Green AI, Gur RE, Kahn R, McEvoy J, Perkins D, Sharma T, Zipursky R, Wei H, Hamer MM (HGDH Study Group): Comparative efficacy and safety of atypical and conventional antipsychotic drugs in first-episode psychosis: a randomized, double-blind trial of olanzapine versus haloperidol. Am J Psychiatry 2003; 160:1396–1404; correction, 160:1901 9. Schooler N, Rabinowitz J, Davidson M, Emsley R, Harvey PD, Kopala L, McGorry PD, Van Hove I, Eerdekens M, Swyzen W, De Smedt G (Early Psychosis Global Working Group): Risperidone and haloperidol in first-episode psychosis: a long-term randomized trial. Am J Psychiatry 2005; 162:947–953 10. McEvoy JP, Hogarty GE, Steingard S: Optimal dose of neuroleptic in acute schizophrenia: a controlled study of the neuroleptic threshold and higher haloperidol dose. Arch Gen Psychiatry 1991; 48:739–745 11. Zhang-Wong J, Zipursky RB, Beiser M, Bean G: Optimal haloperidol dosage in first-episode psychosis. Can J Psychiatry 1999; 44:164–167 12. Malla A, Norman R, Scholten D, Townsend L, Manchanda R, Takhar J, Haricharan R: A comparison of two novel antipsychotics in first episode non-affective psychosis: one-year outcome on symptoms, motor side effects, and cognition. Psychiatry Res 2004; 129:159–169 13. Montes JM, Ciudad A, Gascon J, Gomez JC; EFESO Study Group: Safety, effectiveness, and quality of life of olanzapine in firstepisode schizophrenia: a naturalistic study. Prog Neuropsychopharmacol Biol Psychiatry 2003; 27:667–674 14. Good KP, Kiss I, Buiteman C, Woodley H, Rui Q, Whitehorn D, Kopala L: Improvement in cognitive functioning in patients with first-episode psychosis during treatment with quetiapine: an interim analysis. Br J Psychiatry 2002; 43(suppl):S45–S49 15. Tauscher-Wisniewski S, Kapur S, Tauscher J, Jones C, Daskalakis ZJ, Papatheodorou G, Epstein I, Christensen BK, Zipursky RB: Quetiapine: an effective antipsychotic in first-episode schizophrenia despite only transiently high dopamine-2 receptor blockade. J Clin Psychiatry 2002; 63:992–997 16. Ohlsen RI, O’Toole MS, Purvis RG, Walters JT, Taylor TM, Jones HM, Pilowsky LS: Clinical effectiveness in first-episode patients. Eur Neuropsychopharmacol 2004; 14(suppl 4):S445–S451 17. Kay SR, Fiszbein A, Opler LA: The Positive and Negative Syndrome Scale (PANSS) for schizophrenia. Schizophr Bull 1987; 13:261–276 18. Guy W (ed): ECDEU Assessment Manual for Psychopharmacology: Publication ADM 76-338. Washington, DC, US Department of Health, Education, and Welfare, 1976, pp 217–222 19. Spitzer RL, Williams JBW, Gibbon M, First MB: Structured Clinical Interview for DSM-IV (SCID). New York, New York State Psychiatric Institute, Biometrics Research, 1995 20. Addington D, Addington J, Schissel B: A depression rating scale for schizophrenics. Schizophr Res 1990; 3:247–251 21. Heinrichs DW, Hanlon TE, Carpenter WT Jr: The Quality of Life Scale: an instrument for rating the schizophrenic deficit syndrome. Schizophr Bull 1984; 10:388–398 22. Barnes TRE: A rating scale for drug-induced akathisia. Br J Psychiatry 1989; 154:672–676

Received Sept. 13, 2005; revisions received May 25 and Sept. 19, 2006, and Jan. 16, 2007; accepted Feb. 1, 2007. From the Duke University Medical Center, Durham, N.C.; the Department of Psychiatry, College of Physicians and Surgeons, Columbia University, New York; the Department of Psychiatry, University of North Carolina School of Medicine, Chapel Hill, N.C.; AstraZeneca Pharmaceuticals LP, Wilmington, Del.; SUNY Downstate Medical Center, Brooklyn, New York; and the University of Cincinnati College of Medicine, Cincinnati, Ohio. Address correspondence and reprint requests to Dr. McEvoy, Clinical Research Service, John Umstead Hospital, 1003 12th St., Bldg. 32, Butner, NC 27509; jpmcevoy@duke.edu (e-mail). Dr. McEvoy has received research funding or speaking fees from AstraZeneca, Eli Lilly, and Janssen. Dr. Lieberman has received research funding from Acadia, Bristol-Myers Squibb, GlaxoSmithKline, Janssen, Merck, Organon, and Pfizer and holds a patent related to work with Repligen. He has also served without remuneration as a consultant or on advisory boards for AstraZeneca, Eli Lilly, GlaxoSmithKline, Lundbeck, Organon, and Pfizer. Dr. Perkins has received research funding from AstraZeneca, Bristol-Myers Squibb, Otsuka, Eli Lilly, Janssen, and Pfizer and consulting and educational fees from AstraZeneca, Bristol-Myers Squibb, Eli Lilly, Janssen, GlaxoSmithKline, Forest Labs, Pfizer, and Shire. Dr. Hamer has served in an advisory, consulting, or data monitoring capacity for or has been involved in a contract agreement between University of North Carolina and the following: Wyeth, Allergan, AstraZeneca, Corcept Pharmaceuticals, Epix Pharmaceuticals, GlaxoSmithKline, Johnson & Johnson, Eli Lilly, Pfizer, SAS Institute, Schwartz, Solvay, and Somerset Pharmaceuticals. He or his wife holds shares of stock from Amgen, Bristol-Myers Squibb, Eli Lilly, Genentech, Proctor & Gamble, and Sepracor. Dr. Weiden has received grant support or speaker or consulting fees from AstraZeneca, Bristol-Myers Squibb (Otsuka), Janssen, Langeloth Foundation, NIMH, Pfizer, Shire, Solvay, and Vanda. Dr. Strakowski has received grant support or speaker or consulting fees from Abbott, AstraZeneca, DiMedix, Eli Lilly, France Foundation, Janssen, Forest, Nutrition 21, Pfizer, and Repligen. Dr. Lazarus is an employee of AstraZeneca and holds stock options with the firm. Dr. Sweitzer is an employee of AstraZeneca, owns shares of AstraZeneca, and holds stock options with the firm. Ms. Olexy is an employee of AstraZeneca. Dr. Gu reports no competing interests. The Comparison of Atypicals in First Episode of Psychosis research program was coordinated by the University of North Carolina. Funding for this program was provided by AstraZeneca Pharmaceuticals LP (5077IL/0114). The authors acknowledge the assistance of Sandra Woolson for statistical programming. This study is registered at www.ClinicalTrials.gov under the title “CAFE: Comparison of Atypicals in First Episode of Psychosis” (govIdentifier: NCT00034892, Study ID Numbers: 5077IL/0114). All criteria as stated in the Clinical Trial Registration policy have been met.

References
1. Lieberman J, Jody D, Geisler S, Alvir J, Loebel A, Szymanski S, Woerner M, Borenstein M: Time course and biologic correlates of treatment response in first-episode schizophrenia. Arch Gen Psychiatry 1993; 50:369–376 2. Robinson DG, Woerner MG, Alvir JMJ, Geisler S, Koreen A, Sheitman B, Chakos M, Mayerhoff D, Bilder R, Goldman R, Lieberman JA: Predictors of treatment response from a first episode of schizophrenia or schizoaffective disorder. Am J Psychiatry 1999; 156:544–549 3. Gaebel W, Moller HJ, Buchkremer G, Ohmann C, Riesbeck M, Wolwer W, Von Wilmsdorff M, Bottlender R, Klingberg S: Pharmacological long-term treatment strategies in first episode schizophrenia: study design and preliminary results of an ongoing RCT within the German Research Network on Schizophrenia. Eur Arch Psychiatry Clin Neurosci 2004; 254:129–140 4. Lieberman JA: Is schizophrenia a neurodegenerative disorder? a clinical and neurobiological perspective. Biol Psychiatry 1999; 46:729–739

Am J Psychiatry 164:7, July 2007

ajp.psychiatryonline.org

1059

EFFICACY AND TOLERABILITY OF ATYPICAL ANTIPSYCHOTICS
23. Simpson GM, Angus JWS: A rating scale for extrapyramidal side effects. Acta Psychiatr Scand Suppl 1970; 212:11–19 24. Blackwelder WC: “Proving the null hypothesis” in clinical trials. Controlled Clin Trials 1982; 3:345–353 25. Zhong KX, Lieberman JA, Hamer RM, Sweitzer D: Comparison of quetiapine and risperidone in the treatment of schizophrenia: a randomized, double-blind, flexible-dose, 8-week study. J Clin Psychiatry 2006; 67:1093–1103 26. Lieberman JA, Stroup TS, McEvoy JP, Swartz MS, Rosenheck RA, Perkins DO, Keefe RS, Davis SM, Davis CE, Lebowitz BD, Severe J, Hsiao JK; Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) Investigators: Effectiveness of antipsychotic drugs in patients with chronic schizophrenia. N Engl J Med 2005; 353:1209–1223 27. McEvoy JP, Lieberman JA, Stroup TS, Davis SM, Meltzer HY, Rosenheck RA, Swartz MS, Perkins DO, Keefe RSE, Davis CE, Severe J, Hsiao JK, CATIE Investigators: Effectiveness of clozapine versus olanzapine, quetiapine, and risperidone in patients with chronic schizophrenia who did not respond to prior atypical antipsychotic treatment. Am J Psychiatry 2006; 163:600–610 28. Stroup TS, Lieberman JA, McEvoy JP, Swartz MS, Davis SM, Rosenheck RA, Perkins DO, Keefe RSE, Davis CE, Severe J, Hsiao JK, CATIE Investigators: Effectiveness of olanzapine, quetiapine, risperidone, and ziprasidone in patients with chronic schizophrenia following discontinuation of a previous atypical antipsychotic. Am J Psychiatry 2006; 163:611–622 29. Bradford DW, Perkins DO, Lieberman JA: Pharmacological management of first-episode schizophrenia and related nonaffective psychoses. Drugs 2003; 63:2265–2283

1060

ajp.psychiatryonline.org

Am J Psychiatry 164:7, July 2007

ATTACHMENT B

ATTACHMENT C

ATTACHMENT D

ATTACHMENT E

EXCERPT
1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 STATE OF MINNESOTA DISTRICT COURT COUNTY OF HENNEPIN FOURTH JUDICIAL DISTRICT Court File No. 62 CO 06 11934 -------------------------------Mary Weiss, on her own behalf, and as the next of kin and Trustee of the estate of Dan Markingson, deceased, Plaintiff, vs. Board of Regents for the University of Minnesota; Dr. Stephen Olson; Dr. Charles Schulz; Institutional Review Board for the University of Minnesota; Astrazeneca Pharmaceuticals LP, Astrazeneca LP and Zeneca, Inc., Defendants. ---------------------------------AUDIO VISUAL DEPOSITION OF STEPHEN OLSON, M.D. MAY 1, 2007 VERBATIM COURT REPORTING 763-493-4535 2 The following is the deposition of STEPHEN OLSON, M.D.; Mari Skalicky, Court Reporter, Notary Public, pursuant to Notice of Taking Deposition, at 701 Xenia Avenue South, Suite 600, Minneapolis, Minnesota, commencing at approximately 9:00 a.m., MAY 1, 2007.

3 4 5 6 7 8

25 Q. (BY DR. BARDEN) Was the court ever informed 1 that Dan would be going into a research study? 2 MR. ALSOP: Object on the basis of 3 foundation. Go ahead, Doctor.

4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Q. (BY DR. BARDEN) Let me ask you this. Did you ever inform the court that Dan was actually going into a research study? A. Not directly, although, the county case manager who was responsible -Q. Objection, move to strike. Go ahead and finish your answer and we'll ask for the time back the next day. Go ahead. MR. ALSOP: We're not coming back a next day. DR. BARDEN: Let me just make for the record here. If the witness rambles on and if we add up the time and at the end of the day he's rambled on, obviously not answering my question and he's wasted a half an hour of our time, we will seek that time back, period. MR. ALSOP: You can go ahead and seek it. He's answered the question. Q. (BY DR. BARDEN) So please, go ahead and answer your -- give the answer that you think is responsive to my question. A. Would you ask it again? VERBATIM COURT REPORTING 763-493-4535

79 1 MR. ALSOP: Doctor, answer the question 2 the way you feel is most responsive. He's not 3 rambling, and he will answer the question as he 4 sees fit. 5 A. Will you ask the question again? 6 Q. Would you read that back. 7 (Record read back.) 8 A. I didn't inform the court directly but the 9 county case manager is required to make a 10 report to the court that the patient was 11 meeting the terms of the stay of commitment 12 that he seek appropriate care. And I had 13 discussions with Mr. Pettit numerous times that 14 the participation of the CAFE was one way that 15 he could receive appropriate care, although, 16 there were certainly other options had Dan 17 chosen not to participate in the study. 18 Q. Objection. Move to strike the entire response 19 after the first sentence. Did you follow-up 20 and make sure that the court was informed by

21 Mr. Pettit that Dan had not gone into treatment 22 but in fact had gone into a clinical research 23 study? 24 A. No. 25 Q. Other than Dan, were there any other subjects VERBATIM COURT REPORTING 763-493-4535 80 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 in your study who were court-ordered into treatment, yet ended up in your research study? MR. ALSOP: It's repetitious, but go ahead. A. I don't recall but there may have been. Q. (BY DR. BARDEN) But no one, the IRB, Dr. Schulz, the University, no one raised any questions about that, correct? MR. ALSOP: It's repetitious. Talk about wasting time with multiple questions, that's been asked and answered multiple times. Go ahead one more time. A. Repeat the question. (Record read back.) A. Not at the time and not after the conduct of the study was investigated following Dan's death.

ATTACHMENT F

ATTACHMENT G

ATTACHMENT H

EXCERPT
1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 STATE OF MINNESOTA DISTRICT COURT COUNTY OF HENNEPIN FOURTH JUDICIAL DISTRICT Court File No. 62 CO 06 11934 -------------------------------Mary Weiss, on her own behalf, and as the next of kin and Trustee of the estate of Dan Markingson, deceased, Plaintiff, vs. Board of Regents for the University of Minnesota; Dr. Stephen Olson; Dr. Charles Schulz; Institutional Review Board for the University of Minnesota; Astrazeneca Pharmaceuticals LP, Astrazeneca LP and Zeneca, Inc., Defendants. ---------------------------------AUDIO VISUAL DEPOSITION OF STEPHEN OLSON, M.D. MAY 1, 2007

3 Q. These are Dan's journals marked March 9 and 4 March 23rd of '04. These are in order in Dan's 5 journal, and they are in the order for March 6 '04, one is 3/09 but it says '05 but it's in 7 the order of '04. And then the next one is 8 03/23/04 these are Dan's journals. And I will 9 just read this to you, Doctor. You've been 10 given to observation on truth today, Dan. 11 MR. ALSOP: Which entry are we reading? 12 Q. 3/09/05. That's what it's marked 3/09/04. 13 You've been given observation on truth today, 14 Dan. If someone makes an assumption in asking 15 for a confirmation, such as you've seen Friday,

16 right? You may recast. The question by 17 saying, Have I seen Friday or more generously, 18 if you think the assumption might possibly be 19 all right to make concerning the average 20 person, like you've seen Star Wars, right. You 21 may answer the question, then say is it good 22 enough to assume that somebody's seen it. 23 Sounds like the writings of a psychotic, 24 doesn't it, Dr. Olson? 25 A. It could be. VERBATIM COURT REPORTING 763-493-4535 234 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Q. That's a yes or no. Then the next page 3/23/04, world walking, you were at a farm house and we're getting presents from dogs who had presents fastened in plastic bags to their snouts. And it goes on to say, in the gloaming and breening, you were thinking of naming it gloaming and greening or gloam-green. That was someone brings a snowslide in summer or mid summer. It has been left behind, et cetera. Classic psychotic writing, isn't it, Dr. Olson? A. Well, it certainly could be but this is an English major, and I, you know, need to know what his state of mind at the time, what was his intention. If this is just stream of consciousness writing, you know, it could be Joyce. Q. Any undergraduate psychology student reading this would assume that this patient is psychotic and being ineffectively, incompetently untreated, isn't that correct? MR. GROSS: Objection, argumentative, MR. ALSOP: Multiple question. Argumentative. Repetitious. Go ahead, Doctor. MS. SVITAK: Objections. Q. Next. VERBATIM COURT REPORTING 763-493-4535 235 1 A. It certainly is thought disordered and, yes, 2 the untrained observer would look at this and

3 say, this person is not well. Had I been aware 4 of this writing, then I might have thought 5 differently. I would have asked Dan what this 6 meant. 7 Q. So if an international expert in schizophrenia 8 says that this is a clear indication that Dan 9 was improperly or incompetently treated, you 10 would not agree with that? 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 MR. ALSOP: Object, assumes facts not in evidence. It's vague and ambiguous and argumentative. You can answer, Doctor. Q. (BY DR. BARDEN) That's a yes or no. A. Well, clearly his -Q. Our time is limited, so I'll ask yes or no questions. MR. ALSOP: If you can answer yes or no. If you can't, Doctor, tell him. A. I can't answer yes or no. Q. The next exhibit we're going to go to is Bates number PSY 000021 dated 3/25/04. (Exhibit No. 21 was marked for identification by the Court Reporter). Q. (BY DR. BARDEN) Again, these are all from

VERBATIM COURT REPORTING 763-493-4535 236 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 March '04. (Exhibit No. 21 was marked for identification by the Court Reporter). Q. Okay. This is Exhibit No. 21. Doctor, do you see this? Do you notice this as the clinical record for Dan? A. Yes. Q. 3/25/04, he says he's leaving for California as soon as court order expires. Do you see that? A. Yes. Q. But that would cost you money, wouldn't it? Be much better to recommit him so you could continue to draw funding from Astrazeneca until the end of the study period, correct? MR. ALSOP: Object as argumentative. Misstatement of the evidence and as to form.

17 MS. SVITAK: Same objection. 18 MR. GROSS: Same objection. 19 Q. It says -20 A. I didn't want him to go to California because I 21 didn't think he was capable of living 22 independently. Had nothing to do with whether 23 he stayed in the study or not. 24 Q. You didn't think he was competent to make 25 treatment decisions then, right, that's why you VERBATIM COURT REPORTING 763-493-4535 237 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 wanted his commitment extended? A. No. Q. See, this is the fatal flaw in your plan, Doctor, because by trying to get him his extension of his commitment, you had to argue that he was incapable of consenting to your study, so he should have been reconsented at that point, correct? MR. ALSOP: It's argumentative, object as to form. It's multiple in nature. Go ahead, Doctor. MS. SVITAK: Same objection. A. No. His judgment about -- his judgment about his future plans, his finances was quite impaired and his willingness to continue in the study didn't -- he didn't reflect any lack of understanding of what staying in the CAFE study meant in terms of his treatment. Q. Oh, so, his judgment for his plans and his finances and his travels and his future life are completely incompetent, but his judgment to stay in your personally profitable research study is just fine, is that what you just testified, Dr. Olson? A. No, I disagree with your characterization.

VERBATIM COURT REPORTING 763-493-4535 238

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

MR. ALSOP: It's argumentative. Q. (BY DR. BARDEN) It says, he plans to take a Greyhound bus until it runs out and then hitchhike the rest of the way to California. How many people have done that before, Doctor? MR. ALSOP: Object as vague as to doing what. Go ahead. A. I don't know. Q. (BY DR. BARDEN) I will withdraw the question. It says he has no place to live, no one to stay with, no money saved, not been making car insurance payments, no question how this has been maintained. Had you read that in March of 2004? A. Yes. Jeanie -Q. That's a yes or no question. We're running out of time, Doctor. We're going to try to end here at 4:30. MR. ALSOP: Doctor, if you can answer yes or no, fine. If you can't -Q. I'm asking yes or no questions for a specific reason. MR. ALSOP: Well, if he says he can't answer yes or no he's not going to. Q. (BY DR. BARDEN) Question was, had you seen VERBATIM COURT REPORTING 763-493-4535

239 1 this in March of '04? If he can't answer that 2 yes or no, then we'll have a competency 3 hearing. Did you see this document in March of 4 '04? Yes or no? 5 A. Yes. 6 MR. ALSOP: Doctor, listen. 7 A. Yes, I saw it and I had a discussion with 8 Jeanie even before she wrote this note about 9 the -- what was going on and, consequently, our 10 intervention was that we were recommending an 11 extension of his stay of commitment as a way of 12 trying to keep him in the State of Minnesota, 13 and then while he was here, to try to move him 14 into independent living to see if he could 15 manage on his own. 16 Q. Objection, move to strike everything after the 17 word "yes" and ask for the time back. Doctor,

18 if you look down the middle of the note there, 19 it says he has cancelled several therapy 20 appointments and therapists reports, he is not 21 talking in sessions. Had you read that in 22 March of '04? 23 A. Yes. 24 Q. Next document is the Workbook Sheet No. 5. 25 Also dated March 31st, 2004. You don't have VERBATIM COURT REPORTING 763-493-4535 240 1 any copies of this? Okay. We only have one 2 copy. We'll make it an exhibit and we can see 3 it then. If you can just identify that for 4 everyone in the room, Doctor. 5 A. This is part of the psychoeducation materials 6 that was provided and taught to Dan by either 7 Jeanie or one of the other research staff 8 that's dated March 31st. It shows a 9 bell-shaped curve with a patient starting at 10 the prodrome, going through a point of recovery 11 and then down the other side, not taking 12 medications, getting worse, relapse symptoms, 13 relapse of symptoms. So it's the course of a 14 recovery with a relapse. 15 Q. And this is marked Exhibit what? 16 MR. ALSOP: She hasn't had had a chance 17 to mark it yet. 18 MS. SVITAK: Is there Bates numbers on 19 there? 20 Q. (BY DR. BARDEN) Let me look real quick. Bates 21 number CS 000776. All right. Now, Doctor, 22 this is in March 31st, 2004. Do you see that? 23 This is the same time period we're talking 24 about where the Wisconsin Card Sort says not 25 taking meds. We read from his journal and

VERBATIM COURT REPORTING 763-493-4535 241 1 2 we've just read that he is not going to therapy and he is not talking in therapy and he has

3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

plans to go to California, correct? MR. ALSOP: That's a misstatement of the evidence. Q. This is march of 'O4. A. No. That's incorrect. We thought he was taking his meds. That statement on the Wisconsin Card Sort is an oversight by the psychometrist filling in the screen, the face screen on the test. I didn't have access to his journal. I didn't even know he had a journal, and although he was -Q. Let me ask you a quick question about this. A. Let me finish my answer. MR. ALSOP: Let him finish his answer. A. You're alleging that I knew that he was deteriorating and I'm telling you that we didn't know that, that we thought he was taking his medications, that he was in day treatment, he was at the group home, and no one in either of those places observed that he was becoming more psychotic. Q. Except that his mother kept telling you that he was deteriorating rapidly and was going to VERBATIM COURT REPORTING 763-493-4535 242 commit suicide and nobody listened to her, correct? MR. ALSOP: That's argumentative. Q. She said that repeatedly. MR. ALSOP: Argumentative and vague. Go ahead, Doctor. A. She said that in a letter to Dr. Schulz. I don't know that she said it repeatedly. Q. It's in the clinical notes several times, isn't it, Doctor? Jean Kenney wrote it down. What has to happen before anyone pays attention, does he have to kill himself, quote, Mary Weiss, correct? A. That's correct. Q. Yes. Now, let me ask you a question about that. This is in March of 2004, and on this recovery curve, what has Dan checked? He's checked full recovery with no symptoms and functioning well in life. Do you see that?

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

20 A. (Nods head in the affirmative). 21 Q. This is exactly the point where he was at 22 December -- way back in November 14th when you 23 wrote he lacked capacity because he was 24 delusional and he had no insight, and he's 25 exactly where he is at that point in March of VERBATIM COURT REPORTING 763-493-4535 243 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 2004, and no one in your study is noticing it because they have a financial interest not to notice it, isn't that correct? MR. ALSOP: Doctor, it's argumentative. It's multiple, vague and as to form. You can answer. A. No. His condition was different in March and April of 2004. He was living in a group home. He was successfully completing a day treatment program and was moving to -- after initially planning to go to California after we had that meeting, he agreed to work on independent living and see how that went before he made any decision to return. Q. This is a yes or no question. During the time you were treating Dan, did you see the Eagan Counseling Clinic notes of March 29th, 2004 stating that he is showing slightly more disorganization and thought and stream of speech and risk to self low with plan. Did you see a note indicating that Dan -A. Repeat that, the risk to self. Q. With plan. A. No, I did not see that nor was I informed that he had told his counselor that he had a plan. VERBATIM COURT REPORTING 763-493-4535 244 1 Q. But you signed 1572 taking responsibility for 2 the subjects in your study, correct? 3 A. Right, but I'm not responsible for the 4 counselor in Eagan informing whether or not 5 that therapist informs anyone of those

6 findings. I wasn't aware of that. 7 Q. Next exhibit is 4/15/04, clinical note Bates 8 stamped PSY 000144. Do you have that 144? 9 Page 144. This will be marked as Exhibit No. 10 23. It's also Bates No. PSY 000022. 11 (Exhibit No. 22-23 were marked for 12 identification by the Court Reporter). 13 Q. (BY DR. BARDEN) This is a yes or no question. 14 Dr. Olson, as Dan's study psychiatrist, 15 treating psychiatrist and someone who signed 16 1572, during the time of April '04, had you 17 read this note which says, quote, Mary again 18 reiterating Dan's meds are not working. He is 19 totally out of control. He says things such 20 as, quote, are you asking me or telling me, 21 unquote. She also stated, quote, do we have to 22 wait until he kills himself or someone else 23 before anyone does anything, unquote. Had you 24 seen that? Yes or no? 25 A. Yes.

VERBATIM COURT REPORTING 763-493-4535 245 1 Q. Next, exhibit -- let me ask you this as a yes 2 or no question. Had you seen any notes in the 3 occupational therapy records that in April of 4 2004 that said that over time client has become 5 more isolated. He seems to have no interest in 6 interacting with his peers. Personal 7 appearance, disheveled, isolated and withdrawn, 8 poor insight and self awareness. Plan to 9 become an actor in California continues. 10 Delusions seems fixed. Did you as the 11 psychiatrist for Dan and the person that signed 12 1572, had you seen that note in April of '04 or 13 not? 14 MR. ALSOP: Object, speculative and 15 lacking in foundation without seeing it. Go 16 ahead, Doctor, if you can recall. 17 A. Is that from occupational therapy at Fairview 18 day treatment program? 19 Q. Len Bennatti, have you ever heard of Len 20 Bennatti O.T.?

21 22 23 24 25

A. Well, I asked you whether it's from the Fairview Day Treatment Records or not. Q. Progress notes from FBS. A. Fairview Behavioral Services? Q. Yes.

VERBATIM COURT REPORTING 763-493-4535 246 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 A. No. I hadn't seen that specific note but I -Q. It's just a yes or no question. A. I regularly met with -MR. ALSOP: No. You can answer the question. Q. It's a yes or no question. MR. ALSOP: Doctor, if you can't answer yes or no, tell him. Q. Fine. And I'll get the time back. MR. ALSOP: You won't. Go ahead, Doctor. Q. We're going to have to come back anyway for the record, so go ahead. MR. ALSOP: If you can answer yes or no, do it. If you can't, tell him that. A. No, I didn't see the specific note but I met regularly with the day treatment staff and when they discussed Dan's case, we all agreed that he had poor insight, that he was showing some worsening of negative symptoms, but that he also was discharged from the day treatment program the week of his suicide, which was an indication that they felt that his progress had been -- that he had successfully completed the program. Q. So yes or no, even though he was disheveled,

VERBATIM COURT REPORTING 763-493-4535 247 1 isolated, withdrawn, had fixed delusions in 2 your mind at this time, he still was competent 3 to be a subject in your research study? 4 A. I disagree that his plan to become a actor was

5 6 7 8

a delusion. I think that was -- I think that was a poor -- that was poor judgment on his part, but I don't think that was a delusion on the order of thinking his mother is a lizard.

ATTACHMENT I

EXCERPT
1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 STATE OF MINNESOTA DISTRICT COURT COUNTY OF HENNEPIN FOURTH JUDICIAL DISTRICT Court File No. 62 CO 06 11934 -------------------------------Mary Weiss, on her own behalf, and as the next of kin and Trustee of the estate of Dan Markingson, deceased, Plaintiff, vs. Board of Regents for the University of Minnesota; Dr. Stephen Olson; Dr. Charles Schulz; Institutional Review Board for the University of Minnesota; Astrazeneca Pharmaceuticals LP, Astrazeneca LP and Zeneca, Inc., Defendants. ---------------------------------AUDIO VISUAL DEPOSITION OF STEPHEN OLSON, M.D. MAY 1, 2007

0 Q. What kind of -- as someone who is doing 11 research involving human subjects, and as 12 someone trained by the University of Minnesota 13 and supervised by the IRB and Dr. Schulz, what 14 kind of authorization to use health care 15 information do you need from human subjects in 16 your research? 17 MR. GROSS: Objection, form. 18 MR. ALSOP: I'll join in that. 19 A. What kind of authorization do I need for what? 20 Q. (BY DR. BARDEN) To use health care records. 21 A. To use health care records? 22 Q. Yes. 23 MR. ALSOP: I'll object as vague and 24 ambiguous.

25 A. I don't understand if it's the hospital's VERBATIM COURT REPORTING 763-493-4535 142 1 health care records, then I'm involved in 2 writing and forming. I don't need anyone's 3 permission to document what's going on. If I 4 want to release that information to someone 5 else, I need the patient's permission. 6 Q. And someone conducting research with human 7 subjects and trained by the University of 8 Minnesota and supervised by the IRB and Dr. 9 Schulz, how many signatures and what kinds of 10 signatures do you need to get a study subject 11 enrolled? 12 MR. GROSS: Objection, form. 13 MR. ALSOP: It's vague and ambiguous 14 also. Go ahead. 15 A. I think the patient needs to sign the consent 16 form, the informed consent document, and 17 currently patients also sign a separate HIPA 18 document that contains language about protected 19 health information and how it will be used, but 20 I don't believe at the time that Dan entered 21 the CAFE study that a separate document was 22 required. I believe that that protected health 23 information was incorporated into the informed 24 consent document. 25 Q. Isn't it in fact the case that Dan was supposed VERBATIM COURT REPORTING 763-493-4535 143 1 2 3 4 5 6 7 8 9 10 to sign a HIPA authorization and he never was asked to do so and never did so in your study? MR. ALSOP: It's a multiple question, vague and ambiguous. Go ahead, Doctor. A. I'm not aware that there were any documentation failures like that, no. Q. (BY DR. BARDEN) Did anyone in your CAFE study sign a HIPA authorization form? A. I don't know. Q. So you don't know if Dan was the only one who

11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

did not sign a HIPA authorization? A. I don't know. Q. Do you know what the University rules are with regard to conducting research when a subject has not signed a HIPA authorization form? A. No, I don't know what the consequences are or the rules. Q. Do you think that would be a good thing to know before you do research with human subjects? MR. ALSOP: Object as argumentative. It's also irrelevant. Go ahead, Doctor. A. Yes. I always try to follow the appropriate documentation rules. And if at the time that Dan entered the study we were required to have his signature on a separate HIPA document and VERBATIM COURT REPORTING 763-493-4535

144 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 we failed to do so, I would have expected that when the IRB and the FDA inspected the program, I would have been informed of that. Q. And it's your understanding that Dan never during the entire time he was in your study, he never did sign a HIPA document? A. That's not my understanding. I don't know. I don't recall offhand what he signed or not. Q. Well, there is not in any of the documents we've been disclosed, in fact, there is a blank form. So if there was a blank form in the file for a HIPA authorization that was never signed by anybody, would that help you to understand when you needed to do that? MR. ALSOP: Object as speculative, lacking foundation. A. Could you repeat the question? Q. If there was a blank HIPA form in the CAFE documents that were disclosed to us with a place for people to sign that was never signed, would that help you understand when you needed to get a HIPA signature? A. Well, that would suggest that we in fact should have had a -- he should have signed that, yes. Q. Isn't that required by federal law and the

VERBATIM COURT REPORTING 763-493-4535 145 1 University of Minnesota policy to have a study 2 specific HIPA form approved by the IRB and 3 signed by each subject? 4 MR. ALSOP: Object to foundation. Go 5 ahead, Doctor, if you know. Also multiple. 6 A. Yeah, I believe that's correct. 7 Q. (BY DR. BARDEN) Did you violate that rule in 8 the CAFE study? 9 A. I don't -- I don't have any -- I don't have any 10 personal knowledge that we violated that. 11 Again, if I had -- if that were the case, I 12 would be surprised that it hadn't been brought 13 to my attention before, and maybe it has and I 14 forgot about it.

ATTACHMENT J

AstraZeneca Was Aware of Seroquel Risks in 2000, Records Show
By Joe Schneider and Margaret Cronin Fisk - December 5, 2008 00:01 EST Dec. 5 (Bloomberg) -- AstraZeneca Plc., facing more than 15,000 consumer claims alleging the antipsychotic drug Seroquel causes diabetes, knew about the risk as far back as 2000, according to company documents shown in federal court. AstraZeneca Global Safety Officer Wayne Geller concluded there was “reasonable evidence to suggest Seroquel therapy can cause” diabetes and related conditions, according to documents presented yesterday in federal court in Tampa, Florida. Geller drew his conclusions following a review of available studies and internal trials, according to the documents. The internal documents were shown publicly for the first time during a hearing over the qualifications of expert witnesses the plaintiffs plan to use at trial. They are to testify in a lawsuit over the drug’s effects when the proceeding begins in February. While portions of the documents were shown in court, the filings remain sealed at the request of the London-based pharmaceutical company. Seroquel, used to treat bipolar disorder, brought in $4.03 billion last year, making it AstraZeneca’s second-biggest seller after its ulcer treatment Nexium. The company is trying to broaden the medicine’s use to offset revenue lost as pricing pressure in the U.S. reduces demand for Nexium. AstraZeneca is facing claims over the drug in state and federal courts in the U.S. The company has had an additional 2,243 cases dismissed or dropped. The initial trial, set to begin Feb. 2, is based on a complaint filed by Linda Guinn of Palm Bay, Florida. Developed Diabetes Guinn claimed she developed diabetes after taking Seroquel. The drug is part a class of newer antipsychotic drugs including Eli Lilly & Co.’s Zyprexa and Johnson & Johnson’s Risperdal. Thousands of consumers have sued the companies claiming they hid the risks of the drugs and marketed them for unapproved purposes. Lilly has paid $1.2 billion to settle 31,000 claims by individuals. The U.S. Food and Drug Administration, in a Jan. 8 letter, asked AstraZeneca to provide an analysis of all its clinical trials relating to Seroquel, Paul Pennock, the plaintiffs’ lawyer, told U.S. District Judge Anne Conway yesterday, showing her the letter.

AstraZeneca responded June 26 with an analysis that showed 2.4 percent of people who began treatment with normal levels of sugar in the blood became technically diabetic after 52 weeks. Normal levels are less than 100 milligrams per 10 deciliters. Someone is labeled diabetic with a reading of 126 milligrams, said Laura Massey Plunkett, a human risk assessment specialist, who read in court from the company report. 1.4 Percent That result, compared to 1.4 percent of patients given a placebo who showed the same increase in blood-sugar levels, Plunkett testified, makes it almost 70 percent more likely that people taking Seroquel would develop diabetes than people who weren’t taking the drug. “It’s clear this compound Seroquel can cause diabetes,” she said. “I don’t think there is any real controversy about that.” AstraZeneca lawyer Jane Thorpe said that was just a part of the company’s analysis and data showed no significant increases when all people who had a sugar reading of less than 125 milligrams were included. Thorpe also questioned Plunkett’s conclusions, saying she relied only on studies that supported her views and ignored those that didn’t show an increased tendency for people taking Seroquel to develop diabetes. Studies have shown that Seroquel and similar medications known as atypical antipsychotics are associated with an increased risk of diabetes. These studies prompted the FDA to require AstraZeneca and other drugmakers to warn doctors of the risks in September 2003. AstraZeneca doesn’t intend to settle these lawsuits, company spokesman Tony Jewell said in an interview. “We intend to litigate these cases on the merits,” he said. The case is In Re Seroquel Products Litigation, 06-MD- 01769, U.S. District Court, Middle District of Florida (Tampa). To contact the reporter on this story: Joe Schneider in Tampa, Florida, at jschneider5@bloomberg.net and; Margaret Cronin Fisk in Southfield, Michigan, at mcfisk@bloomberg.net. To contact the editor responsible for this story: David E. Rovella at drovella@bloomberg.net.

AstraZeneca Seroquel Studies ‘Buried,’ Papers Show (Update3)
By Jef Feeley and Margaret Cronin Fisk - February 27, 2009 15:16 EST Feb. 27 (Bloomberg) -- AstraZeneca Plc “buried” unfavorable studies on its antipsychotic drug Seroquel, according to an internal e-mail unsealed as part of litigation over the medicine. The drugmaker failed to publicize results of at least three clinical trials of Seroquel and engaged in “cherry picking” of data from one of those studies for use in a presentation, an AstraZeneca official said in a December 1999 e-mail unsealed yesterday under an agreement between the company and lawyers for patients. The London-based company faces about 9,000 lawsuits claiming it failed to properly warn users that Seroquel can cause diabetes and other health problems. “The larger issue is how we face the outside world when they begin to criticize us for suppressing data,” John Tumas, an AstraZeneca publications manager, told colleagues in the e-mail. More than 15,000 patients have sued AstraZeneca, claiming the company withheld information of a connection between diabetes and Seroquel use from doctors and patients. Many of the lawsuits also claim AstraZeneca promoted Seroquel, approved by the U.S. Food and Drug Administration for schizophrenia and bipolar disorder, for unapproved uses. AstraZeneca fell 114 pence, or 4.8 percent, to 2,243 pence in London trading. The shares have declined 20 percent this year. “AstraZeneca has studied Seroquel extensively and shared all relevant and required data with the FDA -- both before and after the agency approved it as safe and effective,” Tony Jewell, AstraZeneca’s spokesman, said in an e-mailed statement. ‘Acted Responsibly’ “None of the documents can obscure the fact that AstraZeneca acted responsibly and appropriately as it developed and marketed Seroquel,” Jewell said. Seroquel, which generated sales of $4.45 billion in 2008, is the company’s secondbiggest seller after the ulcer treatment Nexium. AstraZeneca has denied wrongdoing in its handling of the drug and is vowing to fight the lawsuits in court. The company said today that the FDA asked for more information on its application to expand the use of Seroquel XR to include adults with generalized anxiety disorder,

following a similar request in December seeking additional data on the drug for major depression in adults. The agency has scheduled an advisory panel meeting on April 8 to review the drug’s safety and effectiveness, AstraZeneca said on Feb. 25. The U.K. drugmaker agreed to release more than 100 files with information about the drug yesterday after Bloomberg News filed a motion this month to unseal records in the case. Unsealed Documents The company and lawyers for former Seroquel users agreed to unseal thousands of documents just before appearing before U.S. Magistrate Judge David A. Baker in Orlando, Florida, for a hearing on that motion. All federal-court lawsuits over the drug have been consolidated in Orlando. The drugmaker is urging the judge to continue the confidential designation on nine other files, including some that involve what AstraZeneca told foreign regulators about Seroquel and sales representatives’ notes on doctors’ meetings. The company’s U.S. headquarters is in Wilmington, Delaware. “The court has a duty to make sure things are not inappropriately kept secret,” Baker said at yesterday’s hearing. One unsealed document showed AstraZeneca officials considered Trial 15, one of the studies Tumas described as ‘buried,” to be a problem because it didn’t produce favorable results on the issue of weight gain for patients taking the drug. Gaining weight can be a factor in the development of diabetes. ‘Smoke-And-Mirrors Job’ Richard Lawrence, an AstraZeneca official, said in a February 1997 e-mail that the company had engaged in a “great smoke-and-mirrors job” in dealing with U.S. and Canadian investigators on the trial’s results. “Adopting the approach Don has outlined should minimize (and dare I venture to suggest) could put a positive spin (in terms of safety) on this cursed study,” Lawrence said in the e- mail. It isn’t clear from the e-mail which person Lawrence is referring to. In his December 1999 e-mail, Tumas said that AstraZeneca had “buried trials 15, 31, 56,” and was considering a study listed as COSTAR. Details of the trials weren’t specified in the e-mail. He also noted that “there is growing pressure from outside the industry to provide access to all data from clinical trials conducted by industry.”

Tumas chastised colleagues for using favorable data produced by Trial 15 without disclosing the full study results, according to his unsealed e-mail. “There has been a precedent set regarding ‘cherry picking’ of data,” which was used in “the recent Velligan presentations,” he said. “Thus far, I am not aware of any repercussions regarding interest in the unreported data.” ‘Ethical Behavior’ The publications manager indicated that AstraZeneca had a favorable reputation for engaging in “ethical behavior” when it came to disclosing study results on its drugs. “We must decide if we wish to continue to enjoy this distinction,” Tumas wrote. “The reporting of the COSTAR results will not be easy,” he added. “We must find a way to diminish the negative findings. But, in my opinion, we cannot hide them.” AstraZeneca knew Seroquel was linked to weight gain and diabetes and failed to properly disclose its findings to doctors or patients, plaintiffs’ lawyer Ed Blizzard said at a press conference today. “We believe that Study 15 was buried because of the weight-gain findings,” Blizzard said. ‘Don’t Look Good’ Among the unsealed documents was a March 2000 e-mail from Tumas in which he noted a study comparing Seroquel to Risperdal, Johnson & Johnson’s rival antipsychotic drug, and Haloperidol, an older, generic antipsychotic medicine, produced data “that don’t look good.” The results showed that Seroquel failed to best Risperdal in at least five different categories and only outperformed the placebo used in the study, according to data made public yesterday. Those categories included mood, anxiety and hostility, the documents showed. Tumas warned colleagues in an e-mail that AstraZeneca officials seemed to be “highlighting the only good stuff” about the study when the overall results showed that Seroquel was “less effective than haloperidol and our competitors.” AstraZeneca also unsealed documents showing that Wayne Geller, the drugmaker’s global safety officer, acknowledged in 2000 that researchers had found some links between diabetes and Seroquel. Geller, who wrote the company’s “safety position paper” on Seroquel, said in a filing with Dutch authorities that there was “reasonable evidence to suggest that Seroquel therapy can cause impaired glucose regulation including diabetes mellitus in certain individuals.”

‘Fairly Sizeable’ The drugmaker also found a “fairly sizeable” number of cases where Seroquel users’ blood-sugar levels had been affected by the drug, Geller noted in the report, according to the unsealed documents. In August 2000, AstraZeneca said in a response to the FDA that “preclinical data has provided no evidence that Seroquel treatment in man may be associated with diabetes,” according to an unsealed document produced in the lawsuits. In a May deposition that was unsealed yesterday, Geller said the Seroquel safety paper wasn’t his final position on the drug’s links to diabetes, but rather a reflection of discussions in 2000 about the drug’s attributes. The statement about “reasonable evidence” on Seroquel’s possible links to diabetes was “an artifact of an earlier discussion document” and not a conclusion, Geller said in the deposition. Zyprexa Settlements Seroquel is part of class of newer antipsychotic drugs, including Risperdal and Eli Lilly & Co.’s Zyprexa, which studies have linked to an increased risk of diabetes. That research prompted the FDA in September 2003 to require AstraZeneca and other drugmakers to warn doctors of the risks. Lilly has agreed to pay at least $1.2 billion to settle lawsuits filed by about 31,000 patients who used Zyprexa. The Indianapolis-based drugmaker said last month it would pay an additional $1.42 billion to resolve claims by state and federal officials that it marketed the drug for unapproved uses. Lilly also agreed to plead guilty to a criminal charge. AstraZeneca officials have vowed to defend the Seroquel cases “on their individual merits” and have refused to settle any claims. Lawyers for former Seroquel users contend that AstraZeneca knew of Seroquel’s risks years before the FDA required a stronger warning. The first two lawsuits set for trial in federal court in Orlando were dismissed last month when U.S. District Judge Anne Conway determined the plaintiffs couldn’t prove Seroquel contributed to their development of diabetes. First Trial The first Seroquel trial is set for April in state court in Delaware. AstraZeneca officials have criticized lawyers for former Seroquel users for pushing to have documents in the case unsealed.

“It is not surprising that plaintiffs’ lawyers have resorted to these tactics to distract attention from their lack of success on the merits of the claims,” Jewell said in his statement. The case is In Re Seroquel Products Litigation, 06-MD-01769, U.S. District Court, Middle District of Florida (Orlando). To contact the reporters on this story: Margaret Cronin Fisk in Southfield, Michigan, at mcfisk@bloomberg.net; Jef Feeley in Wilmington, Delaware, at jfeeley@bloomberg.net. To contact the editor responsible for this story: David Rovella at drovella@bloomberg.net.

E-Mail: AstraZeneca Knew in 1997 that Seroquel Caused Weight Gain
By Jim Edwards | March 3, 2009 AstraZeneca knew as far back as 1997 that Seroquel put patients at risk of weight gain, according to the company’s own internal memos. The documents also appear to show that some AZ execs developed strategies to “neutralize” information that suggested Seroquel caused weight gain or diabetes, even after the FDA asked the company to warn patients about Seroquel’s diabetes side effect. The memos were produced in the ongoing Seroquel litigation. The contents of some of them have been reported on before, but Philip Dawdy at Furious Seasons has created a page on which they can all be downloaded, here. The documents consist of AZ’s internal emails, draft papers and promo materials. Some of them concentrate on whether the data on Seroquel was showing signs of weight gain or diabetes. And the Clinical Psychology and Psychiatry blog has a lengthy dissection of just one study that seems to show that Seroquel isn’t as good as Haldol. What follows, however, is a selection of quotes from memos written by AZ’s staff, all on the subject of weight gain, diabetes, and whether the company knew that Seroquel was causing these side effects. The memos form a timeline of who knew what, and when. Feb 12, 1997, memo from Richard Lawrence: I am not 100% comfortable with this data being made publically available at the present time … however I understand we have little choice … Lisa has done a great ’smoke-andmirrors’ job! Aug. 13, 1997, memo from Lisa Arvanitis to her colleagues: 1. Is there a competitive advantage for SEROQUEL re-weight gain which we can articulate in posters/talks/vis aids? We know we have weight gain but is it limited to the short term …? … I was really struck by how consistent the data was. Feb 24, 1999, memo from Nick Hough, regarding a single study that appeared to find weight loss among users: We must not get too carried away with ‘weight loss’ when we know the rest of our data appears to point in the other direction…

Dec. 6, 1999, email from John Tumas: The larger issue is how do we face the outside world when they begin criticizing us for suppressing data. 1998 draft paper intended for 11th ECNP conference in Paris: Clinically significant weight gain, that is more than 7 percent increase in body weight, was seen more with quetiapine [Seroquel] than placebo - 24 percent compared with four percent … Clin Drug Invest Journal, Aug 18, 1998, Zeneca sponsored study, “Effect of Clozapine-Quetiapine Combination Therapy on Weight and Glycaemic Control”: All 65 patients showed weight loss … 2000 study in International Journal of Psychiatry in Clinical Practice, written by AZ: In these patients, there was no overall effect on weight across the body mass index … March 23, 2000, memo from John Tumas: The data don’t look good. In fact, I don’t know how we can get a paper out of this. June 22, 2000, discussion document: Safety data … suggest the possibility of an association between SEROQUEL use and impaired glucose regulation including reports of new onset diabetes mellitus… the number of reports is fairly sizeable. August 2000 memo from AZ to the FDA: Thus very few cases of diabetes mellitus (and related complications), hyperglycaemia, and weight gain have been reported. AstraZeneca believes that the current US Seroquel label accurately describes patient experiences to date of these conditions. Oct. 26, 2000, memo from Dominic Aked: I agree we need to be able to tell a convincing story to our internal and external customers. I’m sure we can do this…A promotional claim ‘Seroquel is weight neutral during the long-term treatment['] should help make this distinction. [By July 2001, AZ had received at least 46 reports of diabetes in patients taking Seroquel, 21 cases of ketoacidosis or acidosis, and 11 deaths. By the end of 2003 AZ had received at least 23 more.] Headline on promotional piece with a 2001 copyright: Distinct advantages of a favorable weight profile Sept. 15, 2002, FDA letter requiring label changes re diabetes: …we have concluded that the product labelling for all atypical antipsychotics should be updated…

Nov. 26, 2002, “Objection handler” for sales reps: The company’s safety database has reassuring data concerning Seroquel’s diabetic potential… [The FDA asked AZ to send out a dear doctor letter about diabetes and Seroquel on Sept. 11, 2003. AZ did not send out the letter until January 2004.] March 4, 2003: Sales rep detail flow: Key Communications: Seroquel, unlike some other antipsychotics, is not associated with meaningful weight gain, either in the short or long term or across the recommended dosing range. Oct. 15, 2003, memo from Russell Katz to FDA: AstraZeneca completed a comprehensive internal analysis of existing data and concluded that the available data do not establish a causal link between diabetes and Seroquel. [On April 24, 2004, AZ was forced to send out a revised “Dear Doctor” letter due to the fact that the first one was misleading, as it potentially downplayed the need to continually monitor a patient’s blood sugar levels.] “VM 08 15 05″ Our objective is to neutralize customer objections to SEROQUEL’s weight and diabetes profile. Nov. 16, 2006, FDA writes a ‘Warning Letter’ regarding a sales aid: This piece is false or misleading because it minimizes the risk of hyperglycaemia and diabetes mellitus … Dec. 18, 2008, memo from FDA: The weight gain signal is significant for both adult and pediatric populations and should be elevated to the Warnings and Precautions section [of the drug's label].
 

BNET’s previous coverage of AZ and Seroquel: Exec Warned AZ on Negative Seroquel Results: “We Cannot Hide Them”; Info Later “Buried” AstraZeneca’s Sex-for-Studies Seroquel Scandal: Did Research Chief Bias the Science? Seroquel Trial: Allegations of Sex-for-Secrets; AZ Wants Papers Sealed FDA Asks AZ to Tighten Seroquel Label; 9,000 Lawsuits Start Trials Next Week The History of AstraZeneca’s Mismarketing of Seroquel FDA: J&J’s Risperdal and Lilly’s Zyprexa Are Over-Used in Kids AstraZeneca Allegedly Recruits Tigger and Eeyore to Sell Seroquel

     

ATTACHMENT K

EXCERPT
1 STATE OF MINNESOTA DISTRICT COURT 2 COUNTY OF HENNEPIN FOURTH JUDICIAL DISTRICT Case Type: Wrongful Death 3 ----------------------------Court File No. 4 Mary Weiss, on her own behalf, 62-CO-06-11934 and as the next of kin and trustee 5 of the Estate of Dan Markingson, deceased, 6 Plaintiff, 7 -vs8 Board of Regents for the University of 9 Minnesota; Dr. Stephen Olson; Dr. Charles Schulz; Institutional, 10 Review Board for the University of Minnesota; Astrazeneca Pharmaceuticals LP; 11 Astrazeneca LP and Zeneca, Inc., 12 Defendants. 13 - - - - - - - - - - - - - - - - - - - - - - - - - - - - 14 Deposition of: CHARLES SCHULZ, M.D. 15 16 17 Date: 18 Commencing at: 9:05 a.m. 19 June 22, 2007 Taken at: Gislason & Hunter, LLP 701 Xenia Avenue South Suite 500 Minneapolis, Minnesota

25 Q 10 11 A 12 Q 13 14 15 16

You see here's an e-mail from Amy VandenEynden at Quintiles to Jeannie Kenney. Do you see that? Yes, I do. And it says, "It sounds like you have a good amount of leads/potentials. Hopefully your hard work will start to pay off soon!" Exclamation point. "The specialty inpatient unit sounds like it will help out a lot with recruitment!!" Exclamation point, exclamation point.

17 18 19 20 21 22 23 24 25

Did you set up that specialty inpatient unit to make it particularly easy to recruit psychotic patients? MR. ALSOP: It's a multiple question, it's vague, but go ahead. THE WITNESS: I talked with the administration at University of Minnesota Medical Center to indicate to them that I thought having a unit that specialized in the care of people with serious psychiatric illness would improve the quality of care to 195

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 1 2 3 4 5 6 7 8

schizophrenic patients in the Twin Cities; and in addition to that, I discussed, when I made my original recommendations, that it would foster the academic goals of the medical school and our department. BY DR. BARDEN: Q You see down here it says, "Take care and try not to get too frustrated!" Trying to recruit can be frustrating sometimes. Isn't that right, Doctor? A Yes. Q Okay. Let's look at the next e-mail. Bottom of 14. Again, this is -MR. ALSOP: What page number? DR. BARDEN: Fourteen, at the bottom there. It's UM CAFE 2959. MR. ALSOP: Okay. BY DR. BARDEN: Q It says Jeannie Kenney to Amy VanderEynden at Quintiles. It says in the e-mail, "The last young woman who was referred from inpatient would have benefitted so much from this study. I put in a lot of time, meeting with the case manager and attended the discharge planning meeting. I thought the parents were interested. I do have her scheduled with Dr. Olson for outpatient med management so perhaps we can gain their trust and offer it as an option again." 196 Is this the kind of e-mail that goes back and forth between the drug company financing running the study and the people recruiting, to the best of your knowledge? MS. AHMANN: Objection, as to misstates who this is to and from, and also lack of foundation. DR. BARDEN: Back and forth? This is from Jeannie Kenney to Amy VandenEynden at Quintiles, right?

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

MS. AHMANN: Yeah, but you said it was from the drug company. DR. BARDEN: I don't think I said drug company. THE WITNESS: Yes, you did. MR. ALSOP: I think you did. DR. BARDEN: Okay. Excellent. BY DR. BARDEN: Q From the company paid by the drug company to conduct the study? MS. AHMANN: Lack of foundation. BY DR. BARDEN: Q So is this the kind of letter you see? MR. ALSOP: Objection as lacking in foundation. BY DR. BARDEN: Q Do they share information about patients and who they're 197

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

trying to recruit and whether they've spoken to the parents? A I haven't seen e-mails like this in my experience, no. Q Trying to gain their trust and -- okay. It says, "the hospital is opening a psychosis specialty inpatient unit first of April. Dr. Olson and I are very much involved in the planning so will have closer contact with first episode people who come in. We are really hoping to promote state-of-the-art care and potential for research is part of the programming. In fact, it will be discussed in the interviewing of staff so it will be nice to have a treatment team fully on board with what we are doing in the department." Have you ever seen that e-mail before today? A No. Q Okay. Next, next, this is the e-mail back. "Hi--I can understand that you are frustrated!!" Exclamation point, exclamation point. "Do what you can with subject 002. Please call if the subject gets out of window, because there is always the possibility of granting exceptions. If I am not available, contact Jennifer Frantz. As for recruitment, have you had any luck with any of the area facilities? I thought you mentioned you'd go to area sites and give them information. Is there anything we can [do to] provide to help you with this? Amy." Have

198 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 you ever seen this before today? A No, I did not. Q Okay. Next e-mail, Jeannie Kenney to Amy VanderEynden, Quintiles. "Hi Amy. Having trouble with Subject 002. His sister just died, his father has terminal cancer and now the grandmother is sick. He missed a visit and now just missed the next one. I hope I can get him in for the allowable window for the visit he missed Friday. I understand the situation is really difficult but am afraid we might be losing him when his family thinks he probably needs this treatment now more than ever. Have had another person show interest from inpatient and then the parent put the pressure on and said," quote, "'NO,'" unquote. Paragraph, parentheses, "(3rd time this has happened)," close parentheses. "Have tried to ask about concerns, etc. but usually just get a NO. So, some frustration here because we really need to get more enrollees. We've had none for January and that concerns me a lot." Did you ever see that letter before today? A No. Q The date of that is January 27, 2003. "Parent put the pressure on and said, 'NO". (3rd time this has happened)." Does that help us understand why nobody listened to Mary Weiss? 199 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 MR. ALSOP: You don't have to answer that question. There is no question. Just wait. Ask -BY DR. BARDEN: Q Based on your knowledge of the CAFE study and what's in this letter, does that help us understand why Dr. Olson ignored Mary Weiss's complaints about the study and complaints about her son's deteriorating mental health and warnings that he was going to commit suicide? MR. ALSOP: It's argumentative and a misstatement of the facts. Go ahead, Doctor. And lacking in foundation. If you have any idea, go ahead. MS. AHMANN: Same, join in the objection. THE WITNESS: I don't see the link between this e-mail and what I know of Dan's case. BY DR. BARDEN: Q Okay. Please look at page 18 on the bottom right.

17 18 19 20 21 22 23 24 25

A Q A Q

Again, this is from Jennifer I. Frantz at Quintiles, November 2002, to Jeannie Kenney. "Hi Jeannie, I made a blunder when I reviewed your consent and am really really sorry." Goes on to say, "My hope is that your IRB already caught it or that it has not been reviewed yet." I'm sorry, where are we now? We're just following right along here. I guess we kind of skipped this part? Yeah. I'm going to ask you questions about particular

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

200 sentences. "I don't think this should hold up any patients at your site as long as you explain the procedures correctly and indicate it on the form that they sign while notifying the IRB of the mistake. I am so sorry for any inconvenience this causes you. Thanks, Jen." Did you as a coinvestigator ever see this e-mail before today? A No. Q Okay. To Amy VandenEynden from Jeannie Kenney, and this is September 29, 2002, UM CAFE 2989. "Hello, I know that our site is on hold/probation for CAFE recruitment but I have some things I want to make sure I have updated." As a coinvestigator, were you ever aware that your site was on hold/probation for CAFE recruitment? A Yes. Q And what was that for, to the best of your knowledge? A My understanding was that Quintiles felt, or indicated that it is expensive to run multiple sites, and that if a site is unable to enter subjects in the study, then what they would prefer to do is work with sites that can. Q And how long were you on hold/probation? A I don't know.

ATTACHMENT L

April 27, 2010

For $520 Million, AstraZeneca Settles Case Over Marketing of a Drug
By DUFF WILSON

AstraZeneca has completed a deal to pay $520 million to settle federal investigations into marketing practices for its blockbuster schizophrenia drug, Seroquel, the Attorney General, Eric Holder, said at a news conference Tuesday afternoon. “AstraZeneca paid kickbacks to doctors as part of an illegal scheme to market drugs for unapproved uses,” Kathleen Sebelius, secretary of health and human services, said at the event in Washington. She said the company promoted drugs for unapproved uses by children, the elderly, veterans and prisoners. Glenn Engelmann, AstraZeneca’s U.S. general counsel, released a statement saying the company denies the allegations but settled the investigation with the payment. “It is in the best interest of AstraZeneca to resolve these matters and to move forward with our business of discovering and developing important, life-changing medicines — while avoiding the delay, uncertainty, and expense of protracted litigation,” Mr. Engelmann said. AstraZeneca becomes the fourth pharmaceutical giant in the last three years to pay to settle federal investigations into illegal marketing of antipsychotic drugs, a lucrative category of medications that have quickly risen to the top of United States sales charts. AstraZeneca agreed to sign a corporate integrity agreement with the federal government over its marketing of Seroquel for unapproved uses, but will not face criminal charges, company and federal officials said. The company, based in London, has been accused of misleading doctors and patients by playing up favorable research and not adequately disclosing studies that show Seroquel increases the risk of diabetes. AstraZeneca still faces more than 25,000 civil lawsuits filed on behalf of patients contending that the company did not disclose the drug’s risks. The deal would make formal an agreement in principle the company reached last October with the United States attorney in Philadelphia. At that time, AstraZeneca said in a filing

with the Securities and Exchange Commission that it had set aside $520 million in respect to the investigation. The company was facing two federal investigations and two whistle-blower lawsuits involving Seroquel sales and marketing practices. One of the investigations related to physicians who had participated in clinical trials. The other inquiry involved sales staff. As a result of aggressive marketing, Seroquel has been increasingly used for children and elderly people for indications not approved by the Food and Drug Administration. The drugs have caused rapid weight gain in children, and side effects including deaths have prompted warnings against giving the drugs to elderly patients for dementia. Although doctors are permitted to prescribe any approved drug for off-label uses, it is illegal for drug makers to promote medications for any purpose not specifically approved by the F.D.A. AstraZeneca, which reported $4.9 billion in Seroquel sales in 2009, plans to report its first-quarter financial results on Thursday. The company will join a series of American pharmaceutical companies that have paid to settle federal investigations into illegal marketing. In the largest such case, Pfizer paid $2.3 billion last September, including $1.3 billion in the biggest criminal fine of any type in United States history, for off-label marketing of the painkiller Bextra and other drugs. Bextra was withdrawn from the market in 2005. The Pfizer fine included $301 million for off-label marketing of its antipsychotic drug Geodon. Eli Lilly paid $1.4 billion in January 2009 to settle investigations into illegal marketing of its antipsychotic drug Zyprexa. Lilly’s settlement included a $515 million criminal fine, which until the Pfizer case was the largest such fine ever imposed on a corporation. In 2007, Bristol-Myers Squibb and a subsidiary paid $515 million to settle federal and state investigations into marketing of its antipsychotic drug Abilify. The newer generation of antipsychotics has surpassed cholesterol-lowering drugs to become the nation’s top-selling category of medications, accounting for $14.6 billion of the nation’s $300 billion in drug spending last year, according to the research firm IMS Health. Seroquel, a pill usually taken once or twice a day that sells for more than $4 each, was the fifth-best-selling drug in the United States last year, IMS said. As with other antipsychotics, much of that spending is by the federal government, through the Medicaid and Medicare programs.

AstraZeneca, with American headquarters in Wilmington, Del., has previously denied wrongdoing in the Seroquel investigations. It has paid $656 million to defend itself in court against more than 25,000 civil lawsuits, the company said in an S.E.C. filing in January. Those cases are only recently beginning to reach trial. The company has argued that people who were found to have diabetes after taking Seroquel already had diabetes or had existing conditions that made them at high risk of the disease. According to company e-mail messages unsealed in civil lawsuits, AstraZeneca “buried” — a manager’s term — a 1997 study that showed Seroquel users gained 11 pounds a year, while publicizing a study that asserted users lost weight. Company e-mail messages also refer to doing a “great smoke-and-mirrors job” on unfavorable studies. Gardiner Harris contributed reporting.

AstraZeneca buried study on antipsychotic Seroquel
A study indicated long ago that the pricey medication caused significant weight gain, documents show, but the drug maker declined to publicize the findings.
March 18, 2009|Washington Post

The study would come to be called "cursed," but it started out merely as Study 15. It was a long-term trial of the antipsychotic drug Seroquel. The common wisdom in psychiatric circles was that newer drugs were far better than older drugs, but Study 15's results suggested otherwise. As a result, newly unearthed documents show, Study 15 suffered the same fate as many industry-sponsored trials that yield data drug makers don't like: burial. It took eight years before a taxpayer-funded study rediscovered what Study 15 had found -- and raised serious concerns about an entire new class of expensive drugs. Study 15 was silenced in 1997, the same year Seroquel was approved by the Food and Drug Administration to treat schizophrenia. The drug went on to be prescribed to hundreds of thousands of patients around the world and has earned billions for London-based AstraZeneca International. The results of Study 15 were never published or shared with doctors, even as less rigorous studies that came up with positive results for Seroquel were published and used in marketing campaigns. The results of Study 15 were provided only to the FDA, which says that it does not have the authority to make such studies public. AstraZeneca spokesman Tony Jewell defended the Seroquel research and said the company had disclosed the drug's risks. Since 1997, the drug's labeling has noted that weight gain and diabetes were seen in study patients, although the company says the data are not definitive. The label states that the metabolic disorders may be related to patients' underlying diseases. The FDA, Jewell added, had access to Study 15 when it declared Seroquel safe and effective. The trial, which compared patients taking Seroquel and an older drug called Haldol, "did not identify any safety concerns," AstraZeneca said in an

e-mail. Jewell added, "A large proportion of patients dropped out in both groups, which the company felt made the results difficult to interpret." Details of Study 15 have emerged through lawsuits, which allege that Seroquel caused weight gain, hyperglycemia and diabetes in thousands of patients. The Houston-based law firm Blizzard, McCarthy & Nabers, one of several that have filed about 9,210 lawsuits over Seroquel, publicized the documents, which show that the patients taking Seroquel in Study 15 gained an average of 11 pounds in a year -- alarming company scientists and marketing executives. A Washington Post analysis found that about four out of five patients quit taking the drug in less than a year, raising doubts about its effectiveness. An FDA report in 1997, moreover, said Study 15 did offer useful safety data. The FDA said the study showed that patients taking higher doses of the drug gained more weight. In approving Seroquel, the agency said that 23% of patients taking the drug in all available studies experienced significant weight gain, compared with 6% of control-group patients taking sugar pills. In 2006, the FDA warned AstraZeneca against minimizing metabolic problems in sales pitches. In the years since, taxpayer-funded research has found that newer antipsychotic drugs such as Seroquel, which are 10 times more expensive, offer little advantage over older ones. The older drugs cause involuntary muscle movements known as tardive dyskinesia, and the newer ones have been linked to metabolic problems. Far from dismissing Study 15, internal documents show that company officials were worried because 45% of the Seroquel patients had experienced what AstraZeneca physician Lisa Arvanitis called "clinically significant" weight gain. In an e-mail dated Aug. 13, 1997, Arvanitis reported that across all patient groups and treatment regimens, regardless of how numbers were crunched, patients taking Seroquel gained weight. In a separate note, company strategist Richard Lawrence praised AstraZeneca's efforts to put a "positive spin" on "this cursed study" and said of Arvanitis: "Lisa has done a great 'smoke and mirrors' job!" Two years after those exchanges, in 1999, the documents show that the company presented different data at an American Psychiatric Assn. conference and at another European meeting. The conclusion: Seroquel helped psychotic patients lose weight. The claim was based on a company-sponsored study by a Chicago psychiatrist, who reviewed the records of 65 patients who switched their medication to Seroquel. But documents show that AstraZeneca held the doctor in light regard and had concerns about his methods.

St. Paul Pioneer Press (Minnesota) March 18, 2009 Wednesday U doctor scrutinized over drug research BYLINE: By Jeremy Olson jolson@pioneerpress.com SECTION: BREAKING; Education; News; Health LENGTH: 1435 words A top University of Minnesota psychiatrist's ties to a drug maker have come under scrutiny because he reported that the company's blockbuster antipsychotic, Seroquel, was significantly superior to other drugs -- despite evidence to the contrary. Two months after an internal analysis by the company, AstraZeneca, found Seroquel was no better than an older, cheaper antipsychotic, Dr. S. Charles Schulz used much of the same data to publicly report that the company's drug was "more effective." The disconnect between the company's private findings in March 2000 and the psychiatrist's optimistic report to the American Psychiatric Association in May 2000 are further evidence to critics that the drug industry can shape, revise or even conceal negative research. It also feeds concerns that drug companies are paying noted doctors such as Schulz, the U's chair of psychiatry, for research results that advance their marketing agendas. Schulz has received $112,000 in consulting fees and university grants from 2002 through 2007 from AstraZeneca according to state records, and nearly $450,000 from rival drug maker Eli Lilly. "I hope that our findings help physicians better understand the dramatic benefits of newer medications like SEROQUEL," Schulz said in an AstraZeneca news release on his 2000 report, "because, if they do, we may be able to help ensure patients receive these medications first." Fresh questions have surfaced about Schulz's old research because his name came up in documents AstraZeneca released late last month to attorneys who have sued the company. The release included the startling AstraZeneca analysis and e-mails about whether to forward this negative information to Schulz for his presentation. "The data don't look good," one executive wrote. "In fact, I don't think we can even get a paper out of this." An Internet psychiatry blog first raised questions March 2 about the research Schulz presented at the APA conference and why it lacked any of the company's findings.

"It raises troubling questions when an independent academic author presents results that are in direct opposition to the underlying data," wrote the blogger, an anonymous academic. In an interview with the Pioneer Press last week, Schulz defended his research and presentation of Seroquel as accurate and ethical. However, he acknowledged the corporate press release from his APA presentation might have exaggerated in calling Seroquel "significantly superior." "You know," he said, "I can't disagree with that." Sales of Seroquel rocketed over the past decade, turning it from a second-choice antipsychotic to a first-choice drug with one of the highest market shares in its class. But enthusiasm for Seroquel has ebbed after U.S. Food and Drug Administration required the makers of second-generation antipsychotics in 2003 to include warnings about the increased risk of diabetes. A national comparative study also reported in 2005 that four newer antipsychotics, including Seroquel, were no better than Trilafon, an older drug. CONFLICTING FINDINGS The outlook was different in 2000, when Seroquel was new and psychiatrists hoped it could match the effectiveness of Haldol but cause fewer side effects. While Haldol's arrival in the 1950s helped revolutionize the treatment of schizophrenia, the drug also asked Schulz to conduct a "meta-analysis," a review of four existing studies that compared the effectiveness of Seroquel with Haldol and placebo pills. His initial look created optimism -- showing in three of the four trials that more patients responded favorably to treatment with Seroquel than Haldol. The results compelled AstraZeneca to look deeper into the data for more positives. What they got were negatives. On a rating scale, patients in these same four studies showed a greater reduction in behavioral symptoms when taking Haldol instead of Seroquel, according to the analysis. Company officials were caught off guard, according to copies of their e-mails. "My guess is that we all (including Schulz) saw the good stuff, i.e. the meta analyses of responder rates that showed we were superior to placebo and haloperidol and then thought that further analyses would be supportive and that a paper was in order," emailed John Tumas, a head of AstraZeneca'spublications team, at the time. Schulz in an interview said his APA study included the initial "response rate" analysis, and not the additional findings from the company. He later published a paper in 2003 that was similar to the AstraZeneca analysis, measuring changes in patient behaviors and finding that Seroquel was "at least as good as" Haldol.

However, even this 2003 study has drawn skepticism. Schulz used data from five clinical trials to produce the paper, including the four already used in the AstraZeneca analysis. Despite the similar sources of data, the results in these two reports are notably different. The AstraZeneca document found that Haldol, on average, improved patient scores on the behavioral rating scale by 3.31 more points than Seroquel. In Schulz's paper, Haldol was only 0.12 points better. Likewise, AstraZeneca found that Haldol was slightly better at reducing depression and anxiety, but Schulz reported that Seroquel was slightly better. "These two things, they don't go together," said Glen Spielmans, an assistant professor of psychology at Metropolitan State University, who has reviewed the documents. "Either the (company) analysis was wrong or Schulz's presentation was wrong." The added fifth study showed little difference between Haldol and Seroquel overall, and certainly couldn't account for the three-point difference between Schulz's report and conclusions. "Somehow, three points vanish," Spielmans said. "It's not just the spin. The numbers are different, too." Schulz said it probably wasn't just the fifth study but updates to the other four studies that made the difference. He said he can "understand the suspicions" of his work, but in the end his 2003 report only said that Seroquel and Haldol were equivalent. Other wellaccepted studies at the time had reached the same conclusion. The strategies in this instance seem textbook for pharmaceutical company research, said Dr. Erick Turner, an Oregon researcher who previously reviewed psychotic drugs for the FDA. His study last year in the New England Journal of Medicine found that positive drug studies are almost always published, whereas negative studies are either not published, distorted to emphasis any positives, or combined with positive studies to dilute their impact. Meta-analysis reports can also be distorted, he said, by mixing a few negative studies with more positive ones to forecast a desired outcome. "They just blanket the universe with these glowing reports," he said, "and the negative stuff is in there a little bit." CHANGING PRACTICES Schulz's research is not a central focus of the lawsuits against AstraZeneca, which accuse the London-based company of withholding information about Seroquel's elevated risk of

diabetes and obesity. Attorneys for patients who suffered these side effects are trying to prove a pattern of cover-ups by the company, though. Schulz said he was never asked by AstraZeneca to sell or spin his results, nor has he ever agreed to withhold negative research. Case in point: Schulz published an industry-funded study in 2008 that found that another antipsychotic, Zyprexa, was no better than a placebo pill in treating borderline personality disorder. Schulz said he has cut back his role with drug companies -- mostly forgoing paid lectures to clinics -- because of the perception of bias. University of Minnesota leaders are also considering tighter rules about how and when industry money can be accepted. "As these issues have begun to unfold, I decided I didn't want to do anything that would hurt my institution," he said. "I've curtailed activities for that reason." Schulz does have ongoing research projects funded by industry. An Eli Lilly grant allowed him to examine the potential of medical imaging to analyze brain activity in mental disorders and to then select the best treatment. A U spokesman said that the dean of the medical school, Dr. Deborah Powell, is aware of the controversy over Schulz's research and has offered him her full support. Dr. Carl Elliot, a U bioethics professor, said the conflicting AstraZeneca studies are suspicious, but it's difficult to know if Schulz is at fault. "Was Schulz fooled?" Elliott asked. "Or was he complicit?"

Schizophrenia Research 62 (2003) 1 – 12 www.elsevier.com/locate/schres

The efficacy of quetiapine vs. haloperidol and placebo: a meta-analytic study of efficacy
S. Charles Schulz a,*, Ruth Thomson a, Martin Brecher b
a

Department of Psychiatry, University of Minnesota Medical School, F282/2A West 2450 Riverside Avenue, Minneapolis, MN 55454, USA b SeroquelR, AstraZeneca, Wilmington, DE, USA Received 16 August 2002; received in revised form 19 November 2002; accepted 25 November 2002

Abstract Introduction: Atypical antipsychotics form a new class of treatment for psychotic disorders that offers advantages over conventional antipsychotics, such as haloperidol. Among these advantages is a lower risk of side effects—in particular movement disorders. The atypical antipsychotics that are currently commercially available are clozapine, risperidone, olanzapine, quetiapine, and ziprasidone. The focus of this report is on the efficacy of quetiapine. Data Sources/Study Selection: A meta-analysis was performed on three placebo- and five haloperidol-controlled clinical trials of quetiapine. Efficacy was assessed using the Brief Psychiatric Rating Scale (BPRS), the Positive and Negative Syndrome Scale (PANSS), the Clinical Global Impression (CGI), and the Scale for the Assessment of Negative Symptoms (SANS). In addition, a responder analysis was performed assessing patients who demonstrated a 40% improvement on the BPRS total score. Results: The results showed that quetiapine was significantly ( p < 0.05) superior to placebo in improving psychotic symptoms. In addition, quetiapine was not significantly different from haloperidol on measures of efficacy measured by BPRS change score, but was superior to haloperidol in terms of response rate using observed case analysis (but not using last observation carried forward analysis). Conclusion: This meta-analysis supports the use of quetiapine as a front-line treatment for schizophrenia. D 2003 Elsevier Science B.V. All rights reserved.
Keywords: Atypical antipsychotics; Quetiapine; Movement disorders; Schizophrenia

1. Introduction Pharmacologic treatment of the psychotic symptoms of schizophrenia emerged in the early 1950s with the observation that chlorpromazine was not only calming, but also reduced hallucinations and delusions (for a review of early studies, see Angrist and

* Corresponding author. Tel.: +1-612-273-9820; fax: +1-612273-9817. E-mail address: scs@umn.edu (S.C. Schulz).

Schulz, 1990). The substantial optimism for this new treatment was later tempered as both transient/reversible and persistent movement disorder side effects were observed and demonstrated to be associated with neuroleptic use. Furthermore, as time went by there was an agreement that the acceptability of the neuroleptics was low—a frustrating situation for patients and physicians as these agents (e.g., chlorpromazine, haloperidol, thiothixene) were the only class of medications that were demonstrated to be efficacious. Therefore, investigators moved their focus to compounds that could reliably reduce psychotic symptoms

0920-9964/03/$ - see front matter D 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0920-9964(02)00522-4

Charles Schulz under scrutiny for Seroquel study suicide
Is U of M department of psychiatry chair in the pocket of AstraZeneca?
By Andy Mannix Wednesday, Feb 2 2011 Mary Weiss knew something wasn't right with her son. Only a year before, Dan Markingson had seemed perfectly normal. But his latest letter from Los Angeles suggested a troubled mind. He claimed he was about to become famous. He was at a crossroads in his life, and would soon have more free time. He even had a big movie premiere in the works. "I knew then that something was wrong," says Weiss. "I knew that there wasn't a premiere, and when he said he was going to have a lot more free time, I thought he was quitting his job." Weiss immediately jumped in her car and drove to California. When she arrived, she found her son far worse off than she'd feared. He was talking nonsense and couldn't be reasoned with. Weiss tried to convince Markingson to come back to Minnesota, where she could look after him. But he had a stipulation: He would only return home if his dead grandmother Daisy told him to. Weiss went to an internet cafe down the street and created an email account under the name "GuardianAngelDaisy." Pretending to be her own deceased mother, she urged Markingson to return to Minnesota. Eventually, he agreed. He was home for only 10 days before he decided to return to California. Weiss pleaded with him to stay, but he refused. She could either drive him to the airport, or never see him again. Weiss followed him to Los Angeles, where she again tried to urge her son to go back to Minnesota. But this time, his grandmother's emails weren't enough. Markingson wanted to talk to a higher authority: Michael the Archangel. Weiss created another fake email account as Archangel Michael. The two exchanged emails for more than a week before Markingson finally agreed to fly home.

Once he was back, Weiss called the South St. Paul police. An officer came to her home to evaluate her son. During the interview, Markingson casually mentioned he would soon be attending a devil-worshipping event in Duluth, and might be ordered to kill people. That triggered a trip to Fairview University Medical Center, where Markingson was diagnosed with psychosis and placed on a 72-hour hold. In order to be released, Markingson agreed to a stay of commitment, which would allow him to leave the hospital as long as he followed a treatment plan. The plan involved Markingson enrolling in a study called Comparison of Atypicals in First Episode, or CAFE. The research was sponsored by AstraZeneca, maker of Seroquel, one of the antipsychotic drugs being investigated. When Weiss found out her son was a human guinea pig, she was furious. She called the hospital and tried to pull her son out of the treatment plan, to no avail. Although Markingson was mentally unfit, he was somehow able to consent to the drug trial. Over the next few months, Markingson's condition only worsened, Weiss says. His doctor wouldn't return her calls, so she tried writing a letter to the head of the department, Dr. S. Charles Schulz. He didn't reply. It wasn't until April 28, after Weiss's third letter, that she received a cursory response, in which Schulz wrote, "it was not clear to me how you thought the treatment team should deal with this issue." Ten days later, on May 8, Markingson sat in the bathtub of the halfway house where he was staying and stabbed himself to death with a box cutter. "I left this experience smiling!" read the suicide note.

MARKINGSON'S SUICIDE HAS cast a harsh spotlight on the University of Minnesota psychiatry department. The Federal Drug Administration, the Attorney General's Office, and the college's Internal Review Board all wanted to know how a 26year-old research subject ended up dead. So did his mom. After a year of combing through studies and public records, Weiss filed a malpractice suit against Schulz and the U of M, accusing them of putting Big Pharma's bottom line ahead of her son's mental health. In December, a group of eight bioethicists at the U of M wrote a letter to the college's Board of Regents, demanding the appointment of an independent board to investigate whether lapses in ethics and judgment led to Markingson's suicide.

"This goes beyond everything and anything, and this should have brought the house down on the university," says Vera Sharav, president of the Alliance for Human Research Protection, a patient-advocacy group. "There has to be zero tolerance, because a lot hangs on it, including lives." The issue will soon come to a head. The U of M has been investigating a complaint about Schulz's connections to Big Pharma and is expected to issue the results in a matter of weeks. "If there's any question that the investigation was superficial, it ought to be by an independent group that can determine what the facts are," says Jerome P. Kassirer, former editor of the New England Journal of Medicine, who is familiar with the circumstances surrounding the Markingson case. "It looks worrisome to me."

DR. SCHULZ DIDN'T start out on a traditional path to psychiatry. As an undergrad at the University of Southern California, he majored in history even though he planned on becoming a doctor. Years later, as a medical student at the University of California-Los Angeles, Schulz still didn't know what field he would specialize in. He thought of practicing general medicine, but a visit to the Camirillo State Mental Hospital brought his life into focus. At the time, conditions like schizophrenia and autism were generally considered to be untreatable. Camirillo was a progressive facility, and one of the few hospitals to offer hope. "I...was stunned by how seriously ill the schizophrenic patients were and worried about who was going to care for them," Schulz explains in an email. It was then that he decided to pursue psychiatry. By the time he graduated from college and finished his residency, Schulz had begun to narrow his focus to schizophrenia in teenagers. Thorazine, Haldol, and Stelazine were the only drugs available at the time, and Schulz was troubled by their long-term side effects, such as impaired motor movement. So when a second generation of antipsychotics known as Atypicals hit the market, Schulz plunged in with both feet. Throughout the '80s and '90s, Schulz earned headlines in highprofile scientific journals studying the new drugs. In 1999, Schulz received an offer for a prestigious position at one of the best medical schools in the country. He accepted the offer to become the head of psychiatry at the University of Minnesota.

AROUND THIS TIME, AstraZeneca introduced a new pill to the world of antipsychotics: Seroquel. According to AstraZeneca's marketing campaign, Seroquel was the next big thing in treating schizophrenia. It would make older drugs like Haldol and Thorazine look primitive. But Seroquel wasn't exactly flying off the shelves. When Schulz came to the U of M, the drug was in its third year on the market and its sales paled in comparison to that of competitors. In an effort to change this, AstraZeneca commissioned studies to prove Seroquel's superiority to the existing drugs on the market. Internal emails detail a plan to place the results of these studies in a scientific journal. The man who would help make it happen was the University of Minnesota's new head of psychiatry: Charles Schulz. He was scheduled to present an analysis of the research at the upcoming American Psychiatric Association conference. But in March 2000, all was not going according to plan. "The data don't look good," AstraZeneca publications manager John Tumas wrote on March 23, 2000. "What seems to be the case is that we were highlighting the only good stuff, and that our own analysis supports the 'view out there' that we are less effective than Haloperidol and our competitors." Yet Schulz painted a much rosier picture at the APA conference just two months later. Despite AstraZeneca's findings that Seroquel did not perform as well as older antipsychotics, Schulz declared it "significantly superior" to haloperidol, a competing medication. "I hope that our findings help physicians better understand the dramatic benefits of new medications like Seroquel because, if they do, we may be able to help ensure patients receive these medications first," Schulz said in an AstraZeneca press release sent out that day. That wasn't the last time Schulz did the heavy lifting to promote Seroquel. A review published in Science Direct claimed a generic version of Seroquel called quetiapine was superior to haloperidol in some measures and should be used as a "front-line treatment for schizophrenia." While Schulz says he designed the study, AstraZeneca clearly had a strong hand. The third person listed as an author is Martin Brecher, the executive director of medical

science for the drug company. A footnote at the end of the study credits the pharmaceutical company's role in preparing the article. That's not the only questionable decision. A graph in one of the seven studies Schulz cites is actually incorrect, says Dr. Glen Spielmans, a psychiatry professor at Metro State University who analyzed Schulz's work for City Pages. The article references an earlier trial conducted by five Japanese psychiatrists comparing the two drugs. Schulz's article shows a graph of quetiapine slightly outperforming haloperidol. However, the original study shows the two performed exactly the same, Spielmans says. "How do you come up with this?" asks Spielmans. "Sometimes honest mistakes happen. I'm not at all implying that this was intentional cooking of the books or anything, but something's not right."

OVER THE YEARS, Schulz continued to conduct research funded by AstraZeneca. In 2001, he agreed to be the co-investigator for Study 41, designed to prove that Seroquel SR—a new, longer-lasting version of the drug—was more effective than a placebo. AstraZeneca set up 45 study sites in the United States and four in Canada. For six weeks at a time, psychiatrists interviewed patients every few days. But after a year of running the tests, it was clear that the study was a failure—Seroquel SR was barely more effective than a sugar pill. Instead of publishing the results, however, AstraZeneca decided to keep a lid on the data, according to internal emails. "This information should be kept in the strictest confidence," warned Scott French, AstraZeneca's study delivery manager. A month later, Seroquel's brand manager, Simon Hagger, also emphasized the importance of suppressing the negative data. "You will all be aware that we are under clear instruction from the highest level within AstraZeneca at this time not to discuss the details surrounding trial 41 with any external customers," Hagger wrote in an email to fellow employees. A few years later, AstraZeneca decided to re-run the trial in the hope of achieving more favorable results. Now named Study 132, the design was almost exactly like Study 41, save for a few minor variable changes. This time, however, the results demonstrated longer-lasting Seroquel's positive effects.

Study 132 was published in the Journal of Clinical Psychiatry in June 2007. Schulz is listed as the second author, even though all of the trial sites were held out of the country. Six of the other seven authors of the study are employees of AstraZeneca or on the company's advisory board. Studies 41 and 132 represent a widespread controversy in clinical trials controlled by pharmaceutical companies. Industry-designed studies almost always produce favorable results or never see the light of day, explains Dr. Eric Campbell, a Harvard Medical School professor who has studied this trend. "If a study is funded by industry, it's very likely—if not certain—that by the time that study reaches publications, it's going to favor that company's products and services," says Campbell. "At the end of the day, what matters for the drug companies is getting new drugs to market. And convincing physicians to use them, regardless of whether they're better." Schulz says more than marketing interests were at stake. "Study 132 was designed for the approval of the long-acting quetiapine," he maintains, adding that all the data were properly submitted. "It was designed in order to receive FDA approval."

MARKINGSON HAD COLLECTED a strange assortment of items in the few months he lived in the halfway house: a pile of clothes clearly too big to fit, a couple of books, and an old raincoat. Weiss stepped into the bathroom; she was visiting within two hours of finding out her son had died. No one had bothered to clean up after the coroner came to collect Markingson's body—or if they did, hadn't done a very thorough job. A few inches of black fluid still ringed the bottom of the bathtub. "I was in shock," says Weiss. "I was just numb." In Markingson's wallet, Weiss found a $20 check from the CAFE study. The next few weeks went by like she was watching someone else's life. Weiss couldn't accept that she would never see her son again. She wanted to understand why it had happened. Weiss began digging up documents and filed a formal request with the University of Minnesota, obtaining a file of her son's medical records. She was stunned by what she found. It turned out that when Markingson arrived at Fairview, his doctor, Stephen Olson, had approached him about a treatment plan involving the CAFE study. Olson was also the

principal investigator for the study, and has received more than $100,000 from AstraZeneca over the years. More alarming was that Schulz—the man to whom Weiss had appealed for help getting her son out of the program—was a co-investigator on the CAFE study. "He had the opportunity to take Dan out of the study," Weiss says of Schulz. "He had the opportunity and he didn't." In summer 2005, Weiss hired two attorneys to represent her in a malpractice lawsuit against Schulz, Olson, the University of Minnesota, and AstraZeneca. One of the lawyers was Dr. Chris Barden. As both a psychologist and an attorney, he had the expertise to dissect the complicated case. Barden subpoenaed Schulz for two depositions over the course of the next year. On video camera, Barden pressed the professor on his long history of taking money from pharmaceutical companies and whether it influenced his judgment in the CAFE study. "From '99 through the time that Dan Markingson was a subject in the CAFE study—that is through 2004—how much money have you received from drug firms?" Barden asked. "I can only make an estimate," Schulz replied, pegging the number somewhere between $150,000 and $180,000. In a separate deposition, Barden read an excerpt from a bioethics book arguing for the importance of informing patients about a doctor's financial ties to drug companies. "Do you agree or disagree with that statement?" asked Barden "I don't agree with that statement," replied Schulz, arguing that disclosing this information could "confuse" the situation. "Have you had any training in biomedical ethics?" pressed Barden. "I've taken the courses at the University of Minnesota that are required for us to participate in clinical research." "And isn't this part of that training?" "I'm not aware," said Schulz. "I don't recall that." A cornerstone of Barden's case was Markingson's monetary value to the U of M as a patient. The university stood to make $15,648 for every subject who completed the trial—a total of $327,000.

But the judge ruled that there wasn't enough evidence to prove culpability. Olson settled for $75,000—an amount Weiss says didn't even cover her legal fees. Schulz, AstraZeneca, and the U of M got off scot-free.

ON A WEDNESDAY morning in March 2009, Weiss sat at a long, rectangular conference table in the Minnesota Capitol building. It had been five years since her son's suicide, and her anger had not ebbed one bit. "Right now, what I want to see is that no other family has to go through what I went through, and has to lose a child totally unnecessarily," Weiss told the legislators. She had contacted Rep. Karla Bigham, who had in turn authored a bill that would make it illegal for a patient under a stay of commitment to be enrolled in a clinical study. "I thought this was just wrong," Bigham says. "If somebody would have listened to her, and if these safe measures were there, it might not have happened." Weiss's story made an impression. After she testified, the Minnesota Legislature unanimously approved the bill, and Dan's Law went into effect in August 2009. But that didn't mean that Weiss had finished her quest. She and family friend Mike Howard had continued investigating Schulz and the U of M psychiatry department in an effort to expose their role in the tragedy. The deeper Weiss and Howard looked, the more the evidence they discovered about Schulz's relationship with Big Pharma. From 2005 to 2010, he had received at least $522,000 from pharmaceutical companies for research, consulting fees, and other compensation, according to public disclosure reports. He'd been paid more than $86,000 by AstraZeneca alone. Yet for all the boxes of carefully annotated studies, Weiss didn't have a forum for her case. Then Howard found the University of Minnesota's code of conduct. Weiss may no longer have legal standing, but she could still file a grievance with Schulz's employer. The complaint accuses Schulz of allowing drug companies to use his name on studies he didn't write or research, hiding research results that didn't come out favorably to drug company interests, and violating the U of M's ethics policies by not disclosing his financial relationships to patients.

For his part, Schulz says he's done nothing wrong, and maintains that he hasn't violated any U of M policies. The FDA, state Attorney General's Office, and the U of M's Internal Review Board all found no link between Markingson's suicide and the CAFE study. Schulz also points out that he started looking into quetiapine before it was even branded as Seroquel by AstraZeneca. "I haven't been particularly focused on quetiapine, in my opinion, but on the best ways to help schizophrenic people," Schulz wrote in an email. A college spokesman confirmed the probe is "pending," but wouldn't give any further details. A decision is expected in the next couple of weeks. In the meantime, the outrage has only grown. Eight U of M bioethicists wrote a letter to the Regents asking that an independent board look into the Markingson case. Other faculty members and the Graduate and Professional Student Assembly have formally joined the push for an outside investigation. "Schulz should be fired," Weiss says. "And if he's not, I will certainly, until the end of my life, make his life miserable."

ATTACHMENT M

making a killing

54 m o t h e r j o n e s |

s e pte m b e r / o c to b e r 2 0 1 0

artcredit tk

Clinical trials have become marketing exercises for Big Pharma – and cash-strapped universities are helping make the sale. Too bad for Dan Markingson.
by carl elliott illustrations by sam weber

Making a Killing
it’s not easy to work up a good feeling about the institution that destroyed your life, which may be why Mary Weiss initially seemed a little reluctant to meet me. “You can understand my hesitation to look other than with suspicion at anyone associated with the University of Minnesota,” Mary wrote to me in an email. In 2003, Mary’s 26-year-old son, Dan, was enrolled against her wishes in a psychiatric drug study at the University of Minnesota, where I teach medical ethics. Less than six months later, Dan was dead. I’d learned about his death from a deeply unsettling newspaper series by St. Paul Pioneer Press reporters Jeremy Olson and Paul Tosto that suggested he was coerced into a pharmaceutical-industry study from which the university stood to profit, but which provided him with inadequate care. Over the next few months, I talked to several university colleagues and administrators, trying to learn what had happened. Many of them dismissed the story as slanted and incomplete. Yet the more I examined the medical and court records, the more I became convinced that the problem was worse than the Pioneer Press had reported. The danger lies not just in the particular circumstances that led to Dan’s death, but in a system of clinical research that has been thoroughly co-opted by market forces, so that many studies have become little more than covert instruments for promoting drugs. The study in which Dan died starkly illustrates the hazards of market-driven research

s e pte m b e r / o c to b e r 2 0 1 0 | m o t h e r j o n e s

55

making a killing

and the inadequacy of our current oversight system to detect them. Mary Weiss is a slight, white-haired woman in her late sixties who smiles ruefully at any question, no matter how painful. She is the sort of Minnesota liberal who volunteers for political campaigns and signs her email with flowers. When we first met at a coffee shop in St. Paul, she was wearing an Obama pin on her sweater. Mary raised Dan alone, working a job at the postal service. Old photographs show Dan growing into his good looks; according to Mary, he was also a gifted student. In high school, Dan got a perfect score on the verbal portion of his sat. He graduated from the University of Michigan in 2000 with an English degree, and that fall he moved to Los Angeles, hoping to become a screenwriter or an actor. To support himself, he got a job as a celebrity-tour bus driver. When Mary went out to Los Angeles for a visit in the summer of 2003, it was clear Dan had changed. He’d adopted a new last name, Markingson. His behavior was bizarre. “He said, ‘You haven’t told me when the event is going to be,’” Mary said. She had no idea what he was talking about. The next day, he took her to his apartment. He’d encircled his bed with wooden posts, salt, candles, and money, which he said would protect him from evil spirits. He showed her a spot on the carpet that he said the aliens had burned. I asked Mary how she’d reacted to all of this. “I panicked. I called 911,” she replied. But when the police arrived, Dan was able to convince them she had overreacted. “He said, ‘Oh, my mother just drove from Minnesota and she’s very tired,’” she recalled. Worried that Dan was seriously ill, she tried to convince him to return to St. Paul. He visited her in August, returned briefly to California, and then came back to St. Paul in October. Dan grew convinced that the Illuminati were orchestrating an event in Duluth, Minnesota—a “storm” in which he would be called upon to murder people, including Mary. Some of his emails from late September 2003 suggest the extent of his delusions: “I’m aware that people can cast spells that can hurt you at a distance. I’m aware that some people can read minds. I’m aware that some people might actually be ‘hybrids’ and not altogether human.”
56 m o t h e r j o n e s |
s e pte m b e r / o c to b e r 2 0 1 0

In another email, Dan wrote: “I’m especially eager to attend this storm and SLAY those who deserve slaying. I will choose victims immediately… I HAVE NO EMOTIONAL ATTACHMENTS. I KILL FOR FUN!!” On November 12, Dan said he would kill Mary if called upon to do so. She called the police. Dan was taken to Regions Hospital in St. Paul. But the hospital had no psychiatric beds available, so after a few hours Dan was transferred to Fairview University Medical Center, a teaching hospital for the University of Minnesota Academic Health Center. He was treated by Dr. Stephen C. Olson, an associate professor in the university’s psychiatry department, who prescribed Dan Risperdal (risperidone), an antipsychotic drug often prescribed for patients with schizophrenia or bipolar disorder. (In Minnesota, doctors are allowed to give antipsychotic drugs to mentally incompetent patients without their consent for up to 14 days, but only to prevent serious,

medical treatment. Instead, he proposed that Dan take part in an industry-funded study of antipsychotic drugs. The university’s study coordinator, Jean Kenney, had Dan sign a consent form when Mary wasn’t present, and on November 21, he was enrolled in the study. On the surface, the study appeared benign. Its purpose was to compare the effectiveness of three “atypical” antipsychotic drugs, each of which had already been approved by the fda: Seroquel (quetiapine), Zyprexa (olanzapine), and Risperdal (risperidone.) The study was designed and funded by AstraZeneca, the manufacturer of Seroquel, and it called for 400 subjects experiencing their first psychotic episode to take one of the three drugs for a year. AstraZeneca called it the “cafe” study, which stood for “Comparison of Atypicals in First Episode.” The management of the cafe study had been outsourced to Quintiles, a contract research organization, which was conducting it at 26 different sites, including

How much of a risk to human subjects is justified in a study whose aim is to “generate commercially attractive messages”?
immediate physical harm to the patient or others.) Olson believed Dan was psychotic and dangerous, and lacked the ability to make decisions regarding his treatment; on November 14 he signed a document that recommended Dan be committed involuntarily to a state mental institution, noting that he “lacks the capacity to make decisions regarding such treatment.” Three days later, a clinical psychologist also recommended involuntary commitment, reiterating that Dan had threatened to slit his mother’s throat. In Minnesota, patients who have been involuntarily committed are given another option: a “stay of commitment.” Patients can avoid being confined to a mental institution as long as they agree to comply with the treatment program laid out by their psychiatrist. On November 20, Olson asked for a stay of commitment. The court granted the stay for six months, stipulating that Dan had to follow the recommendations of his treatment team. Olson, however, did not simply recommend standard the University of Minnesota. (For more on cros, see “Trial by Hire,” page 60.) Yet the cafe study was not without risks. It barred subjects from being taken off their assigned drug; it didn’t allow them to be switched to another drug if their assigned drug was not working; and it restricted the number of additional drugs subjects could be given to manage side effects and symptoms such as depression, anxiety, or agitation. Like many clinical trials, the study was also randomized and double-blinded: Subjects were assigned a drug randomly by a computer, and neither the subjects nor the researchers knew which drug it was. These restrictions meant that subjects in the cafe study had fewer therapeutic options than they would have had outside the study. In fact, the cafe study also contained a serious oversight that, if corrected, would have prevented patients like Dan from being enrolled. Like other patients with schizophrenia, patients experiencing their first psychotic episode are at higher risk of killing themselves or other people. For

this reason, most studies of antipsychotic drugs specifically bar researchers from recruiting patients at risk of violence or suicide, for fear that they might kill themselves or someone else during the study. Conveniently, however, the cafe study only prohibited patients at risk of suicide, not homicide. This meant that Dan—who had threatened to slit his mother’s throat, but had not threatened to harm himself—was a legitimate target for recruitment. When Mary found out that Dan had been recruited into the cafe study, she was stunned. “I do not want him in a clinical study,” she told Olson. Just a few days earlier, Olson indicated in a petition to the court that Dan was both dangerous and mentally incapable of consenting to antipsychotic medication. How could he now be capable of consenting to a research study with the very same antipsychotics—especially when the alternative was commitment to a state mental institution? After Dan was enrolled, he stayed at Fairview for about two more weeks. By that point, Olson thought Dan’s symptoms were under control, but Mary was still very worried by his erratic behavior. She recalls meeting with the doctor: “Olson came in and sat down and opened his file and said, ‘Oh, Dan is doing so well.’ And I said, ‘No, Dr. Olson, Dan is not doing well.’ I think he was taken aback.” Even so, on December 8, 2003, Dan was transferred to Theo House, a halfway house in St. Paul. He was required to sign an agreement confirming that he understood he could be involuntarily committed if he didn’t continue taking his medication and keeping his cafe study appointments. At the halfway house, Dan often stayed in his room for days. On March 26, 2004 nearly four months after his discharge from Fairview, his thoughts were still “delusional and grandiose,” according to a social

Dan and his mom, Mary Weiss

worker’s note. An occupational-therapy report from April 30 detailed Dan’s condition: “Personal appearance disheveled. Isolated and withdrawn. Poor insight and self-awareness.” Entries in a personal journal that Dan kept during this period don’t show any obvious changes, suggesting that he was improving little, if at all. Mary felt he was becoming angrier. “He was so tense, with this ready-to-explode quality.” Olson saw things differently. “I disagree that he had significant deterioration,” he testified in a 2007 deposition. However, it’s unclear whether Olson actually saw Dan enough to make an informed judgment about his condition. Records suggest most of Dan’s care was managed by social workers. In his deposition, Olson said he saw Dan approximately six times from the date he was admitted in November until he committed suicide in May. Whatever the doctor thought, his actions don’t suggest that he

felt Dan was improving. In late April 2004, as Dan’s stay of commitment was about to expire, Olson recommended extending it for another six months—the duration of the cafe study. He noted that Dan still had “little insight into his mental disorder” and might “place himself at risk of harm if he were to terminate his treatment.” Mary tried to get Dan out of the study or have his treatment changed. She called Olson and tried to see him. She wrote long, detailed letters expressing concerns about everything from Dan’s diet and sleep habits to his medications. In total, she sent five letters to Olson and Dr. Charles Schulz—the chairman of the university’s psychiatry department and a co-investigator on the cafe study—communicating her alarm about Dan’s condition, especially his inner rage. She received only one reply, dated April 28, from Schulz, who wrote that “it was not clear to me how you thought the treatment team should deal with this issue.” Around that time, Mary left a voice message with Jean Kenney, the study coordinator, asking, “Do we have to wait until he kills himself or someone else before anyone does anything?” Before dawn on the morning of May 8, a police officer and a Catholic priest knocked on Mary’s door. Mike Howard, a family friend who lives at her house, answered. Later, in a deposition, Howard described what happened next: “Mary jumped out of her bed and went into the kitchen and stood there, and the priest extended his hand out and said, ‘Mary, I’m here to tell you that Dan passed away.’ And Mary just literally fell down to her knees and started to shriek and cry, and just started begging, ‘Please, no, no, don’t let this happen.’” Dan had stabbed himself to death in the bathtub with a box cutter, ripping open his abdomen and nearly decapitating himself. His body was discovered in the early hours
s e pte m b e r / o c to b e r 2 0 1 0 | m o t h e r j o n e s

courtesy mary weiss

57

of the morning by a halfway-house worker, along with a note on the nightstand that said, “I left this experience smiling!” Later, when the blind on the study was broken, researchers found that Dan was being treated with Seroquel, the drug manufactured by the study sponsor, AstraZeneca. For most of the past half-century, physicians have considered antipsychotic drugs to be among the most unpleasant chemicals in the medicine closet. Thorazine (chlorpromazine), the first antipsychotic, was developed in 1950, and while it could relieve some of the worst symptoms of schizophrenia, that relief came at a serious cost. Not only do antipsychotics often make patients feel sedated and sluggish (they used to be called “major tranquilizers”), they can also cause irreversible “extrapyramidal” symptoms, such as the shuffling gait, rigid muscles, and involuntary lip-smacking sometimes seen in patients who have been taking the drugs for years. The antipsychotics can also cause akathisia, a type of driven, agitated restlessness that ranges from unpleasant to excruciating. Until recently, psychiatrists reserved the drugs for patients with very severe mental illnesses. Over the past decade or so, however, antipsychotics have undergone an extraordinary rehabilitation. By 2008, they were the most lucrative class of drugs in America. Seroquel alone had nearly $4 billion in sales, making it the country’s fifth most profitable drug. The transformation began in the mid-’90s, when pharmaceutical companies began pitching atypical antipsychotics such as Risperdal, Zyprexa, and Seroquel as more effective than older antipsychotics, but relatively free of their ugly side effects. The drugs were also very expensive—one study pegged the cost at 70 to 100 times that of an older drug—but if they didn’t produce extrapyramidal symptoms, their enormous expense seemed justifiable. By the mid-2000s, atypicals were being prescribed not just for schizophrenia but also for anxiety, agitation, insomnia, attention deficit hyperactivity disorder, and depression. The most remarkable upswing came for patients diagnosed with bipolar disorder, which used to be seen as a rare illness. Once bipolar disorder could be treated with atypicals, rates of diagnoses rose dramatically, especially in children. According to a recent
58 m o t h e r j o n e s |
s e pte m b e r / o c to b e r 2 0 1 0

Columbia University study, the number of children and adolescents treated for bipolar disorder rose 40-fold between 1994 and 2003. Another study found that nearly one in five children who visited a psychiatrist came away with a prescription for an antipsychotic drug, despite early reports of alarming side effects. Recent years have seen a backlash. The most damaging blow to the atypicals was an authoritative 2005 study funded by the National Institute of Mental Health—the so-called catie study—which found that the atypical antipsychotics worked no better than a much older antipsychotic called Trilafon (perphenazine), which was developed in the 1950s. The catie study also found that, contrary to the way the drugs had been marketed, side-effect profiles of the atypicals were Dan at age 10 generally no better than the older drug. Other research showed that atypicals were associated with significant weight gain, increased risk of diabetes, and greater possibility of death in patients with dementia. After another large analysis in The Lancet found that most atypicals actually performed worse than older drugs, two senior British psychiatrists penned a damning editorial that ran in the same issue. Dr. Peter Tyrer, the editor of the British Journal of Psychiatry, and Dr. Tim Kendall of the Royal College of Psychiatrists wrote: “The spurious invention of the atypicals can now be regarded as invention only, cleverly manipulated by the drug industry for marketing purposes and only now being exposed.” The cleverest manipulation has been with the clinical trials themselves. For years, critics have charged that pharmaceutical companies massage trials to make their own drugs look better than they really are. One common tactic is to suppress unfavorable data. A notorious example came in the 1990s, when a Wyeth safety officer overwrote the company’s computer files, erasing evidence indicating that its diet drug, fen-phen, caused valvular heart disease. A less risky strategy is simply not to publish potentially damaging trials. In 2004, the

Canadian Medical Association Journal described a leaked document indicating that GlaxoSmithKline had deliberately hidden two studies from regulators showing that its antidepressant, Paxil (paroxetine), could increase the risk of suicide in children. The company has paid nearly a billion dollars in legal settlements over Paxil, including $390 million for suicides and attempted suicides related to the drug. Evidence of manipulation has also emerged in many of the highprofile pharmaceutical scandals of the past decade, from Merck’s pain drug Vioxx to the recent Senate investigation into GlaxoSmithKline’s diabetes drug Avandia. Something similar has happened with the atypicals. A 2006 study in The American Journal of Psychiatry, which looked at 32 head-to-head trials of atypicals, found that 90 percent of them came out positively for whichever company had designed and financed the trial. This startling result was not a matter of selective publication. The companies had simply designed the studies in a way that virtually ensured their own drugs would come out ahead— for instance, by dosing the competing drugs too low to be effective, or so high that they would produce damaging side effects. Much of this manipulation came from biased statistical analyses and rigged trial designs of such complexity that outside reviewers were unable to spot them. As Dr. Richard Smith, the former editor of the British Medical Journal, has pointed out, “The companies seem to get the results they want not by fiddling the results, which would be far too crude and possibly detectable by peer review, but rather by asking the ‘right’ questions.” Initially, the controversy over atypical antipsychotics was focused largely on Eli Lilly, the manufacturer of Zyprexa. In early 2009, it settled litigation for a record-breaking $1.4 billion for illegal marketing and allegedly hiding the risks of the drug. More recently, however, the scandal has spread to Seroquel. In April 2010, AstraZeneca agreed to pay $520 million to settle two federal investigations and two whistleblower lawsuits alleg-

courtesy mary weiss

making a killing

ing that it had marketed Seroquel illegally and concealed its health risks. The company faces more than 25,000 civil suits. Documents unsealed in related civil suits suggest an alarming pattern of deception. Sales reps were instructed to tell doctors that Seroquel doesn’t cause diabetes, even though the company knew about the link to diabetes as early as 1997. Internal correspondence reveals company officials discussing how to hide or spin potentially damaging studies. “Thus far, we have buried trials 15, 31, 56,” wrote a publications manager in 1999. “The larger issue is how do we face the outside world when they begin to criticize us for suppressing data.” One of those potentially damaging studies led back to the University of Minnesota. In the late 1990s, a clinical trial known as Study 15 unexpectedly failed to show that Seroquel was any better than Haldol, a generic antipsychotic that’s been on the market since the 1960s. In fact, on the main measures, Seroquel performed worse than Haldol. The study also showed that Seroquel increased the risk of weight gain and diabetes. Internal correspondence repeatedly refers to Study 15 as a “failed study,” and company officials discuss possible ways to spin or bury it. “I am not 100% comfortable with this data being made publicly available at the present time,” wrote Richard Lawrence, a senior AstraZeneca official, in 1997. “However I understand that we have little choice…Lisa [Arvanitis, a company physician] has done a great ‘smoke-and-mirrors’ job.” Lawrence referred approvingly to a strategy that he said would “put a positive spin (in terms of safety) on this cursed study.” Later, apparently hoping to find a way to present Seroquel in a better light, the “commercial support team” performed an analysis of a number of other studies, but even that did not show Seroquel to be better than Haldol. Yet when a summary of the AstraZeneca data was presented at the American Psychiatric Association annual conference in 2000, the author claimed Seroquel was “significantly superior” to Haldol. That author was Dr. Charles Schulz, the University of Minnesota psychiatry department chair—and a well-compensated consultant for AstraZeneca. In a press release claiming Seroquel’s superiority over Haldol, Schulz praised it enthusiastically as a “first-choice antipsychotic.”

Although the documents unsealed in the Seroquel litigation do not specifically mention the cafe study in which Dan was enrolled, they do suggest that AstraZeneca planned to establish Seroquel as the “atypical of choice in first-episode schizophrenia,” according to a 2000 “Seroquel Strategy Summary.” A later document titled “Seroquel PR Plan 2001” discusses the agenda for an advisory panel meeting in Hawaii. Among the potential topics were the marketing of Seroquel to first-episode patients, adolescents, and the elderly. The document refers to these populations as “vulnerable patient groups.” Even more alarming are internal documents suggesting that AstraZeneca was

found “safe and effective” by the fda and that it stands behind the cafe study and the rest of its clinical research.) Many clinical studies place human subjects at risk—at a minimum, the risk of mild discomfort, and at worst, the risk of serious pain and death. Bioethicists and regulators spend a lot of time and energy debating the degree of risk that ought to be permitted in a study, how those risks should be presented to subjects, and the way those risks should be balanced against the potential benefits a subject might receive. What is simply assumed, without much consideration at all, is that the research is being conducted to produce scientific knowledge. This assumption is codified in a number

“We have buried trials 15, 31, 56,”wrote an AstraZeneca official. “How do we face the outside world when they begin to criticize us for suppressing data?”
designing clinical trials as a covert method of marketing Seroquel. In 1997, when Dr. Andrew Goudie, a psychopharmacologist at the University of Liverpool, asked AstraZeneca to fund a research study he was planning, a company official replied that “R&D is no longer responsible for Seroquel research—it is now the responsibility of Sales and Marketing.” The official also noted that funding decisions would depend on whether the study was likely to show a “competitive advantage for Seroquel.” Another set of documents from 2003 describes a glucose metabolism study apparently designed to fend off the charge that Seroquel causes patients to gain weight and become diabetic. One slide describes two purposes for the study: a “regulatory” purpose and a “commercial” purpose. The regulatory purpose was to “produce data that will help us defend the Seroquel label.” The commercial purpose was to “produce data that will enable us to generate commercially attractive and competitive messages in relation to diabetes and weight.” The document suggests several possible names for the study, including “Flexible Dose Approach Trial for Atypical Responses to Metabolism,” which could be usefully shortened to the acronym fatfarm. (When I contacted AstraZeneca, a spokesperson would say only that Seroquel has been of foundational ethics documents, such as the Nuremberg Code, which was instituted following Nazi experiments on concentration camp victims. The Nuremberg Code stipulates that an “experiment should be such as to yield fruitful results for the good of society,” and “the degree of risk to be taken should never exceed that determined by the humanitarian importance of the problem to be solved by the experiment.” But what if a research study is not really aimed at producing genuine scientific knowledge at all? The documents emerging in litigation suggest that pharmaceutical companies are designing, analyzing, and publishing trials primarily as a way of positioning their drugs in the marketplace. This raises a question unconsidered in any current code of research ethics. How much risk to human subjects is justified in a study whose principal aim is to “generate commercially attractive messages”? In January 2005, the fda began investigating the circumstances of Dan’s suicide. In a report issued that July, before the larger pattern of Seroquel research had begun to emerge, Sharon L. Matson, the fda investigator, exonerated the university. She wrote, “I did not find any evidence of misconduct, significant violation of the protocol, or regulations governing clinical
s e pte m b e r / o c to b e r 2 0 1 0 | m o t h e r j o n e s

59

Trial by Hire What happens when profit margins drive clinical research?
A decAde Ago, when the inspector general of the department of Health and Human Services (hhs) investigated the recruitment practices of pharmaceutical trials, researchers complained that research sponsors were demanding unrealistically tight deadlines to enroll subjects. Asked by the ig what sponsors were looking for in trial sites, one researcher replied, “Number one–rapid enrollment. Number two–rapid enrollment. Number three–rapid enrollment.” Many researchers attributed the unrelenting pressure to the fact that trials were being managed by businesspeople, not clinicians. over the past 20 years, medical research has become a largely privatized, and thoroughly Taylorized, business. Two-thirds of clinical trials are now privately run. Many trials are advertised by patient recruitment specialists, carried out by “contract researchers,” approved by for-profit ethics boards, and written up for publication by commercial medical education agencies. The largest of the new private industries are contract research organizations (cros), which range from small niche agencies to multinational corporations that manage all aspects of clinical trials, from ethics approval and subject recruitment to the submission of clinical data to the fda. Quintiles, the company that managed the study in which dan Markingson was enrolled, is the largest, with 14 percent of the $11.4 billion global market. cros save money for pharmaceutical companies by deploying the principles of industrial management: breaking trials down into narrow, discrete steps, which can be carried out with maximum efficiency by specialized workers who can be paid relatively low wages. According to Vanderbilt University social scientist Jill Fisher, author of Medical Research for Hire, very little experience is required to be a cro “monitor” —a middle manager, often a nurse, who coordinates the various sites involved in a study. Monitors usually make less than their counterparts at universities or pharmaceutical companies, and job turnover is very rapid. Fisher says, “The goal of many monitors is to be hired by the pharmaceutical industry.” In contrast, the private physicians paid to
60 m o t h e r j o n e s |
s e pte m b e r / o c to b e r 2 0 1 0

supervise clinical trials are often very wellcompensated. A part-time contract researcher conducting four or five clinical trials a year can expect to earn an average of $300,000 in extra income. Yet they generally have little if any research training. They do not generate original scientific ideas, design studies, or analyze the results. Their main role is to help recruit subjects and oversee their trial participation. Research subjects are the most highly prized commodities in the clinical trials industry. Four out of five clinical trials are delayed because of difficulties recruiting subjects. These delays can be costly, as the patent clock on new drugs starts ticking as soon as the patent is filed. As cros have discovered, many research subjects can be persuaded to enroll because they have no health insurance or because they are too poor to afford medication. (According to Fisher, a common cro term for these patients

is “ready-to-recruit.”) In the cafe study, for instance, a Quintiles study monitor suggested that each of the cafe study site coordinators try recruiting subjects at homeless shelters. Nevertheless, early on the University of Minnesota trial site was apparently struggling to keep subjects in the study. “Having trouble with Subject 002,” Jean Kenney, the university’s cafe study coordinator, wrote in a January 2003 email to Quintiles. “His sister just died, his father has terminal cancer and now the grandmother is sick. He missed a visit and now just missed the next one.” Another issue was slow recruitment. “Have had another person show interest from inpatient and then the parent put pressure on and said ‘No’ (third time this has happened),” Kenney wrote. The Quintiles study monitor was consistently upbeat and encouraging. “Try not to get too frustrated!” she advised. “Hopefully your hard work will start to pay off soon!”

making a killing

Many cros market their ability to locate subject populations that are “treatmentnaïve,” meaning patients who have not yet been treated for their illness or who are not taking any other medications. often treatment-naïve subjects are easier to find in poorer countries, where trials can also be conducted with less oversight from the fda. In 2008, according to hhs, 78 percent of subjects enrolled in clinical trials lived outside the United States, including 13,000 subjects in Peru, where the fda conducted no inspections. cros have been involved in some notable clinical trial scandals. In the 1990s, Pharmaceutical Product development, or ppd, one of the largest, was implicated in a notorious fraud scheme carried out by dr. Robert Fiddes, who used his Southern california Research Institute to falsify records and invent patients while conducting trials for nearly every major pharmaceutical company. In 2006, at a trial site at a hospital near London, six healthy subjects nearly died after the cro Parexel paid them 2,000 pounds each to become the first humans to test an experimental compound. In 2005, Bloomberg News reporters discovered that sfbc International Inc. was paying undocumented immigrants to serve as drug guinea pigs in a converted Holiday Inn. The Miami motel was subsequently demolished for fire and safety violations, and the company changed its name to PharmaNet. In 2009, PharmaNet was acquired by jll Partners, a New York hedge fund. Today, if cash-strapped academic centers want to compete for the revenue generated by industry-sponsored trials, they must play by new rules. Academic institutional review boards must approve trials quickly to compete with for-profit irbs (see page 63), and academic study coordinators must recruit subjects quickly to compete with private trial sites. The competition is even stiffer for academic physicians, many of whom must generate part of their own university salaries by obtaining grants and contracts from external funding sources. If academic physicians want to do clinical trials for the pharmaceutical industry, they must compete with contract researchers, who offer little to the body of science but carry out industry-tailored trials efficiently. Such arrangements often reduce academic physicians to little more than industry helpers, collecting data according to a company protocol. All these factors, it seems, were at play in the study that dan Markingson was enrolled in when he died. —C.E.

investigators or irbs”— the university institutional review board charged with reviewing studies to ensure that they measure up to recognized ethical standards. Matson specifically dismissed the suggestion that Dan was mentally incompetent to consent to the study, writing that “there was nothing different about this subject than others enrolled to indicate that he couldn’t provide voluntary, informed consent.” (The fda refused my request to speak with Matson and would not answer questions about the case, citing privacy concerns.) Mary Weiss eventually sued the University of Minnesota, AstraZeneca, Olson, and Schulz, but her case did not even get to trial. District Court Judge John L. Holahan dismissed the suit in 2008 with

screening during autopsy had not found any Seroquel in his bloodstream, which suggested that Dan may not have been taking his medication. After the judgment, however, Mary discovered that Seroquel would not be detected in an ordinary drug screening; a special test is required. In the spring of 2008, she called the coroner’s office in hopes of getting a special screening for Seroquel. To her surprise, she found that her lawyers and the defendants had already obtained one. The report was dated several days after the summary judgment was issued. It showed 73 nanograms per milliliter of Seroquel in his blood, suggesting that Dan was almost certainly taking the drug, although he may have missed the last scheduled dose before he died.

AstraZeneca paid the University of Minnesota $327,000 for the Seroquel trial. Dan’s two doctors received $261,364 in consulting fees and grants.
a partial summary judgment. He ruled that in approving the cafe study, the university irb was performing the type of “discretionary function” that is protected from liability under the state’s Tort Claims Act. The malpractice suit against Schulz was also dismissed, and the suit against Olson was eventually settled—for $75,000, which Mary says wasn’t enough to cover the fees of the expert witnesses her attorneys hired. (Both Schulz and Olson declined to speak about the specifics of the clinical trial or the resulting suit. University spokesman Nick Hanson would say only, “To date, there has been no finding of wrongdoing from any of the investigations or reviews done by the university on this issue.”) The judge also dismissed the case against AstraZeneca. He blasted Mary’s lawyers, saying that they had failed to establish that AstraZeneca had a duty to put the interests of research subjects over the interests of the company and the researchers. But he also lamented the lack of case law about clinical trials, saying on this particular point, “Try as it may, this Court’s independent research has unearthed not a single case or statute to evidence or support such an alleged duty.” The judge further ruled that Mary’s lawyers hadn’t shown a causal link between Seroquel and Dan’s suicide: An initial drug Although Mary’s lawsuit was unsuccessful, it revealed some disturbing financial arrangements at the university. As a patient on public assistance, Dan’s treatment would have normally generated little income for the university. Under its arrangement with AstraZeneca, however, the psychiatry department earned $15,648 for each subject who completed the cafe study. In total, the study generated $327,000 for the department. In fact, during the months before Dan was enrolled, the department was apparently feeling pressure from Quintiles, the cro that managed the study, to step up recruitment. According to emails written by Jean Kenney, the university’s study coordinator, the site had been placed on probation for its recruitment problems, and they were still “struggling to get patients.” In November 2002, Olson had managed to recruit only one subject in six months. That began to change in April 2003, when the psychiatry department established a specialized inpatient unit at Fairview hospital called Station 12, in which every patient could be evaluated for research. By December, Olson had recruited 12 more subjects, including Dan, and Olson had been featured in a cafe study webcast for “turning an underperforming site into a well-performing site.”(Quintiles refused to give comment on the case.)
s e pte m b e r / o c to b e r 2 0 1 0 | m o t h e r j o n e s

61

Olson had another financial reason to maintain good relations with AstraZeneca. According to a disclosure statement for a 2006 conference, he was a member of the AstraZeneca “speaker’s bureau,” giving paid talks for the company. He had similar arrangements with Eli Lilly and Janssen, the makers of the other atypicals being tested in the cafe study, as well as Bristol-Myers Squibb and Pfizer. In addition, Olson was working as a paid consultant for Lilly, Janssen, Bristol-Myers Squibb, and Pfizer. Although Olson is not required to disclose how much industry money he received, a public database maintained by the Minnesota pharmacy board indicates that Olson received a total of $240,045 from the pharmaceutical industry between 2002 and 2008, with $149,344 coming from AstraZeneca. Dr. Charles Schulz, his co-investigator and department chair, received an even greater sum: more than $571,000 from the industry, with $112,020 coming from AstraZeneca. The database does not reliably distinguish between payment by drug companies for consulting and speaking, which usually goes directly into a physician’s pocket, and research grants, which go to the university and are used to help underwrite the salaries of the grant recipients. (Many academic physicians are required by their universities to generate a substantial portion of their salaries by obtaining research grants.) In the US, the primary bodies charged with protecting research subjects are known as institutional review boards. (Read how irbs are becoming privatized, next page.) According to the University of Minnesota, the purpose of its irb is to “protect the rights and welfare of human research subjects.” However, when the university’s irb officials were deposed under oath, they refused to admit that protecting subjects was their responsibility. “So it’s not the institutional review board’s purpose to protect clinical trial subjects, is that what you’re saying?” asked Gale Pearson, one of the attorneys representing Mary Weiss. “That’s true,” replied Moira Keane, the director of the irb. Astonished, Pearson kept returning to the question, to make sure that she understood it correctly. Keane refused to budge. Instead, she claimed that the role of the irb was to make sure that Olson and the trial sponsor had a plan to protect subjects. (If this were true, it would render irbs worthless: The sponsor and in62 m o t h e r j o n e s |
s e pte m b e r / o c to b e r 2 0 1 0

vestigator are the ones that the irb is supposed to protect subjects from.) The University of Minnesota doesn’t exactly have a stellar record of investigating internal misconduct. In 1994, the director of child and adolescent psychiatry, Dr. Barry Garfinkel, was sentenced to federal prison for five felonies related to research fraud involving the Ciba-Geigy drug Anafranil (clomipramine). The research assistant who blew the whistle in 1989 lost her job, and under the terms of a secret agreement struck with Garfinkel, the university kept the fraud secret for four years, until he was finally indicted. In 1995, the university was sanctioned by the National Institutes of Health after revelations that the head of

recorded for the 400 subjects in the study were an alleged homicide and five suicide attempts, including two successful suicides, both by patients taking Seroquel. (One of these patients, of course, was Dan Markingson.) According to the study authors—three AstraZeneca employees and seven academic physicians, many of whom also consulted for the company—the suicides occurred “despite the close attention provided in clinical research aftercare programs.” The authors claimed that the cafe study showed Seroquel to be of “comparable effectiveness” to Zyprexa and Risperdal for first-episode patients. According to some experts, the study could hardly have shown otherwise, be-

“R&D is no longer responsible for Seroquel research—it is now the responsibility of Sales and Marketing,” an AstraZeneca official told a doctor.
transplant surgery, Dr. John Najarian, had generated millions of dollars for the university by illegally manufacturing and selling an immunosuppressant drug without fda approval; an investigation by the Minneapolis Star Tribune revealed that the university had known of the illegal activity for years. Still more scandals have recently emerged, including a Senate investigation of the chairman of spinal surgery, Dr. David Polly, for failing to disclose $1.2 million he had been paid to consult for the device manufacturer Medtronic, and a series of investigative reports in the New York Times about the industry ties of Minnesota physicians, including some connected to the university. When the scandals began to escalate several years ago, Dr. Deborah Powell, then the dean of the university’s medical school, appointed a task force to devise a new conflict-of-interest policy. The policy was discarded after the Star Tribune revealed that the co-chair of the task force, Dr. Leo Furcht, had funneled $500,000 of university grant money into his own private company, which he later sold for $9.5 million. Furcht remains chairman of the laboratory medicine and pathology department at the university. In 2007, the American Journal of Psychiatry published the results of the cafe study. Among the 18 “serious adverse events” cause it was designed to produce a good result for Seroquel. When I showed the published study to Dr. Peter Tyrer, the editor of the British Journal of Psychiatry, he said, “I would have major problems accepting a manuscript of that nature.” According to Tyrer, the main problem is the small sample size. Of the 400 subjects enrolled, all but 119 stopped taking the drug before the yearlong study was finished. With so few subjects, the cafe study was statistically underpowered and thus unlikely to detect any difference in effectiveness between the three drugs. The failure to detect a difference allowed AstraZeneca to claim that Seroquel was as good as the other drugs (or in the language of the study, “non-inferiority”). Tyrer told me, “In scientific terms this study is of very little value.” That’s not the only problem. The cafe study was supposedly designed to test the effectiveness of the three antipsychotics, but the way it did this was by measuring the rate of “all-cause treatment discontinuation,” or the percentage of subjects who stopped taking their drug. That is, the cafe study counted an antipsychotic as “effective” if a subject kept taking it until the end of the study. On the face of it, this type of measurement seems highly misleading; simply because a patient continues to take an antipsychotic does not mean that it is

making a killing

working. Many psychiatrists defend treatment discontinuation as a “pragmatic” way of measuring a drug’s overall acceptability, but even by “pragmatic” standards the cafe study presents a problem. More than 70 percent of subjects in the cafe study stopped taking their assigned drug, and the most common reason was simply coded as “patient decision.” According to Dr. John Davis, the Gillman Professor of psychiatry at the University of Illinois-Chicago, the authors of the cafe study obscured their results by failing to say why patients decided to stop taking the drug—whether patients felt the side effects of the drug were too severe, for example, or if they felt the drug was not working. “It is the hiding of the critical outcomes that gives me pause,” he says. “It does not make scientific sense to do a study and not measure one of the most important outcomes.” Yet another problem with the cafe study is its failure to compare Seroquel to any older antipsychotics. “It’s quite a marketing exercise to put all patients in the cafe study on atypical antipsychotics,” says Dr. Glen Spielmans, an associate professor of psychology at Minnesota’s Metropolitan State University. “It removes the older drugs from the discussion.” One reason AstraZeneca may have done this, he suggests, is that Study 15 had already shown Seroquel to be

inferior to the older antipsychotic, Haldol. The bluntest assessment of the study came from Dr. David Healy, a senior psychiatrist at Cardiff University in Wales. Healy is a former consultant to AstraZeneca, among other pharmaceutical companies, and a prominent critic of the industry. “This is a non-study of the worst kind,” he said. “It is designed not to pick up a difference between the three drugs. It looks like an entirely marketing-driven exercise.” If these experts are right, then the study in which Dan Markingson committed suicide was not simply a matter of inadequate informed consent, or financial conflicts of interest, or even failure to monitor a subject’s care. The ethical breach was built into the study from the start. It is one thing to ask people to take risks for science, or the common good, or to help other people. It is another thing entirely to ask them to risk their lives for the marketing goals of AstraZeneca. Mary Weiss is a quiet woman, but her experience has left her angry and bitter. It’s not hard to see why. In the years since she lost her son, she has written letters and filed complaints to one oversight body after the other, and so far she’s gotten little but form letters, rejections, and dismissals. “Well, I don’t think the loss can ever be

replaced,” her friend Mike Howard said in his deposition. “There is probably not a day in Mary’s life that she hasn’t thought about her son, and there is probably not a week goes by that she doesn’t shed tears.” Mary told me that until she and I had coffee last year in St. Paul, no one at the university had ever apologized or expressed regret for her son’s death. In fact, after Dan died, Mary received a plant with a card from the cafe study team. In words that echoed the bizarre, grisly message in Dan’s suicide note, the card read, “We will miss his smile.” Of all the ways in which Mary Weiss has been damaged by the University of Minnesota, there is one episode that still brings a sting of shame to my face. When the lawsuit over Dan’s death was dismissed, the university filed a legal action against Mary, demanding that she pay the university $57,000 to cover its legal expenses. Gale Pearson, one of Mary’s attorneys, says that while such suits are technically permissible, she had never seen one filed in her previous 14 years of legal practice. The university agreed to drop the lawsuit against Mary only when she agreed not to appeal the judge’s decision. “Maybe they want to chill anyone who might think of challenging the university, even if her child had died,” Pearson said. “It gave me a sick feeling.” n

Poor Reviews Profit pressures gut guinea pigs’ only safeguard: institutional review boards.
eSTABLISHed IN the 1970s in response to scandals such as the Tuskegee syphilis experiment, institutional review boards are the primary means of protecting research subjects in the United States. Until recently, most irbs were volunteer committees of clinicians and researchers in the teaching hospitals and medical schools where the research in question was being conducted. But as clinical research began to enter the private sector, a new type of irb emerged: independent boards that review studies in exchange for a fee, promising a faster review. There are about 40 for-profit irbs operating in the US, generating more than $100 million in annual revenue. Some for-profit irbs are professional and serious-minded, while others present a more entrepreneurial face. Take Liberty irb, a forprofit irb in Florida that boasts on its website that it is the winner of the 2008 “Make Mine a Million $ Business” competition, a contest described as “a cross between The Apprentice and American Idol.” Paid by the companies whose protocols they review, for-profit irbs have a direct interest in keeping their clients happy. If one for-profit irb rejects a study as too dangerous, the sponsor can simply send it to another one. defenders argue that companies have an interest in getting a strict ethical review, if only to ward off potential litigation. But recent events suggest otherwise. In March 2009, the government Accountability office revealed the results of a sting operation it conducted on coast irb, a colorado outfit with more than $9 million in revenue in 2008. The gao set up a phony company testing an obviously dangerous “bogus medical device” in a research protocol so “excessively vague” no reputable irb should approve it: The protocol lacked results from animal studies and didn’t reveal where the study would take place, or what institution would carry it out. The principal investigator listed had an expired medical license; the only contact information was a post office box and a cell phone number. Yet coast irb approved the product unanimously and deemed it “probably very safe.” According to the gao, coast irb had reviewed 356 research studies in five years and rejected only one. So are the old academic irbs any better? Located within academic health centers that now compete with contract research organizations for clinical trials, they also face pressure to approve trials quickly. As trials have become more complex, academic irbs have become expensive, costing an average of nearly $750,000 per year—with some costing more than $4 million. As a result, some universities are outsourcing reviews to for-profit irbs. others have decided to shift the cost back to the pharma companies, like the University of Minnesota, which charges $2,500 to review industry-sponsored studies. —C.E.
s e pte m b e r / o c to b e r 2 0 1 0 | m o t h e r j o n e s

63

ATTACHMENT N

Critics say drug firms' payments to doctors are conflict of interest
What they spend: A look at drug company spending in Minnesota — on top specialties and select psychiatrists. By Jeremy Olson and Paul Tosto Pioneer Press Updated: 05/23/2008 05:27:26 PM CDT Drug companies have given $88 million in gifts, grants and fees to Minnesota doctors and caregivers since 2002, according to state payment records, including $782,000 to the two University of Minnesota psychiatrists who oversaw Dan Markingson's participation in a clinical drug trial. A lawsuit over Markingson's suicide, which happened during the drug trial, accused Dr. Stephen Olson and Dr. S. Charles Schulz, chairman of the U's psychiatry department, of coercing the schizophrenic Markingson into the study. The lawsuit, brought by Markingson's mother, Mary Weiss, charged that the doctors were under pressure to recruit patients such as Markingson to maximize payments from AstraZeneca and gain prestige by participating in the drug company's national study. Both doctors said in court depositions that their roles were appropriate and that the money didn't influence their decisions over Markingson — including when his mother argued that he wasn't getting better in the study and should be withdrawn. Schulz was dismissed from the lawsuit in February; Olson settled this spring for an amount a university official described as little more than court costs. Federal reviews of the death didn't result in any penalties against the doctors or the university. The case nonetheless offered an inside look at the kind of financial payments to doctors that some health policy experts and congressional representatives say should be restricted or at least fully disclosed to the public. It also scrutinized the ethics of drug company funding of research — something that has received less public attention and criticism than the free lunches, dinners and trips that drug companies have provided to doctors to promote their drugs. Markingson, 27, killed himself May 8, 2004, in the bathroom of a West St. Paul halfway house. He had been enrolled for more than five months in the university's "CAFE" study, which compared three antipsychotic drugs. Weiss sued the university and the psychiatrists. In an interview, she said doctors have a conflict of interest when they are financially benefiting from studies and caring for patients in those studies at the same time.

"I think they lose sight that these are people," she said, "not their own special little guinea pigs." Minnesota is unique in requiring drug companies to report how much money they give to each doctor, but the reporting system has limitations. It doesn't always distinguish between money for a doctor's travel expenses and money for a research trial, nor does it distinguish money that was in a doctor's name but was passed directly to a research institution. U.S. Sen. Chuck Grassley, R-Iowa, is urging a national reporting system. Grassley held a hearing last year in which two doctors said their colleagues have become trapped by the lures and pressures of drug company money. "Physicians face a difficult choice," testified Dr. Greg Rosenthal, an Ohio eye specialist. "One path is to go along. With drug company money, you can increase your income, prestige, build your practice or fund a department, research or professorships. The middle ground is to simply look away. The hard choice is to fight back." Olson received $220,000 from six companies since 2002, including $149,000 from AstraZeneca, according to the state records. Schulz received $562,000, including $112,000 as a researcher and consultant to AstraZeneca. Olson said his AstraZeneca money went straight to the U but did support his salary. Markingson's full participation in the yearlong study meant up to $15,000 for the university. The amounts aren't unusual, according to the payment records collected by the Minnesota Board of Pharmacy. The records, which were updated this month to include 2007 figures, show 167 Minnesota doctors who have received $100,000 or more since 2002. One in four psychiatrists has received funding from pharmaceutical companies, averaging about $50,000 over the six years. Greater awareness of drug company payments has prompted tighter rules among some Minnesota health care organizations. The Mayo Clinic prohibits its doctors from being paid by drug companies to serve on their speaker's bureaus. Doctors in speaker's bureaus give lectures to other doctors about the company's medications. The St. Mary's clinic system in Duluth recently banned pens, mugs or other freebies bearing drug company logos. There have been fewer steps to restrict drug company funding of research, though most medical journals long ago required doctors to disclose the funding source of any research results they publish. Some health officials are now questioning the drug companies' use of "ghostwriters" to revise articles about research results to promote the drugs they sell.

Many universities view industry-sponsored research as a necessity amid tightening state and federal science budgets. Drug company funding makes up less than 7 percent of the psychiatry department budget at the University of Minnesota, but Schulz said it is needed as the U tries to move up the list of top-funded U.S. research institutions. Since Olson was recruited in 2001 to boost the university's expertise in schizophrenia, he has led the U's efforts in three drug trials funded by AstraZeneca. He also took part in the federally funded "CATIE" trial, which suggested that older antipsychotic drugs were as effective as AstraZeneca's Seroquel and other newer drugs. A growing body of research suggests that drug company money has an influence on study outcomes. One analysis found that industry-funded research was four to five times more likely to produce positive outcomes for a paying company's drug than federally funded research. A report last year found that drug company-funded studies of cholesterol medications were much more likely to produce results that favored their own drugs as well. The CAFE results didn't show that AstraZeneca's Seroquel offered much benefit over two competitors — Zyprexa and Risperdal. Patients gained control over schizophrenic symptoms and tended to stop taking the medications at the same rate, regardless of which drug they took. The level of unhealthy weight gain was comparable, too, albeit slightly higher among the Zyprexa patients. Weiss sued AstraZeneca as well, though the company also was dismissed from the lawsuit. Her attorneys argued that AstraZeneca's goal with the CAFE study was to gain a marketing edge and that the company used selective information from the study to promote Seroquel. The attorneys cited internal documents, which have been sealed under court order, in which AstraZeneca discussed its use of ghostwriters and strategies to present CAFE results in a way that "sells" Seroquel. AstraZeneca declined to discuss documents from the case, but brand corporate affairs manager Abigail Baron said the company's financial arrangements with doctors are necessary to improve health through drug discovery. "That mission cannot be fulfilled," she said, "without close partnership with those on the front lines of patient care and ... research." Jeremy Olson can be reached at 651-228-5583 or jolson@pioneerpress.com. Paul Tosto can be reached at 651-228-2119 or ptosto@pioneerpress.com.

Payments to Minnesota doctors and other caregivers
This search engine contains public information from the Minnesota Board of Pharmacy on pharmaceutical company payments to doctors and other caregivers from 2002 through 2008. The Pioneer Press assembled the 2005, 2006, 2007 and 2008 data, and the nonprofit organization Public Citizen assembled and provided the earlier years. The database has limitations, including that it cannot distinguish between physicians with the same name. Some of it also was derived from paper records, and the transfer process may have resulted in unintended spelling errors or ommisions. Some doctors are also listed under more than one city.

Your search returned the following records of payments to STEPHEN OLSON. Start a new search. Doctor Name Practice Name Paying Company Paym Description ent Year Amoun t

City

OLSON , STEPH EN

MINNEAP ELI LILLY OLIS

2002

REASONABLE HONORARIA AND PAYMENT OF THE REASONABLE EXPENSES OF A PRACTITIONER WHO SERVES ON THE FACULTY $1,500. OR PROVIDES OTHER 00 SERVICES AT A PROFESSIONAL OR EDUCATIONAL CONFERENCE OR MEETING REASONABLE HONORARIA AND PAYMENT OF THE REASONABLE EXPENSES OF A PRACTITIONER WHO SERVES ON THE FACULTY $1,000. OR PROVIDES OTHER 00 SERVICES AT A PROFESSIONAL OR EDUCATIONAL CONFERENCE OR MEETING HONORARIUM/CONSULTIN G FEE IN CONNECTION $3,000. WITH THREE SPEAKER 00 PROGRAMS REASONABLE HONORARIA $7,606. AND PAYMENT OF THE 10

OLSON , STEPH EN

MINNEAP ELI LILLY OLIS

2003

OLSON , STEPH EN OLSON ,

JANSSEN MINNEAP PHARMACEU OLIS TICA INC MINNEAP ELI LILLY OLIS

2003

2004

STEPH EN

REASONABLE EXPENSES OF A PRACTITIONER WHO SERVES ON THE FACULTY OR PROVIDES OTHER SERVICES AT A PROFESSIONAL OR EDUCATIONAL CONFERENCE OR MEETING MINNEAP ELI LILLY OLIS REIMBURSEMENT, HONORARIA OR CONSULTING SERVICES $6,204. 00

OLSON , STEPH EN D OLSON , STEPH ENC OLSON , STEPH ENC OLSON , STEPH ENC OLSON , STEPH ENC OLSON , STEPH ENC OLSON , STEPH ENC

2005

FAIRVIE ASTRAZENEC W UNIVER MINNEAP A 2006 PHARMACEU SITY OLIS TICALS MEDICA L CTR BRISTOLMINNEAP MYERS OLIS SQUIBB CO BRISTOLMINNEAP MYERS OLIS SQUIBB CO BRISTOLMINNEAP MYERS OLIS SQUIBB CO JOM MINNEAP PHARMACEU OLIS TICAL SERVICES BRISTOLMINNEAP MYERS OLIS SQUIBB CO

ANNUAL AGGREGATE COMPENSATION $149,34 PROFESSIONAL/CONSULTI 4.00 NG/EXPENSE REIM

2006

SPEAKER FEES & RELATED $1,507. EXPENSES 62

2006

SPEAKER FEES & RELATED $1,500. EXPENSES 00

2006

SPEAKER FEES & RELATED $1,506. EXPENSES 70

2006

CONSULTING FEE & EXPENSES 3 SPEAKING ENGAGEMENTS

$3,000. 00

2007

SPEAKING

$1,509. 00

Olson, Stephen

MINNEAP ELI LILLY OLIS

2007

REASONABLE HONORARIA AND PAYMENT OF THE REASONABLE EXPENSES OF A PRACTITIONER WHO SERVES ON THE FACULTY $13,122 OR PROVIDES OTHER .83 SERVICES AT A PROFESSIONAL OR EDUCATIONAL CONFERENCE OR MEETING

Olson, Stephen OLSON , STEPH EN

MINNEAP JANSSEN OLIS MINNEAP ELI LILLY OLIS

2007

Consulting Fee in connection with 1 Speaking Engagment HONORARIA AND EXPENSES FOR

$1,000. 00 $19,752 .76

2008

our search returned the following records of payments to STEPHEN OLSON. Start a new search. Doctor Name Practi ce Name Payme Description nt Year

City

Paying Company

Amount

OLSON, STEPHE NC

ROSEVIL LE

ELI LILLY

2002

REASONABLE HONORARIA AND PAYMENT OF THE REASONABLE EXPENSES OF A PRACTITIONER WHO SERVES ON THE $3,649.6 FACULTY OR PROVIDES 0 OTHER SERVICES AT A PROFESSIONAL OR EDUCATIONAL CONFERENCE OR MEETING HONORARIUM/CONSULT ING FEE IN $2,000.0 CONNECTION WITH 0 TWO SPEAKER PROGRAMS REIMBURSEMENT, HONORARIA OR CONSULTING SERVICES $15,425. 67

OLSON, STEPHE N OLSON, STEPHE N OLSON, STEPHE N

ROSEVIL LE

JANSSEN PHARMACEUTI 2003 CA INC

ROSEVIL LE ROSEVIL LE

ELI LILLY

2005

ELI LILLY

2006

SERVICE PROFESSIONAL $7,416.9 OR EDUCATIONAL 0 CONFERENCE MTG

EXCERPT
1 STATE OF MINNESOTA 2 COUNTY OF HENNEPIN 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 DISTRICT COURT FOURTH JUDICIAL DISTRICT

Court File No. 62 CO 06 11934 -------------------------------Mary Weiss, on her own behalf, and as the next of kin and Trustee of the estate of Dan Markingson, deceased, Plaintiff, vs. Board of Regents for the University of Minnesota; Dr. Stephen Olson; Dr. Charles Schulz; Institutional Review Board for the University of Minnesota; Astrazeneca Pharmaceuticals LP, Astrazeneca LP and Zeneca, Inc., Defendants. ---------------------------------AUDIO VISUAL DEPOSITION OF STEPHEN OLSON, M.D. MAY 1, 2007

VERBATIM COURT REPORTING 763-493-4535 2 1 2 3 4 5 6 7 8 9

The following is the deposition of STEPHEN OLSON, M.D.; Mari Skalicky, Court Reporter, Notary Public, pursuant to Notice of Taking Deposition, at 701 Xenia Avenue South, Suite 600, Minneapolis, Minnesota, commencing at approximately 9:00 a.m., MAY 1, 2007. *********

10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

APPEARANCES For the Plaintiff: R. Christopher Barden, PhD, JD Ben Griffith, Esquire Gale D. Pearson, Esquire 100 South Fifth Street Suite 1025 Minneapolis, MN 55402

VERBATIM COURT REPORTING 763-493-4535 3

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

For the Defendants: David Alsop, Esquire Gislason & Hunter LLP 701 Xenia Avenue South, Suite 500 Minneapolis, MN 55416 Linda S. Svitak, Esquire Feagre and Benson 2200 Wells Fargo Center Minneapolis, MN 55402 Chuck Gross, Esquire Geraghty, O'Louglin & Kenney 1400 Ecolab University Center St. Paul, MN 55102

18

19 20

ALSO PRESENT: Mary Weiss Ruth Flynn

12 Q. Now, I would like to ask you about your history 13 of your relationship with Astrazeneca. When 14 did you first become a paid consultant or 15 worker of some kind for Astrazeneca? 16 MR. ALSOP: Well, I'll object as vague 17 but go ahead and answer. 18 MS. SVITAK: Same objection. 19 Q. (BY DR. BARDEN) I'm looking for any financial 20 relationship whatsoever you've had with 21 Astrazeneca, and I'm looking for the timeline 22 of that. 23 A. In the early '90s I was a co-investigator 24 with -- at the Ohio State University on a study 25 of Seroquel when it was in phase -- probably,

VERBATIM COURT REPORTING 763-493-4535 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 three, two or three trials. In the sense that I provided some of the care for patients who were hospitalized on an acute psychosis protocol. Q. How much money did you receive from Astrazeneca during that time period? A. I don't know. We probably had maybe two or three patients, and I wasn't the principal investigator. The chairman of my department was. I was the -- I was the inpatient attending. Q. Move to strike as non-responsive. Question was, how much money did you receive from Astrazeneca during that time period? A. Nothing. Q. What was the first date that you received money from Astrazeneca, and how much money was that? A. I don't -- I don't recall if back in the time that I was at Ohio State University if I gave any lectures on behalf of Astrazeneca, but the time that I first recall would be when I first

22 23 24 25

arrived at Minnesota, Dr. Schulz was doing a study of extended release Serequel and he asked me to take over that study and I did. And we didn't recruit anybody for the study so we -- I VERBATIM COURT REPORTING 763-493-4535 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

don't think we got paid anything for doing that study. Q. And that would have been what year? A. That would have been in 2001. Q. So 2001 you took over study, you couldn't recruit any subjects so you were not paid any money, is that correct? A. That's correct. Q. It's difficult to recruit subjects for a lot of research projects; isn't that right? A. Yes. Q. In fact, it's very difficult, isn't it? MR. ALSOP: Object as vague and repetitive, but go ahead. A. In some cases it could be. Q. (BY DR. BARDEN) Let's move along now to the next time you worked for Astrazeneca. What year was that and how much were you paid? A. To the best of my recollection, I may have gone to one speaker's training or consultant meeting for Astrazeneca sometime in the time 2001 to 2002 or three. I don't recall ever being asked to give a talk on behalf of Astrazeneca, and I only recall going to one meeting. I also may have met somewhere in 2004 or five with some VERBATIM COURT REPORTING 763-493-4535 26

1 2 3 4 5 6

representatives from Astrazeneca to do a -what we call the preceptorship, be a training pharmaceutical representatives and understanding the nature of the treatment of psychotic disorders and how to go about that. I would have been paid probably $1,000 to

7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

attend a consultant meeting and probably $500 or $750 to do a training, you know, in-house at my office. That's the extent of my consulting relationship with Astrazeneca. Q. So prior to your research study that we're going to discuss today in which Dan was enrolled, how much money total had you received from Astrazeneca? A. $1,000 to the best of my knowledge. Q. And did they send you 1099's for that or how was that reported to the IRS? A. Well, they would have sent me 1099's but in some cases the -- I checked my tax records and have no 1099's from Astrazeneca. Sometimes these consulting meetings or I'm giving a lecture is through an intermediary, and I can't identify on the basis of the 1099s, you know, who the actual pharmaceutical company that sponsors the whatever activity it was so -VERBATIM COURT REPORTING 763-493-4535 27 Q. When did you check your tax records and go back and look to see how much you've been paid by Astrazeneca? A. Oh, probably a couple months ago. Whenever there was a -- there were some -MR. ALSOP: Just asked when it was. Couple months ago? A. A month or two ago. Q. (BY DR. BARDEN) Why did you do that a couple months ago? MR. ALSOP: That's privileged, discussions with me. Don't answer the question. If there is another reason beyond that, you can answer the question. Q. (BY DR. BARDEN) Yeah, as we've discussed I'm never going to ask you about anything you've discussed with your lawyers. Okay. So let's talk specifically about this study. Okay. You were paid some funding from Astrazeneca to do this study, right? A. (Nods head in the affirmative). Q. The CAFE study is what it's called?

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

23 A. That's correct. 24 Q. So the CAFE study was a personally profitable 25 study for you, right? VERBATIM COURT REPORTING 763-493-4535 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 A. No. Q. You didn't -- you weren't paid any money for doing the CAFE study? A. The University-sponsored projects administration had a contract with Astrazeneca to be paid a certain amount of money per subject per visit as is typical with all clinical trials. And that money went to the University and how that money -- how that money was allocated, you know, to the department, I don't know all of the details, but I didn't personally receive any money beyond my regular salary. Q. Okay. So the money, you didn't get a check directly from Astrazeneca? A. No, absolutely not. Q. The money went to the University of Minnesota from Astrazeneca for your work on the CAFE study, correct? A. Yes. Q. And that money went into what kind of a fund? Some medical schools have a pot of money that is divided among the faculty. Some it goes into some kind of travel fund, I mean, there is all kinds of ways that it's done. How is it VERBATIM COURT REPORTING 763-493-4535 29 1 done at the University? 2 MR. ALSOP: Object on the basis of 3 foundation. If you know, go ahead. 4 A. Well, I don't know all of the details. It's 5 not my responsibility to divide up the money as 6 it comes. Some of it is indirect cost that 7 goes to the University-sponsored project

8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

administration for overhead. Some of it is -some of the monies that come back are paid -are into a budget that I have some control over that would allocate money for payment of salaries of the staff as well as a percentage that would go for -- to the department to help cover my salary. We paid subjects. We paid for their transportation. We paid rent for the space that we used, so -Q. So some of your salary from the University of Minnesota would be paid for with these funds, correct? A. Yes. Q. You just don't know the percentage of your salary that was paid for with these funds? A. Not off the top of my head. That information undoubtedly could be tracked back through the budget of that particular account. VERBATIM COURT REPORTING 763-493-4535 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

Q. So part of your salary during these years was paid by Astrazeneca, correct? MS. SVITAK: Objection. MR. ALSOP: I'll join. Misstatement of his testimony. Q. (BY DR. BARDEN) Part of the funding for your salary came from Astrazeneca, correct? A. Yes. Q. And you're saying you don't know how much of that? A. I don't know exactly how much. Q. Can you give me a ballpark figure? A. I would say that no more than 10 percent of my salary over that entire period of time would have come from Astrazeneca. Q. To the best of your knowledge, was Dr. Schulz receiving any money from Astrazeneca? MR. ALSOP: Object on the basis of foundation. Go ahead. You can answer. A. Dr. Schulz has -- also has research -Dr. Schulz also has research projects and may have other arrangements with Astrazeneca that I

23

don't have any direct knowledge of.

24 Q. Have you never talked to Dr. Schulz about what 25 kind of work he does with and for Astrazeneca? VERBATIM COURT REPORTING 763-493-4535 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 A. I'm aware that he's a consultant to Astrazeneca but I don't know anything about the specific responsibilities that he has or his financial arrangements with the company, no. Q. As a consultant for Astrazeneca, is it your understanding that he is paid money by Astrazeneca? MR. ALSOP: Object as speculative, lacking foundation. MS. SVITAK: And foundation. MR. ALSOP: If you know, go ahead. A. Yes, I would assume that he's paid money. Q. Have you never discussed this with him? A. No. Q. You mentioned that some of the money went into a pot that you had control over to pay salaries, is that correct? A. That went into a research account administered by the University. Q. Who paid the salary for Jean Kenney? Did she work with you on this study? A. Her salary was partly paid from the CAFE account, yes. Q. So you were in charge of some -- of paying some of her salary, correct?

VERBATIM COURT REPORTING 763-493-4535 32 1 A. Yes, she was also an employee of University of 2 Minnesota Physicians, that she was a licensed 3 clinical social worker and provided clinical 4 care in the outpatient clinic for which she 5 billed services directly to patients or their 6 insurance companies.

7 Q. Other than what you've disclosed on the record 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 so far, are there any other financial relationships and ties and incentives between Astrazeneca and you, Dr. Schulz and Jean Kenney? A. In regards to the CAFE study? Q. In -- with regard to anything. A. Well, I have others -- I have other studies. I have a currently operating study called Clear study, which is also sponsored by Astrazeneca for which the same kind of arrangements are made. Compensated for patient visits and budget goes to the University and to an account out of which we pay for the expenses of the research. Q. So this is another study where another percentage of your salary would be paid by Astrazeneca? A. Yes.

Payments to Minnesota doctors and other caregivers
This search engine contains public information from the Minnesota Board of Pharmacy on pharmaceutical company payments to doctors and other caregivers from 2002 through 2008. The Pioneer Press assembled the 2005, 2006, 2007 and 2008 data, and the nonprofit organization Public Citizen assembled and provided the earlier years. The database has limitations, including that it cannot distinguish between physicians with the same name. Some of it also was derived from paper records, and the transfer process may have resulted in unintended spelling errors or ommisions. Some doctors are also listed under more than one city.

Your search returned the following records of payments to SELLMAN SCHULZ. Start a new search. Payme nt Year 2002

Doctor Name SCHULZ, SELLMANCHAR LES SCHULZ, SELLMANCHAR LES

Practice Name

City

Paying Company ASTRAZENECA PHARMACEUTIC ALS ASTRAZENECA PHARMACEUTIC ALS

Description

Amount

MINNEAPO LIS MINNEAPO LIS

HONORARIUM

$18,920.9 7

2002

HONORARIUM

$3,500.00

SCHULZ, SELLMANCHAR LES

MINNEAPO LIS

ELI LILLY

2002

REASONABLE HONORARIA AND PAYMENT OF THE REASONABLE EXPENSES OF A PRACTITIONER WHO SERVES ON THE $10,363.7 FACULTY OR 3 PROVIDES OTHER SERVICES AT A PROFESSIONAL OR EDUCATIONAL CONFERENCE OR MEETING REASONABLE HONORARIA AND PAYMENT OF THE $6,231.22 REASONABLE EXPENSES OF A PRACTITIONER WHO SERVES

SCHULZ, SELLMANCHAR LES

MINNEAPO LIS

ELI LILLY

2003

ON THE FACULTY OR PROVIDES OTHER SERVICES AT A PROFESSIONAL OR EDUCATIONAL CONFERENCE OR MEETING REASONABLE HONORARIA AND PAYMENT OF THE REASONABLE EXPENSES OF A PRACTITIONER WHO SERVES ON THE $106.78 FACULTY OR PROVIDES OTHER SERVICES AT A PROFESSIONAL OR EDUCATIONAL CONFERENCE OR MEETING COMPENSATIO N FOR SERVICES $3,100.00

SCHULZ, SELLMANCHAR LES

MINNEAPO LIS

ELI LILLY

2003

SCHULZ, SELLMANCHAR LES

MINNEAPO LIS

ASTRAZENECA PHARMACEUTIC ALS

2004

SCHULZ, SELLMANCHAR LES

MINNEAPO LIS

ELI LILLY

2004

REASONABLE HONORARIA AND PAYMENT OF THE REASONABLE EXPENSES OF A PRACTITIONER WHO SERVES ON THE $6,572.08 FACULTY OR PROVIDES OTHER SERVICES AT A PROFESSIONAL OR EDUCATIONAL CONFERENCE OR MEETING COMPENSATIO N FOR SERVICES REIMBURSEME NT, HONORARIA OR CONSULTING SERVICES $34,833.3 1

SCHULZ, SELLMANCHAR LES

MINNEAPO LIS

ASTRAZENECA PHARMACEUTIC ALS

2005

SCHULZ, SELLMANCHAR LES

MINNEAPO LIS

ELI LILLY

2005

$273,535. 43

SCHULZ, SELLMANCHAR LES

MINNEAPO LIS

ELI LILLY

2005

REIMBURSEME NT, HONORARIA OR CONSULTING SERVICES ANNUAL AGGREGATE COMPENSATIO N EDUCATIONAL SVCS/EXPENSE REIM

$992.77

SCHULZ, SELLMANCHAR LES

UNIVERSI TY OF MINNESO TA

MINNEAPO LIS

ASTRAZENECA PHARMACEUTIC ALS

2006

$35,997.0 0

SCHULZ, SELLMANCHAR LES

MINNEAPO LIS

ELI LILLY

2006

SERVICE PROFESSIONAL OR $55,045.0 EDUCATIONAL 7 CONFERENCE MTG Honoraria and Expenses for Speaking Honoraria and Expenses for Speaking Honoraria and Expenses for Speaking Honoraria and Expenses for Speaking Honoraria and Expenses for Speaking Advisory Board Consulting $2,165.09

Schulz, SELLMANCHAR LES Schulz, SELLMANCHAR LES Schulz, SELLMANCHAR LES Schulz, SELLMANCHAR LES Schulz, SELLMANCHAR LES Schulz, SELLMANCHAR LES

MINNEAPO LIS MINNEAPO LIS MINNEAPO LIS MINNEAPO LIS MINNEAPO LIS MINNEAPO LIS

ASTRAZENECA PHARMACEUTIC ALS ASTRAZENECA PHARMACEUTIC ALS ASTRAZENECA PHARMACEUTIC ALS ASTRAZENECA PHARMACEUTIC ALS ASTRAZENECA PHARMACEUTIC ALS ASTRAZENECA PHARMACEUTIC ALS

2007

2007

$2,090.38

2007

$2,535.73

2007

$3,027.19

2007

$2,000.00

2007

$3,850.00

Schulz, SELLMANCharles

MINNEAPO LIS

ELI LILLY

2007

REASONABLE HONORARIA AND PAYMENT OF THE REASONABLE EXPENSES OF A PRACTITIONER WHO SERVES ON THE $88,986.7 FACULTY OR 2 PROVIDES OTHER SERVICES AT A PROFESSIONAL OR EDUCATIONAL CONFERENCE OR MEETING

SCHULZ, SELLMANCHAR LES

MINNEAPO LIS

ELI LILLY

2008

HONORARIA AND EXPENSES $9,546.31 FOR SPEAKING

ATTACHMENT O

ATTACHMENT P

Hastings Center Bioethics Forum
Carl Elliott, 03/03/2011

How the FDA Got the Markingson Case Wrong
During the years after Dan Markingson committed suicide in an industry-sponsored drug study at the University of Minnesota, his mother looked for justice anywhere she could find it. Mary Weiss wrote countless letters and e-mails; she filed a lawsuit against the university and the study sponsor; along with her friend, Mike Howard, she sent out complaints to any agency or government body that might possibly take an interest, no matter how tenuous its connection to the protection of research subjects. Most of these efforts went nowhere. That is no surprise; there is simply very little protection for human subjects in privatelysponsored clinical trials. More surprising is the way that Ms. Weiss’ desperate efforts at justice are now being used against her. As calls to investigate her son’s suicide have escalated, University of Minnesota officials have repeatedly refused to look into the case, arguing that the university has already been exonerated by other agencies, especially the FDA. It is true that the FDA investigated the case, in response to a complaint filed by Mike Howard. But the inspection report, issued in 2005, hardly qualifies as an exoneration. In fact, the report is a very puzzling document: occasionally misleading, remarkably uncritical, and often simply baffling. When I was investigating Markingson’s suicide for an article in Mother Jones, FDA officials refused to answer my questions about the report. But if University of Minnesota officials plan to use the FDA report as a shield against criticism, the report is worth a very careful look. First, though, a brief review of the case. Dan Markingson was a 26 year-old man from St. Paul and a recent graduate of the University of Michigan who began showing signs of mental illness in the summer of 2003. His thoughts became paranoid and delusional, and he became convinced that he was part of a vast cult, which was calling on him to murder people, including his mother. On November 12, 2003, Markingson was taken to Fairview Hospital in Minneapolis, where he was seen by Dr. Stephen Olson, a psychiatrist at the University of Minnesota. Olson recommended involuntary commitment, and a court agreed. Later, despite objections by Ms. Weiss, Olson recruited Markingson into a clinical trial of antipsychotic drugs. The clinical trial, which was known as the CAFÉ study (an acronym for Comparison of Atypicals in First-Episode Schizophrenia), was sponsored by AstraZeneca, the manufacturer of Seroquel (quetiapine), and managed by Quintiles, a Contract Research Organization. The CAFÉ study compared three different atypical antipsychotic drugs, each of which was already being marketed: Zyprexa (olanzapine), Risperdal (risperidone)

and Seroquel (quetiapine.) The University of Minnesota was one of 26 sites for the double-blinded trial, which was to last one year. After Markingson was enrolled in the CAFÉ study, he spent about two weeks in Fairview Hospital before being discharged against his mother’s wishes to a halfway house. Over the next five months Ms. Weiss repeatedly expressed her concerns about her son’s medical condition, especially his increasing agitation and rage. Her warnings were largely ignored. Finally, in desperation, she warned the study coordinator that her son might kill himself. On May 7, 2004, Markingson mutilated himself with a box cutter so violently that he nearly decapitated himself. His body was found by halfway house workers in the shower, along with a suicide note that said, “I went through this experience smiling.” Markingson had been taking Seroquel. Was Markingson in any position to give informed consent for this study? After all, he had just been involuntarily committed – not just because he was psychotic, but because he was potentially homicidal. The Dakota County court had issued a legal order requiring Markingson to follow the treatment recommendations of his psychiatrist. If Markingson refused those recommendations, he would be confined to a mental institution. So when his psychiatrist recommended a clinical trial, it is no surprise that Markingson consented. But was that consent valid? Yes, it was, according to the FDA inspector, who does not even consider the possibility that Markingson’s consent might have been influenced by the threat of involuntary commitment. One might have thought that a commitment order would violate the Common Rule, which requires researchers to minimize the “the possibility of coercion or undue influence” when recruiting subjects for research. Certainly the Minnesota legislature thought so in 2009, when it banned researchers from recruiting most psychiatric patients into a drug trial if they are under a commitment order. But the FDA inspector concluded otherwise, writing, “There was nothing different about this subject than the others enrolled to indicate that he couldn’t provide voluntary informed consent” (p. 9). So much for coercion. What about Markingson’s decision-making capacity? He was, after all, being recruited into a trial designed specifically for patients experiencing an acute psychotic episode. The FDA inspector summarily dismisses the suggestion that his psychosis might be a problem, noting that the study coordinator, Jeanne Kenney, had certified Markingson as competent to consent. In fact, the inspector implies that such an assessment was not even ethically necessary, pointing out that it was “neither required nor mentioned in the study protocol” (p. 9). What the FDA inspector failed to note is that during the period leading up to his enrollment, Markingson had been repeatedly judged incapable of consenting to neuroleptic drugs. On November 14, 2003 Dr. Olson signed a commitment document

stating that Markingson “lacks the ability to make decisions regarding such treatment.” On November 17, a pre-petition screening team recommended commitment, noting Markingson’s bizarre beliefs and his refusal to acknowledge his mental illness. On November 19, a court-appointed clinical psychologist confirmed those assessments, writing that Markingson “is believed not to have the capacity to make decisions regarding neuroleptic medication.” Yet on November 21, when Markingson was asked to consent to the CAFÉ study, this assessment of his mental state was reversed and he was judged competent. Is it possible that a psychotic patient’s mental capacity could have improved so dramatically in two days? Yes. But there are some good reasons to doubt this happened. First, the final competence assessment was made not by an independent party, but by the study coordinator for the CAFÉ study, Jean Kenney, whose job it was to recruit subjects for the study. This is hardly an impartial, disinterested assessment. Second, Kenney was a social worker, not a psychiatrist or psychologist trained to make competence assessments. Third, even after Mr. Markingson had been judged competent to consent to the CAFÉ study, his involuntary commitment order was not lifted. This suggests that his mental state had not changed quite so dramatically. And finally, one of the most persistent features of Markingson’s psychosis was a lack of insight into his condition. Markingson did not believe he had a mental illness. How could he be competent to consent to a study comparing treatments for his mental illness when he would not even acknowledge that he was mentally ill? What the FDA inspection does address, indirectly but in some detail, is another controversy at the heart of the case: should Olson have dropped Markingson from the CAFE study? The answer to this question might seem obvious. It is hard to imagine an outcome much worse than the one that resulted from leaving him in. And nobody disputes that Mary Weiss believed her son was doing very poorly in the CAFÉ study, and that she warned the CAFÉ study team that she thought he might kill himself. Yet the obvious question goes unasked. Why didn’t Olson take these warnings seriously? Instead, the FDA inspector goes to great lengths to refute the allegation that Markingson’s condition worsened in the study. Not an easy job to do for a subject who has stabbed himself to death, but there are notes to review, charts to examine, and clinical rating scales to evaluate. That record shows a subject who, by the most generous interpretation possible, failed to improve over a period of five and a half months; whose own psychiatrist acknowledged so little improvement that his stay of commitment could not be lifted; whose mother was convinced that his condition was spiraling dangerously downward; and whose life ended in a grisly suicide. The FDA’s conclusion? The record “does not appear to indicate a significant decline or deterioration” (page 10). The FDA inspector dismisses the suggestion that Olson violated the court order for Markingson to get treatment by recruiting him into the CAFÉ study. Part of her rationale appears to be that Markingson’s caseworker, David Pettit, approved of this enrollment.

The FDA inspector did not actually interview Pettit to confirm this, but she cites a chart entry where Pettit apparently expressed support. What she fails to note is that Pettit was not even assigned to the case until three days after Markingson signed the consent form for the CAFÉ study. What is more, there is no evidence that the judge who issued the commitment order ever even knew about the CAFÉ study. In his deposition, Olson admitted that he did not inform the court that he had recruited Markingson into a clinical trial (pp. 78-79.) Of crucial importance here is whether Markingson’s care was compromised by enrollment in the CAFÉ study. The FDA inspector appears to believe not. Summarizing Olson’s remarks uncritically, she cites the advantages of being in a clinical trial rather than getting standard therapy (p. 8). She does not mention the disadvantages. If Markingson had not been enrolled in a double-blinded study, for example, both he and Olson would have known what drug he was being given. Olson would have had the freedom to change Markingson to another drug when it was clear that he was not improving. Also, he would have been able to use drug combinations not permitted by the protocol. Almost as alarming as what the inspection report includes is what it leaves out. Most trials of antipsychotic drugs exclude subjects who are at risk of suicide or violence, to minimize the possibility that they will harm themselves or others in the trial. The CAFÉ study excluded subjects at risk of suicide, but for reasons that remain unclear, it did not exclude subjects at risk of violence, and the University of Minnesota IRB did not object. If the exclusion criteria had excluded potentially violent subjects, Markingson could not have been enrolled in the trial; he had been involuntarily committed precisely because he was potentially homicidal. In fact, at the time of his enrollment, there seems to have been broad agreement that he was at high risk of acting out his delusions. In the years since the FDA report was issued, more alarming revelations about the trial have emerged. In May 2008, Paul Tosto and Jeremy Olson published a series of articles in the St. Paul Pioneer Press that revealed, among other things, financial conflicts of interest involving Olson and his co-investigator, Dr. Charles Schulz, and as well as the development of a locked unit for severely psychotic patients in Fairview Hospital, where patients are apparently targeted for recruitment. In 2010, AstraZeneca agreed to pay $520 million to settle two federal investigations and two whistleblower lawsuits alleging that it had marketed Seroquel illegally and concealed its health risks. One of those investigations involved physicians who had conducted clinical trials. As I reported in Mother Jones in September 2010, the documents unsealed in that case implicated Dr. Charles Schulz, the Chair of the Department of Psychiatry and a co-investigator for the CAFÉ study. Last month, City Pages journalist Andy Mannix reported even more evidence that Schulz was involved in manipulating trial results for AstraZeneca. The response of University of Minnesota officials to these reports has been remarkably consistent. The suicide was tragic, they say, but the university has already been cleared of

any blame. That uniform message has been issued by the Office of the General Counsel, the Vice-President for Research, the Communications Office, and the Dean of the Medical School. In December 2010, eight faculty members associated with the Center for Bioethics at the University of Minnesota wrote a public letter to the Board of Regents requesting an independent investigation. That letter was followed by a second letter from the Faculty for the Renewal of Public Education. On February 7, 2011, the Board of Regents refused that request. It cited, among other reasons, the fact that the FDA had already reviewed the case. Carl Elliott, a Hastings Center Fellow, is a professor at the Center for Bioethics at the University of Minnesota. His most recent book is White Coat, Black Hat: Adventures on the Dark Side of Medicine.

ATTACHMENT Q

ATTACHMENT R

ATTACHMENT S

X-Original-To: ellio023@ruby.tc.umn.edu Delivered-To: ellio023@ruby.tc.umn.edu X-Umn-Remote-Mta: [N] lsoft.ahc.umn.edu [128.101.246.151] #+LO+HN+TR X-Umn-Classification: local X-Umn-Remote-Mta: [N] x-160-94-186-116.surg.umn.edu [160.94.186.116] #+LO+TS+AU+HN X-Mailer: QUALCOMM Windows Eudora Version 7.1.0.9 Approved-By: Aaron Friedman <ahccomm@TC.UMN.EDU> Date: Tue, 8 Feb 2011 15:45:40 -0600 Reply-To: Aaron Friedman <ahccomm@tc.umn.edu> From: Aaron Friedman <ahccomm@tc.umn.edu> Subject: Response to Regents' Statement To: ALL-AHC-ALL@LSOFT.AHC.UMN.EDU Sender: owner-all-ahc-all@LSOFT.AHC.UMN.EDU Dear colleagues, Yesterday afternoon, the University's Board of Regents responded to a call from some of our faculty members to further investigate the events surrounding the tragic suicide of patient Dan Markingson five years ago. The Regents' response included a detailed analysis by General Counsel Mark Rotenberg of the specific charges of the late November letter, and can be reviewed here. There has been extensive external review of the events at both a state and federal level, as well as within our University, and I know there will remain opinions surrounding this tragic incident. I see the Regents' statement as the end of the University's review of this specific patient's case. As a result of this case, our department of psychiatry has experienced significant scrutiny and withering criticism over the past five years, and through it all, the faculty of the department have performed remarkably well in fulfilling its mission. When Chuck Schulz was recruited as chair ten years ago, he came to the University with a strong reputation in place and promised to build a solid program in education, research, and patient care. That he has done. His leadership has led to the recruitment of seventeen productive faculty members engaged in a wide range of mental health clinical care and leading research. The number of University graduates choosing the field of psychiatry for their residency and future careers has jumped from three to fifteen, increasing the number of new physician psychiatrists to care for patients. And the number of NIH grants to the department has more than doubled, with a tripling of the funding amount. These are all marks of a successful program under able leadership in the Medical School. In addition, I also want to note that Stephen Olson continues to dedicate much of his work to treatment-resistant psychosis, and continues to treat some of the most complex psychiatric cases within the Fairview health system. There is one issue underlying these ongoing accusations that I would like to address here,

and that is the idea that somehow, engaging in research funded by industry is a practice to always avoid. When our physicians, dentists, nurses, pharmacists, veterinarians, and public health professionals discover new ideas, new knowledge, or new treatments, it's our job to ensure that knowledge becomes widely available. That’s our mission on behalf of the public good. In academia, we publish those results in peer-reviewed journals, and sometimes, those results are picked up by existing or emerging businesses. That's a good outcome and something we promote as the state's land grant institution with an expectation for the public’s benefit. We all work within a university that sees the value of working with industry to further our research and service missions. And yes, we need to carefully oversee those relationships to ensure any conflicts are recognized and managed, and the new University policies in place have been developed to do just that. I am certain there will be further discussion of private support of research at a public university, and I look forward to a thoughtful airing of those issues. Aaron Friedman, M.D. Vice President for Health Sciences and Dean of the Medical School

Sign up to vote on this title
UsefulNot useful