You are on page 1of 127

EVIDENCE BASED

MEDICINE

Self-Instructional Manual

Noel L. Espallardo, MD, MSc


 
Department of Clinical Epidemiology 
UP College of Medicine and Philippine General Hospital 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Copyright, 2008 
TABLE OF CONTENTS
 
   
CONTRIBUTORS  i 
 
   
ACKNOWLEDGEMENT  i 
 
   
EVIDENCE BASED MEDICINE  1 
      Introduction   
 
   
SEARCHING THE MEDLINE  6 
      Session Briefing 
      Searching the MEDLINE Database 
 
   
DIFFERENTIAL DIAGNOSIS  15 
      Session Briefing 
      Clinical Decision on Differential Diagnosis 
      Workshop Briefing 
      Appraisal Sheet 
 
   
DIAGNOSTIC TEST  25 
      Session Briefing 
      Clinical Decision on a Diagnostic Test 
      Workshop Briefing 
      Appraisal Sheet  
 
   
THERAPY OR PREVENTION  35 
      Session Briefing 
      Clinical Decision on Therapy or Prevention 
      Workshop Briefing 
      Appraisal Sheet 
 
TABLE OF CONTENTS (CONT’D)
 
   
HARMFUL EFFECT  48 
      Session Briefing 
      Clinical Decision on Harm 
      Workshop Briefing 
      Appraisal Sheet 
 
   
PROGNOSIS  58 
      Session Briefing 
      Clinical Decision on Prognosis 
      Workshop Briefing 
      Appraisal Sheet 
 
   
HEALTH ECONOMIC ANALYSIS  69 
      Session Briefing 
      Clinical Decision on Health Economic Analysis 
      Workshop Briefing 
      Appraisal Sheet 
 
   
SYSTEMATIC REVIEW OR META‐ANALYSIS  80 
      Session Briefing 
      Clinical Decision on Systematic Review or Meta‐analysis 
      Workshop Briefing 
      Appraisal Sheet 
 
   
CLINICAL PRACTICE GUIDELINES  90 
      Session Briefing 
      How to Use a Clinical Practice Guideline 
      Workshop Briefing 
      Appraisal Sheet 
 
 
 
 
Evidence-based Medicine: Self-instructional Manual

CONTRIBUTORS
 
Dr. Leilani Apostol and Dr. Ma Elinore Alba of the Department of Family and 
Community Medicine, UP College of Medicine and Philippine General Hospital 
 
They were dedicated students in the Master of Science program of the 
Department of Family and Community Medicine, UP College of Medicine and 
Philippine General Hospital. Their knowledge and experience in conducting 
evidence‐based medicine training in Family Medicine have been very helpful in 
designing this package. 
 
 
ACKNOWLEDGEMENT
 
Dr. Noel Juban and the Faculty of the Department of Clinical Epidemiology,  
UP College of Medicine 
 
The department is considered as the center of evidence‐based medicine in the 
Philippines. The staff provided essential advice and feedback in the application of 
evidence‐based. 
 
 

Page i
Evidence-based Medicine Self-Instructional Manual

EVIDENCE BASED MEDICINE


INTRODUCTION
 
Medicine is a dynamic endeavor. Everyday challenging problems arise, new 
modalities of treatment are promoted, disease management done under 
minimal or far from ideal conditions. Suppose a patient who consulted for cough 
productive of yellowish phlegm and asked for a prescription of an antibiotic. Will 
you prescribe it or not? These are common problems that may escape our 
attention and diminish the quality of care we give if we make inappropriate 
decisions. 
 
In the old practice faced with this question, a physician will just ask a colleague 
or an expert for the answer or rely on his/her prior knowledge of the disease. He 
may also prescribe a drug because of the promotional lecture sponsored by the 
manufacturer.   
 
In evidence‐based medicine (EBM) a new paradigm is introduced. Before he 
makes a decision, the physician will first try to retrieve his latest article about the 
topic that he kept from his file, appraise the article then makes a decision. Later, 
he evaluates the effectiveness of his decision. This loop ensures improvement in 
the quality of care.  
 
 
DEFINITION OF EBM
 
Evidence‐based medicine is defined as the process of systematically finding, 
appraising, and using contemporaneous research findings as the basis for clinical 
decisions (NLM, 2008). Evidence‐based medicine follows four steps:  
 
• formulate a clear clinical question from a patient's problem 
• search the literature for relevant clinical articles 
• evaluate (critically appraise) the evidence for its validity and usefulness 
• implement useful findings in clinical practice.  
 
 
THE PARADIGM SHIFT
 
The traditional method of answering clinical problems was based on the 
following assumptions (User’s Guide, 1992):  

Page 1
Evidence-based Medicine Self-Instructional Manual

• clinical experience is the way of building and maintaining one's 
knowledge  
• basic mechanisms of disease is a sufficient guide for clinical practice 
• traditional medical training and common sense is sufficient to allow one 
to evaluate new tests and treatment.  
• clinical experts are a sufficient to generate valid guidelines for clinical 
practice 
 
The traditional assumptions are now being questioned with the new paradigm. 
The assumptions of the new paradigm are (User’s Guide, 1992):  
• clinical experience are crucial but in the absence of systematic 
observation one must be cautious in the interpretation of information 
derived from clinical experience for it may at times be misleading.  
• understanding of basic mechanisms of disease are necessary but 
insufficient guides for clinical practice  
• understanding certain rules of evidence is necessary to correctly interpret 
literature on causation, prognosis, diagnostic tests, and treatment 
strategy 
 
The new paradigm makes learning new things more self‐directed and less reliant 
on teachers. Students can gain the skills to make independent assessments of 
evidence, and thus evaluate the credibility of opinions being offered by experts. 
The purpose of this course is to introduce the concept of evidence based 
medicine and the use of these concepts to improve the quality of his/her own 
practice. 
 
 
WHY TEACH EBM
 
Medical education faces a problem in a present setting: too much information, 
too little time, too many students in crowded rooms, and exams that 
discouraged real life‐long learning (Rangachari, 2007). There is a need to make 
students asks questions about useful information and try to seek the answer for 
themselves. 
 
The term "evidence based medicine" was coined at McMaster Medical School in 
Canada in the 1980's to label an “active clinical learning” strategy, which people 
at the school had been developing for over a decade (NLM, 2008). Randomized 
controlled trials have shown that evidence‐based medicine learning is more 
effective than didactic learning among medical interns in family medicine. They 
provide better care in terms of providing treatment to patients with 
hypertension (Espallardo, 2006). 
 

Page 2
Evidence-based Medicine Self-Instructional Manual

 
MEDICAL STUDENTS LEARNING EBM
 
Currently most physicians report a moderate amount of exposure to EBM. But 
physicians in clinical research careers were more favorable towards EBM than 
those in the clinical practice careers (Luebbe, 2007). But among those who were 
already exposed to EBM there was mismatch perceived competence and their 
actual performance. This suggests that better education in EBM is needed (Caspi, 
2006). Starting the training in the undergraduate program (medical students) is 
the logical approach. Early experience helps medical students learn, helps them 
develop appropriate attitudes towards their studies and future practice and 
orientates medical curriculums towards society's needs (Littlewood, 2005). 
 
 
OBJECTIVES
 
After going through the readings and workshops of this manual, the participants 
should be able to: 
 
• define and describe the steps in applying evidence‐based medicine into 
his/her own clinical practice 
• appraise and use randomized controlled trials and other types of studies 
in solving clinical problems in clinical practice 
• make an efficient literature search and identify problems and solutions in 
the application of evidence based medicine   
 
 
METHODS
 
This is a series of self‐reading materials and group discussions. Allot a fix time for 
you to read the reading assignment in the manual. Conduct the group discussion 
with at least 5 of your colleagues. Assign a facilitator and a co‐facilitator/scribe 
for each discussion. Observe the rules enumerated in the succeeding section. 
Monitor your progress by reviewing the checklist provided before each section. 
 
The pace of learning depends on your time schedule. However, I suggest that 
you allot one day a week for the reading time and group discussion. You can also 
try to apply critical appraisal by yourself with other topics of interest. 
 
 
 
 
 

Page 3
Evidence-based Medicine Self-Instructional Manual

RULES OF DISCUSSION
 
The ground rules that are encouraged to be observe during the group discussion 
are as follows: 
 
• Honor the established time limits. 
• Allow one person to talk at a time. 
• Focus on the topic. Avoid sideline conversation. 
• Listen to what others have to say. All ideas are valued.  
• Encourage participation in the discussion for all participants. 
• Critique on ideas and thoughts, not on the person. 
 
 
ROLES AND RESPONSIBILITIES
 
FACILITATOR 
 
The facilitator serves as the process facilitator. He/she is also responsible for the 
content and final outcome of the discussion. His/her responsibilities are to: 
 
• Provide the process to achieve the objectives and desired outcomes. 
• Pose probing but non‐threatening questions to provoke thought, clarify 
discussion and bring insight on some points. 
• Provide balance. Facilitate rather than lead the discussion. 
• Remain neutral on content and avoid evaluation and decisions on ideas. 
• Encourage equal participation among group members. 
 
CO‐FACILITATOR/SCRIBE 
 
The co‐facilitator helps the facilitator to achieve his/her objectives. He/she may 
join the group discussion, but must bear in mind of his/her other functions: 
 
• Be a timekeeper to ensure progress. 
• Contribute ideas to the topic being discussed. 
• Meet with the facilitator during the break to discuss the process and 
ideas on how to proceed. 
• Record essential information (content) and observation (group process) 
for post‐discussion processing and evaluation. 
 
PARTICIPANTS 
 
The participants are encouraged to actively participate in the discussion. He/she  

Page 4
Evidence-based Medicine Self-Instructional Manual

working agreement set by the group. The amount of learning you will get from 
this course is proportional to the degree of your participation. 
 
 
REFERENCES
 
Caspi O, McKnight P, Kruse L, Cunningham V, Figueredo AJ, Sechrest L. Evidence‐based medicine: 
discrepancy between perceived competence and actual performance among graduating medical 
students. Med Teach. 2006 Jun;28(4):318‐25. 
 
Espallardo NL. Effectiveness of Critical Appraisal Workshop as a Method for Disseminating a 
Clinical Practice Guideline on Hypertension. Fil Fam Phys 2006; 44 (2): 54‐60. 
 
Littlewood S, Ypinazar V, Margolis SA, Scherpbier A, Spencer J, Dornan T. Early practical 
experience and the social responsiveness of clinical education: systematic review. BMJ. 2005 Aug 
13;331(7513):387‐91. 
 
Luebbe AM, Radcliffe AM, Callands TA, Green D, Thorn BE. Evidence‐based practice in 
psychology: perceptions of graduate students in scientist‐practitioner programs. J Clin Psychol. 
2007 Jul;63(7):643‐55. 
 
National Library of Medicine. Medical Subject Headings. www.ncbi.nlm.nih.gov/sites/entrez 
(May 27, 2008). 
 
Rangachari PK. Back to the future? Active learning of medical physiology in the 1900s. Adv 
Physiol Educ. 2007 Dec;31(4):283‐7. 
 
Users' Guides to Evidence‐based Medicine. JAMA. 1992;268(17):2420‐5. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Page 5
Evidence-based Medicine Self-Instructional Manual

READING ASSIGNMENT
SEARCHING THE MEDLINE
(SESSION BRIEFING)
 
 
CHECKLIST
 
Have you read the previous two topics? 
 Yes          No 
 
Did I get your interest now? 
 Yes          No 
 
Are you ready to work with your group or colleagues? 
 Yes          No 
 
If your answer is yes to all of the questions, congratulations! You have the 
potential of being a quality health care provider. 
 
If your answer is yes to only one of the questions, don’t worry you can always go 
back and read the previous two topics. 
 
 
ANOTHER CHECKLIST
 
Have you formed a group? 
 Yes          No 
 
Has the group elected a facilitator? 
 Yes          No 
 
Has the group elected a co‐facilitator/scribe? 
 Yes          No 
 
If not yet, what are you waiting for!  
 
 
 
ORGANIZE NOW!
 
 
 

Page 6
Evidence-based Medicine Self-Instructional Manual

OBJECTIVES
 
The purpose of the reading assignment is to introduce to the participants how to 
formulate a question and search the MEDLINE for the answer.  
 
At the end of the reading session, you should be able to search the MEDLINE 
with skill and confidence that you were able to retrieve the right answer to the 
problem.  
 
 
INSTRUCTIONS
 
Read the assignment for an article on MEDLINE search. Focus on the use of key 
terms and how to use the operant words AND or OR.   
 
After reading the paper, proceed to the library or internet and try searching in 
the MEDLINE.   
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Page 7
Evidence-based Medicine Self-Instructional Manual

READING ASSIGNMENT
SEARCHING THE MEDLINE
DATABASE
CLINICAL SCENARIO
 
Supposing a patient with cough came in to your clinic and asks for an anti‐biotic 
drug because he wants to be relieved of his cough right away. Will you prescribe 
it?  
 
Patients usually come in to the clinic for problems. Unfortunately these 
problems are vague and sometimes not clearly stated. To state the problem 
clearly, you must bear in mind that there are only three important elements that 
the patient want to know: 
 
• What their diseases are 
• What treatment they should be given 
• What is the expected outcome of the treatment 
 
These three important elements in clinical research are basically: 
 
• the patient (P) 
• the intervention/exposure (I) 
• the outcome (O) 
 
Sometimes the researcher can add 
 
• method (M) 
 
Going back to our scenario, my clearly stated problem will be: “Among patients 
with cough (P) will anti‐biotic (I) provide symptom relief faster?” 
 
 
TRADITIONAL METHOD OF LITERATURE SEARCH
 
A recent survey of important knowledge sources that influence clinical practice 
was conducted among faculty members, fellows and residents of a large 
teaching tertiary care hospital. The results showed that the most important 
resources were English journals, text books and experience (Yousefi‐Nooraie, 
2007). This dominance of the traditional information resources and experience‐

Page 8
Evidence-based Medicine Self-Instructional Manual

based medicine may be one of the barriers to the dissemination of evidence‐
based medicine.  
 
Thus educational programs to develop skills of efficiently searching the research 
literature need to be developed. Brief (two‐hour) instructional intervention on 
EBM‐based techniques for searching Medline for evidence related to a clinical 
problem provided to the students have been shown to be effective. With this 
training, students had fewer search errors and correspondingly higher quality 
searches. The most common search errors were a lack of Medical Subject 
Headings (MeSH) explosion, missing MeSH terms, lack of appropriate limits, 
failure to search for best evidence, and inappropriate combination of all search 
concepts (Gruppen, 2005).  
 
 
PUBMED, ENTREZ AND MEDLINE
 
PubMed was developed by the National Center for Biotechnology Information 
(NCBI) at the National Library of Medicine (NLM), located at the U.S. National 
Institutes of Health (NIH). PubMed provides access to bibliographic information 
that includes MEDLINE, as well as:  
 
• out‐of‐scope citations (e.g., articles on plate tectonics or astrophysics) 
from certain MEDLINE journals, primarily general science and chemistry 
journals 
• citations that precede the date that a journal was selected for MEDLINE 
indexing.  
• additional life science journals that submit full text to PubMed Central 
and receive a qualitative review by NLM.  
 
Entrez is the text‐based search and retrieval system used at NCBI for services 
including PubMed.   
 
MEDLINE is the largest component of PubMed and is the freely accessible online 
database of biomedical journal citations and abstracts created by the U.S. 
National Library of Medicine (NLM). Approximately 5,200 journals published in 
the United States and more than 80 other countries have been selected and are 
currently indexed for MEDLINE. A distinctive feature of MEDLINE is that the 
records are indexed with NLM's controlled vocabulary, the Medical Subject 
Headings (MeSH). 
 
In the internet, you can access MEDLINE through PUBMED or GRATEFULMED or 
through other organizations.  
 

Page 9
Evidence-based Medicine Self-Instructional Manual

PARTS OF PUBMED HOMEPAGE


 
Side bar 
 
Page header 
 
Entrez databases 
Query box 
Features tab 
 
 
 
 
 
 
 
BASIC SEARCH STRATEGY
 
The first step in using PubMed is to first develop a search strategy, a plan that 
helps you look for the information you need. This can be done by doing the 
following steps: 
 
• Identify the key concepts (should include the PIOM discussed earlier) 
• Determine alternative terms for these concepts (can be facilitated with 
MESH term search) 
• Refine your search (use limits like publication dates, study subjects, study 
designs, patient age, etc) 
 
Our clinical question in the previous scenario was “Among patients with cough 
(P) will anti‐biotic (I) provide symptom relief faster?” The key concepts in my 
search terms are: 
• Cough 
• Antibiotics 
• Relief of symptoms 
 
When I type the key term “cough” in the search box the yield was 27,751 articles 
and it will be impossible for me to browse these articles. When I typed “cough 
AND antibiotics" the yield was 2,175 and when I typed “cough AND antibiotics 
AND relief” the yield was 15 articles. Now I can browse through these articles. 
How did this happen? 
 
 
 

Page 10
Evidence-based Medicine Self-Instructional Manual

THE BOOLEAN PRINCIPLE


Searching MEDLINE
THE BOOLEAN LOGIC
A B
1 10 11
2 3 9
4 5 8 12
6 7 13
14
15 16

C
A OR B (A AND B) OR C
A AND B A AND (B OR C)
A AND B AND C A AND (B AND C)
 
 
In the Boolean principle, elements labeled such as 1, 2, 3, etc., can belong to set 
A, B, C, etc. Some of these elements can belong to two or more sets, and some 
of these sets may contain no elements. In the above example, if you want to 
combine elements in two sets you use the word OR, i.e. the elements in set A OR 
set B are 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13 and 14. If you want the elements 
common to the two sets you use the word AND, i.e. the elements that belong to 
both set A AND set B are 7, 8 and 9. 
 
In the MEDLINE, articles are indexed together as set of articles based on their 
key words. Just like the Boolean principle the main operators in the MEDLINE are 
also the words AND and OR. As a beginner this may be enough for you. 
 

Searching MEDLINE
THE BOOLEAN LOGIC
TB RCTs
1 10 11
2 3 9
4 5 8 12
6 7 13
14
15 16

MENINGITIS
TB OR RCTs
TB AND RCTs
TB AND RCT AND MENINGITIS

Page 11
Evidence-based Medicine Self-Instructional Manual

In the example above, if I want articles about tuberculosis (TB), I will type 
“tuberculosis” in the search window and the result will give me 9 articles. But if 
my concern is only to read randomized controlled trials (RCT) on tuberculosis I 
will type in the search window “tuberculosis AND randomized controlled trial” 
and it will give me 3 articles. Reading 3 articles instead of 9 will save me a lot of 
time.  
 
What will you type in the search box if you want randomized controlled trials on 
tuberculous meningitis?  
 
 
 
 
What articles will you get if you typed “tuberculosis AND randomized controlled 
trial AND meningitis”? 
 
 
 
 
THE ADVANCED SEARCH STRATEGY
 
The advance search page can be accessed by clicking the link Advanced Search 
(beta) on the right side of the query box. I used the advance search option and 
got the results shown below. When I type cough AND antibiotic, the yield was 2, 
175 articles and when I type cough AND antibiotic relief, the yield was 15 
articles. If I have no time to browse through 15 articles, I can limit this further by 
checking other boxes for relevance such as date of publication, type of studies, 
age of subjects etc. This can also be done in the basic search but it will take 
several steps. 
 

 
 
 

Page 12
Evidence-based Medicine Self-Instructional Manual

 
 
 

Page 13
Evidence-based Medicine Self-Instructional Manual

REFERENCES
 
Gruppen LD, Rana GK, Arndt TS. A controlled comparison study of the efficacy of training medical 
students in evidence‐based medicine literature searching skills. Acad Med. 2005 Oct;80(10):940‐
4. 
 
National Library of Medicine. PubMed OVerview. www.ncbi.nlm.nih.gov/sites/entrez (May 27, 
2008). 
 
Yousefi‐Nooraie R, Shakiba B, Mortaz‐Hedjri S, Soroush AR. Sources of knowledge in clinical 
practice in postgraduate medical students and faculty members: a conceptual map. J Eval Clin 
Pract. 2007 Aug;13(4):564‐8. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Page 14
Evidence-based Medicine Self-Instructional Manual

READING ASSIGNMENT
CRITICAL APPRAISAL OF AN
ARTICLE ABOUT DIFFERENTIAL
DIAGNOSIS
(SESSION BRIEFING)
 
 
CHECKLIST
 
Have you read the topic on MEDLINE search? 
 Yes          No 
 
Have you gone to library? 
 Yes          No 
 
Are you ready to work with your group or colleagues? 
 Yes          No 
 
If your answer is yes to all of the questions, congratulations!  
 
You are all set for an exciting learning experience! 
 
 
OBJECTIVES
 
The purpose of the reading assignment is to introduce to the participants the 
concept of medical decision making using an article about differential diagnosis.  
 
At the end of the reading session, you should be able to answer the user guides 
questions for the workshop. 
 
 
INSTRUCTIONS
 
Read the assignment for an article about a differential diagnosis. Focus on the 
critical appraisal questions, why they are asked and how to get the answers from 
the paper. After reading the paper you can proceed to conduct the group 
workshop.  
 
 

Page 15
Evidence-based Medicine Self-Instructional Manual

READING ASSIGNMENT
CLINICAL DECISION ON
DIFFERENTIAL DIAGNOSIS
CLINICAL SCENARIO
 
Suppose a 35‐year old female patient came in to your clinic for fever and 
abdominal pain for a week. There was neither diarrhea nor dysuria. You read in 
the papers that there was an increased incidence of dengue fever in children. 
How will you optimize (request only for what is essential) the diagnostic 
laboratory tests for this patient? 
 
Naturally you can do that if you already have an initial diagnosis in mind. 
Unfortunately there might be many of them. This session will help you trim 
down the differential diagnosis and request only for the laboratory tests that are 
essential. Differential diagnosis is the method of limiting the possible causes of 
the patient’s symptoms before making a final diagnosis. Identifying the right 
differentials will make patient management more focused and efficient. 
 
Differential diagnosis can be arrived at by using the anatomic approach i.e. 
considering the possibilities based on organs that may be affected within the 
proximity of the symptom like chest pain may have differential diagnosis like 
herpes zoster (skin), costochondritis (ribs), pneumonia (lings) or angina (heart).  
If the symptom is systemic like fever, the differentials can by be pathophysiology 
i.e. vascular, inflammatory/infectious, neoplastic/neurologic, degenerative, 
intoxication/idiopathic, congenital, allergic/autoimmune, trauma and endocrine 
(VINDICATE) (Friedland, 1998). With these approaches however the frequency or 
probability of each differential will not be known.. 
 
If we consider all known causes equally possible (the ‘possibilistic’ approach), 
then the patient will have unnecessary diagnostic tests performed on them. 
Instead, we must considering first those that are more common (a ‘probabilistic’ 
approach), or more more serious if left undiagnosed and untreated (a 
‘prognostic’ approach) or more responsive to treatment (Richardson, 1999). And 
they are important because the probabilities of the individual differential will 
help us focus our diagnostic strategies as shown in the table below. 
 
The disease probabilities can be taken from population prevalence statistics or 
from original research. Research studies focus more directly on the frequency of 
diseases that cause symptoms (Kroenke, 1997) are preferred over population 
survey because they are more associated with presenting symptoms. 

Page 16
Evidence-based Medicine Self-Instructional Manual

Probabilities of Differential Diagnosis and Recommended Diagnostic Strategies  
 
Differential Diagnosis  Diagnostic Tests  Treatment 
   
Working diagnosis –  Choose test(s) with high Start empiric 
specificity and LR+ much
most possible cause  larger than one. 
treatment 
that should be ruled in 
 
   
Alternative diagnosis –  Choose test(s) with high Start supportive 
sensitivity and LR- much
other possibilities that  smaller than one. 
treatment 
should be ruled‐out 
 
     
Remote diagnosis   None  None 
 
 
 
 
SEARCH
 
You found an unpublished retrospective study in the archives of your 
department written by a previous resident Santos AR, entitled “Differential 
diagnosis of typhoid fever in the emergency room”.   
 
 
CRITICAL APPRAISAL
 
RELEVANCE QUESTION 
 
• Is the objective of the article on differential diagnosis similar to your 
clinical dilemma? 
 
To answer this question, look at the objective of the study. It is important that 
your article is relevant to the question you have raised in order for you to make 
maximum use of the results of the study and be able to apply it to decision 
making that influences patient care. For differential diagnosis it is important that 
the focus is to find the cause of symptoms, clinical and laboratory presentation 
among patients similar to your patient or case scenario. 
 
 
 
 

Page 17
Evidence-based Medicine Self-Instructional Manual

VALIDITY GUIDES 
 
• Did the study patients represent the full spectrum of patients who 
present with this clinical problem? 
 
Study designs that answer clinical questions like differential diagnosis can be a 
cross‐sectional study or cohort study. An important element with these designs 
is how the subjects are recruited so they can represent other patients who may 
also have the same symptoms i.e. representativeness.  
 
The definition of the clinical problem under study describes the population to 
which the study will be applied. The problem usually is a symptom or an 
abnormal physical examination such as headache or abdominal mass or a 
combination of symptoms and abnormal physical findings like headache and 
facial asymmetry. This is usually defined in the inclusion and exclusion criteria of 
the study.  
 
With the symptom already defined, the other strategies that can assure 
representativeness are any of the following: 
• Random selection – not always possible in clinical setting 
• Consecutive patient recruitment – most feasible 
• Recruitment in defined setting – must always be done 
 
The Article by Santos included patients consulting for fever in the emergency 
room. Although the inclusion criteria were fever alone as the chief complaint, 
there was a subgroup analysis of patients with fever and abdominal complaint. 
The total number of patients included in the study was 235. This coincides with 
your case scenario.  
   
• Were the criteria for each final diagnosis explicit and credible?  
 
Determination of final diagnosis must be clearly described, may not necessarily 
be based on the ultimate reference standard. However the criteria must be 
explicit enough to make sure that different clinician will arrive at the similar 
diagnosis (inter‐rater reproducibility). 
 
The final diagnosis in Santos’s paper was based on clinical syndromes and 
criteria. Blood cultures, ultrasound and other tests were not done to establish 
the final diagnosis in only 48% of the cases.  
 
 
 
 
 

Page 18
Evidence-based Medicine Self-Instructional Manual

• Was the diagnostic work‐up comprehensive and consistently applied? 
 
The set of diagnostic work‐up should be thorough to come up with an accurate 
diagnosis. Then a minimum set of diagnostic work‐up that includes a thorough 
history and physical examination and a few initial laboratory tests should have 
been applied consistently for all patients. This can be answered when the study 
described a prospective approach in identifying patients in the study. 
Retrospective approach is usually limited because, records cannot guarantee a 
standard diagnostic approach for everybody. 
 
The diagnostic tests done for 85% of patients in the Santos study were CBC, 
urinalysis and stool examination. Temperature was measured using a mercury 
type thermometer and records with “febrile” as reported documentation of 
fever were excluded. 
   
• For initially undiagnosed patients, was follow‐up to come up with a 
diagnosis sufficiently long and complete? 
 
Sometimes the diagnosis at the early stage of the disease is really difficult and 
the patient may be classified as not having the disease or undetermined. To 
assure ourselves with the eventual diagnosis of undetermined cases, we may 
have to observe them over time. 
 
The Santos study observed the patients for 24 to 48 hours in the emergency 
room. In most patients antibiotics were started and the patients were sent home 
without fever. 
 
OVERALL, IS THE STUDY VALID? 
 
Although the study was a retrospective study, you decided that you can use this 
article because it is the only available study in your setting. 
 
WHAT ARE THE RESULTS? 
 
What were the diagnoses and their probabilities? How precise are the estimate 
of the probabilities? 
 
The probabilities of the differential diagnosis are reported as either incidence or 
prevalence with their 95% confidence interval. 
 
In the Santos study, the following top three diseases were the most common 
diagnosis given to adult patients with fever: a) typhoid fever, 34%; b) urinary 
tract infection, 32%; and c) acute gastroenteritis, 29%. No confidence intervals 
were reported. 

Page 19
Evidence-based Medicine Self-Instructional Manual

 
CAN THE RESULTS HELP ME IN CARING FOR MY PATIENTS? 
 
• Are the study patients similar to my own? 
 
For a study on differential diagnosis be applied to your patient you have to be 
assured that the characteristics of your patient is similar to the study’s inclusion 
criteria.  
 
In the Santos study, they included patients consulting in the emergency room 
but were eventually sent home. The cases seen in this study seemed to be the 
milder cases similar to your patient. 
 
• Do you think the disease probabilities in the study still apply today? 
 
Disease prevalence and incidence change across time. Old disease can be 
controlled because of effective treatment. Thus a paper on differential diagnosis 
may still include smallpox for patients with fever and skin lesions in the 1950’s, 
the probability is almost zero today. The probability of Dengue fever may differ 
in different times of the year. A little knowledge on epidemiology of disease 
across time may be necessary to have an accurate answer to this question. 
However if the disease in question does not vary over time then this is not a 
problem. 
 
The study of Santos was a three‐year retrospective study from January 1988 to 
December 1991. Seasonal variation may have been accounted for but the study 
is already 9 years old. Unfortunately you cannot find a more recent one. 
 
RESOLUTION OF THE PROBLEM IN THE SCENARIO 
 
After appraising the study of Santos you decided that your diagnostic tests will 
focus on ruling in or ruling out typhoid fever, urinary tract infection and 
gastroenteritis. 
 
 
REFERENCES
 
Friedland ed. Evidence‐based Medicine: A framework for clinical practice. Appleton and Lange, 
1998. 
 
Kroenke K. Symptoms and science: the frontiers of primary care research [Editorial]. J Gen Intern 
Med 1997; 12: 509 ‐ 510. 
 
Richardson WS, Wilson MC, Guyatt GH, Cook DJ, Nishikawa J, and the Evidence Based Medicine 
Working Group. JAMA, 1999 Apr 7; 281(13):1214‐9. 

Page 20
Evidence-based Medicine Self-Instructional Manual

WORKSHOP
CRITICAL APPRAISAL OF
AN ARTICLE ABOUT
DIFFERENTIAL DIAGNOSIS
(SESSION BRIEFING)
 
 
OBJECTIVES
 
The purpose of the workshop is to introduce to the participants the concept of 
medical decision making about differential diagnosis. Another objective is to 
introduce concepts of critical appraisal of an article regarding differential 
diagnosis focusing on the following: 
 
• validity 
• interpretation of the results 
• applicability of the results 
 
 
INSTRUCTIONS
 
Divide the participants into groups of six to ten persons per group. Assign a case 
scenario to each group and formulate an answerable problem from the scenario. 
Establish initial group consensus on how to proceed with the scenario. 
 
Ask the group to read the article retrieved to answer the problem in the 
scenario. Focus on the abstract, methods and results section. Again establish a 
group consensus on how to proceed with the scenario. Note any change in 
decisions. 
 
Critically appraise the article using the appraisal sheet provided. Answer validity 
questions, analyze the results and determine the applicability of the results. 
Establish another group consensus and note any change in decision. 
 
Process the exercise. Focus on barriers and solution to the application of the 
exercise in usual clinic practice. 
 
 
 
 

Page 21
Evidence-based Medicine Self-Instructional Manual

TIME ALLOTTED
 
The time allotted for this workshop is two hours. The recommended break‐up is: 
 
• 15 minutes to analyze the scenario and develop consensus 
• 15 minutes to read the article 
• 45 minutes to appraise the validity 
• 15 minutes to analyze results 
• 20 minutes to establish applicability 
• 10 minutes to summarize the process 
 
 
DESIRED OUTCOME
 
The participants should make a clinical decision on the scenario based on the 
critical appraisal of the evidence. 
 
 

Page 22
Evidence-based Medicine Self-Instructional Manual

Appraisal Sheet 
CLINICAL DECISION ON DIFFERENTIAL DIAGNOSIS 
 
   
CLINICAL SCENARIO OR     
QUESTION   
 
     
SEARCH                 
 
CRITICAL APPRAISAL   
   
RELEVANCE  Is the objective of the article on differential diagnosis 
similar to your clinical dilemma? 
 
 
 
   
PRIMARY VALIDITY  Did the study patients represent the full spectrum of 
GUIDES  patients who present with this clinical problem? 
Definition of the clinical problem or the patient whom the 
study will be applied. 
   
 
 
 
 
Were the criteria for each final diagnosis explicit and 
credible?  
Determination of final diagnosis must be clearly described, 
may not necessarily be the ultimate reference standard.  
 
 
 
 
 
 
   
SECONDARY VALIDITY  Was the diagnostic work‐up comprehensive and 
GUIDES  consistently applied? 
 
   
 

Page 23
Evidence-based Medicine Self-Instructional Manual

 
For initially undiagnosed patients, was follow‐up to come 
up with a diagnosis sufficiently long and complete? 
 
   
 
 
 
OVERALL, IS THE STUDY   
VALID?   
   
WHAT ARE THE  What were the diagnoses and their probabilities? How 
RESULTS?  precise are the estimates of probabilities? 
 
 
 
             
 
 
 
 
   
CAN THE RESULTS HELP  Are the study patients similar to my own? 
ME IN CARING FOR MY  Inclusion criteria, exclusion criteria, clinical definition 
PATIENTS?     
 
 
 
 
 
Do you think the disease probabilities in the study still 
apply today? 
Is the study recent? Could the probabilities change since the 
study publication? 
  
 
 
 
 
 
   
RESOLUTION OF THE   
PROBLEM IN THE   
SCENARIO   

Page 24
Evidence-based Medicine Self-Instructional Manual

READING ASSIGNMENT
CRITICAL APPRAISAL OF
AN ARTICLE ABOUT A
DIAGNOSTIC TEST
(SESSION BRIEFING)
 
 
CHECKLIST
 
Have you read the previous topic on differential diagnosis? 
 Yes          No 
 
Have you undergone the workshop on differential diagnosis with your group? 
 Yes          No 
 
Did you enjoy the workshop? 
 Yes          No 
 
If your answer is no to the last question please state the reasons below and 
share it to the group before starting the next workshop.  
 
 
OBJECTIVES
 
The purpose of the reading assignment is to introduce to the participants the 
concept of medical decision making using an article about a diagnostic test. At 
the end of the reading session, you should be able to answer the user guides 
questions for the workshop. 
 
 
INSTRUCTIONS
 
Read the reading assignment for an article about a diagnostic test. Focus on the 
critical appraisal questions, why they are asked and how to get the answers from 
the paper.  
 
After reading the paper you can proceed to conduct the group workshop. 
 
 
 

Page 25
Evidence-based Medicine Self-Instructional Manual

READING ASSIGNMENT
CLINICAL DECISION ON A
DIAGNOSTIC TEST
 
 
CLINICAL SCENARIO
 
A 70 years old female patient came in to the clinic complaining of forgetfulness. 
She’s afraid that she has dementia just like her sister. She does not want to be 
subjected to MRI or CT scan. You referred her to the psychiatric resident and she 
suggested that you perform the Mini‐mental State Examination (MMSE), but you 
doubt her decision. 
 
The next weekend you went to the library and try to learn more about the 
MMSE. 
 
 
SEARCH
 
After searching in the MEDLINE you found the article by Mulligan et al entitled “ 
A comparison of alternative methods of screening for dementia in clinical 
settings” published in the Archive of Neurology, June 1996. Luckily the full text 
was also available. 
 
 
CRITICAL APPRAISAL
 
RELEVANCE  
 
• Was the objective of the paper relevant to your clinical question? 
 
Most of the time we read journal articles because the topic is interesting.  
Because of this application to clinical practice is not ensured.  We can only 
ensure that the results of the article are applied to practice if the objectives of 
the article are relevant to the clinical problems we see in clinical practice. Thus 
the objective of the study must determine the accuracy (outcome) of the 
contemplated diagnostic test (intervention/exposure) among patients 
(population) similar to your case scenario. 
 
 
 

Page 26
Evidence-based Medicine Self-Instructional Manual

VALIDITY GUIDES 
 
• Was there an independent comparison with a reference standard? 
 
There are two elements in this guide question i.e. use of a reference standard 
and independent comparison. A reference standard for a diagnostic test is the 
test that gives the information nearest to the “truth”. Thus the accuracy of the 
test should be compared against the standard. If the diagnostic test 
approximated the standard, that means the test also approximates the “truth”. 
An independent comparison means that the reader of the reference standard 
did not know the result of the diagnostic test being evaluated (Jaeschke, 1994). 
Awareness of the initial test result may lead to increase confirmation with 
reference standard leading to bias on the accuracy of the diagnostic test being 
evaluated. Thus the first question you should answer is whether there was a 
comparison with the reference standard and whether the reference standard 
used was acceptable to your setting. The second is whether the reader of the 
reference standard was blinded to result of the diagnostic test being evaluated. 
 
In the study by Mulligan et al the reference standard used was the diagnosis of 
dementia based on the DSM‐III‐R.  
   
• Did the patient sample include an appropriate spectrum of patients to 
whom the test will be used?  
 
The accuracy of a diagnostic test among patients with low risk for the disease is 
different from patients with high risk of the disease. The clinical utility of a test 
can be seen when used among persons who are healthy, patients who are very 
sick and mostly those in‐between because these are the patients who will be 
requiring the test. Patients consulting in family practice usually belong to the 
healthy and in‐between groups while patients consulting in the hospitals are 
those in the in‐between and more severe groups. The in‐between groups may 
give an underestimate of the accuracy of the test (but it is the accuracy value to 
whom the test will be used) while the healthier and more severe may give an 
overestimate of the accuracy. If all groups are equally represented the average 
accuracy will be obtained. 
 
The study of Mulligan et al included elderly patients consulting in a geriatric 
hospital and memory clinic. The elderly age group is the population with the 
highest risk of dementia thus the results from this study may be an 
overestimate.  
 
 
 
 

Page 27
Evidence-based Medicine Self-Instructional Manual

• Was the reference standard done regardless of the result of the 
diagnostic test being evaluated? 
 
In some studies, the accuracy of a diagnostic test is examined retrospectively 
(chart review of actual practice). In actual practice however, physicians request 
to perform the reference standard based on the initial result of the diagnostic 
test. The reference standard is used to verify the initial finding i.e. when positive. 
When this happens most of the data available will be those positive for the 
diagnostic test and will likely be positive in the reference standard. This will 
increase the accuracy of the test. This is called verification bias (Jaeschke, 1994). 
To avoid this, the study must show that the reference standard was done 
regardless of the result of the diagnostic test being evaluated. 
 
It was mentioned in the Mulligan et al study that the comparison with the DSM‐
III‐R was blind and independent. 
 
• Were the methods for performing the test described in sufficient detail 
to permit replication? 
 
This is necessary so that the reader will be able to duplicate the test in his/her 
own setting and get the same valid result. Description should include 
preparation for the patient such as diet, drugs to avoid, precautions, ideal 
conditions for performing the diagnostic test and a step by step description of 
how the diagnostic test is done and interpreted. 
 
There are a lot of papers dealing with instructions on how to administer the 
MMSE and its interpretation. 
 
OVERALL, IS THE STUDY VALID? 
 
Yes. You accepted the validity since most of the questions were answered 
adequately. 
 
WHAT ARE THE RESULTS? 
 
• What were the likelihood ratios for the different possible test results? 
 
Likelihood ratio indicates by how much a given test result increases the pre‐test 
probability of the disease. A likelihood ratio of 1 means that the post‐test 
probability is similar to the pre‐test probability. A likelihood ratio of greater than 
1 increase the chance that the disease is present, and the greater the likelihood 
ratio the greater is the increase in chance. 
 

Page 28
Evidence-based Medicine Self-Instructional Manual

Some papers give the sensitivity and specificity values rather than the likelihood 
ratio. The formula for computing the likelihood ratio from sensitivity and 
specificity is shown below: 
 
Likelihood ratio of a positive test            LR (+) = Sn/1‐Sp 
Likelihood ratio of a negative test           LR (‐) = 1‐Sn/Sp 
 
A rough guide in evaluating LR values:  
• LRs >10 or < 0.1 generate large, and often conclusive changes from pre‐ 
to post‐test probability;  
• LRs of 5‐10 and 0.1‐0.2 generate moderate shifts in pre‐ to post‐test 
probability;  
• LRs of 2‐5 and 0.5‐0.2 generate small (but sometimes important) changes 
in probability; and  
• LRs of 1‐2 and 0.5‐1 alter probability to a small (and rarely important) 
degree. 
 
In the Mulligan study the LR (+) MMSE is 2.46 and the LR (‐) is 0.14. The 
alternative clinical test is the Antisaccadic Eye Movement Test (AEMT) but did 
not do very well compared with the MMSE. 
 
CAN THE RESULTS HELP ME IN CARING FOR MY PATIENTS? 
 
• Will the reproducibility of the test result and its interpretation be 
satisfactory in my setting? 
 
Even if the test was described very well, the reproducibility of the interpretation 
is necessary for the test to be adequately applied in your setting. The paper 
should report measures of agreement between interpreters or raters.  
 
If agreement is not reported, decide for yourself. Is the interpretation simple 
enough? Is the basis of the interpretation clear and specific? 
 
The MMSE questionnaire also contains instruction on how to administer the test 
and interpret the result. 
   
• Are the results applicable to my patient? 
 
If your setting is somewhat similar to that in the study, and the inclusion criteria 
include characteristics of your patient, then you can apply the results to your 
patient. Sometimes your clinical judgment is required. 
 
 

Page 29
Evidence-based Medicine Self-Instructional Manual

• Will the results change my management? 
 
The usefulness of a diagnostic test result is whether it will help the clinician 
manage his/her patient. If the diagnostic test will lead the doctor decide to give 
treatment or not, then the test is helpful and may be requested. However if after 
applying the LR value to determine the post‐test probability and this did not help 
in the decision, then you don’t have to perform the test. 
 
RESOLUTION OF THE PROBLEM IN THE SCENARIO 
 
After looking at the Mulligan et al study, you agreed with the psychiatry resident 
and even asked her help to administer the MMSE to the patient yourself. 
 
 
REFERENCES
Jaeschke R, Guyatt G, Sackett D, and the Evidence Based Medicine Working Group. How to Use
an Article About a Diagnostic Test: Validity Guides. JAMA, 1994; 271(5):389-391.

Jaeschke R, Guyatt G, Sackett D, and the Evidence Based Medicine Working Group. How to Use
an Article About a Diagnostic Test: Results and Applicability. JAMA, 1994; 271(9):703-707.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Page 30
Evidence-based Medicine Self-Instructional Manual

WORKSHOP
CRITICAL APPRAISAL OF AN
ARTICLE ABOUT A
DIAGNOSISTIC TEST
(SESSION BRIEFING)
 
 
OBJECTIVES
 
The purpose of the workshop is to introduce to the participants the concept of 
medical decision making about the usefulness of a diagnostic test. Another 
objective is to introduce concepts of critical appraisal of an article regarding a 
diagnostic test focusing on the following: 
 
• Validity 
• Reference standard 
• Sensitivity and specificity 
• Likelihood ratios 
• Applicability of the results 
 
 
INSTRUCTIONS
 
Divide the participants into groups of six to ten persons per group. Assign a case 
scenario to each group and formulate an answerable problem from the scenario. 
Establish initial group consensus on how to proceed with the scenario. 
 
Ask the group to read the article retrieved to answer the problem in the 
scenario. Advice them to focus on the abstract, methods and results section. 
Again, establish a group consensus on how to proceed with the scenario. Note 
any change in decisions. 
 
Critically appraise the article using the appraisal sheet provided. Answer validity 
questions, analyze the results and determine the applicability of the results. 
Establish another group consensus and note any change in decision. 
 
Process the exercise. Focus on barriers and solution to the application of the 
exercise in usual clinic practice. 
               
 

Page 31
Evidence-based Medicine Self-Instructional Manual

TIME ALLOTTED
 
The time allotted for this workshop is two hours. The recommended break‐up is: 
 
• 15 minutes to analyze the scenario and develop consensus 
• 15 minutes to read the article 
• 45 minutes to appraise the validity 
• 15 minutes to analyze results 
• 20 minutes to establish applicability 
• 10 minutes to summarize the process 
 
 
DESIRED OUTCOME
 
The participants should make a clinical decision on the scenario based on the 
critical appraisal of the evidence.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Page 32
Evidence-based Medicine Self-Instructional Manual

Appraisal Sheet 
CLINICAL DECISION ON A DIAGNOSTIC TEST 
 
   
CLINICAL SCENARIO OR     
QUESTION   
 
     
SEARCH                 
 
CRITICAL APPRAISAL   
   
RELEVANCE  Is the objective of the study relevant to your clinical 
question? 
 
 
 
   
VALIDITY GUIDES  Was there an independent and blind comparison with a 
reference standard? 
What was the reference standard. Were they assessed 
independently? 
   
 
 
 
 
 
Did the patient sample include an appropriate spectrum of 
patients to whom the test will be used?  
 
 
 
 
 
 
   
  Was the reference standard done regardless of the result 
of the diagnostic test being evaluated? 
 
 
   
 
 

Page 33
Evidence-based Medicine Self-Instructional Manual

 
Were the methods for performing the test described in 
sufficient detail to permit replication? 
 
   
 
   
 
 
OVERALL, IS THE STUDY   
VALID?   
 
   
WHAT ARE THE  What were the likelihood ratios for the different possible 
RESULTS?  test results? 
 
   
 
 
 
 
   
CAN THE RESULTS HELP  Will the reproducibility of the test result and its 
ME IN CARING FOR MY  interpretation be satisfactory in my setting? 
PATIENTS?   
   
 
 
 
Are the results applicable to my patient? 
 
 
 
 
 
Will the results change my management? 
 
 
 
 
 
RESOLUTION OF THE   
PROBLEM IN THE   
SCENARIO 

Page 34
Evidence-based Medicine Self-Instructional Manual

READING ASSIGNMENT
CRITICAL APPRAISAL OF AN
ARTICLE ABOUT THERAPY OR
PREVENTION
(SESSION BRIEFING)
 
 
CHECKLIST
 
Have you read the previous topic on diagnostic tests? 
 Yes          No 
 
Have you undergone the workshop on diagnosis with your group? 
 Yes          No 
 
Did you enjoy the workshop? 
 Yes          No 
 
If your answer is no to the last question please state the reasons below and 
share it to the group before starting the next workshop. 
 
 
OBJECTIVES
 
The purpose of the reading assignment is to introduce to the participants the 
concept of medical decision making using an article about therapy or prevention.  
 
At the end of the reading session, you should be able to answer the user guides 
questions for the workshop. 
 
 
INSTRUCTIONS
 
Read the reading assignment for an article about therapy or prevention. Focus 
on the critical appraisal questions, why they are asked and how to get the 
answers from the paper.  
 
After reading the paper you can proceed to conduct the group workshop. 
 
 

Page 35
Evidence-based Medicine Self-Instructional Manual

READING ASSIGNMENT
CLINICAL DECISION ON
THERAPY OR PREVENTION
 
 
CLINICAL SCENARIO
 
You are working at the out‐patient clinic in your institution when a 55 year‐old, 
male bank executive came to your clinic for constricting chest pain located at 
mid‐sternum precipitated by exertion and relieved by rest.  His blood pressure 
was 130/80, heart rate was 85/minute and the respiratory rate was 20/minute.  
An ECG was done revealing lateral wall ischemia.  You sent the patient home on 
oral nitrates and aspirin.  However, further work‐ups revealed 
hypercholesterolemia with a level of 6.5 mol/liter.  You gave dietary advice but 
the patient claimed that he has been on a high fiber diet with low fat intake for 
the past 6 months. 
 
He is now inquiring if he should take a drug for his elevated cholesterol. 
 
 
SEARCH
 
You decided to translate this clinical dilemma to an answerable question.  You 
also decided that the article should include a population of patients that has an 
elevated cholesterol who have already undergone dietary therapy for at least 6 
months to be consistent with your patient’s profile.  The drug intervention 
should be compared with placebo.  The article must report clinically important 
outcomes, such as reduction in cardiovascular deaths.  Finally, you wanted an 
article that employed randomization. 
 
The article you finally retrieved included 4444 patients randomized to either 
simvastatin or placebo.  However, before applying the results to your patient, 
you decided to appraise the article using the following guides. 
 
 
RANDOMIZED CONTROLLED TRIALS
 
Randomized controlled trials are the standards design to prove effectiveness of 
drugs or other forms of intervention. When done properly, it can provide the 
best evidence of effectiveness. In this type of design, individuals are randomly 
assigned (randomization) to either of the two or more groups, one with the 

Page 36
Evidence-based Medicine Self-Instructional Manual

intervention the other without the intervention being tested or another 
intervention. Randomization tries to make the two groups similar for both 
known and unknown factors that may affect the outcome other than the 
intervention being tested. Then they are observed forward in time and their 
outcome compared. The outcome can be the cure of a disease, relief of 
symptoms or improvement in quality of life (Espallardo, 2000). 
 
 
CRITICAL APPRAISAL
 
RELEVANCE 
 
• Is the objective of the article comparing therapeutic interventions 
similar to your clinical dilemma? 
 
Before going any further, first ascertain if the objective of the study addresses 
the clinical problem you face.  Appraising an irrelevant article would not be 
helpful to your clinical dilemma and will be a waste of time.  Below are a few tips 
that will help you decide on relevance: 
 
• Population of the study (P )– should be similar to the characteristic of 
your patient. 
• Intervention/comparative intervention/exposure (I) – should include the 
therapeutic intervention you want to test.   
• Outcome of the study (O) – one of the outcomes measured should be the 
goal you and your patient wish to work for.   
 
VALIDITY GUIDES 
 
• Was the assignment of patients to treatment randomized? 
 
In order to answer this question, the reader is advised to look into the article’s 
abstract or methodology section.  For the 4S study, randomization was done and 
was written both in the abstract and methods section. 
 
The strength of randomization is that if the sample size is sufficiently large, it 
assures that both known and unknown determinants of outcome are evenly 
distributed between the treatment and control groups.  In the absence of 
randomization, these factors might be difficult to control and might be the one 
strongly influencing outcome rather than the treatment itself (Guyatt, 1994). 
 

Page 37
Evidence-based Medicine Self-Instructional Manual

At times though because of the rarity of the disease and small patient sample 
size, randomization might not be feasible.  In these cases, a clinician must rely on 
weaker studies but should be aware of its potentials for errors. 
 
Were all patients who entered the trial properly accounted for and attributed at 
its conclusion?  
 
• Was follow‐up complete? 
 
This is best checked by looking at the number of patients enrolled at the outset 
and comparing this with the number of patients reported in the results table. 
 
Every patient who entered the trial should be accounted for at its conclusion.  If 
substantial numbers are “lost to follow‐up”, the validity of the conclusions are 
open to question.  A drop‐out rate of 20% or more is usually declared as 
substantial.  If the number lost to follow‐up is less than this the reader can 
decide if this affects the conclusion by assuming a “worst case scenario”.  This 
means that the numbers lost in the treatment group are assumed to have bad 
outcomes and the numbers lost in the control group are assumed to have been 
cured and if the conclusions differ, a substantial number was lost to follow‐up. 
 
Another way of deciding whether follow‐up was complete is to check whether an 
intention to treat analysis was done.  If this is reported one can safely assume 
that follow‐up was complete. 
 
• Were patients analyzed in the groups to which they were randomized? 
 
It simply means that all those belonging to the control group or treatment group 
are analyzed from beginning to end in this same grouping including those who 
were dropped or withdrawn or changed treatment.  No crossing over treatment 
modalities were done as this would likely lead to biased results. 
 
Excluding non‐compliant patients from the analysis leaves behind those who 
may be destined to have a better outcome and destroys the unbiased 
comparison provided by randomization. This principle of attributing all patients 
to the group to which they were randomized results in an "intention‐to‐treat" 
analysis. This strategy preserves the value of randomization: prognostic factors 
that we know about, and those we don't know about, will be, on average, 
equally distributed in the two groups, and the effect we see will be just that due 
to the treatment assigned (Guyatt, 1994).  
 
 
 

Page 38
Evidence-based Medicine Self-Instructional Manual

• Were patients, their clinicians, and study personnel "blind" to 
treatment? 
 
To answer this question the reader is again advised to look into the abstract or 
the methodology section. 
 
Blinding is the process by which the intervention being given is concealed from 
the patient, the clinicians and the one who analyzes the data.  Patients, clinicians 
and data analysts are likely to have an opinion regarding the experimental 
treatment.  These opinions, whether optimistic or pessimistic, can systematically 
distort reporting of treatment outcomes.  As to avoid these “reporter and 
observer” bias, blinding is necessary. 
 
• Were the groups similar at the start of the trial? 
 
To answer this question, one should look for a report of the comparison of the 
baseline characteristics of the experimental and control group.  For most studies, 
this is labeled as table 1.  In the 4S trial, baseline characteristics were similar. 
 
For reassurance about the study’s validity, readers would like to be informed 
that the treatment and control groups were similar for all the factors that 
determine clinical outcomes of interest save for the experimental therapy.  The 
greater the similarity between known prognostic factors for the control and 
experimental group, the more likely that the results can be attributed to the 
intervention, rather than due to the differences in these factors. 
 
• Aside from the experimental intervention, were the groups treated 
equally?  
 
Interventions other than the treatment under study, when differentially applied 
to the treatment and control groups, are called “co‐interventions”.  This might 
distort the results since they in themselves might cause changes in reported 
outcomes. 
 
OVERALL, IS THE STUDY VALID? 
 
If you want to be strict about it, you should answer yes in all 5 questions.  
However, you as the user of the journal can make the decision.  A simple rule 
might be to answer yes to at least, one primary guide and two secondary guides. 
 
The 4S study yielded yes to all the appraisal questions, hence you decided that 
over‐all the study was valid.  You now proceed to analyze the results. 
 
 

Page 39
Evidence-based Medicine Self-Instructional Manual

WHAT ARE THE RESULTS? 
 
• How large was the treatment effect? 
 
Most randomized controlled trials report outcomes either as treatment success 
or treatment failures and adverse effects.  Examples of outcomes include cure 
rates, side effects or death.  Patients either do or do not suffer these events and 
the article frequently reports the proportion of patients who develop such 
events.  In the 4S study, 11.5% died in the placebo group and 8.2% died in the 
simvastatin group.  By eyeballing these figures, simvastatin seems better in 
reducing deaths.  But how else could this figures be compared?  The following 
simple computations could help: 
 
Risk in Control (Rc) = Death in control/N patients in the control 
 
Risk in Treatment (Rt) = Death in treatment/N patients in treatment 
 
Absolute Risk Reduction (ARR) = Rc – Rt = 0.115 ‐ 0.082 = 0.033 
 
Relative Risk (RR) = Rt/Rc = 0.082/0.115 = 0.71 
 
Relative Risk Reduction (RRR) = 1 – RR = 1‐ 0.71 = 0.29 (29%) 
 
Absolute risk reduction is the absolute difference between the proportion who 
died in the placebo group compared to the simvastatin group.  Relative risk is the 
risk of events in the simvastatin group or new treatment relative to the placebo 
or control group.  The most useful measure to use in explaining the benefit of 
treatment to patients is the relative risk reduction.  In this case, simvastatin 
treatment reduces deaths 29% more than placebo. 
 
• How precise was the estimate of treatment effect? 
 
To decide regarding precision, one should look at the reported 95% confidence 
interval. The closer these values, the more precise your estimates.  If this is not 
reported check the p‐value, anywhere from </= .10 is acceptable. 
 
 Ideally, the reported minimum and maximum values of this interval should all be 
positive or all be negative for the absolute risk reduction, all above below one for 
the relative risk and relative risk reduction to be considered precise. 
 
 
 
 
 

Page 40
Evidence-based Medicine Self-Instructional Manual

CAN THE RESULTS HELP ME IN CARING FOR MY PATIENTS? 
 
• Can the results be applied to my patient care? 
 
If your patient meets all of the inclusion criteria and none of the exclusion 
criteria, the applicability of the study’s results to your patient is without 
question.  It is rare however that we get a patient who conforms to all the 
characteristics of the study subjects.  In these cases, we should decide if the 
reason is compelling enough not to apply the results of our study to our 
particular patient. 
 
• Were all clinically important outcomes considered? 
 
              Clinically important outcomes may range form decreasing mortality, 
morbidity, improving quality of life.  These are outcomes that are important to 
the patients and will lead directly to reducing symptoms or decreasing death.  
Some studies might report improvement in cholesterol levels, improvement in 
PFTs but these are what might be labeled as surrogate endpoints.  That is, the 
researchers have substituted these physiologic measures for important 
outcomes we have just mentioned.  For reduction in these laboratory 
parameters does not always translate into decrease in morbidity and mortality. 
 
A dramatic example of the danger of substitute endpoints was found in the 
evaluation of the usefulness of clofibrate as anti‐cholesterol drug. It shown to 
decrease serum cholesterol but was shown to increase all‐cause mortality. 
Similar findings were noted in an anti‐arrhythmia trial hen the investigators had 
to stop the trials when they discovered that mortality was substantially higher in 
patients receiving antiarrhythmic treatment than in those receiving placebo.  
  
• Are the likely treatment benefits worth the potential harm and costs? 
 
Computation for cost effectiveness and checking for side effects might be done 
to check if the treatment benefits are worth the potential harm and costs. 
 
RESOLUTION OF THE PROBLEM IN THE SCENARIO 
 
The article was valid and giving simvastatin reduces all deaths by 29% compared 
to placebo.  However when we look into costs, giving this drug treatment is fairly 
expensive.  However due to the marked cardiac risks for this patient you decide 
to give simvastatin since your patient could also afford drug treatment. 
 
At this point, we hope that you are encouraged to adapt this new paradigm in 
your medical decision making.  The steps are relatively simple.  First, define the 
problem clearly.  Second, use one of several search strategies to come up with 

Page 41
Evidence-based Medicine Self-Instructional Manual

relevant articles.  Third, appraise the article.  Assess the results and the 
applicability of your article to your patient.  Guided by this, you then decide on 
the action to take. 
 
This may sound like a tedious process but as they say practice makes perfect.  
And if each of us will continue asking, searching, we will continue learning and 
hopefully improving our patient care.   
 
 
REFERENCES
 
Espallardo, NL. Research Protocol Development for Resident Physicians. Family Medicine 
Research Group, Inc. Manila, 2000. 
 
Guyatt GH, Sackett D, Cook DJ, for the Evidence Based Medicine Working Group. How to Use an 
Article About Therapy or Prevention: Validity. JAMA, 1993;270(21):2598‐2601. 
 
Guyatt GH, Sackett D, Cook DJ, for the Evidence Based Medicine Working Group. How to Use an 
Article About Therapy or Prevention: Results and Apllicability. JAMA, 1994;271(1):59‐63.   
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Page 42
Evidence-based Medicine Self-Instructional Manual

WORKSHOP
CRITICAL APPRAISAL OF
AN ARTICLE ABOUT THERAPY
OR PREVENTION
(SESSION BRIEFING)
 
 
OBJECTIVES
 
The purpose of the workshop is to introduce to the participants the concept of 
medical decision making about treatment. Another objective is to introduce 
concepts of critical appraisal of an article regarding treatment focusing on the 
following: 
 
• validity 
• randomization 
• intention‐to‐treat 
• interpretation of the results 
• applicability of the results 
• clinically relevant endpoints 
• cost‐effectiveness 
 
 
INSTRUCTIONS
 
Divide the participants into groups of six to ten persons per group. Assign a case 
scenario to each group and formulate an answerable problem from the scenario. 
Establish initial group consensus on how to proceed with the scenario. 
 
Ask the group to read the article retrieved to answer the problem in the 
scenario. Focus on the abstract, methods and results section. Again, establish a 
group consensus on how to proceed with the scenario. Note any change in 
decisions. 
 
Critically appraise the article using the appraisal sheet provided. Answer validity 
questions, analyze the results and determine the applicability of the results. 
Establish another group consensus and note any change in decision. 
 
Process the exercise. Focus on barriers and solution to the application of the 
exercise in usual clinic practice. 

Page 43
Evidence-based Medicine Self-Instructional Manual

TIME ALLOTTED
 
The time allotted for this workshop is two hours. The recommended break‐up is: 
 
• 15 minutes to analyze the scenario and develop consensus 
• 15 minutes to read the article 
• 45 minutes to appraise the validity 
• 15 minutes to analyze results 
• 20 minutes to establish applicability 
• 10 minutes to summarize the process 
 
 
DESIRED OUTCOME
 
The participants should make a clinical decision on the scenario based on the 
critical appraisal of the evidence. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Page 44
Evidence-based Medicine Self-Instructional Manual

Appraisal Sheet 
CLINICAL DECISION ON THERAPY OR PREVENTION 
 
   
CLINICAL SCENARIO OR     
QUESTION   
 
     
SEARCH                 
 
CRITICAL APPRAISAL   
   
RELEVANCE  Is the objective of the article comparing therapeutic 
  interventions similar to your clinical dilemma? 
 
 
 
   
VALIDITY GUIDES  Was the assignment of patients to treatment randomized? 
Randomization vs. random selection 
   
 
 
 
 
Were all patients who entered the trial properly accounted 
for and attributed at its conclusion?  
 
Was follow‐up complete? 
Dropouts, withdrawals 
 
 
 
 
 
 
Were patients analyzed in the groups to which they were 
randomized? 
Intention‐to‐treat analysis 
 
 
 
 
 

Page 45
Evidence-based Medicine Self-Instructional Manual

   
  Were patients, their clinicians, and study personnel "blind" 
to treatment? 
 
   
   
 
 
 
OVERALL, IS THE STUDY   
VALID?   
 
   
WHAT ARE THE  How large was the treatment effect? 
RESULTS?  Risk in control, risk in treatment, relative risk, relative risk 
reduction, absolute risk reduction 
 
 
 
 
               
 
How precise was the estimate of treatment effect? 
95% confidence interval, p value 
 
 
 
   
CAN THE RESULTS HELP  Can the results be applied to my patient care? 
ME IN CARING FOR MY  Inclusion criteria, exclusion criteria 
PATIENTS?     
 
 
 
 
 
Were all clinically important outcomes considered? 
Outcome, results 
  
 
 
 
 
 

Page 46
Evidence-based Medicine Self-Instructional Manual

Are the likely treatment benefits worth the potential harm 
and costs? 
Side effects, NNT, costs 
 
 
 
 
 
 
   
RESOLUTION OF THE   
PROBLEM IN THE 
SCENARIO 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Page 47
Evidence-based Medicine Self-Instructional Manual

READING ASSIGNMENT
CRITICAL APPRAISAL OF AN
ARTICLE ABOUT HARM
(SESSION BRIEFING)
 
 
CHECKLIST
 
Have you read the previous topic on treatment? 
 Yes          No 
 
Have you undergone the workshop on treatment with your group? 
 Yes          No 
 
Did you enjoy the workshop? 
 Yes          No 
 
If your answer is no to the last question please state the reasons below and 
share it to the group before starting the next workshop.  
 
 
OBJECTIVES
 
The purpose of the reading assignment is to introduce to the participants the 
concept of medical decision making using an article about harm.  
 
At the end of the reading session, you should be able to answer the user guides 
questions for the workshop. 
 
 
INSTRUCTIONS
 
Read the reading assignment for an article about harm. Focus on the critical 
appraisal questions, why they are asked and how to get the answers from the 
paper.  
 
After reading the paper you can proceed to conduct the group workshop. 
 
 
 
 

Page 48
Evidence-based Medicine Self-Instructional Manual

READING ASSIGNMENT
CLINICAL DECISION ON HARM
 
 
CLINICAL SCENARIO
 
You are the resident physician covering for the consultant in his clinic. You met 
one of his patient, 45 year old male who is on anti‐cholesterol medication. You 
noted that his cholesterol is now only 4.35 mmol/L but the patient was still 
prescribed with anti‐cholesterol drug from his last week visit. You approach the 
consultant and mentioned it to him. He told you its okay; it’s the high cholesterol 
that we should be concerned with anyway. 
 
Still doubtful, you went to the library and look for the harmful effect of very low 
cholesterol. 
 
 
SEARCH
 
You found the article by Zureik et al, entitled “Serum cholesterol concentration 
and death from suicide in men: Paris prospective study 1” published in the 
British Medical Journal, September 1996. 
 
 
COHORT AND CASE-CONTROL STUDIES
 
Harmful effects of drugs or other exposure can best be studied with cohort or 
case‐control study designs. A cohort is any group of individuals who share the 
same characteristics. In a cohort study, selection of subjects starts with 
identifying individuals who have the same characteristics or presence or absence 
of a particular cause or exposure. They are then divided into two groups, those 
with the characteristics or causes and those without the characteristics. They are 
then observed forward in time and determine who among them develop the 
outcome or effect (Espallardo, 2000).  
 
Cohort studies can also be prospective or retrospective depending on the 
manner of patient recruitment. If recruitment is being done forward in time it is 
prospective, but if the cohort already existed in the past and data gathering is 
being done by reviewing existing clinical records then it is retrospective 
(Espallardo, 2000). 
 

Page 49
Evidence-based Medicine Self-Instructional Manual

In a case‐control study, inclusion of subjects starts with defining or selecting 
those who have the outcome or effect. These are considered as cases. Then this 
group is compared with subjects who don’t have the outcome or effect. These 
are considered as the controls. Both the cases and the control should be taken 
from within the same population. Then the two groups are investigated as to the 
presence or absence of hypothesized causes or risk factors for the outcome 
(Espallardo, 2000).  
 
Case‐control study can be prospective or retrospective depending on the 
manner of patient recruitment. If recruitment is being done as cases develop 
forward in time it is prospective, but if the cases have already developed in the 
past and patient recruitment is being done by reviewing existing clinical records 
then it is retrospective (Espallardo, 2000). 
 
 
CRITICAL APPRAISAL
 
RELEVANCE 
 
• Is the objective of the article on harm similar to your clinical dilemma? 
 
Your formulated clinical question must be addressed by the objective of the 
study. For a decision to stop the harmful exposure the article must be designed 
to determine the harmful effect (outcome) of the a drug, chemical or 
environmental substances (intervention/exposure) to the patient (population). 
This must be clearly stated by the objective of the study. 
 
VALIDITY GUIDES 
 
• Were there clearly identified comparison groups? 
 
Just like studies of effectiveness of treatment, studies of harm require that a 
control group for comparison should be done. Unfortunately, randomized 
controlled trials to prove harm is not ethical, so studies of weaker design like 
cohort or case‐control studies may be relied upon. Both designs has a control 
group for comparison, in the former controls are chosen by absence of exposure 
and in the latter by absence of the outcome or disease. 
 
In a cohort study, a group of patients are observed. They are divided into those 
with the exposure and those without the exposure. Then the outcome is 
observed forward in time. In a case‐control study, patients with the outcome are 
gathered. Then a group of patient without the outcome matched to the 
preceding group for certain characteristics other than the exposure is also 

Page 50
Evidence-based Medicine Self-Instructional Manual

gathered. The presence of the exposure in both groups is then ascertained 
(Levine, 1994).  
 
In a cross‐sectional study, patients are grouped and analyzed with respect to 
their outcome and exposure. This study design can also be a basis for 
establishing harmful effect, but the temporal relationship of the exposure 
occurring before the outcome cannot be established. 
 
The study by Zureik observed 6,728 men who had measurements of serum 
cholesterol. They grouped and categorized cholesterol levels into low (<4.78 
mmol/L), normal (4.78 to 6.21 mmol/L) and high (>6.21 mmol/L). The changes 
were also categorized. These were the comparison groups. 
 
• Were the exposures and outcomes measured in the same way in the 
groups compared? 
 
Measurement of outcomes must be similar in both groups. In cohort study, the 
investigators must show that diligent observation for the outcome was done in 
the groups with the exposure (high risk) as well as the groups without the 
exposure (low risk). When the patient with the exposure were observed more 
diligently, there will be a higher detection rate of the outcome leading to 
increase incidence of the disease in the exposed group. This is called surveillance 
bias (Levine, 1994). This must be avoided. This bias may also occur in case‐
control studies, when the detection of exposure is more diligent in the group 
with the disease or outcome. 
 
Suicide data were taken from the national databases and death certificates in all 
groups in the Zureik study. 
 
• Was follow‐up sufficiently long and complete? 
  
The length of follow‐up must be sufficiently long enough to detect the outcome. 
If follow‐up is short, the chance of underestimating the effect of the exposure is 
high. When the relation between asbestos and lung cancer was being 
investigated the relative risk was only 1.4 in the early years of observation 
compared to the subsequent relative risk of 18.2 when the years of observation 
was extended to 15 years and beyond.  
 
The Zureik study observed their patients for 4 years after enrolment. They had 
95% follow‐up. 
 
 
 
 

Page 51
Evidence-based Medicine Self-Instructional Manual

• Is the temporal relationship between the exposure and outcome correct 
and dose response gradient present? 
 
For an exposure to cause an effect, two important criteria may be considered. 
First the exposure must be present before the outcome and second there must 
be a dose response gradient, i.e. the higher the dose the higher is the probability 
of the outcome. Cross‐sectional studies usually cannot establish temporal 
relationship but it can establish a dose response gradient and a comparison 
between groups.  
 
The Zureik study measured serum cholesterol before the event of suicide so the 
temporal relationship was correct.  
 
OVERALL, IS THE STUDY VALID? 
 
Since all the answers to the validity guides, the study can be considered valid. 
 
WHAT ARE THE RESULTS? 
 
• What is the magnitude of the association between exposure and 
outcome? Was the estimate of the risk precise? 
 
The relative risk is the incidence of the adverse effect in the group with the 
exposure divided by the incidence of adverse effect in the group without the 
exposure. If the relative risk is more than 1, then the exposure is causing harm 
and if less than 1 the exposure reduces harm. Relative risk is usually computed 
when the design is a cohort study. 
 
In a case control study, the odds ratio is computed. The odds ratio approximates 
the relative risk, especially when the disease is rare.  
 
The values of 95% confidence interval should be greater than 1 to say that the 
exposure really causes harm. If one value of the 95% confidence interval is less 
than 1 and the other is more than 1, then the effect of the exposure is uncertain.  
 
The Zureik study showed that those with low cholesterol had increased risk of 
suicide with a relative risk of 3.16 with a 95% CI of 1.38 to 7.22. The analysis was 
also adjusted for age, smoking and mean corpuscular volume. 
 
 
 
 
 
 

Page 52
Evidence-based Medicine Self-Instructional Manual

CAN THE RESULTS HELP ME IN CARING FOR MY PATIENTS? 
 
• Are the study patients similar to my own? 
 
Just like in an article about a beneficial intervention, for the harmful effect to be 
extrapolated to your patient you have to be assured that the characteristics of 
your patient is similar to the study’s inclusion criteria.  
 
The Zureik study recruited men between 43 to 52 years old with similar 
demographic characteristics with our patient. 
 
 
• Should I attempt to stop the exposure?  
 
In answering this question you should consider the following: 
• How large and precise is the risk of harm? 
• What are the consequences if I withdraw the exposure? 
• Do I have any alternative for the exposure? 
 
Decision is simple when the answers to these questions are clear. For example 
cigarette smoking has been associated with increase incidence of lung cancer 
and cardiac deaths, but withdrawing smoking may lead to “decrease quality of 
life” for smokers. But recently an alternative like nicotine patch has been shown 
to decrease withdrawal discomfort and eventually improve smoking cessation. 
So the decision to withdraw smoking for every patient consulting in the clinic is 
warranted. 
 
RESOLUTION OF THE PROBLEM IN THE SCENARIO 
 
Based on the appraisal you decided to go back to your consultant and inform 
him about the study that you found. He thanked you for the information and 
promised to withdraw the anti‐cholesterol drug when the patient comes back. 
 
 
REFERENCES
 
Espallardo, NL. Research Protocol Development for Resident Physicians. Family Medicine 
Research Group, Inc. Manila, 2000. 
 
Levine M, Walter S, Lee H, Haines T, Holbrook A, Moyer V, for the Evidence Based Medicine 
Working Group. How to Use an Article about Harm. JAMA, 1994;271(20):1615‐1619. 
 
 
 
 

Page 53
Evidence-based Medicine Self-Instructional Manual

WORKSHOP
CRITICAL APPRAISAL OF AN
ARTICLE ABOUT HARM
(SESSION BRIEFING)
 
 
OBJECTIVES
 
The purpose of the workshop is to introduce to the participants the concept of 
medical decision making about harmful effect of a substance or drug. Another 
objective is to introduce concepts of critical appraisal of an article regarding a 
harmful effect of a substance or drug focusing on the following: 
 
• Validity 
• Cohort study, case‐control study, case series and case report 
• Interpretation of the results 
• Applicability of the results 
 
 
INSTRUCTIONS
 
Divide the participants into groups of six to ten persons per group. Assign a case 
scenario to each group and formulate an answerable problem from the scenario. 
Establish initial group consensus on how to proceed with the scenario. 
 
Ask the group to read the article retrieved to answer the problem in the 
scenario. Focus on the abstract, methods and results section. Again, establish a 
group consensus on how to proceed with the scenario. Note any change in 
decisions. 
 
Critically appraise the article using the appraisal sheet provided. Answer validity 
questions, analyze the results and determine the applicability of the results. 
Establish another group consensus and note any change in decision. 
 
Process the exercise. Focus on barriers and solution to the application of the 
exercise in usual clinic practice. 
 
 
 
 
 

Page 54
Evidence-based Medicine Self-Instructional Manual

TIME ALLOTTED
 
The time allotted for this workshop is two hours. The recommended break‐up is: 
 
• 15 minutes to analyze the scenario and develop consensus 
• 15 minutes to read the article 
• 45 minutes to appraise the validity 
• 15 minutes to analyze results 
• 20 minutes to establish applicability 
• 10 minutes to summarize the process 
 
 
DESIRED OUTCOME
 
The participants should make a clinical decision on the scenario based on the 
critical appraisal of the evidence. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Page 55
Evidence-based Medicine Self-Instructional Manual

Appraisal Sheet 
CLINICAL DECISION ON HARM 
 
   
CLINICAL SCENARIO OR     
QUESTION   
 
     
SEARCH                 
 
CRITICAL APPRAISAL   
   
RELEVANCE  Is the objective of the article on harm similar to your 
clinical dilemma? 
 
 
 
   
PRIMARY VALIDITY  Were there clearly identified comparison groups? 
GUIDES   
 
 
 
 
 
 
Were the exposures and outcomes measured in the same 
way in the groups compared? 
 
 
 
 
 
 
 
Was follow‐up sufficiently long and complete? 
 Is the temporal relationship between the exposure and 
outcome correct and dose response gradient present? 
 
 
 
 
 
 

Page 56
Evidence-based Medicine Self-Instructional Manual

   
OVERALL, IS THE STUDY   
VALID? 
 
   
WHAT ARE THE  What is the magnitude of the association between 
RESULTS?  exposure and outcome? Was the estimate of the risk 
precise? 
 
 
 
 
 
 
 
 
   
CAN THE RESULTS HELP  Are the study patients similar to my own? 
ME IN CARING FOR MY   
PATIENTS?   
 
 
 
 
 
Should I attempt to stop the exposure?  
 
 
 
 
 
 
 
   
RESOLUTION OF THE   
PROBLEM IN THE 
SCENARIO 
 
 
 
 
 
 
 

Page 57
Evidence-based Medicine Self-Instructional Manual

READING ASSIGNMENT
CRITICAL APPRAISAL OF AN
ARTICLE ABOUT PROGNOSIS
(SESSION BRIEFING)
 
 
CHECKLIST
 
Have you read the previous topic on harm? 
 Yes          No 
 
Have you undergone the workshop on harm? 
 Yes          No 
 
Did you enjoy the workshop? 
 Yes          No 
 
If your answer is no to the last question please state the reasons below and 
share it to the group before starting the next workshop.  
 
 
OBJECTIVES
 
The purpose of the reading assignment is to introduce to the participants the 
concept of medical decision making using an article about prognosis.  
 
At the end of the reading session, you should be able to answer the user guides 
questions for the workshop. 
 
 
INSTRUCTIONS
 
Read the reading assignment for an article about prognosis. Focus on the critical 
appraisal questions, why they are asked and how to get the answers from the 
paper.  
 
After reading the paper you can proceed to conduct the group workshop. 
 
 
 
 

Page 58
Evidence-based Medicine Self-Instructional Manual

READING ASSIGNMENT
CLINICAL DECISION ON
PROGNOSIS
 
 
CLINICAL SCENARIO
 
Your brother consulted you because of what happened to his wife lately. She just 
had a miscarriage (after 6 months of pregnancy) last week and she had not taken 
her meals lately. They are already in their five years of marriage and they don’t 
have a child yet. Your brother is asking you if they should take a vacation despite 
his being very busy at work and he is trying to save for the house mortgage. 
What will you advice him? 
 
If you advice him to take a vacation, they might loss their house or even his job. 
If you advice to continue working and let time heal his wife’s grief, the chance of 
a psychological problem worsening is great. 
 
Since you have attended a workshop on evidence based family practice, you 
formulated the question “What is the chance that my brother’s wife will go 
further into grief after the miscarriage considering that they still don’t have any 
child yet?” and went to the library. 
 
 
SEARCH
 
You typed the combination of the terms, pregnancy loss and prognosis and grief 
and you were able to get across the article of Janssen et al. study entitled “A 
prospective study of risk factors predicting grief intensity following pregnancy 
loss” publish in the Archive of General Psychiatry last January 1997. 
 
Now you proceed to see if the information in this study can be used to answer 
your brother’s question. 
 
 
WHAT IS PROGNOSIS
 
Prognosis refers to the development of possible “outcome” of disease i.e. death 
in patient with cancer. Prognostic factors are characteristics of a particular 
patient can be used to predict that patient's eventual outcome i.e. patients 
advanced TNM cancer stage may have more death than those with less advance 

Page 59
Evidence-based Medicine Self-Instructional Manual

TNM cancer stage. Prognostic factors need not necessarily cause the outcomes 
but just predict their development. Thus prognosis is a prediction of the 
probable outcome of a disease based on a individual's condition and the usual 
course of the disease as seen in similar situations. Risk factors on the other hand 
are patient characteristics associated with the development of the disease rather 
than the outcome of the disease.  
 
The study designs for prognostic and risk factors are cohort study and case‐
control study. Cross‐sectional studies do not give valid conclusions about 
prognostic or risk factors because temporal relationship between factors and 
outcome is not established.  
 
A cohort study follows one or more groups (cohorts) of individuals who have not 
yet suffered an adverse event and monitor the number of outcome events over 
time. An ideal cohort study consists of well defined sample of subjects 
representative of the population of interest, and uses objective outcome criteria.  
 
Investigators can also collect "cases" of individuals who have already suffered 
the outcome event (death due to cancer) and compare them to "controls" who 
have not (cancer patients who are alive). In these "case‐control" studies the 
investigators count the number of individuals in each group with a particular 
prognostic factor (advance or less advance TNM cancer stage).  
 
To be valid studies on prognosis, these observational studies must be conducted 
and reported with information that address this appraisal guide (Von Elm, 2007) 
 
 
CRITICAL APPRAISAL
 
RELEVANCE 
 
• Is the objective of the article on prognosis similar to your clinical 
dilemma? 
 
Your formulated clinical question must be addressed by the objective of the 
study. The PIO can still be applied in this type of article. The objective of the 
study must clearly state that it is determining the prognosis (outcome) of some 
patients with the prognostic factor (intervention/exposure) among patients with 
the disease being studied (population). 
 
 
 
 

Page 60
Evidence-based Medicine Self-Instructional Manual

VALIDITY GUIDES 
 
• Was there a representative sample of patients without the outcome at 
the start of observation? 
 
The authors must specify how they defined or diagnosed the patients included in 
the study. The authors should also specify at what stage of the disease they 
started observing their patients. If these were not done bias can distort the 
result of the study. If the study included patients who are more severe, the 
prognosis will naturally be poor and if they include patients who are mild, the 
prognosis will be good. However if you mix these patients in one study without 
subgroup analysis, the results will be mixed and biased. 
 
If this is not explicit in the study, you can look at the inclusion criteria or examine 
the setting where the study was done. The inclusion criteria may give the basis 
for the diagnosis, and the setting may give the stage of the disease i.e. 
outpatient setting may have included patients in the earlier stage and hospital 
setting may be patients in the late stage. 
 
In the Janssen et al study, 221 women were recruited through a magazine add. 
They had a stable marriage and reported a recent pregnancy loss. So these 
women may have been recruited at a relatively similar stage. 
 
• Was follow‐up sufficiently long and complete?  
 
Just like in the paper about harm, the length of follow‐up must be sufficiently 
long enough to detect the outcome. If follow‐up is short, the chance of arriving 
at a good prognosis is high because few will develop the outcome resulting to 
false hopes for the patient. If it is too long, the prognosis will be poor because 
everybody will eventually die in the long term. Measuring prognosis over a given 
period is usually acceptable i.e. 5 year survival for chronic diseases, 6‐24 months 
survival for cancer, 30 days survival after ICU admission etc. 
 
The number of lost to follow‐up will also lead to bias results especially when the 
outcome is unknown. If patients were lost to follow‐up because they felt bad 
about the outcome, prognosis will be better if they are excluded in the analysis. 
If they were lost to follow‐up because they felt better and the investigators 
assumed the worse scenario, the prognosis will look bad. 
 
In the Janssen et al study, follow‐up was 94%, a relatively high rate. 
 
 
 
 

Page 61
Evidence-based Medicine Self-Instructional Manual

• Were the criteria for determining the prognostic factor and outcome 
explicit and credible? 
 
The criteria for determining the outcome in study about prognosis are usually 
straightforward I.e. mortality. Mortality or survival can be taken from death 
certificates and other medical records; morbidity can be taken from 
hospitalization records, etc. In some cases outcomes are recurrence of disease or 
disease progression in which case this must be clearly defined. Definitions can be 
taken from the NLM MESH definitions or ICD 10 classification of the WHO. 
 
The Janssen et al study, measured the following prognostic factors; a) Symptom 
Checklist‐90, b) Dutch Personality Inventory, and c) information about quality of 
partnership, education, religion, social support etc. using existing records and 
surveys. The outcomes were measured using the Perinatal Grief Scale 
immediately after pregnancy loss and at 6, 12 and 18 months. 
 
• Was there adjustment for other prognostic factors? 
 
Age and sex are factors that can affect prognosis but something we cannot do 
about. Thus many prognostic studies look at the effect of other modifiable 
prognostic factors by adjusting for age and sex. Doing subgroup analysis does 
this. In subgroup analysis, the results of the study are presented for each 
subgroup i.e. age and sex. Thus the prognosis of different TNM stage for cancer 
can be presented in different age group or in different sex. 
 
Another method is through multivariate analysis or regression model approach. 
In this approach the basic variables included in the model are the prognostic 
factor and the outcome. To adjust for age and sex, these variables are included 
into the model. This is usually described in the analysis section of the 
methodology. 
 
In the Janssen et al study, multivariate analysis was done. 
 
OVERALL, IS THE STUDY VALID? 
 
Since all the validity questions were fulfilled, the study can be considered to be 
valid. 
 
 
 
 
 
 
 

Page 62
Evidence-based Medicine Self-Instructional Manual

WHAT ARE THE RESULTS? 
 
• How large is the likelihood of outcome to occur in those with the 
prognostic factor in a specified period of time? Was it statistically 
significant? 
 
The relative risk is the incidence of the outcome in the group with the prognostic 
factor divided by the incidence of the outcome in the group without the 
prognostic factor. If the outcome being measured is death and the relative risk is 
more than 1, then the factor results into poor prognosis and if less than 1 the 
factor causes good prognosis. Relative risk is usually computed when the design 
is a cohort study. In a case control study, the odds ratio is computed. The odds 
ratio approximates the relative risk, especially when the disease is rare.  
 
If the relative risk or odds risk is 1.11, it means the chance of developing the 
outcome is just slightly higher if the patient has the prognostic factor. If the 
relative risk or odds risk is 1.99 it means the chance of developing the outcome is 
almost two times (2x) and if the relative risk or odds risk is 9.89 the chance is 
almost ten times (10x). The question of “how large is the chance” involves 
preferential judgment from the patient and the physician. 
 
To be statistically significant, the upper and lower values of 95% confidence 
interval should be greater than 1 to say that the factor gives a bad prognosis 
when the outcome is death. If one value of the 95% confidence interval is less 
than 1 and the other is more than 1, then the effect of the prognostic factor is 
uncertain. In some cases, studies report the p value for statistical significance i.e 
p <0.05 as significant. 
 
The Janssen et al study reported that grief intensity was higher for a) women 
who had been pregnant longer, b) pre‐loss neurotic personalities, c) pre‐loss 
psychiatric symptoms, and d) did not have any living children. All these factors 
were significant at p <0.05. 
 
CAN THE RESULTS HELP ME IN CARING FOR MY PATIENTS? 
 
• Are the study patients similar to my own? 
 
Just like in an article about harm, for the prognostic factor to be extrapolated to 
your patient you have to be assured that the characteristics of your patient is 
similar to the study’s inclusion criteria. The setting may also be important. 
Patients being observed in setting where the facilities are advanced and 
complete may have better prognosis than among patients who are being 
observed in resource poor setting even though they have the same prognostic 
factor. 

Page 63
Evidence-based Medicine Self-Instructional Manual

 
The subjects in the Janssen et al study were women who reported recent 
pregnancy loss with stable marital relationship. The subjects were similar to the 
sister‐in‐law’s case. 
 
• Can I use the results to decide on the intervention or reassure my 
patient? 
 
Prognostic data should be used in decisions about therapy. Knowing the 
probability of the outcome based on the prognostic factors present in the 
patient should influence the decision to give or withhold treatment. For example 
surgical excision for cancer with the hope of improving survival should be 
withheld in favor of palliation treatment if the prognosis of the patient is very 
poor.  
 
Prognosis data may also be helpful in reassuring anxious patients about their 
outcome. For example some patients with dyspepsia may become too worried 
about the chronic epigastric symptom and can be reassured and counseled 
about the low prognosis of dyspepsia leading to bleeding ulcer or cancer. 
 
 
RESOLUTION OF THE PROBLEM IN THE SCENARIO 
 
Based on the Janssen et al study, you would rather advice your brother to take a 
vacation, because his wife’s grief may even intensify based on the results of 
Janssen et al study. 
 
 
REFERENCES
 
Andreas Laupacis, George Wells, W. Scott Richardson, Peter Tugwell for the Evidence‐Based 
Medicine Working Group. How to Use an Article about Prognosis. JAMA. 1994;272(3):234‐237. 
 
National Library of Medicine. Medical Subject Headings. www.ncbi.nlm.nih.gov/sites/entrez 
(May 27, 2008). 
 
Von Elm E, Altman D, Egger M, Pocock S, Gøtzsche P, Vandenbroucke J. STROBE Initiative. 
Strengthening the reporting of observational studies in epidemiology (STROBE) statement: 
guidelines for reporting observational studies. BMJ 2007; 335: 806‐808. 
 
 
 
 
 
 

Page 64
Evidence-based Medicine Self-Instructional Manual

WORKSHOP
CRITICAL APPRAISAL OF AN
ARTICLE ABOUT PROGNOSIS
(SESSION BRIEFING)
 
 
OBJECTIVES
 
The purpose of the workshop is to introduce to the participants the concept of 
medical decision making using an article about prognosis. Another objective is to 
introduce concepts of critical appraisal of an article regarding prognosis: 
 
• Validity 
• Representative sample 
• Interpretation of the results 
• Applicability of the results 
 
 
INSTRUCTIONS
 
Divide the participants into groups of six to ten persons per group. Assign a case 
scenario to each group and formulate an answerable problem from the scenario. 
Establish initial group consensus on how to proceed with the scenario. 
 
Ask the group to read the article retrieved to answer the problem in the 
scenario. Focus on the abstract, methods and results section. Again, establish a 
group consensus on how to proceed with the scenario. Note any change in 
decisions. 
 
Critically appraise the article using the appraisal sheet provided. Answer validity 
questions, analyze the results and determine the applicability of the results. 
Establish another group consensus and note any change in decision. 
 
Process the exercise. Focus on barriers and solution to the application of the 
exercise in usual clinic practice. 
 
 
TIME ALLOTTED
 
The time allotted for this workshop is two hours. The recommended break‐up is: 
 

Page 65
Evidence-based Medicine Self-Instructional Manual

• 15 minutes to analyze the scenario and develop consensus 
• 15 minutes to read the article 
• 45 minutes to appraise the validity 
• 15 minutes to analyze results 
• 20 minutes to establish applicability 
• 10 minutes to summarize the process 
 
 
DESIRED OUTCOME
 
The participants should make a clinical decision on the scenario based on the 
critical appraisal of the evidence. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Page 66
Evidence-based Medicine Self-Instructional Manual

Appraisal Sheet 
CLINICAL DECISION ON PROGNOSIS 
 
   
CLINICAL SCENARIO OR     
QUESTION   
 
     
SEARCH                 
 
CRITICAL APPRAISAL   
   
RELEVANCE  Is the objective of the article on harm similar to your 
clinical dilemma? 
 
 
 
   
PRIMARY VALIDITY  Was there a representative sample of patients without the 
GUIDES  outcome at the start of observation? 
 
 
 
 
 
 
Was follow‐up sufficiently long and complete?  
 
 
 
 
 
   
  Were the criteria for determining the prognostic factor and 
outcome explicit and credible? 
 
 
 
 
Was there adjustment for other prognostic factors? 
 
   
 
 

Page 67
Evidence-based Medicine Self-Instructional Manual

OVERALL, IS THE STUDY   
VALID?   
 
   
WHAT ARE THE  How large is the chance of the outcome to occur in a 
RESULTS?  specified period of time? How precise were they? 
 
 
 
 
 
 
 
 
   
CAN THE RESULTS HELP  Are the study patients similar to my own? 
ME IN CARING FOR MY   
PATIENTS?   
 
 
 
Can I use the results to decide on the intervention or 
reassure my patient? 
 
 
 
 
 
 
 
   
RESOLUTION OF THE   
PROBLEM IN THE 
SCENARIO 
 
 
 
 
 
 
 
 
 
 

Page 68
Evidence-based Medicine Self-Instructional Manual

READING ASSIGNMENT
CRITICAL APPRAISAL OF AN
ARTICLE ABOUT HEALTH
ECONOMIC ANALYIS
(SESSION BRIEFING)
 
 
CHECKLIST
 
Have you read the previous topic on prognosis? 
 Yes          No 
 
Have you undergone the workshop on prognosis with your group? 
 Yes          No 
 
Did you enjoy the workshop? 
 Yes          No 
 
Please state the reasons below and share it to the group before starting the next 
workshop.  
 
 
OBJECTIVES
 
The purpose of the reading assignment is to introduce to the participants the 
concept of medical decision making using an article about a health economic 
analysis.  
 
At the end of the reading session, you should be able to answer the user guides 
questions for the workshop. 
 
 
INSTRUCTIONS
 
Read the reading assignment for an article about a health economic analysis. 
Focus on the critical appraisal questions, why they are asked and how to get the 
answers from the paper.  
 
After reading the paper you can proceed to conduct the group workshop. 
 

Page 69
Evidence-based Medicine Self-Instructional Manual

READINGS
CLINICAL DECISION USING AN
ARTICLE ON HEALTH
ECONOMIC ANALYSIS
 
 
CLINICAL SCENARIO
 
Your father is the mayor of one of the town in the Visayas. His budget for health 
care is minimal but he wants to start a program for pregnant women. Although 
most of the pregnancies in the area were low risk, most patients still go to the 
nearest obstetrician who is in the city 50 kilometers away. He is planning to set‐
up a maternity hospital with a trained obstetrician in your hometown. He is now 
asking you for a recommendation.  
 
What is your recommendation? 
 
 
SEARCH
 
Having attended an EBM workshop you thought that the best way to convince 
your father is to show cost‐effectiveness of handling low risk deliveries with 
different types of trained physicians. You were able to get an article by Ratcliffe 
and Tucker entitled “The costs of alternative types of routine antenatal care for 
low‐risk women: obstetrician‐led shared care vs care by general practitioners 
and midwives” published in the Journal of Health Services and Policy, 1996. 
 
Appraising the article and conveying the information to your father seemed to 
be a good strategy, so you proceeded in appraising the article. 
 
 
HEALTH ECONOMIC ANALYSIS
 
Health economic analysis is a formal, quantitative methods used to compare 
alternative strategies with respect to their cost and their expected outcomes 
(Eisenberg, 1989). Its purpose is to inform decisions on resource allocation. It can 
potentially inform decisions in institutions like hospitals and in regional or 
national health policy (Russell, 1996). In this design the cost of a particular 
intervention is estimated. Estimation include direct and indirect costs. There are 
three types of economic analysis depending on the type of outcome. If the 

Page 70
Evidence-based Medicine Self-Instructional Manual

outcome being considered is effectiveness of treatment, it is called cost‐
effectiveness analysis. If the outcome is savings in terms of monetary units it is 
called cost‐benefit analysis. If the outcomes are equal and the cost is the only 
one being compared it is called cost minimization. 
 
 
CRITICAL APPRAISAL
 
RELEVANCE 
 
• Is the objective of the article on economic analysis similar to your 
clinical dilemma? 
 
Your scenario must be addressed by the objective of the study. The perspective 
or “point of view” in economic analysis usually refers to the one who will pay for 
the intervention. Often, the point of view of the economic analysis is stated in 
the objective and this is important to determine its relevance to your case 
scenario. If you are using an economic analysis for policy decision like the 
government pay for this kind of drug or facility, then the point of view must be 
the societal point of view. If your scenario is to assist a patient to make a 
decision on an intervention in a “pay‐for‐service” setting, then the perspective 
must be from the patient or “payer” perspective. Health insurance perspective is 
also a payer perspective. 
 
VALIDITY GUIDES 
 
• Did the analysis provide a full economic comparison of health care 
strategies? 
   
Physicians usually choose between two alternatives. The range of alternative 
strategies examined must include at least the currently accepted standard and 
the new alternative (O’Brien, 1997). Another alternative is the “do nothing” 
alternative but may not be realistic in some cases because of ethical issues.  
 
When we compare the cost of giving each of the two alternatives, this is cost 
analysis. When we use this to make a decision, we only give the alternative with 
the lowest cost that may not be necessarily effective. When we use comparison 
of effectiveness such as a randomized controlled trial in making a decision, we 
give an effective alternative that the patient may not be able to afford. Thus it 
makes sense to consider cost and effectiveness when making clinical decisions. 
 
 A full economic analysis compares not only the cost of the two alternatives but 
also integrate information about efficacy of the alternatives. Thus cost and 

Page 71
Evidence-based Medicine Self-Instructional Manual

outcomes should both be analyzed for each alternative strategies being 
compared. This can be achieved if the study design is a cost‐benefit or a cost 
effectiveness approach. 
 
The Ratcliffe and Tucker study was a full economic valuation of the cost of tests, 
investigations, personnel, cost incurred by the patients of pregnancies delivered 
by obstetricians vs family physicians. The study included patients enrolled in a 
randomized controlled trial to answer the effectiveness outcome. 
 
• Were the costs and outcomes properly measured and valued?  
 
Cost pertains to resources used and this must be differentiated from charges or 
prices of commodities. The point of view of costing refers to the one who will 
pay for the cost and this may differ. For example cost to government hospital or 
funder may be different from the point of view of the patient. What may be cost 
saving for the funder may actually be an increased cost for the patient. The ideal 
point of view is from the society’s view, but this is difficult to measure. Thus 
proper measurement of cost may differ from the health care system. In a system 
where the payment is a fee‐for‐service set‐up, an economic analysis on the point 
of view of the paying patient may be a good basis for making decision. In a 
system where health care is being paid for by the government, an economic 
analysis from the point of view of society may be a good basis. Thus in measuring 
cost, the point of view of the analysis must be established (O’Brien, 1997).  
 
Outcomes in health care must be an outcome that is of value to the patient and 
society. It should be something that can be appreciated by the patient. For 
example in making a decision about the treatment for hypertension, outcomes 
like decrease in incidence of mortality or stroke, decrease in hospitalization or 
myocardial infarction instead of just the lowering of blood pressure should be 
the outcome to be considered. In addition these outcomes must be measured in 
the best possible designs i.e. randomized controlled trials for treatment, 
controlled comparison for complex intervention etc. Systematic reviews or meta‐
analysis of these interventions are better methods for establishing outcomes.  
 
Lastly, the cost and outcome must be expressed as a ratio i.e. cost per outcome 
(cost per life‐year gained, or cost per death avoided etc.). This expression of 
result will give the most relevant information for decision making. 
 
In the Ratcliffe and Tucker study costs were extracted from clinical data that 
came from a randomized controlled trial. They included cost per patient, staffing 
cost, non‐health services cost and mean societal cost. Cost data was available in 
about 94% of subjects included in the trial. 
 
 

Page 72
Evidence-based Medicine Self-Instructional Manual

• Was appropriate allowance made for uncertainties in the analysis? 
 
Economic analysis usually depends on analysis of secondary data and since not 
all data are available for each alternative, some assumptions need to be made. 
There are uncertainties to these assumptions. A good economic analysis is one 
that recognizes these uncertainties and look at how it affects their findings. This 
method is called sensitivity analysis (O’Brien, 1997).  
 
In sensitivity analysis the estimates for key variables in costs are changed in 
order to assess their impact they on the results. The changes can be based on 
variation in price index, opportunity costs, geographical price differences etc.  In 
terms of outcomes, the variation can be from confidence intervals or from 
lowest and highest effect noted from the studies that were reviewed. 
 
OVERALL, IS THE STUDY VALID? 
 
Since the study was a full economic comparison and wide perspective of cost 
was considered you decided that the study was valid. 
 
WHAT ARE THE RESULTS? 
 
• What were the costs and outcomes of each strategy? 
 
Economic analysis papers should have tables that report the costs of resources 
used in the alternative intervention such as drugs, personnel services, facilities, 
supplies etc. It should also contain tables about the outcome of each alternative. 
Lastly, this is expressed as a cost‐effectiveness ratio or cost‐benefit.  
 
The table below is a good guide to help decide the alternative to choose. “C” will 
be the best choice since it is more effective and less cost. “B” is not a good 
choice because it is less effective but more expensive. “A” is more effective but 
more expensive as well. 
 
  Effectiveness 
High  Low 
Cost  High  A  B 
Low  C  D 
 
 
Sometimes, an alternative may be more effective but also more expensive. To 
make a decision in this scenario, an incremental analysis should be done. 
Incremental cost is the amount we pay for the added effectiveness of the 

Page 73
Evidence-based Medicine Self-Instructional Manual

alternative. Usually the availability of resources and personal judgment may be 
needed to decide whether the incremental cost is worth it.  
 
Ratcliffe and Tucker showed that the total societal mean cost for GP or midwife 
care was lower by P 2,178 and was statistically significant.  
 
• How much does allowance for uncertainty change the result? 
 
Looking at how uncertainties affect the results is called sensitivity analysis. A 
direct approach for doing this is by computing for the cost effectiveness ratio 
using the lower and upper limit of the 95% confidence interval of the 
effectiveness outcome. If the cost‐effectiveness values are reversed with 
sensitivity analysis then the results are considered to be soft and its reliability is 
less.  
 
CAN THE RESULTS HELP ME IN CARING FOR MY PATIENTS? 
 
• Could my patients expect similar outcomes? 
   
The inclusion criteria of the cited clinical trial or other studies reviewed to 
determine the outcome and the setting from which the trial was done can be 
duplicated in your setting will play an important factor. If the patients in the 
study are similar to your patient and setting for which the intervention was given 
can be duplicated, then you can expect the same outcome (O’Brien, 1997).  
 
• Could my patients expect similar costs? 
 
The health care system may be different from the setting where the economic 
analysis was done. This difference may lead to difference in costing once applied 
for decision making in your setting. Cost data may be different for two reasons: 
1) clinical practice vary in resource consumption associated with the treatment 
and 2) prices for resources differ from those used in the study (O’Brien, 1997). 
This can only be answered if the analysis presented the detailed cost so the 
reader can decide whether the costing in the study can also be applied in his/her 
own setting. 
 
Countries may differ with respect to the value they place on health benefits. If 
$50,000 per life‐year is an acceptable cost‐effectiveness threshold for the US it 
may not be affordable in the Philippines. Countries vary in their willingness to 
pay for health care (O’Brien, 1997).  
 
 
 
 

Page 74
Evidence-based Medicine Self-Instructional Manual

RESOLUTION OF THE PROBLEM IN THE SCENARIO 
 
Since most pregnancies in your hometown were low risk, you showed to your 
father that the town will save more money if they hire family physicians or 
midwives rather than an obstetrician. You called up your father one month later 
and the town council opted to hire more midwives. 
 
 
REFERENCES
 
Espallardo, NL. Research Protocol Development for Resident Physicians. Family Medicine 
Research Group, Inc. Manila, 2000. 
 
Eisenberg JM. Clinical economics. A guide to the economic analysis of clinical practices. JAMA, 
1989; 262:2879‐86.  
 
O'Brien B, Heyland D, Richardson WS, Levine M, Drummond M, for the Evidence‐Based Medicine 
Working Group. How to use an Article on Economic Analysis of Clinical Practice: Validity Guides. 
JAMA, 1997; 277(19):1552‐1557. 
 
O'Brien B, Heyland D, Richardson WS, Levine M, Drummond M, for the Evidence‐Based Medicine 
Working Group. How to use an Article on Economic Analysis of Clinical Practice: Results and 
Applicability. JAMA, 1997; 277(22):1802‐1806.  
 
Russell LB, Gold MR, Siegel JE, Daniels N, Weinstein MC. The role of the cost‐effectiveness 
analysis in health and medicine. JAMA, 1996; 276(1)‐1172‐7. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Page 75
Evidence-based Medicine Self-Instructional Manual

WORKSHOP
CRITICAL APPRAISAL OF AN
ARTICLE ON HEALTH
ECONOMIC ANALYSIS
(SESSION BRIEFING)
 
 
OBJECTIVES
 
The purpose of the workshop is to introduce to the participants the concept of 
medical decision making about an economic analysis. Another objective is to 
introduce concepts of critical appraisal of an article regarding an economic 
analysis focusing on the following: 
 
• Validity 
• Costs in health care 
• Interpretation of the results 
• Incremental costs 
• Applicability of the results 
 
 
INSTRUCTIONS
 
Divide the participants into groups of six to ten persons per group. Assign a case 
scenario to each group and formulate an answerable problem from the scenario. 
Establish initial group consensus on how to proceed with the scenario. 
 
Ask the group to read the article retrieved to answer the problem in the 
scenario. Focus on the abstract, methods and results section. Again, establish a 
group consensus on how to proceed with the scenario. Note any change in 
decisions. 
 
Critically appraise the article using the appraisal sheet provided. Answer validity 
questions, analyze the results and determine the applicability of the results. 
Establish another group consensus and note any change in decision. 
 
Process the exercise. Focus on barriers and solution to the application of the 
exercise in usual clinic practice. 
 
 

Page 76
Evidence-based Medicine Self-Instructional Manual

TIME ALLOTTED
 
The time allotted for this workshop is two hours. The recommended break‐up is: 
 
• 15 minutes to analyze the scenario and develop consensus 
• 15 minutes to read the article 
• 45 minutes to appraise the validity 
• 15 minutes to analyze results 
• 20 minutes to establish applicability 
• 10 minutes to summarize the process 
 
 
DESIRED OUTCOME
 
The participants should make a clinical decision on the scenario based on the 
critical appraisal of the evidence. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Page 77
Evidence-based Medicine Self-Instructional Manual

Appraisal Sheet 
CLINICAL DECISION ON ECONOMIC ANALYSIS 
 
   
CLINICAL SCENARIO OR     
QUESTION   
 
     
SEARCH                 
 
CRITICAL APPRAISAL   
   
RELEVANCE  Is the objective of the article on economic analysis similar 
to your clinical dilemma? 
 
 
 
   
VALIDITY GUIDES  Did the analysis provide a full economic comparison of 
health care strategies? 
   
 
 
 
 
 
Were the costs and outcomes properly measured and 
valued?  
 
 
 
 
 
   
  Was appropriate allowance made for uncertainties in the 
analysis? 
 
   
 
Are estimates of costs and outcomes related to the baseline 
risk in the treatment? 
 
   
 

Page 78
Evidence-based Medicine Self-Instructional Manual

OVERALL, IS THE STUDY   
VALID?   
 
   
WHAT ARE THE  What were the incremental costs and outcomes of each 
RESULTS?  strategy? 
 
 
 
               
 
Do incremental costs and outcomes differ between 
subgroups? 
 
 
 
 
 
How much does allowance for uncertainty change the 
result? 
  
 
 
 
 
   
CAN THE RESULTS HELP  Could my patients expect similar outcomes? 
ME IN CARING FOR MY     
PATIENTS?   
 
 
 
Could my patients expect similar costs? 
 
 
 
 
 
 
RESOLUTION OF THE   
PROBLEM IN THE   
SCENARIO 
 
 

Page 79
Evidence-based Medicine Self-Instructional Manual

READING ASSIGNMENT
CRITICAL APPRAISAL OF AN
ARTICLE ABOUT SYSTEMATIC
REVIEWS OR META-ANALYSIS
(SESSION BRIEFING)
 
 
CHECKLIST
 
Have you read the previous topic on economic analysis? 
 Yes          No 
 
Have you undergone the workshop on economic analysis with your group? 
 Yes          No 
 
Did you enjoy the workshop? 
 Yes          No 
 
Please state the reasons below and share it to the group before starting the next 
workshop.  
 
 
OBJECTIVES
 
The purpose of the reading assignment is to introduce to the participants the 
concept of medical decision making using an article about systematic reviews or 
meta‐analysis.  
 
At the end of the reading session, you should be able to answer the user guides 
questions for the workshop. 
 
 
INSTRUCTIONS
 
Read the assignment for an article about systematic reviews or meta‐analysis. 
Focus on the critical appraisal questions, why they are asked and how to get the 
answers from the paper.  
 
After reading the paper you can proceed to conduct the group workshop. 
 

Page 80
Evidence-based Medicine Self-Instructional Manual

READING ASSIGNMENT
CLINICAL DECISION ON
SYSTEMATIC REVIEW OR
META-ANALYSIS
 
 
CLINICAL SCENARIO
 
Your grandmother had a history of fall a week ago after taking a bath. She was 
brought to the hospital for treatment of minor bruises in her knees. The x‐rays 
were normal. Your mother asked you if she needs a walker or cane to prevent 
falls and subsequent injury.  
 
 
SEARCH
 
You ask a colleague from Rehabilitation Medicine and she gave you an article 
from the NHS Center for Reviews and Dissemination she got from the internet 
entitled “Preventing falls and subsequent injury in older people”. 
 
 
REVIEWS, SYSTEMATIC REVIEWS AND META-
ANALYSIS
 
A review is secondary study design that integrates findings of two or more 
studies that discuss similar topic usually defined by PIO. There may be some bias 
when the reviewer subjectively decides which studies to include or exclude in 
the review. A systematic review is similar to review but has a way to 
systematically search the literature searching and has systematic rules in 
combining the studies to be reviewed. The results of systematic reviews are 
more often objective than a review. Meta‐analysis is like a systematic review but 
applies some statistical analysis to the results. 
 
A meta‐analysis is a procedure that integrates and combine the results of two or 
more primary studies that are similar in the population enrolled the intervention 
used and the outcome measured. The pooled result is then subjected to a 
statistical analysis. A well conducted meta‐analysis allows a more objective 
appraisal of the existing evidence about a problem than a traditional review or 
systematic review. Meta‐analysis may also be biased owing to the inclusion or 
exclusion of some irrelevant or relevant studies respectively (Espallardo, 2000).  

Page 81
Evidence-based Medicine Self-Instructional Manual

 
 
Sample of meta‐analysis result 
 
 
CRITICAL APPRAISAL
 
RELEVANCE 
 
• Is the objective of the article on meta‐analysis similar to your clinical 
dilemma? 
 
Your formulated clinical question must be addressed by the objective of the 
study. The objective of an appropriately done meta‐analysis is often a focused 
clinical objective with PIO and the method being clearly defined. This is often 
used for the systematic literature search and basis for inclusion or exclusion. 
 
VALIDITY GUIDES 
 
• Did the review address a focused clinical problem? 
 
Systematic reviews of the medical literature try to summarize publications 
related to a similar topic. Because several articles are combined together, 
sometimes the purpose of the review is not clear or very broad. It therefore 
becomes difficult to determine what the review is trying to achieve.  
 

Page 82
Evidence-based Medicine Self-Instructional Manual

In order to know what the objective of the review is, the clinical problem must 
be focused. There must be a clear description of the patient and its relation with 
an exposure or an outcome.  
 
The NHS study specifically stated that they tried to identify strategies that 
prevent falls and subsequent injury in older people. The focused objective 
seemed to apply to your problem. 
 
• Were the criteria for searching and selecting articles for inclusion and 
exclusion explicit and credible?  
 
In conducting systematic reviews, a systematic search and appraisal of the 
literature should be done in order to: 
• ensure that no relevant articles were missed 
• studies were included because they are good studies and not because 
they agree with the authors opinion 
• studies were excluded because they are bad studies and not because 
they disagree with the author’s opinion. 
 
Thus paper should describe the method of searching for the medical literature. 
Statements like “an electronic search of published articles in the MEDLINE using 
the terms . . . from 1966 to 2000 was done” must be found somewhere in the 
methodology section. This assures the readers that the findings of the study 
were based on a wide range of literature source and represent the most current 
and complete information about the clinical problem.  
 
The paper should also describe how they include or exclude retrieved articles. 
The paper must contain statements like “all retrieved abstracts were reviewed 
by three independent reviewers and articles that were randomized controlled 
trial on . . . using the intervention . . .” somewhere in the method section. This 
assures the readers that the articles used in the study were objectively chosen 
and not because they agree with the authors opinion. 
 
The NHS study identified trials published in computerized databases like Social 
Science Citation Index, PSYCHLIT, EMBASE, RCN database, AMED and UNCOVER. 
The citations also identified reviews and peer contribution from reviewers and 
other experts in the field. 
 
• Was the validity of included studies appraised and the appraisal 
reproducible? 
 
Even if all included studies are randomized controlled trials, there may be some 
small differences among different trials that might affect the results of the study. 
Thus a standard appraisal of each article must be done. Peer review may not be 

Page 83
Evidence-based Medicine Self-Instructional Manual

a reliable method of appraisal. Differences in peer perception and interest may 
lead to differences in the result of the appraisal. 
 
While there are no agreed standards to evaluate validity, the review must have 
at least developed a checklist of criteria that focused on the methodology of the 
study being appraised.  
 
The NHS study included only randomized controlled trials that evaluated 
strategies to prevent falls.  
 
OVERALL, IS THE STUDY VALID? 
 
Overall the study is valid. 
 
WHAT ARE THE RESULTS? 
 
• What are the overall results of the systematic review?  
 
When looking at the results of a systematic review or meta‐analysis, you should 
look for clinically relevant presentation like lower mortality rates in one group 
compared to the other, or the difference in quality of life scores between the 
two groups. Sometimes subgroup analysis i.e. patients with high risk and 
patients with low risk, to see the different effect of an exposure may also be 
helpful.  
 
In a meta‐analysis, the confidence interval of the overall results can also be 
computed and this can provide information about the precision of the results.  
 
Thirty‐six randomized controlled trials were included in the meta‐analysis. The 
results showed that 10‐24 weeks of exercise including balance training showed 
an effective risk reduction by as much as 37% with an adjusted fall incidence 
ratio of 0.90 and with a 95% CI of 0.81 to 0.99. 
 
CAN THE RESULTS HELP ME IN CARING FOR MY PATIENTS? 
 
• Are the study patients similar to my own? 
 
In a systematic review, the patient’s characteristics are varied because they 
came from different studies. Application to patients therefore becomes wider. 
When variation in patient inclusion may influence the effect, subgroup analysis 
between different patient characteristics may also help decide what kind of 
patient will benefit from the intervention or will be affected by the exposure.  
 

Page 84
Evidence-based Medicine Self-Instructional Manual

This information can be seen in the methodology section where the researchers 
describe the type of patients in the literature search and inclusion criteria of the 
studies. 
 
The NHS meta‐analysis included only studies done on elderly. 
 
• Are the results of the review relevant to my patient? 
 
When the clinical question of the review is focused, the answer to this question 
becomes evident. You only need to focus on the outcome measured and decide 
whether this is relevant to your patient.  
 
When the outcome differs between studies, they are combined and this is 
reported in systematic reviews or meta‐analysis as effect size. This is a difficult 
situation because the outcomes are combined and you cannot easily decide 
whether the outcome is relevant or not. In this case you can look at the results of 
individual studies and choose studies that give relevant outcomes and use them 
to make a decision. 
 
RESOLUTION OF THE PROBLEM IN THE SCENARIO 
 
Based on the review, balance training rather than walker devices will help 
prevent further fall and subsequent injury. 
 
 
REFERENCES
 
Espallardo, NL. Research Protocol Development for Resident Physicians. Family Medicine 
Research Group, Inc. Manila, 2000. 
 
Oxman A, Cook D, Guyatt G, for the Evidence Based Medicine Working Group. How to Use an 
Overview. JAMA, 1994;272(17):1367‐71.  
 
 
 
 
 
 
 
 
 
 
 
 
 

Page 85
Evidence-based Medicine Self-Instructional Manual

WORKSHOP
CRITICAL APPRAISAL OF AN
ARTICLE ABOUT A SYSTEMATIC
REVIEW OR META-ANALYSIS
(SESSION BRIEFING)
 
 
OBJECTIVES
 
The purpose of the workshop is to introduce to the participants the concept of 
medical decision making using an article about a systematic review or meta‐
analysis. Another objective is to introduce concepts of critical appraisal of an 
article regarding differential diagnosis focusing on the following: 
 
• Validity 
• Review, systematic review or overview, meta‐analysis 
• Interpretation of the results 
• Applicability of the results 
 
 
INSTRUCTIONS
 
Divide the participants into groups of six to ten persons per group. Assign a case 
scenario to each group and formulate an answerable problem from the scenario. 
Establish initial group consensus on how to proceed with the scenario. 
 
Ask the group to read the article retrieved to answer the problem in the 
scenario. Focus on the abstract, methods and results section. Again, establish a 
group consensus on how to proceed with the scenario. Note any change in 
decisions. 
 
Critically appraise the article using the appraisal sheet provided. Answer validity 
questions, analyze the results and determine the applicability of the results. 
Establish another group consensus and note any change in decision. 
 
Process the exercise. Focus on barriers and solution to the application of the 
exercise in usual clinic practice. 
 
 
 

Page 86
Evidence-based Medicine Self-Instructional Manual

TIME ALLOTTED
 
The time allotted for this workshop is two hours. The recommended break‐up is: 
 
• 15 minutes to analyze the scenario and develop consensus 
• 15 minutes to read the article 
• 45 minutes to appraise the validity 
• 15 minutes to analyze results 
• 20 minutes to establish applicability 
• 10 minutes to summarize the process 
 
 
DESIRED OUTCOME
 
The participants should make a clinical decision on the scenario based on the 
critical appraisal of the evidence. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Page 87
Evidence-based Medicine Self-Instructional Manual

Appraisal Sheet 
CLINICAL DECISION ON SYSTEMATIC REVIEW OR META‐ANALYSIS 
 
   
CLINICAL SCENARIO OR     
QUESTION   
 
     
SEARCH                 
 
CRITICAL APPRAISAL   
   
RELEVANCE  Is the objective of the article on harm similar to your 
clinical dilemma? 
 
 
 
   
PRIMARY VALIDITY  Did the review address a focused clinical problem? 
GUIDES   
 
 
 
 
 
 
Were the criteria for searching and selecting articles for 
inclusion and exclusion explicit and credible?  
 
 
 
 
 
 
 
Was the validity of included studies appraised and the 
appraisal reproducible? 
 
 
 
 
 
 
 

Page 88
Evidence-based Medicine Self-Instructional Manual

   
OVERALL, IS THE STUDY   
VALID? 
 
   
WHAT ARE THE  What are the overall results of the systematic review?  
RESULTS?   
 
 
 
 
 
 
 
 
 
   
CAN THE RESULTS HELP  Are the study patients similar to my own? 
ME IN CARING FOR MY   
PATIENTS?   
 
 
 
 
 
Are the results of the review relevant to my patient? 
 
 
 
 
 
 
 
 
   
RESOLUTION OF THE   
PROBLEM IN THE   
SCENARIO   
 
 
 
 
 
 

Page 89
Evidence-based Medicine Self-Instructional Manual

READING ASSIGNMENT
CRITICAL APPRAISAL OF AN A
CLINICAL PRACTICE GUIDELINE
(SESSION BRIEFING)
 
 
CHECKLIST
 
Have you read the previous topic on systematic reviews? 
 Yes          No 
 
Have you undergone the workshop on systematic reviews with your group? 
 Yes          No 
 
Did you enjoy the workshop? 
 Yes          No 
 
Please state the reasons below and share it to the group before starting the next 
workshop.  
 
 
OBJECTIVES
 
The purpose of the reading assignment is to introduce to the participants the 
concept of medical decision making using an article about a clinical practice 
guideline.  
 
At the end of the reading session, you should be able to answer the user guides 
questions for the workshop. 
 
 
INSTRUCTIONS
 
Read the reading assignment for an article about a clinical practice guideline. 
Focus on the critical appraisal questions, why they are asked and how to get the 
answers from the paper.  
 
 
 
 
 

Page 90
Evidence-based Medicine Self-Instructional Manual

READING ASSIGNMENT
HOW TO USE A CLINICAL
PRACTICE GUIDELINE
 
 
CLINICAL PRACTICE GUIDELINES
 
"Clinical Practice Guidelines are systematically developed statements to assist 
practitioner decisions about appropriate health care for specific clinical 
circumstances."(Field, 1990). Guidelines were developed to: 
 
• Make evidence‐based management explicit. 
• Make clinical decision making more objective and scientific. 
• Assess professional performance. 
• Educate the patients and practitioners about current "best practice." 
 
If guidelines are to improve practice, they need to be developed by the people 
who are actually going to have to apply them. Guidelines, like so much else in 
healthcare today are no longer as immutable as the Laws of the Medes, you will 
have to periodically check that they are working, that they are getting to the 
people who are going to use them and that they are up to date.  
 
In general, good topics for guidelines are those which: 
 
• contribute a high workload 
• poor treatment risks disastrous outcomes 
• change is practical with the resources you have available 
• there is variation in current management (i.e. there is uncertainty about 
the best management strategy), but some hope of reaching a consensus 
• the changes you wish to implement will be acceptable to patients 
• there is good evidence to back up the protocols 
 
It might be useful to consider the criteria used to select illnesses for screening 
programs (how common, how serious, how preventable, how acceptable?  
 
Clinical practice guidelines, which have been defined as "systematically 
developed statements to assist practitioner and patient decisions about 
appropriate health care for specific clinical circumstances," represent an attempt 
to distill a large body of medical knowledge into a convenient, readily useable 
format. Like overviews, they gather, appraise and combine evidence. Guidelines, 
however, go beyond most overviews in attempting to address all the issues 

Page 91
Evidence-based Medicine Self-Instructional Manual

relevant to a clinical decision and all the values that might sway a clinical 
recommendation. Like decision analyses, guidelines refine clinical questions and 
balance trade‐offs. Guidelines differ from decision analyses in relying more on 
qualitative reasoning and in emphasizing a particular clinical context. 
 
 
CASE SCENARIO
 
At the end of a busy day in your clinic, you are glad to find that your last patient 
is a 45/male previously diagnosed to have hypertension.  He claims that his 
highest blood pressure for the past month was 160/90 and his usual blood 
pressure was 130/90 and he has been taking a beta‐blocker for the past year.  He 
was relatively symptom free save for occasional headache and nape pains during 
BP spikes.  Thinking that this was just another run of the mill hypertensive 
patient you were ready to refill his prescription and give your usual advice 
regarding diet and exercise.  As you were about to do just that, your patient 
began asking you the relative benefits of the alternative drugs available in the 
market.  He was also asking if he needed to have an ECG done together with 
blood chemistries, urinalysis and a 2D‐Echo since his friend who consulted 
another physician was advised to do that.  Since he was not overweight he was 
asking if a regular exercise program would add any additional benefit in 
controlling his symptoms.  You gave him the usual advice you knew based on 
your knowledge of pathophysiology and pharmacokinetics.  However, as you 
were finally closing your clinic you decided that you were not satisfied with the 
answers you gave him and decided to do a search for the best available 
evidence. 
 
 
SEARCH FOR CPG
 
You decide that in order to find relevant answers to a variety of clinical 
questions, a clinical practice guideline would be the best article to retrieve.  You 
initially searched ww.guidelines.gov and found numerous guidelines for 
hypertension.  However, you did not have sufficient time to download and print 
the full text version of these guidelines.  You then remembered a copy of the 
Philippine compendium sent to you by mail 2 years ago and taught of the 
Philippine Clinical Practice Guidelines on the Detection and Management of 
Hypertension. 
 
You then go home, sit at your desk and decide to review this article in order to 
be more prepared for your next patient encounter. 
 
 

Page 92
Evidence-based Medicine Self-Instructional Manual

CRITICAL APPRAISAL
 
RELEVANCE 
 
• Is the objective of the article on clinical practice guideline similar to 
your clinical dilemma? 
 
Your formulated clinical question must be addressed by the objective of the 
clinical practice guideline. Usually guideline objectives are broad i.e. answers 
questions about the best diagnostic test, the recommended treatment, the 
expected outcome or prognosis etc.  
 
VALIDITY GUIDES 
 
• Were all important options and outcomes considered? 
  
Guidelines aid us in our decision making skills and we make better judgment calls 
if we know all the alternative options open to us and the relative harm and 
benefits of each choice.  Guideline developers then should present most of the 
reasonable options seen in practice and their corresponding outcomes. 
 
In the case of the Philippine Clinical practice guideline on the detection and 
management of hypertension, several treatment options were presented 
ranging from beta‐blockers, diuretics, ACE inhibitors, calcium channel blockers 
with recommendations of the best alternative for hypertensive patients with co‐
morbid conditions. The guideline however did not include the newer generation 
anti‐hypertensives. 
 
As important as presenting all the options, the corresponding outcomes such as 
morbidity and mortality data, prevention of complications and other measures 
that improve health related quality of life should be reported.  Inasmuch as all 
these will be helpful and clinically relevant to individual patients. 
 
In this hypertension guideline, mortality and morbidity data together with 
prevention of hypertensive complications were the outcomes given emphasis. 
However costs and side effects were not well mentioned.  
  
• Was an explicit and sensible process used to identify, select, and 
combine evidence? 
 
Guideline developers must allow the reader to know how the evidence has been 
tracked, reviewed, appraised and combined in order to allow them to ascertain 
the validity of the gathered evidence.   Developers should specify a focused 

Page 93
Evidence-based Medicine Self-Instructional Manual

question, search the literature for available evidence, critically appraise this 
evidence and summarize the results in an easy to understand material.   
 
The Philippine Clinical Practice Guidelines for Hypertension was not very clear 
regarding how they searched the literature and what database they used.  
Mention of tracking, retrieving and appraising was done in Phase 1 of the 
introduction section, but the complete way on how this was done was not 
mentioned.  However, summary on how the articles were reviewed and graded 
was provided. 
 
• Is the guideline likely to account for important recent developments? 
 
You should look for two important dates: the publication date of the most recent 
evidence considered and the date on which the final recommendations were 
made. Some authorities also identify important studies in progress and new 
information that could change the guideline. Ideally, these considerations may 
be used to qualify guidelines as "temporary" or "provisional," to specify dates for 
expiration or review, or to identify key research priorities. For most guidelines, 
however, you must scan the bibliography to get an impression of how current a 
particular guideline may be.  
 
Once you are confident that the clinical practice guideline addresses your clinical 
question and is based on a rigorous up‐to‐date assessment of the relevant 
evidence, you can review the recommendations to determine how useful they 
will be in your practice. 
 
The reader is advised to check the bibliography section of the guideline and 
check the dates of the most recent articles included.  Ideally, the evidence 
should be within the last 2 years before the guideline was published.  Since 
medical knowledge rapidly transforms, this will ensure that our 
recommendations will not be outdated. Hence, there is a need to revise 
guidelines periodically. 
 
This guideline on hypertension was released in 1995 and the latest evidence 
upon looking at the bibliography was in that same year.  Being at present in the 
year 2000 and with the rapid developments in antihypertensive medications, the 
guideline should be due for review and revision. 
 
• Has the guideline been subjected to peer review and testing? 
 
People may interpret evidence differently and their values as to what important 
options and outcomes are may differ.  As such, a guideline that has been 
subjected to scrutiny by external reviewers and tested in an actual clinical 
practice setting and found acceptable might be easier to use.   

Page 94
Evidence-based Medicine Self-Instructional Manual

 
OVERALL, IS THE GUIDELINE VALID? 
 
Once you are confident that the guideline meets at least 2 out of the 3 
requirements above, you can review the recommendations and its applicability 
to our individual patients.  At present, the Philippine Clinical Practice Guidelines 
on Detection and Management of hypertension although a little bit outdated at 
the present time will do. 
 
WHAT ARE THE RECOMMENDATIONS 
 
• Are practical, clinically important, recommendations made? 
 
To be useful guidelines should give practical, unambiguous advice addressing a 
particular clinical situation.  Recommendations should be simple and specific at 
the same time comprehensive enough to allow the reader a chance to assess the 
benefits and costs of following the particular recommendation. 
 
In this hypertension guideline, recommendations are divided into every aspect of 
any encounter with a hypertensive patient.  The following 6 questions are 
addressed by the guideline: 
 
• How should blood pressure be measured? 
• How should hypertension be diagnosed? 
• How should hypertension be worked up? 
• What advice should hypertensive patients receive regarding lifestyle 
modification? 
• How should hypertension be treated? 
• How can hypertension be prevented among normotensives? 
 
This then allows the clinician to answer aspects regarding diagnosis, laboratory 
work‐ups, treatment options, non‐pharmacologic advice and preventive 
measures. 
 
• How strong are the recommendations? 
 
The "strength," "grade," "confidence," or "force" of a recommendation should 
be informed by multiple considerations:  
 
• the quality of the investigations which provide the evidence for the 
recommendations 

Page 95
Evidence-based Medicine Self-Instructional Manual

• the magnitude and consistency of positive outcomes relative to negative 
outcomes (adverse effects, burdens to the patient and the health care 
system, costs) 
• the relative value placed upon different outcomes.  
 
Thus, grading of the recommendations are based on the methodological 
soundness of the available evidence, the number of positive outcomes in 
relation to negative ones and the consistency of findings across different 
evidences available.  It is not enough to look into the fact that randomized 
controlled trials were used as evidence but also if findings across different trials 
were consistent.  Inconsistent findings are at times the reason why different 
guideline developers have different recommendations regarding certain clinical 
issues. 
 
It is also important to note that different guideline developers use different 
standards for grading their recommendations and that this should explicitly be 
placed in the guideline for ease of understanding.   
 
The Philippine Clinical Practice Guidelines for Hypertension used a system 
adopted by the Canadian Hypertension Society. Therapy wise, a lot of Grade A 
recommendation meaning that evidence is based from well‐conducted trials was 
made.   
 
WILL THE RECOMMENDATIONS HELP YOU IN CARING FOR YOUR PATIENTS? 
 
• Is the primary objective of the guideline consistent with your 
objectives? 
 
The purpose of the guideline developers for coming up with recommendations 
may vary from your own.  Guidelines may be disseminated to assist physicians in 
decision making (clinical algorithms), to evaluate their practice and the standard 
of care they give to their patient (quality assurance) or to set limits for physician 
choices (reimbursements, recertification).  In any case, in order to find 
recommendations most suited to your needs, the purpose of the guideline 
should be in line with your intended objective. 
 
This hypertension guideline was made to ensure the availability of a local 
guideline for the detection and management of hypertension.  Since your 
questions dealt with management issues, this guideline is appropriate for your 
purpose. 
 
 
 
 

Page 96
Evidence-based Medicine Self-Instructional Manual

• Are the recommendations applicable to your patients? 
  
You must determine if the kind of patients you have are similar to those patients 
targeted by the guideline.  If your patients have a different prevalence or risk of 
disease, if the diagnostic and therapeutic options recommended are not 
available in your area, the guideline might not apply. 
 
The advantage of reviewing and applying a Philippine practice guideline is that it 
takes into account the characteristics of our setting and hopefully allows it to be 
more responsive.  
 
RESOLUTION OF THE PROBLEM IN THE SCENARIO 
 
Having deemed that the aforementioned guideline is valid, you would still opt to 
give this patient with uncomplicated hypertension a beta‐blocker.  In terms of 
diagnostics, you would request for an FBS, Serum Creatinine, Serum potassium 
and urinalysis.  You would request for these tests since they would have an 
effect on the antihypertensive you would choose.  Furthermore they would 
provide the following additional benefits: a)detection and early treatment of 
diabetes, b)detection of asymptomatic renal disease, c)detection of possible 
secondary hypertension.  You would explain to your patient that since he has no 
symptoms of any cardiac disease, performing an ECG and Echo is not routinely 
recommended. 
 
In terms of non‐pharmacologic advice, you would encourage him to go with 
regular aerobic exercise such as walking, jogging or cycling 30 minutes per day 3‐
4x/week since regular physical activity reduces blood pressure and results in a 
decrease in all cause mortality.   
 
A lot of decision making, considerations of options and outcomes came into play 
for a relatively simple case of hypertension.  And as medical knowledge 
improves, more options will be made available.  As such the need for relevant, 
well‐constructed and tested guidelines to improve our clinical decision making 
will always be there.  Again, we as clinicians should be able to adapt guidelines 
that are valid and whose recommendations will be most appropriate and 
feasible in our respective settings. 
 
 
REFERENCES
Hayward R, Wilson M, Tunis S, Bass E, Guyatt G for the Evidence Based Medicine Working
Group. How to Use a Clinical Practice Guideline: Validity. JAMA, 1995;274(7):570-4

Page 97
Evidence-based Medicine Self-Instructional Manual

Hayward R, Wilson M, Tunis S, Bass E, Guyatt G for the Evidence Based Medicine Working
Group. How to Use a Clinical Practice Guideline: Recommendations and Applicability. JAMA,
1995;274(20):1630-2.

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Page 98
Evidence-based Medicine Self-Instructional Manual

WORKSHOP
CRITICAL APPRAISAL OF
CLINICAL PRACTICE GUIDELINE
(SESSION BRIEFING)
 
 
OBJECTIVES
 
The purpose of the workshop is to introduce to the participants the concept of 
medical decision making about multiple problems related to disease 
management. Another objective is to introduce concepts of critical appraisal of a 
clinical practice guideline: 
 
• validity 
• recommended options and outcomes  
• up‐to‐date recommendations 
• relevance and certainty of the recommendations 
• applicability of the recommendations 
 
 
INSTRUCTIONS
 
Divide the participants into groups of six to ten persons per group. Assign a case 
scenario to each group and formulate an answerable problem from the scenario. 
Establish initial group consensus on how to proceed with the scenario. 
 
Ask the group to read the article retrieved to answer the problem in the 
scenario. Focus on the abstract, methods and results section. Again, establish a 
group consensus on how to proceed with the scenario. Note any change in 
decisions. 
 
Critically appraise the article using the appraisal sheet provided. Answer validity 
questions, analyze the recommendations and determine the applicability of the 
recommendations. Establish another group consensus and note any change in 
decision. 
 
Process the exercise. Focus on barriers and solution to the application of the 
exercise in usual clinic practice. 
 
 
 

Page 99
Evidence-based Medicine Self-Instructional Manual

TIME ALLOTTED
 
The time allotted for this workshop is one hour. The recommended break‐up is: 
 
• 5 minutes to analyze the scenario and develop consensus 
• 10 minutes to read the clinical practice guideline 
• 20 minutes to appraise the validity 
• 10 minutes to analyze the recommendations 
• 5 minutes to establish applicability 
• 10 minutes to summarize the process 
 
 
DESIRED OUTCOME
 
The participants should make a clinical decision on the scenario based on the 
critical appraisal of the clinical practice guideline. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Page 100
Evidence-based Medicine Self-Instructional Manual

Appraisal Sheet 
HOW TO USE A CLINICAL PRACTICE GUIDELINE 
 
   
CASE SCENARIO OR 
QUESTION 
 
   
SEARCH  
 
CRITICAL APPRAISAL   
   
RELEVANCE  Is the objective of the article on clinical practice guideline 
  similar to your clinical dilemma? 
 
 
 
   
VALIDITY GUIDES  Were all important options and outcomes considered? 
  Alternatives, expected results 
 
 
 
 
 
Was an explicit and sensible process used to identify, 
select, and combine evidence? 
        
 
 
 
 
Is the guideline likely to account for important recent 
developments? 
Last update 
 
 
 
Has the guideline been subjected to peer review and 
testing? 
 
 
 
 

Page 101
Evidence-based Medicine Self-Instructional Manual

OVERALL, IS THE   
GUIDELINE VALID?    
 
   
WHAT ARE THE  Are practical, clinically important, recommendations 
RECOMMENDATIONS  made? 
   
 
 
 
 
 
 
How strong are the recommendations? 
Grading 
 
 
 
 
 
 
   
WILL THE  Is the primary objective of the guideline consistent with 
RECOMMENDATIONS  your objectives? 
HELP YOU IN CARING         
FOR YOUR PATIENTS?     
   
 
 
 
Are the recommendations applicable to your patients and 
setting? 
  
 
 
 
 
 
   
RESOLUTION OF THE   
PROBLEM IN THE 
SCENARIO 
 

Page 102
Accuracy of MR Imaging of the
Knee in Adolescents
Nancy M. Major 1 OBJECTIVE. A report in the orthopedics literature states that MR imaging for internal de-
L. Neal Beard, Jr. rangement of the knee has a lower accuracy in adolescents than in adults and may even provide
Clyde A. Helms spurious information that alters clinical management. This assertion has not been specifically
addressed in the radiology literature. The purpose of our study was to determine the accuracy of
MR imaging in adolescents with regard to injury of the cruciate ligaments and menisci.
MATERIALS AND METHODS. A database search of our institution’s records from Janu-
ary 1998 to July 2000 yielded 2140 MR examinations of the knee, all of which had been per-
formed with a standard knee protocol on a 1.5-T magnet. Of these 2140 examinations, 156
included patients younger than 18 years. Fifty-nine of these patients underwent surgery, and the
orthopedic surgeons’ operative reports were used as the gold standard with which the MR imag-
ing results were compared. Thirty-four boys and 25 girls who ranged in age from 11 to 17 years
(mean age, 15 years) were examined. The clinical notes for the remaining 97 patients were evalu-
ated for information about management and clinical improvement.
RESULTS. The sensitivity and specificity values for MR imaging of the menisci and cruciate
ligaments in adolescents were as follows: medial meniscus, 92% sensitivity and 87% specificity;
lateral meniscus, 93% sensitivity and 95% specificity; anterior cruciate ligament, 100% sensitiv-
ity and 100% specificity; and posterior cruciate ligament, 0% sensitivity and 100% specificity.
CONCLUSION. Our data suggest that MR imaging of the knee in adolescents is sensi-
tive, specific, and accurate.

I n the adult population, MR imag-


ing has been accepted as the imag-
ing gold standard for detection of
internal derangement of the knee [1–3]. How-
the detection of injuries to the cruciate liga-
ments and menisci.

ever, when we discussed MR imaging of the Materials and Methods


knee of an adolescent patient with the orthope- A retrospective search of the musculoskeletal
dic surgeons at our institution, the surgeons ex- MR imaging database for knee examinations per-
formed from January 1998 to July 2000 was con-
pressed concern that MR imaging would be
ducted. Our search yielded 2140 knee examinations.
less useful in our adolescent patient than in an
Of these 2140 examinations, 156 included patients
adult patient because of the purported lower who were younger than 18 years. Fifty-nine of these
accuracy of MR imaging for the detection of 156 patients proceeded to surgery, and the orthope-
internal derangement of the knee in adoles- dic surgeons’ operative reports were used as the gold
cents. This notion had apparently been ac- standard with which we compared the MR imaging
cepted as truth by our orthopedic surgeons, so results. The study group included 34 boys and 25
Received January 21, 2002; accepted after revision they sometimes forego an MR imaging ex- girls who ranged in age from 11 to 17 years (mean
June 25, 2002. amination because they believe that the infor- age, 15 years). The clinical notes about the remain-
Presented at the annual meeting of the American mation from the study would not be useful ing 97 patients were evaluated for management and
Roentgen Ray Society, Seattle, April–May 2001. improvement. Of the 97 patients, 49 were boys and
and could even be misleading. This assertion
1 48 were girls, with ages ranging from 13 to 17 years
All authors: Department of Radiology, Duke University has been made in the orthopedics literature
Medical Center, Box 3808, Durham, NC 27710. Address (mean age, 15 years).
[4] but, to our knowledge, has not been spe- One of five musculoskeletal radiologists from our
correspondence to N. M. Major.
cifically addressed in the radiology literature. institution had prospectively evaluated each MR im-
AJR 2003;180:17–19
The purpose of our study was to determine aging examination. For each patient, the radiolo-
0361–803X/03/1801–17 the accuracy of MR imaging in adolescents gist’s interpretations of the menisci and cruciate
© American Roentgen Ray Society compared with that in adults with regard to ligaments were compared with the arthroscopic sur-

AJR:180, January 2003 17


Major et al.

gical findings; surgery was performed by various or- of the arthroscopic and MR imaging findings signal in the Hoffa fat pad, one Wrisberg variant
thopedic surgeons. MR imaging data were then yielded the following results. MR evaluation of discoid lateral meniscus, one medial collateral
categorized as true-positive, true-negative, false-pos- of the medial meniscus revealed 11 true-posi- ligament sprain, and one bucket-handle menis-
itive, and false-negative. From these data, sensitivity tives, 41 true-negatives, six false-positives, and cus tear. Of the four patients with anterior cruci-
and specificity for the detection of meniscal and cru-
one false-negative; these values resulted in a ate ligament tears, two refused surgery and two
ciate ligament tears in our study group of adoles-
cents were computed and compared with the same
92% sensitivity and 87% specificity. For the were lost to follow-up. One patient had an ante-
values in adults for the same radiologists (obtained lateral meniscus, the MR interpretations con- rior cruciate ligament tear and a bucket-handle
from another study) over approximately the same sisted of 14 true-positives, 42 true-negatives, meniscus tear. Twenty-five patients had a final
time period. two false-positives, and one false-negative, impression in the dictated report as “signal in ei-
All the MR imaging examinations were per- which resulted in a 93% sensitivity and 95% ther the meniscus or soft tissues not felt to be sig-
formed on a 1.5-T magnet (Signa; General Electric specificity. MR findings for the anterior cruci- nificant.” Therefore, the total number of cases
Medical Systems, Milwaukee, WI), and identical ate ligament yielded 26 true-positives and 33 that were not diagnosed with pathology was 64.
protocols were used for each of the examinations. true-negatives with zero false-positives and
Our standard knee protocol includes axial, sagittal, zero false-negatives, which resulted in a 100%
and coronal fast spin-echo T2-weighted imaging Discussion
sensitivity and specificity. For the posterior
(TR/TE effective, 3500/65) with fat suppression and
sagittal proton density imaging (TR/TE, 2000/20)
cruciate ligament, neither true-positives nor Our results indicate that the accuracy of MR
with fat suppression. The remaining parameters in- false-positives were recorded for the MR im- imaging for the detection of internal derange-
clude a matrix of 256 × 192, 2 excitations, a field of aging findings; there were 58 true-negatives ment of the knee in adolescents is similar to
view of 16 × 16 cm, and a slice thickness of 4 mm/ and one false-negative. These values yielded a that in adults. The idea for this retrospective
0.4 mm. 0% sensitivity and 100% specificity. database study came from repeated discus-
Meniscus tears were identified if linear high sig- The sensitivities and specificities of MR sions with various orthopedic surgeons at our
nal abutting the articular surface or abnormal mor- imaging for the detection of tears in the ado- institution, during which we were told that MR
phology was seen. The anterior cruciate ligament lescent group were essentially the same as imaging of the knee for detection of internal
was identified as torn if the fibers were disrupted and those for the adult group, which included a derangement is less useful in adolescents than
were no longer parallel to the intercondylar notch.
series of 203 patients (Table 1). adults because of a report of decreased accu-
Of the 97 patients who did not undergo ar- racy. The surgeons referred us to an article by
Results throscopy, “normal” was assigned as the diag- Stanitski [4]. In this article, Stanitski compared
The sensitivity and specificity values of MR nosis in 39 patients. Forty-six patients had no clinical examination findings, MR imaging re-
imaging for the detection of internal derange- additional follow-up. Other diagnoses encoun- sults, and arthroscopic findings in 28 children
ment of the knee in adolescents and adults are tered were 10 bone contusions, seven patellar and adolescents (age range, 8–17 years) with
shown in Table 1. dislocations (contusion pattern not counted in knee injuries. Articular surface, anterior cruci-
In the group of adolescents with arthro- previous group), four anterior cruciate liga- ate ligament, and meniscal injuries were re-
scopic correlation, arthroscopy showed 11 me- ment tears, two hematomas, two cases of Os- viewed and the conclusions were as follows:
dial meniscus tears, 14 lateral meniscus tears, good-Schlatter disease, two cases of jumper’s “Overall, magnetic resonance imaging diag-
25 anterior cruciate ligament tears, and one knee, two osteochondral lesions, one posterior noses added little guidance to patient manage-
posterior cruciate ligament tear. Comparison cruciate ligament injury, one case of abnormal ment and at times provided spurious
information.” The data in Table 2 are from
Stanitski’s article. Stanitski reported 75% total
Sensitivity and Specificity of MR Imaging of the Knee in Adolescents
TABLE 1 disagreement between clinical and MR imag-
Versus Adults
ing and 78.5% total disagreement between ar-
Adolescents (n = 59) Adults (n = 203) throscopic findings and MR imaging results.
Location
Sensitivity (%) Specificity (%) Sensitivity (%) Specificity (%) In that study, total agreement was found be-
tween clinical examination and arthroscopy in
Medial meniscus 92 87 90 80
78.5% of the cases. Therefore, Stanitski as-
Lateral meniscus 93 95 78 93 serted that clinical examination and arthroscopy
Anterior cruciate ligament 100 100 100 97 are superior to MR imaging of adolescents. The
Posterior cruciate ligament 0 100 83 98 orthopedics literature includes a second article
written by McDermott et al. [5]; in that article,
the authors state that the accuracy for MR im-
Comparison of Findings from Clinical Examination, MR Imaging, and
TABLE 2
Arthroscopy of the Knee
aging of knee abnormalities in pediatric pa-
tients is not well established. We believe that
Agreement (%) Total Disagreement our data refute the argument that MR imaging
Findings Compared
Total Partial (%) of the knee is less accurate in adolescents than
in adults.
Clinical examination vs MR imaging 14.3 10.7 75
A number of differences between our study
Clinical examination vs arthroscopy 78.5 7.1 14.3 and that conducted by Stanitski [4] exist. First,
MR imaging vs arthroscopy 7.1 14.3 78.5 the latter study included results from only 28
Note.—All data are from [4]. patients, whereas our study included results

18 AJR:180, January 2003


MR Imaging of the Knee in Adolescents

from 59 patients. However, although the num- 95% for medial meniscus tears, and 85–90% In conclusion, we believe that MR imaging
ber of patients in our study is more than double accuracy for lateral meniscus tears [1–3, 6, of the knee is just as useful as a clinical ad-
that in the other study, the total number is still 7], and there is no reason to believe that these junct in adolescents as in adults. Therefore,
small. Therefore, a small sample is a potential numbers should not hold true for general ra- MR imaging of the knee in adolescents can
shortcoming of our study. Possible explana- diologists. Stanitski asserted that sensitivity assist in preventing unnecessary surgery such
tions for the small number of MR imaging and specificity of MR imaging for detecting as diagnostic arthroscopy. In circumstances
studies in adolescents include the reluctance of internal derangements of the knee were infe- in which surgery is deemed necessary, MR
orthopedic surgeons to use MR imaging in rior in adolescents compared with adults. Af- imaging can aid in surgical planning, which
these patients because of the report by Sta- ter evaluating our data, we found that the benefits the orthopedic surgeon as well as the
nitski and the possibility that adolescents are sensitivity and specificity values for MR im- patient because the information provided by
less likely to have internal derangement of the aging of adolescents and adults were essen- MR imaging leads to decreased procedure
knee than adults. tially the same (Table 1). and tourniquet time.
Another difference between our study and Although our primary intention was to
that of Stanitski [4] is that Stanitski used determine the accuracy of MR imaging of
grade 2 meniscal signal abnormality as evi- the knee compared with arthroscopy in ado-
References
dence for meniscus tear in an unspecified lescents, we also assessed the outcomes for
1. Mackenzie R, Palmer CR, Lomas DJ, Dixon AK.
number of patients. It is well known that the 97 patients who did not undergo arthros-
Magnetic resonance imaging of the knee: diag-
grade 2 intrameniscal signal is evidence of in- copy. Forty-six patients did not undergo a nostic performance studies. Clin Radiol 1996;51:
trasubstance degeneration rather than a tear, follow-up examination. A lack of follow-up 251–257
because grade 2 intrameniscal signal does not could indicate that either the symptoms re- 2. Mink J, Levy T, Crues JI. Tears of the anterior cru-
disrupt the articular surface. These cases were solved so clinical follow-up was not needed ciate ligament and menisci of the knee: MR imag-
erroneously diagnosed as tears in that study, or the patient was seen elsewhere for addi- ing evaluation. Radiology 1988;167:769–774
3. De Smet AA, Graf BK. Meniscal tears missed on
therefore reducing the accuracy of MR imag- tional follow-up. Of the remaining patients
MR imaging: relationship to meniscal tear pat-
ing for revealing meniscal abnormalities. In who did undergo follow-up, the visit con- terns and anterior cruciate ligament tears. AJR
addition, Stanitski did not provide the imaging sisted of one-time physical therapy or ortho- 1994;162:905–911
parameters used to evaluate the meniscus. If pedic follow-up without any additional 4. Stanitski CL. Correlation of arthroscopic and
sequences with a long TE were chosen to eval- follow-up or intervention. Four anterior cru- clinical examinations with magnetic resonance
uate the meniscus, tears could have been over- ciate ligament tears were identified, but the imaging findings of injured knees in children and
adolescents. Am J Sports Med 1998;26:2–6
looked. Proper protocols will aid the patients did not undergo surgery at our insti-
5. McDermott MJ, Bathgate B, Gillingham BL, Hen-
radiologist (and surgeon) in accurately assess- tution: two did not want surgery and the nrikus WL. Correlation of MRI and arthroscopic
ing the integrity of the meniscus. other two were lost to follow-up. The patient diagnosis of knee pathology in children and adoles-
In the Stanitski study [4], the accuracy of with the bucket-handle meniscus tear was cents. J Pediatr Orthop 1998;18:675–678
the radiologists’ interpretations of the MR among these four patients. No additional 6. Boeree NR, Watkinson AF, Ackroyd CE, Johnson
images of adults is not known. A potential “surgical lesions” were identified. None of C. Magnetic resonance imaging of meniscal and
shortcoming in our study is that only muscu- the patella dislocations had associated carti- cruciate injuries of the knee. J Bone Joint Surg Br
1991;73:452–457
loskeletal radiologists interpreted MR images lage loss (our surgeons’ indication for oper-
7. Lee JK, Yao L, Phelps CT, Wirth CR, Czajka J,
rather than general radiologists. However, the ating). The osteochondral lesions were Lazman J. Anterior cruciate ligament tears: MR
radiology literature reports 95–100% accu- stable by MR appearance. Surgery was not imaging compared with arthroscopy and clinical
racy for anterior cruciate ligament tears, 90– considered for these two patients. tests. Radiology 1988;166:861–864

AJR:180, January 2003 19


The Ne w E n g l a nd Jo u r n a l o f Me d ic i ne

VITAMIN E SUPPLEMENTATION AND CARDIOVASCULAR EVENTS


IN HIGH-RISK PATIENTS

THE HEART OUTCOMES PREVENTION EVALUATION STUDY INVESTIGATORS*

ABSTRACT sume more than 100 IU of vitamin E a day for more


Background Observational and experimental stud- than two years have lower rates of coronary events10,11
ies suggest that the amount of vitamin E ingested in and lower rates of progression of coronary artery le-
food and in supplements is associated with a lower sions.12 However, observational studies cannot distin-
risk of coronary heart disease and atherosclerosis. guish whether the lower risk of coronary heart disease
Methods We enrolled a total of 2545 women and associated with higher levels of vitamin E consump-
6996 men 55 years of age or older who were at high tion is due to the vitamin or to other associated life-
risk for cardiovascular events because they had car- style factors such as increased exercise and other
diovascular disease or diabetes in addition to one oth- aspects of diet. There have been four randomized,
er risk factor. These patients were randomly assigned
according to a two-by-two factorial design to receive
controlled trials of the relation between vitamin E and
either 400 IU of vitamin E daily from natural sources coronary heart disease,13-16 but their results are con-
or matching placebo and either an angiotensin-con- flicting, perhaps because of the low doses of vita-
verting–enzyme inhibitor (ramipril) or matching pla- min E used in some studies,13,14 the small numbers
cebo for a mean of 4.5 years (the results of the com- of events,15 or the limited duration of treatment.15,16
parison of ramipril and placebo are reported in a We evaluated a high dose (400 IU per day) of vi-
companion article). The primary outcome was a com- tamin E from natural sources, which has high bio-
posite of myocardial infarction, stroke, and death from availability, in a large, five-year, prospective study of
cardiovascular causes. The secondary outcomes in- patients at high risk for cardiovascular events. The
cluded unstable angina, congestive heart failure, re- primary outcome was a composite of myocardial in-
vascularization or amputation, death from any cause,
complications of diabetes, and cancer.
farction, stroke, and death from cardiovascular causes.
Results A total of 772 of the 4761 patients as- The secondary outcomes included death from any
signed to vitamin E (16.2 percent) and 739 of the 4780 cause, hospitalization for unstable angina or conges-
assigned to placebo (15.5 percent) had a primary out- tive heart failure, revascularization or limb amputa-
come event (relative risk, 1.05; 95 percent confidence tion, complications of diabetes, and cancer. The trial
interval, 0.95 to 1.16; P=0.33). There were no signifi- was also designed to evaluate the effects of an angio-
cant differences in the numbers of deaths from car- tensin-converting–enzyme inhibitor, ramipril, on the
diovascular causes (342 of those assigned to vita- incidence of cardiovascular events. After nearly 4.5
min E vs. 328 of those assigned to placebo; relative years of follow-up, the collection of data on cardiovas-
risk, 1.05; 95 percent confidence interval, 0.90 to 1.22), cular disease was stopped in April 1999 on the basis
myocardial infarction (532 vs. 524; relative risk, 1.02;
of a finding by the independent data and safety mon-
95 percent confidence interval, 0.90 to 1.15), or stroke
(209 vs. 180; relative risk, 1.17; 95 percent confidence itoring board that the trial had conclusively demon-
interval, 0.95 to 1.42). There were also no significant strated the benefits of ramipril and a lack of effect of
differences in the incidence of secondary cardiovascu- vitamin E on cardiovascular events. This report pre-
lar outcomes or in death from any cause. There were sents our findings relating to the effects of vitamin E
no significant adverse effects of vitamin E. on the primary and secondary cardiovascular out-
Conclusions In patients at high risk for cardiovas- comes. The study has been continued in the majority
cular events, treatment with vitamin E for a mean of of centers to evaluate the effects of vitamin E on the
4.5 years has no apparent effect on cardiovascular incidence of cancer.
outcomes. (N Engl J Med 2000;342:154-60.)
©2000, Massachusetts Medical Society. METHODS
Study Design
The Heart Outcomes Prevention Evaluation (HOPE) Study is

O
XIDATIVE modification of low-density a double-blind, randomized trial with a two-by-two factorial de-
lipoprotein is an important step in the de-
velopment and progression of atheroscle- Address reprint requests to Dr. Salim Yusuf at the Canadian Cardiovascular
rosis in experimental studies,1,2 and antiox- Collaboration Project Office, Hamilton General Hospital, 237 Barton St.
E., Hamilton, ON L8L 2X2, Canada, or at yusufs@fhs.mcmaster.ca.
idants such as vitamin E have been shown to slow The writing group (Salim Yusuf, D.Phil., Gilles Dagenais, M.D., Janice
atherosclerosis.3-5 An inverse relation has been ob- Pogue, M.Sc., Jackie Bosch, M.Sc., and Peter Sleight, D.M.) assumes re-
served between coronary heart disease and the con- sponsibility for the overall content and integrity of the manuscript.
sumption of fruits, vegetables, and other foods *The investigators are listed in the Appendix of the Heart Outcomes Pre-
vention Evaluation Study Investigators. Effects of an Angiotensin-Convert-
containing vitamins, particularly vitamin E.6-9 Obser- ing–Enzyme Inhibitor, Ramipril, on Cardiovascular Events in High-Risk
vational studies have indicated that persons who con- Patients. N Engl J Med 2000;342:145-53.

154 · Ja nu ar y 2 0 , 2 0 0 0
VITAMIN E SUP P L E ME NTAT ION A ND C A R D IOVASC UL A R EV ENTS IN H IGH -RISK PATIENTS

sign, conducted to evaluate the effects of ramipril and vitamin E


in 9541 patients at high risk for cardiovascular events. The results TABLE 1. BASE-LINE CHARACTERISTICS OF THE PATIENTS.*
of the comparison of ramipril with placebo are reported in a com-
panion article.17 Details of the methods are given in that article17
and in a previously published article.18 Briefly, eligible patients at VITAMIN E GROUP PLACEBO GROUP
high risk were randomly assigned to receive either 400 IU of vi- CHARACTERISTIC† (N=4761) (N=4780)
tamin E from natural sources or an equivalent placebo daily for
Age — yr 66±7 66±7
4 to 6 years (mean, 4.5) and in addition to receive either 10 mg of
Blood pressure — mm Hg 139±20/79±11 139±20/79±11
ramipril or an equivalent placebo daily. Patients were evaluated
every six months for a variety of outcomes. Heart rate — beats/min 69±11 69±11
Body-mass index‡ 28±4 28±4
Outcomes Female sex — no. (%) 1263 (26.5) 1282 (26.8)
History of coronary artery disease 3857 (81.0) 3832 (80.2)
The primary outcome was a composite of myocardial infarction,
— no. (%)
stroke, and death from cardiovascular causes. Deaths classified as Myocardial infarction 2499 (52.5) 2535 (53.0)
due to cardiovascular causes were unexpected deaths presumed to Stable angina pectoris 2653 (55.7) 2668 (55.8)
be due to ischemic cardiovascular disease and occurring within 24 Unstable angina pectoris 1205 (25.3) 1246 (26.1)
hours after the onset of symptoms without clinical or postmor- CABG 1229 (25.8) 1251 (26.2)
tem evidence of another cause; deaths from myocardial infarction PTCA 851 (17.9) 863 (18.1)
or stroke that occurred within seven days after the myocardial in- Stroke or transient ischemic attacks 530 (11.1) 500 (10.5)
farction or stroke; and deaths from congestive heart failure, dys- — no. (%)
rhythmia, pulmonary embolism, or ruptured abdominal aortic an- Peripheral vascular disease — no. (%)§ 2109 (44.3) 2037 (42.6)
eurysm. Deaths for which the cause was uncertain were presumed Hypertension — no. (%) 2219 (46.6) 2222 (46.5)
to be due to cardiovascular disease. Myocardial infarction was diag- Diabetes — no. (%) 1838 (38.6) 1816 (38.0)
nosed when two of the following three criteria were met: typical Known elevated total cholesterol 3109 (65.3) 3171 (66.3)
symptoms, increased cardiac enzyme levels (at least twice the up- — no. (%)
per limit of normal), and diagnostic electrocardiographic changes. Known low HDL cholesterol 893 (18.8) 869 (18.2)
Stroke was defined as a neurologic deficit lasting more than 24 — no. (%)
hours. A computed tomographic or magnetic resonance imaging Current cigarette smoking — no. (%) 665 (14.0) 679 (14.2)
examination was recommended to define the type of stroke. Medications — no. (%)
Secondary and other outcomes were death from any cause; un- Beta-blockers 1901 (39.9) 1870 (39.1)
stable angina, defined as worsening angina or angina at rest requir- Aspirin or other antiplatelet agents 3665 (77.0) 3616 (75.6)
ing hospitalization; hospitalization for heart failure with clinical Lipid-lowering agents 1352 (28.4) 1401 (29.3)
and radiologic signs of congestion; revascularization or limb am- Diuretics 728 (15.3) 717 (15.0)
putation; the development of overt nephropathy or the need for Calcium-channel blockers 2249 (47.2) 2236 (46.8)
dialysis or laser therapy among patients with diabetes; and the de- Left ventricular hypertrophy on ECG 411 (8.6) 382 (8.0)
velopment of heart failure or new or worsening angina regardless — no. (%)
of the need for hospitalization. Microalbuminuria — no. (%) 1012 (21.3) 976 (20.4)

RESULTS *Plus–minus values are means ±SD.


†CABG denotes coronary-artery bypass grafting, PTCA percutaneous
Characteristics of the Patients transluminal coronary angioplasty, HDL high-density lipoprotein, and
ECG electrocardiogram.
The characteristics of the 9541 patients are shown
‡The body-mass index is calculated as the weight in kilograms divided
in Table 1. The rate of compliance with the assigned by the square of the height in meters.
regimen was high throughout the study. The percent- §Peripheral vascular disease included claudication, a history of peripheral
ages of patients who were taking vitamin E in the arterial disease, or a ratio of blood pressure in the ankle to blood pressure
in the arm of less than 0.90.
vitamin E and placebo groups, respectively, were 94.2
percent and 1.0 percent at one year, 93.3 percent and
1.7 percent at two years, 91.3 percent and 2.0 per-
cent at three years, 90.2 percent and 2.7 percent at
four years, and 89.2 percent and 3.4 percent at the (287 vs. 277; relative risk, 1.06), or strokes (209 vs.
final visit. 180; relative risk, 1.17) (Fig. 2 and 3). The total num-
bers of deaths were similar in the two groups (535 vs.
Primary Cardiovascular Outcomes and Deaths 537; relative risk, 1.00). Vitamin E had no signifi-
from Any Cause
cant effect on the primary outcome either among
A total of 772 of the 4761 patients who were as- patients who were receiving ramipril (338 events
signed to receive vitamin E (16.2 percent) and 739 among those who were receiving vitamin E and 313
of the 4780 who were assigned to placebo (15.5 per- events among those who were receiving placebo; rel-
cent) had a primary cardiovascular event (relative risk, ative risk, 1.08) or among patients who were not re-
1.05; 95 percent confidence interval, 0.95 to 1.16; ceiving ramipril (421 and 405 events, respectively;
P=0.33) (Table 2 and Fig. 1). There were no signif- relative risk, 1.05).
icant differences between the groups in the numbers
of deaths from cardiovascular causes (342 in the vi- Secondary Cardiovascular and Combined Outcomes
tamin E group vs. 328 in the placebo group; relative There were no differences between patients as-
risk, 1.05), myocardial infarctions (532 vs. 524; rel- signed to vitamin E and those assigned to placebo in
ative risk, 1.02), deaths from coronary heart disease the number of hospitalizations for unstable angina

Vol ume 342 Numb e r 3 · 155


The Ne w E n g l a nd Jo u r n a l o f Me d ic i ne

TABLE 2. INCIDENCE OF THE PRIMARY OUTCOME AND OF DEATHS FROM ANY CAUSE.

VITAMIN E PLACEBO
GROUP GROUP RELATIVE RISK
OUTCOME (N=4761) (N=4780) (95% CI)* P VALUE†

no. (%)

Myocardial infarction, stroke, or 772 (16.2) 739 (15.5) 1.05 (0.95–1.16) 0.33
death from cardiovascular causes‡
Death from cardiovascular causes§ 342 (7.2) 328 (6.9) 1.05 (0.90–1.22) 0.54
Myocardial infarction§ 532 (11.2) 524 (11.0) 1.02 (0.90–1.15) 0.74
Stroke§ 209 (4.4) 180 (3.8) 1.17 (0.95–1.42) 0.13
Death from any cause 535 (11.2) 537 (11.2) 1.00 (0.89–1.13) 0.99

*CI denotes confidence interval.


†P values were calculated with use of the log-rank test.
‡The number of events among those receiving ramipril did not differ significantly between those
assigned to receive vitamin E and those assigned to placebo (338 vs. 313). Similar results were ob-
served among those who received matching placebo rather than ramipril (421 vs. 405).
§A patient may have had more than one event.

Primary Outcome Subgroup Analyses


There was no heterogeneity of results among sub-
0.20
groups defined according to sex, age, previous car-
Proportion of Patients

Vitamin EF diovascular disease, or use of other drugs with re-


Placebo
0.15 spect to the primary or secondary outcomes (data not
shown). Specifically, there was no significant differ-
ence in the incidence of the primary outcome among
0.10 patients with diabetes (325 of those assigned to vita-
min E vs. 313 of those assigned to placebo; relative
0.05
risk, 1.04) or among smokers (135 vs. 139; relative
risk, 1.02).
Adverse Effects
0.00
0 500 1000 1500 There was no significant difference between groups
Days of Follow-up in the incidence of adverse effects or in the number
of patients who stopped taking the study medication.
Figure 1. Kaplan–Meier Estimates of the Effect of Vitamin E
on the Composite Outcome of Nonfatal Myocardial Infarction,
There was no increase in hemorrhagic stroke associ-
Stroke, or Death from Cardiovascular Causes. ated with vitamin E use (17 of those assigned to vi-
The relative risk of the composite outcome in the vitamin E tamin E had hemorrhagic stroke, as compared with
group as compared with the placebo group was 1.05 (95 percent 13 of those assigned to placebo) or among those who
confidence interval, 0.95 to 1.16; P=0.33). were also taking an antiplatelet agent (11 vs. 8).
DISCUSSION
In our study, vitamin E did not reduce the inci-
(586 vs. 569; relative risk, 1.04), hospitalizations for dence of cardiovascular events, as compared with the
heart failure (160 vs. 144; relative risk, 1.12), or re- incidence among patients assigned to placebo, dur-
vascularizations or limb amputations (848 vs. 787; ing a follow-up period of four to six years. Given the
relative risk, 1.09) (Table 3). There were no signifi- large number of events and the consistent lack of dif-
cant differences in the number of patients with angina ference in all secondary cardiovascular outcomes, it is
of new onset (278 vs. 245; relative risk, 1.15) or mi- very unlikely that vitamin E had any clinically worth-
crovascular complications of diabetes (340 vs. 325; while beneficial effect on cardiovascular disease dur-
relative risk, 1.06). A combined analysis of the pro- ing four or five years of treatment.
portion of patients who had any primary or second- Results have been reported from four randomized
ary event found a nonsignificantly higher rate among trials of the effects of vitamin E on cardiovascular
those assigned to vitamin E (1630 vs. 1576; relative events. In a Chinese study, 29,584 adults from Linxian
risk, 1.05; 95 percent confidence interval, 0.98 to Province, who did not have cardiovascular disease at
1.13; P=0.14). entry, were randomly assigned to receive daily vita-

156 · Ja nu ar y 2 0 , 2 0 0 0
VITAMIN E SUP P L E ME NTAT ION A ND CA R D IOVASC UL A R EV ENTS IN H IGH -RISK PATIENTS

A Myocardial Infarction Death from Any Cause


0.15 0.15

Proportion of Patients
Proportion of Patients

Vitamin EF Vitamin EF
Placebo Placebo

0.10 0.10

0.05 0.05

0.00 0.00
0 500 1000 1500 0 500 1000 1500
Days of Follow-up Days of Follow-up
Figure 3. Kaplan–Meier Estimates of the Effect of Vitamin E on
B Stroke the Incidence of Death from Any Cause.
The relative risk in the vitamin E group as compared with the
0.06 placebo group was 1.00 (95 percent confidence interval, 0.89 to
1.13; P=0.99).
Proportion of Patients

Vitamin EF
Placebo
0.04
vious myocardial infarction at entry, there was a non-
significant increase in the risk of death from coronary
heart disease (relative risk, 1.33; 95 percent confidence
0.02
interval, 0.86 to 2.05; P=0.20). However, a reduc-
tion in the risk of nonfatal myocardial infarction was
documented among men assigned to vitamin E only
0.00 (40 vs. 55; relative risk, 0.62; 95 percent confidence
0 500 1000 1500 interval, 0.41 to 0.96), but not among those receiv-
Days of Follow-up ing the combination of vitamin E and beta carotene,
in comparison with those receiving placebo only.19
Figure 2. Kaplan–Meier Estimates of the Effect of Vitamin E on
the Incidence of Myocardial Infarction (Panel A) and Stroke
In this subgroup, the number of events was small. In
(Panel B). the remaining patients in this study, there was no sig-
The relative risk of myocardial infarction in the vitamin E group nificant effect of vitamin E on nonfatal or fatal my-
as compared with the placebo group was 1.02 (95 percent confi- ocardial infarction, despite large numbers of events
dence interval, 0.90 to 1.15; P=0.74), and the relative risk of stroke (1204 and 907, respectively).20 Thus, in this well-
was 1.17 (95 percent confidence interval, 0.95 to 1.42; P=0.13). conducted trial, vitamin E had no effect on coronary
heart disease. Although the trial used a low dose of
synthetic vitamin E (50 mg per day), the median level
of alpha-tocopherol increased significantly, from 28.5
min E (30 mg), beta carotene, and selenium supple- µmol per liter at base line to 42.5 µmol per liter at
ments or to receive placebo.13 During the 5.2 years of three months.
follow-up, there was a 9 percent decrease in deaths The third trial was the Cambridge Heart Antiox-
from any cause without any significant reduction in idant Study, which randomly assigned 2002 patients
cardiovascular events. The dose of vitamin E in this with coronary atherosclerosis to receive either vita-
study was small, the nutritional status and cardiovas- min E or placebo.15 The mean alpha-tocopherol levels
cular risk of this population were very different from increased from 34.2 to 51.1 µmol per liter in pa-
those of Western populations, and the beneficial ef- tients receiving 400 IU of vitamin E per day and to
fects on overall mortality cannot be attributed only to 64.5 µmol per liter in patients receiving 800 IU per
vitamin E. day. The majority of the patients received 400 IU per
The second trial was the Alpha-Tocopherol, Beta day. After a median follow-up of 1.4 years, a large re-
Carotene Cancer Prevention Study, involving 29,133 duction in the number of patients with nonfatal my-
male smokers who were 50 to 69 years of age.14 Daily ocardial infarction was observed (14 in the vitamin E
treatment with 50 mg of vitamin E for five to eight group vs. 41 in the placebo group; relative risk, 0.53;
years had no effect on the risk of death from coronary 95 percent confidence interval, 0.11 to 0.47; P=
heart disease. In a subgroup of 1862 men with a pre- 0.005), but there was no difference in deaths due to

Vol ume 342 Numb e r 3 · 157


The Ne w E n g l a nd Jo u r n a l o f Me d ic i ne

TABLE 3. INCIDENCE OF SECONDARY AND OTHER OUTCOMES.

VITAMIN E PLACEBO
GROUP GROUP RELATIVE RISK P
OUTCOME (N=4761) (N=4780) (95% CI)* VALUE†

no. (%)

Revascularization or limb amputation 848 (17.8) 787 (16.5) 1.09 (0.99–1.20) 0.07
Hospitalization for unstable angina 586 (12.3) 569 (11.9) 1.04 (0.93–1.17) 0.52
New-onset angina 278 (5.8) 245 (5.1) 1.15 (0.97–1.37) 0.11
Worsening angina 1215 (25.5) 1186 (24.8) 1.02 (0.94–1.11) 0.63
Claudication 762 (16.0) 753 (15.8) 1.02 (0.92–1.13) 0.70
Hospitalization for heart failure 160 (3.4) 144 (3.0) 1.12 (0.90–1.41) 0.32
Heart failure 530 (11.0) 457 (9.6) 1.17 (1.03–1.32) 0.02
Complications of diabetes‡ 340 (7.1) 325 (6.8) 1.06 (0.91–1.23) 0.47

*CI denotes confidence interval.


†P values were calculated with use of the log-rank test.
‡Complications included nephropathy, dialysis, and laser therapy.

TABLE 4. META-ANALYSIS OF THE EFFECTS OF VITAMIN E ON MYOCARDIAL INFARCTION, STROKE,


OR DEATH FROM CARDIOVASCULAR CAUSES IN LARGE TRIALS.*

STUDY DAILY DOSE DURATION OF STUDY VITAMIN E PLACEBO RELATIVE RISK (95% CI)

mg yr no. with events/total no. (%)

ATBC14 50 5.0 1889/14,564 (13.0) 1970/14,569 (13.5) 0.96 (0.90–1.03)


CHAOS15 »400 1.3 41/1035 (4.0) 64/967 (6.6) 0.60 (0.40–0.89)
GISSI16 300 3.5 571/5660 (10.1) 584/5664 (10.3) 0.98 (0.87–1.10)
Current study 400 4.5 772/4761 (16.2) 739/4780 (15.5) 1.05 (0.95–1.16)
Total 3273/26,020 (12.6) 3357/25,980 (12.9) 0.97 (0.92–1.02)†

*CI denotes confidence interval, ATBC Alpha-Tocopherol, Beta Carotene Cancer Prevention Study Group, CHAOS
Cambridge Heart Antioxidant Study, and GISSI Gruppo Italiano per lo Studio della Sopravvivenza nell’Infarto Miocardico.
†Relative risks and confidence intervals were derived by the method of Yusuf et al.21; P=0.27.

cardiovascular causes (27 vs. 23; relative risk, 1.18; 95 placebo group (295 vs. 284; relative risk, 1.02; 95 per-
percent confidence interval, 0.62 to 2.27; P=0.61). cent confidence interval, 0.87 to 1.21), and the num-
In this trial, the number of events was small and there ber of deaths from coronary heart disease was slightly
were imbalances in several base-line characteristics that smaller (227 vs. 249; relative risk, 0.92; 95 percent
call into question whether randomization resulted in confidence interval, 0.77 to 1.11). Neither difference
truly comparable groups. was statistically significant.16
Furthermore, the very large reduction in nonfatal Our study used a high dose of vitamin E (400 IU
myocardial infarction within a relatively short time per day), had high rates of compliance, and involved
(median, 1.4 years) is inconsistent with the results of high-risk patients. The study had a large number of
other interventions, such as lipid-lowering agents or primary outcomes and therefore had high statistical
antihypertensive medications, that reduce cardiovas- power (more than 90 percent power to detect a 13
cular events. It is therefore likely that the results of the percent relative reduction in the risk of the primary
Cambridge Heart Antioxidant Study may have been outcome). Furthermore, a large number of secondary
due to chance. This possibility is supported by the outcomes (e.g., revascularization or limb amputation,
results of a recent Italian trial,16 in which 11,000 pa- unstable angina, worsening angina, and heart failure)
tients who had had myocardial infarctions were ran- were examined. Such data are not available from most
domly assigned to receive 300 IU of vitamin E per trials. Combining the data from all trials of vitamin E
day or placebo for a median of 3.5 years. The num- indicates that such treatment has little effect on the
ber of patients with nonfatal myocardial infarction risk of death or cardiovascular events (Table 4), at least
was slightly higher in the vitamin E group than the over a four-to-six-year period.

158 · Ja nu ar y 2 0 , 2 0 0 0
V ITAMIN E SUP P L E ME NTAT ION A ND C A R D IOVASC UL A R EV ENTS IN H IGH -RISK PATIENTS

Steinberg has hypothesized that unlike agents that We are indebted to W. Whitehill, F. Schutze, N. Bender, B. Ran-
lower cholesterol or blood pressure, antioxidants may goonwala, A. Ljunggren, G. Olsson, J.C. Dairon, J. Ghadiali, B.
Carter, J.P. St. Pierre, W. Schulz, M. Jensen, L. Rios-Nogales, M.
have to be used for more than five years to have a de- Bravo, J. Bourgouin, and C. Vint-Reed for support and to Karin
monstrable benefit, since the primary mechanism of Dearness for secretarial help.
these agents may be the prevention of new lesions.22
Therefore, in a population like the one we studied, it REFERENCES
may take longer than five years to detect an effect on 1. Steinberg D, Parthasarathy S, Carew TE, Khoo JC, Witztum JL. Be-
clinical outcomes. However, the Physicians’ Health yond cholesterol: modifications of low-density lipoprotein that increase its
atherogenicity. N Engl J Med 1989;320:915-24.
Study did not find a benefit of beta carotene (another 2. Witztum JL. The oxidation hypothesis of atherosclerosis. Lancet 1994;
antioxidant with a different action) after 12 years.23 344:793-5.
Similar data are not available for vitamin E, but ob- 3. Carew TE, Schwenke DC, Steinberg D. Antiatherogenic effect of
probucol unrelated to its hypercholesterolemic effect: evidence that antiox-
servational studies that demonstrated a lower rate of idants in vivo can selectively inhibit low density lipoprotein degradation in
coronary heart disease with vitamin E supplementa- macrophage-rich fatty streaks and slow the progression of atherosclerosis
in the Watanabe heritable hyperlipidemic rabbit. Proc Natl Acad Sci U S A
tion suggested that a lower risk should be evident af- 1987;84:7725-9.
ter two years.10,11 In a nested substudy, we are examin- 4. Verlangieri AJ, Bush MJ. Effects of d-a-tocopherol supplementation on
ing whether the thickness of the carotid intima and experimentally induced primate atherosclerosis. J Am Coll Nutr 1992;11:
131-8.
media (an indication of the risk of early atheroscle- 5. Williams RJ, Motteram JM, Sharp CH, Gallagher PJ. Dietary vitamin
rosis) can be favorably altered by vitamin E.24 If so, E and the attenuation of early lesion development in modified Watanabe
Steinberg’s hypothesis may be worth exploring with rabbits. Atherosclerosis 1992;94:153-9.
6. Knekt P, Reunanen A, Jarvinen R, Seppanen R, Heliovaara M, Aromaa
more prolonged follow-up or treatment to assess A. Antioxidant vitamin intake and coronary mortality in a longitudinal
whether such changes in the development of athero- population study. Am J Epidemiol 1994;139:1180-9.
7. Gaziano JM, Manson JE, Branch LG, Colditz GA, Willett WC, Buring
sclerosis would translate into a benefit in terms of clin- JE. A prospective study of consumption of carotenoids in fruits and vege-
ical outcomes. tables and decreased cardiovascular mortality in the elderly. Ann Epidemiol
Although the moderate duration of vitamin E sup- 1995;5:255-60.
8. Kushi LH, Folsom AR, Prineas RJ, Mink PJ, Wu Y, Bostick RM. Die-
plementation (four to six years) and the characteris- tary antioxidant vitamins and death from coronary heart disease in post-
tics of the population may explain our finding of a menopausal women. N Engl J Med 1996;334:1156-62.
lack of benefit of vitamin E, another reason may be 9. Klipstein-Grobusch K, Geleijnse JM, den Breeijen JH, et al. Dietary an-
tioxidants and risk of myocardial infarction in the elderly: the Rotterdam
our use of vitamin E alone, without other antioxidants. Study. Am J Clin Nutr 1999;69:261-6.
In the epidemiologic studies that found an associa- 10. Stampfer MJ, Hennekens CH, Manson JE, Colditz GA, Rosner B,
Willett WC. Vitamin E consumption and the risk of coronary disease in
tion between higher dietary intake of vitamin E and women. N Engl J Med 1993;328:1444-9.
lower rates of coronary heart disease, higher vitamin E 11. Rimm EB, Stampfer MJ, Ascherio A, Giovannucci E, Colditz GA,
consumption was also associated with higher intake Willett WC. Vitamin E consumption and the risk of coronary heart disease
in men. N Engl J Med 1993;328:1450-6.
of a number of other antioxidants and micronutri- 12. Hodis HN, Mack WJ, LaBree L, et al. Serial coronary angiographic
ents.6-9 It is possible that vitamin E supplementation evidence that antioxidant vitamin intake reduces progression of coronary
requires these cofactors to have a beneficial effect.25 artery atherosclerosis. JAMA 1995;273:1849-54.
13. Blot WJ, Li JY, Taylor PR, et al. Nutrition intervention trials in Lin-
Although the existence of interactions between vita- xian, China: supplementation with specific vitamin/mineral combinations,
min E and other vitamins,10 beta carotene,6,14 or cancer incidence, and disease-specific mortality in the general population.
J Natl Cancer Inst 1993;85:1483-92.
selenium13 is not supported by the findings of pro- 14. The Alpha-Tocopherol, Beta Carotene Cancer Prevention Study Group.
spective observational studies or randomized trials, The effect of vitamin E and beta carotene on the incidence of lung cancer
this hypothesis can be tested only in trials in which and other cancers in male smokers. N Engl J Med 1994;330:1029-35.
15. Stephens NG, Parsons A, Schofield PM, Kelly F, Cheeseman K,
combinations of vitamins are given; some such trials Mitchinson MJ. Randomised controlled trial of vitamin E in patients with
are now in progress.26-28 coronary disease: Cambridge Heart Antioxidant Study. Lancet 1996;347:
In conclusion, 400 IU of vitamin E administered 781-6.
16. GISSI-Prevenzione Investigators (Gruppo Italiano per lo Studio della
daily for four to six years had no beneficial effects on Sopravvivenza nell’Infarto Miocardico). Dietary supplementation with n-3
cardiovascular outcomes in a high-risk population of polyunsaturated fatty acids and vitamin E after myocardial infarction: re-
sults of the GISSI-Prevenzione trial. Lancet 1999;354:447-55.
patients who were 55 years of age or older. Vita- 17. The Heart Outcomes Prevention Evaluation Study Investigators. Ef-
min E was well tolerated, with no significant adverse fects of an angiotensin-converting–enzyme inhibitor, ramipril, on cardio-
events as compared with placebo. This finding pro- vascular events in high-risk patients. N Engl J Med 2000;342:145-53.
18. The HOPE Study Investigators. The HOPE (Heart Outcomes Pre-
vides some reassurance for the conduct of large, vention Evaluation) Study: the design of a large, simple randomized trial
longer-term trials to address unanswered questions of an angiotensin-converting enzyme inhibitor (ramipril) and vitamin E in
regarding vitamin E, such as its possible effects in patients at high risk of cardiovascular events. Can J Cardiol 1996;12:127-
37.
preventing cancer. 19. Rapola JM, Virtamo J, Ripatti S, et al. Randomised trial of a-tocoph-
erol and b-carotene supplements on incidence of major coronary events in
men with previous myocardial infarction. Lancet 1997;349:1715-20.
Funded by the Medical Research Council of Canada, Natural Source Vi- 20. Virtamo J, Rapola JM, Ripatti S, et al. Effect of vitamin E and beta
tamin E Association, Negma, Hoechst–Marion Roussel, AstraZeneca, King carotene on the incidence of primary nonfatal myocardial infarction and
Pharmaceuticals, and the Heart and Stroke Foundation of Ontario. Dr. Yusuf fatal coronary heart disease. Arch Intern Med 1998;158:668-75.
was supported by a Senior Scientist Award of the Medical Research Coun- 21. Yusuf S, Peto R, Lewis J, Collins R, Sleight P. Beta blockade during
cil of Canada and a Heart and Stroke Foundation of Ontario Research and after myocardial infarction: an overview of the randomized trials. Prog
Chair. Cardiovasc Dis 1985;27:335-71.

Vol ume 342 Numb e r 3 · 159


The Ne w E n g l a nd Jo u r n a l o f Me d ic i ne

22. Steinberg D. Clinical trials of antioxidants in atherosclerosis: are we 26. Hercberg S, Galan P, Preziosi P, et al. Background and rationale be-
doing the right thing? Lancet 1995;346:36-8. hind the SU.VI.MAX Study, a prevention trial using nutritional doses of a
23. Hennekens CH, Buring JE, Manson JE, et al. Lack of effect of long- combination of antioxidant and minerals to reduce cardiovascular diseases
term supplementation with beta carotene on the incidence of malignant and cancers: SUpplementation en VItamines et Mineraux AntioXydants
neoplasms and cardiovascular disease. N Engl J Med 1996;334:1145-9. Study. Int J Vitam Nutr Res 1998;68:3-20.
24. Lonn EM, Yusuf S, Doris CI, et al. Study design and baseline charac- 27. MRC/BHF Heart Protection Study of cholesterol-lowering therapy
teristics of the Study to Evaluate Carotid Ultrasound Changes in Patients and of antioxidant vitamin supplementation in a wide range of patients at
Treated with Ramipril and Vitamin E: SECURE. Am J Cardiol 1996;78: increased risk of coronary heart disease death: early safety and efficacy ex-
914-9. perience. Eur Heart J 1999;20:725-41.
25. Upston JM, Terentis AC, Stocker R. Tocopherol-mediated peroxida- 28. Jha P, Flather M, Lonn E, Farkouh M, Yusuf S. The antioxidant vita-
tion of lipoproteins: implications for vitamin E as a potential antiathero- mins and cardiovascular disease: a critical review of epidemiologic and clin-
genic supplement. FASEB J 1999;13:977-94. ical trial data. Ann Intern Med 1995;123:860-72.

160 · Ja nu ar y 2 0 , 2 0 0 0
new england
The
journal of medicine
established in 1812 march 17 , 2005 vol. 352 no. 11

Cardiovascular Risk Associated with Celecoxib in a Clinical Trial


for Colorectal Adenoma Prevention
Scott D. Solomon, M.D., John J.V. McMurray, M.D., Marc A. Pfeffer, M.D., Ph.D., Janet Wittes, Ph.D.,
Robert Fowler, M.S., Peter Finn, M.D., William F. Anderson, M.D., M.P.H., Ann Zauber, Ph.D.,
Ernest Hawk, M.D., M.P.H., and Monica Bertagnolli, M.D.,
for the Adenoma Prevention with Celecoxib (APC) Study Investigators*

abstract

background
Selective cyclooxygenase-2 (COX-2) inhibitors have come under scrutiny because of re- From the Cardiovascular Division, Depart-
ments of Medicine (S.D.S., M.A.P., P.F.)
ports suggesting an increased cardiovascular risk associated with their use. Experimen-
and Surgery (M.B.), Brigham and Women’s
tal research suggesting that these drugs may contribute to a prothrombotic state pro- Hospital, Harvard Medical School, Boston;
vides support for this concern. Western Infirmary, University of Glasgow,
Glasgow, Scotland (J.J.V.M.); Statistics Col-
laborative, Washington, D.C. (J.W., R.F.);
methods National Cancer Institute, Bethesda, Md.
We reviewed all potentially serious cardiovascular events among 2035 patients with a (W.F.A., E.H.); and Memorial Sloan-Ketter-
ing Cancer Center, New York (A.Z.). Address
history of colorectal neoplasia who were enrolled in a trial comparing two doses of
reprint requests to Dr. Solomon at the Car-
celecoxib (200 mg or 400 mg twice daily) with placebo for the prevention of colorectal diovascular Division, Brigham and Women’s
adenomas. All deaths were categorized as cardiovascular or noncardiovascular, and Hospital, 75 Francis St., Boston, MA 02115,
or at ssolomon@rics.bwh.harvard.edu.
nonfatal cardiovascular events were categorized in a blinded fashion according to a
prespecified scheme. *Participants in the APC study are listed
in the Appendix.
results
This article was published at www.nejm.
For all patients except those who died, 2.8 to 3.1 years of follow-up data were available. org on February 15, 2005.
A composite cardiovascular end point of death from cardiovascular causes, myocardial
N Engl J Med 2005;352:1071-80.
infarction, stroke, or heart failure was reached in 7 of 679 patients in the placebo group
Copyright © 2005 Massachusetts Medical Society.
(1.0 percent), as compared with 16 of 685 patients receiving 200 mg of celecoxib twice
daily (2.3 percent; hazard ratio, 2.3; 95 percent confidence interval, 0.9 to 5.5) and with
23 of 671 patients receiving 400 mg of celecoxib twice daily (3.4 percent; hazard ratio,
3.4; 95 percent confidence interval, 1.4 to 7.8). Similar trends were observed for other
composite end points. On the basis of these observations, the data and safety monitor-
ing board recommended early discontinuation of the study drug.

conclusions
Celecoxib use was associated with a dose-related increase in the composite end point
of death from cardiovascular causes, myocardial infarction, stroke, or heart failure. In
light of recent reports of cardiovascular harm associated with treatment with other
agents in this class, these data provide further evidence that the use of COX-2 inhibitors
may increase the risk of serious cardiovascular events.

n engl j med 352;11 www.nejm.org march 17, 2005 1071

Downloaded from www.nejm.org on March 18, 2005 . This article is being provided free of charge for use in Philippines.
Copyright © 2005 Massachusetts Medical Society. All rights reserved.
The new england journal of medicine

enomatous polyps in the colon and rectum one year

t he promise of a lower incidence


of gastrointestinal side effects with the use
of selective cyclooxygenase-2 (COX-2) in-
hibitors than with the use of nonselective nonste-
roidal antiinflammatory drugs (NSAIDs) or aspirin
and three years after endoscopic polypectomy. The
trial was led by the Strang Cancer Prevention Cen-
ter (New York) and cosponsored by the National
Cancer Institute and Pfizer. Ninety-one sites par-
has led to a marked increase in prescriptions for ticipated (72 in the United States, 1 in the United
COX-2 inhibitors, despite the fact that they offer Kingdom, 8 in Australia, and 10 in Canada). Partici-
similar degrees of pain relief.1-3 In addition, the pants ranged from 32 to 88 years of age and were
identification of COX-2 as a promoter of intestinal considered to have a clinically significant risk of
tumorigenesis suggested that inhibiting this en- colorectal adenoma on the basis of a history of ei-
zyme could prevent the formation of premalignant ther multiple adenomas or a single adenoma that
colorectal adenomas.4-8 Recently, however, this class was at least 0.5 cm in diameter. All known adeno-
of drugs has come under scrutiny because of clini- mas were removed colonoscopically before drug
cal reports that they were associated with an in- treatment began.
creased risk of serious cardiovascular harm.9-11 The A detailed medical history, including baseline as-
mechanism of this effect is suggested in part by ev- sessment of cardiovascular disease status and risk
idence that selective inhibition of COX-2 can block factors for cardiovascular disease, was obtained for
the production of prostacyclin without affecting each patient. The protocol was reviewed and ap-
the synthesis of thromboxane A2,10 thereby poten- proved by the appropriate institutional review
tially creating a prothrombotic state. boards, and all patients provided written informed
The observation of an increased incidence of consent before enrollment. Patients were randomly
death from cardiovascular causes, myocardial in- assigned to treatment with the use of a computer-
farction, or stroke among patients receiving rofe- generated randomization schedule. At the time of
coxib in the Adenomatous Polyp Prevention on data review, 2035 patients had undergone random-
Vioxx (APPROVe) trial and the associated voluntary ization in a double-blind manner at a 1:1:1 ratio,
withdrawal of this drug from the market prompted after stratification according to the use or nonuse of
the data and safety monitoring board and steering aspirin for cardiovascular prophylaxis and the en-
committee of a similar ongoing trial of celecoxib to rolling center. Enrollment began in November 1999
request a focused reassessment of data on cardio- and concluded in March 2002. Compliance was as-
vascular safety by an independent committee, with sessed by means of both pill counts and standard
the results presented at their scheduled meeting on monitoring of medical records every 6 to 12 weeks.
December 10, 2004. The study was a prospective,
randomized, double-blind, multicenter trial assess- review of cardiovascular safety
ing the efficacy of celecoxib for the prevention of The cardiovascular safety committee developed end-
adenomatous polyps in patients who had under- point definitions as guidelines for adjudication. The
gone endoscopic polypectomy. Because neither committee classified and adjudicated the end points
prior clinical trials nor observational studies had by defining a hierarchy of composite end points
reported a clearly increased risk of cardiovascular (based on clinical importance and the prior findings
events with celecoxib use,2,5,12-16 this longer-term, with rofecoxib). These guidelines were designed
placebo-controlled trial provided an important op- specifically to assess cardiovascular safety (listed
portunity to evaluate the potential association. This in the Supplementary Appendix, available with the
report describes the findings of the independent full text of this article at www.nejm.org). An initial
cardiovascular safety committee. review identified all deaths and potential nonfatal
cardiovascular adverse events. Two experienced in-
methods dependent assessors reviewed these events using
medical records and narratives supplied by site in-
patients vestigators. Myocardial infarction was defined on
The Adenoma Prevention with Celecoxib (APC) the basis of either a clinical presentation character-
study compared the efficacy and safety of 200 mg of ized by typical symptoms, signs, or electrocardio-
celecoxib twice daily, 400 mg of celecoxib twice graphic changes associated with an elevation in
daily, and placebo in reducing the occurrence of ad- the level of a cardiac marker or angiographic evi-

1072 n engl j med 352;11 www.nejm.org march 17 , 2005

Downloaded from www.nejm.org on March 18, 2005 . This article is being provided free of charge for use in Philippines.
Copyright © 2005 Massachusetts Medical Society. All rights reserved.
celecoxib and cardiovascular risk

dence of coronary thrombosis. Stroke was defined available at the time of the original analysis. This
as a persistent focal neurologic event whose onset analysis contains data on three additional cardio-
was sudden and was not due to trauma or a tumor. vascular events that were not included in the origi-
Other cardiovascular events were categorized ac- nal report.
cording to a preplanned schema. When this initial
documentation was insufficient for adjudication, results
additional information was obtained from the in-
vestigative sites. At the time of the analysis, 77 percent of the 2035
The entire cardiovascular safety committee was patients had completed the study, and all of the re-
unaware of the patients’ treatment assignments maining surviving patients had completed at least
throughout the review process. For the purposes of 2.8 years of follow-up (range, 2.8 to 3.1). The base-
this analysis, we evaluated a hierarchy of compos- line characteristics were similar among the three
ite end points, including death from cardiovascular groups (Table 1). The incidence of the prespecified
causes, myocardial infarction, stroke, heart failure, composite cardiovascular end points, analyzed ac-
unstable angina, and the need for a cardiovascular cording to the time to the first event, and the asso-
procedure. ciated hazard ratios are shown in Table 2. As com-
pared with the placebo group, the group given 200
statistical analysis mg of celecoxib twice daily had a hazard ratio for
Randomization codes were provided to Statistics death from cardiovascular causes, myocardial in-
Collaborative (Washington, D.C.). All analyses were farction, stroke, or heart failure of 2.3 (95 percent
performed according to the intention-to-treat prin- confidence interval, 0.9 to 5.5), and the group re-
ciple, with data on each patient analyzed according ceiving 400 mg of celecoxib twice daily had a hazard
to the original randomized treatment assignment. ratio of 3.4 (95 percent confidence interval, 1.4 to
Log-rank tests were used to compare the time to a 7.8). The results for the individual components of
cardiovascular event in the three groups for each the composite end point are shown in Table 3.
composite end point of interest. Cox models, with There were six deaths in the placebo group, six in
the treatment group as the only covariate, were the group given 200 mg of celecoxib twice daily, and
used to estimate hazard ratios for the two celecoxib nine in the group given 400 mg twice daily, and one,
groups as compared with the placebo group. Al- three, and six of the deaths, respectively, were due
though the randomization was stratified according to cardiovascular causes. The Kaplan–Meier curves
to the baseline use or nonuse of aspirin and the cen- for the combined end point of death from cardio-
ter, the Cox models did not include these stratifying vascular causes, myocardial infarction, stroke, or
variables. Censoring was defined by assuming that heart failure in the three groups are shown in Fig-
a patient was followed for 37 months, until death, ure 1. The annualized incidence of death from car-
or until January 6, 2005 (the date defined for this diovascular causes, stroke, myocardial infarction, or
analysis as the common close-out date) — which- heart failure was 3.4 events per 1000 patient-years
ever came first. At the time of this review, we had in the placebo group, 7.8 events per 1000 patient-
follow-up information for more than 97 percent of years in the group given 200 mg of celecoxib twice
the patient-years at risk. Incidence rates were cal- daily, and 11.4 events per 1000 patient-years in the
culated for individual and composite cardiovascu- group given 400 mg twice daily.
lar events by dividing the number of patients with In addition to the increased risk of the prespec-
events by the number of patient-years at risk. ified composite end point of cardiovascular events,
Important subgroups based on baseline charac- the point estimate of the number of venous throm-
teristics were prespecified. To examine whether the boembolic events was also increased (though not
effect of celecoxib varied between subgroups, we significantly) among patients receiving celecoxib:
constructed Cox models with terms for treatment, four in the group given 400 mg of celecoxib twice
subgroup, and the interaction between subgroup daily and three in the group given 200 mg twice dai-
and treatment and evaluated the interaction terms ly, as compared with one in the placebo group (haz-
for statistical significance. ard ratio for the two celecoxib groups combined,
Recommendations to the study’s data and safety 3.5; 95 percent confidence interval, 0.4 to 28.5).
monitoring board were made on the basis of data There was no apparent increase in the risk of un-

n engl j med 352;11 www.nejm.org march 17, 2005 1073

Downloaded from www.nejm.org on March 18, 2005 . This article is being provided free of charge for use in Philippines.
Copyright © 2005 Massachusetts Medical Society. All rights reserved.
The new england journal of medicine

Table 1. Baseline Characteristics of the Patients.*

Placebo Celecoxib, 200 mg Celecoxib, 400 mg


Characteristic (N=679) Twice Daily (N=685) Twice Daily (N=671)

Age — yr 59.7±9.7 59.7±9.4 59.9±9.4


Male sex — no. (%) 473 (69.7) 460 (67.2) 454 (67.7)
History of cardiovascular events — no. (%) 321 (47.3) 335 (48.9) 307 (45.8)
Myocardial infarction 29 (4.3) 22 (3.2) 31 (4.6)
Cerebrovascular disease 14 (2.1) 20 (2.9) 13 (1.9)
Congestive heart failure 14 (2.1) 6 (0.9) 11 (1.6)
Angina 51 (7.5) 50 (7.3) 42 (6.3)
Hypertension 277 (40.8) 287 (41.9) 260 (38.7)
Diabetes — no. (%)† 61 (9.0) 66 (9.6) 64 (9.5)
Current smoker — no. (%) 122 (18.0) 119 (17.4) 96 (14.3)
Aspirin use — no. (%) 213 (31.4) 201 (29.3) 200 (29.8)
Use of lipid-lowering drug — no. (%) 184 (27.1) 188 (27.4) 191 (28.5)

* Plus–minus values are means ±SD. There were no significant differences among the groups.
† Data were missing for one patient in the placebo group.

stable angina, arrhythmia, or the need for a cardio-treatment with 200 or 400 mg of celecoxib to pre-
vascular procedure. The hazard ratios associated vent colorectal adenomas led to a dose-related in-
with celecoxib use decreased when a broader class crease in the risk of serious cardiovascular events,
of cardiovascular events, including unstable angi- including death from cardiovascular causes, myo-
na and the need for a cardiovascular procedure, cardial infarction, stroke, and heart failure. These
was added to the composite end point. The hazard results were consistent among the individual com-
ratio associated with celecoxib was not significant-ponents of the composite end point. Because the
ly affected by any of the baseline characteristics ex-
use of other selective COX-2 inhibitors, including
amined, including aspirin use at baseline (Table 4).rofecoxib, valdecoxib, and parecoxib, has also been
On December 16, 2004, on the basis of these associated with an increased rate of cardiovascular
findings, the advice of the cardiovascular safety events,17,18 our results heighten concern that this
committee, and previous findings with drugs in the class of drug may be associated with increased car-
same class, the data and safety monitoring board diovascular risk. The cardiovascular safety commit-
concluded that continued exposure to celecoxib tee also completed a preliminary review of cardio-
vascular safety in another study, the Prevention of
placed patients at increased risk for serious cardio-
vascular events. On the basis of this recommenda- Spontaneous Adenomatous Polyps (PreSAP) trial,
tion, the steering committee stopped the use of which randomly assigned patients with a history of
study medication among the patients remaining in colorectal adenomas to receive either 400 mg of cele-
the trial. The trial remained blinded, and follow-upcoxib once a day or placebo. The preliminary analy-
for the end point of adenoma continued. Three sis did not show an increase in risk at this dose.
events that were documented after the study was The reason for the apparent increase in cardio-
stopped are included in the present analysis; their vascular risk associated with the use of COX-2 in-
inclusion does not alter the overall conclusions of hibitors is uncertain. One prominent hypothesis
the report issued on December 16, 2004. involves the effects of COX-2 inhibitors on two
key prostanoids, prostacyclin and thromboxane
discussion A2, which have a crucial role in vascular homeo-
stasis.9,19,20 These prostanoids are generated by
In a large, randomized, placebo-controlled, double- the action of the cyclooxygenase-1 (COX-1) and
blind, multicenter trial, we found that twice-daily COX-2 isoenzymes on arachidonic acid.21 Throm-

1074 n engl j med 352;11 www.nejm.org march 17 , 2005

Downloaded from www.nejm.org on March 18, 2005 . This article is being provided free of charge for use in Philippines.
Copyright © 2005 Massachusetts Medical Society. All rights reserved.
Table 2. Incidence of and Hazard Ratios for the Composite End Points in the Celecoxib Groups Relative to the Placebo Group.

Celecoxib, Celecoxib, Both Celecoxib, Celecoxib, Both Celecoxib, Celecoxib, Both


200 mg 400 mg Celecoxib 200 mg 400 mg Celecoxib 200 mg 400 mg Celecoxib
Placebo Twice Daily Twice Daily Groups Placebo Twice Daily Twice Daily Groups Twice Daily Twice Daily Groups
End Point* (N=679) (N=685) (N=671) (N=1356) (N=679) (N=685) (N=671) (N=1356) (N=685) (N=671) (N=1356)
number of patients (percent) rate/1000 patient-years hazard ratio (95% confidence interval)
Death from cardiovas- 1 (0.1) 3 (0.4) 6 (0.9) 9 (0.7) 0.5 1.4 2.9 2.2 3.0 (0.3–28.6) 6.1 (0.7–50.3) 4.5 (0.6–35.5)
cular causes
Death from cardiovas- 4 (0.6) 12 (1.8) 15 (2.2) 27 (2.0) 1.9 5.8 7.4 6.6 3.0 (1.0–9.3) 3.8 (1.3–11.5) 3.4 (1.2–9.7)
cular causes or non-
fatal MI

n engl j med 352;11


Death from cardiovas- 6 (0.9) 15 (2.2) 20 (3.0) 35 (2.6) 2.9 7.3 9.9 8.6 2.5 (1.0–6.4) 3.4 (1.4–8.5) 2.9 (1.2–7.0)
cular causes, non-
fatal MI, or stroke
Death from cardiovas- 7 (1.0) 16 (2.3) 23 (3.4) 39 (2.9) 3.4 7.8 11.4 9.6 2.3 (0.9–5.5) 3.4 (1.4–7.8) 2.8 (1.3–6.3)
cular causes, non-
fatal MI, stroke,

www.nejm.org
or heart failure
Death from cardiovas- 11 (1.6) 18 (2.6) 25 (3.7) 43 (3.2) 5.4 8.7 12.5 10. 6 1.6 (0.8–3.4) 2.3 (1.1–4.7) 2.0 (1.0–3.8)
cular causes, non-
fatal MI, stroke, heart
failure, or angina
celecoxib and cardiovascular risk

Death from cardiovas- 17 (2.5) 26 (3.8) 31 (4.6) 57 (4.2) 8.4 12.7 15.5 14.1 1.5 (0.8–2.8) 1.9 (1.0–3.3) 1.7 (1.0–2.9)

march 17, 2005


cular causes, non-
fatal MI, stroke, heart
failure, or angina
or need for a cardio-
vascular procedure

Copyright © 2005 Massachusetts Medical Society. All rights reserved.


* MI denotes myocardial infarction.

Downloaded from www.nejm.org on March 18, 2005 . This article is being provided free of charge for use in Philippines.
1075
The new england journal of medicine

Table 3. Incidence of Individual Cardiovascular and Fatal Events.

Placebo Celecoxib, 200 mg Celecoxib, 400 mg Both Celecoxib


End Point (N=679) Twice Daily (N=685) Twice Daily (N=671) Groups (N=1356)

number of patients (percent)


Death from any cause 6 (0.9) 6 (0.9) 9 (1.3) 15 (1.1)
Death from cardiovascular causes 1 (0.1) 3 (0.4) 6 (0.9) 9 (0.7)
Death from noncardiovascular causes 5 (0.7) 3 (0.4) 3 (0.4) 6 (0.4)
Nonfatal cardiovascular events
Myocardial infarction 3 (0.4) 9 (1.3) 9 (1.3) 18 (1.3)
Stroke 3 (0.4) 3 (0.4) 5 (0.7) 8 (0.6)
Heart failure 2 (0.3) 1 (0.1) 4 (0.6) 5 (0.4)
Thromboembolic event 1 (0.1) 3 (0.4) 4 (0.6) 7 (0.5)
Resuscitation after sudden 0 0 1 (0.1) 1 (0.1)
cardiac arrest
Hospitalization for unstable 5 (0.7) 4 (0.6) 2 (0.3) 6 (0.4)
angina
Arrhythmia 9 (1.3) 4 (0.6) 7 (1.0) 11 (0.8)
Cardiovascular procedure 7 (1.0) 9 (1.3) 6 (0.9) 15 (1.1)
Other 9 (1.3) 11 (1.6) 14 (2.1) 25 (1.8)

boxane A2, which promotes platelet aggregation, however, were generally short-term studies designed
vasoconstriction, and smooth-muscle proliferation, to assess the use of this class of drug for pain relief
is synthesized primarily in platelets, which express and to evaluate associated adverse gastrointestinal
only COX-1. Conversely, prostacyclin, which has an- events. They included a relatively small proportion
tiaggregative, antiproliferative, and vasodilatory ac- of patients at high risk for cardiovascular events or
tions, is the main prostanoid product of endothelial excluded such patients, despite the fact that many
cells, synthesized as a result of the action of COX-2.22 patients who are taking these drugs or who are con-
Whereas nonselective NSAIDs inhibit both sidered candidates for this therapy are at high car-
COX-1 and COX-2, selective COX-2 inhibitors act diovascular risk.25 Consequently, the studies lacked
primarily on COX-2.9 The selective COX-2 inhibi- adequate statistical power to confirm or refute a car-
tors may therefore suppress vascular production diovascular hazard related to the use of COX-2 in-
of prostacyclin without affecting the synthesis of hibitors.11 The use of active rather than placebo con-
platelet-derived thromboxane A2. This imbalance trols in many of these studies also made the findings
may promote thrombosis and increase the risk of difficult to interpret.
cardiovascular events.10 Nonaspirin, nonselective The results of the Vioxx Gastrointestinal Out-
NSAIDs may also not sufficiently reduce throm- comes Research (VIGOR) trial3 and a subsequent
boxane A2 synthesis long enough to prevent plate- study, APPROVe,26 raised questions about the safety
let aggregation and atherosclerotic events.10 Other of rofecoxib. The VIGOR trial, which compared a
potentially detrimental effects of COX-2 inhibi- nine-month course of 50 mg of rofecoxib per day
tors have been suggested, including elevated blood (a larger dose than that usually recommended for
pressure, though some reports have indicated that the long-term treatment of arthritis) with naproxen
these drugs may have beneficial effects on vascu- in patients with rheumatoid arthritis, reported a
lar health.23 higher risk of myocardial infarction among the pa-
In contrast to our findings, most of the earlier tients receiving rofecoxib.27 Some have attributed
clinical trials of selective COX-2 inhibitors in pa- these findings to the potentially cardioprotective ef-
tients with arthritis did not appear to show an in- fects of naproxen,28,29 although this interpretation
crease in cardiovascular risk.2,5,14,24 These trials, has been a source of contention.18,20

1076 n engl j med 352;11 www.nejm.org march 17 , 2005

Downloaded from www.nejm.org on March 18, 2005 . This article is being provided free of charge for use in Philippines.
Copyright © 2005 Massachusetts Medical Society. All rights reserved.
celecoxib and cardiovascular risk

More recently, the APPROVe trial, a randomized,


placebo-controlled trial designed to evaluate the ef- 0.050
0.045

of Composite End Point


ficacy of rofecoxib for preventing colorectal polyps

Estimated Probability
0.040
Celecoxib, 400 mg
in patients with a history of colorectal adenomas, 0.035
0.030
was terminated early because of an increased risk
0.025
of cardiovascular events.10,26 These results prompt- P=0.01
0.020
Celecoxib, 200 mg
ed voluntary withdrawal of rofecoxib from the mar- 0.015
Placebo
0.010
ket. Topol reported that another controlled trial also 0.005
showed an increased risk of cardiovascular events 0.000
with treatment with 12.5 mg of rofecoxib per day, 0 6 12 18 24 30 36

as compared with nabumetone or placebo.30 Months after First Dose


The results of other studies have aroused con- No. at Risk
Celecoxib, 400 mg 671 669 665 655 651 648 576
cern about the safety of selective COX-2 inhibitors. Celecoxib, 200 mg 685 681 676 675 673 670 595
In a placebo-controlled trial of pain relief after cor- Placebo 679 677 675 672 668 667 585
onary-artery bypass surgery, the use of the paren-
teral COX-2 inhibitor parecoxib followed by oral Figure 1. Kaplan–Meier Estimates of the Risk of the Composite End Point
treatment with its active metabolite valdecoxib, or of Death from Cardiovascular Causes, Myocardial Infarction, Stroke,
or Heart Failure among Patients Who Received Celecoxib (200 mg Twice
treatment with placebo followed by valdecoxib, was
Daily or 400 mg Twice Daily) or Placebo.
associated with a significantly increased risk of car-
The log-rank statistic of 8.73, which has two degrees of freedom, was used
diovascular thromboembolic events.31 In this issue to determine the P value.
of the Journal, Nussmeier et al. report a second trial
showing a significant increase in cardiovascular
events when parecoxib and valdecoxib were used in
the immediate postoperative period after coronary- an increased cardiac risk associated with celecoxib
artery bypass surgery.32 The Therapeutic Arthritis use in non–aspirin users, as compared with those
Research and Gastrointestinal Event Trial (TARGET) taking ibuprofen (but not diclofenac).20 Moreover,
also showed a nonsignificant increase in the risk of the results of a randomized, controlled clinical trial
cardiovascular events with lumiracoxib therapy,10 as of celecoxib in patients with Alzheimer’s disease,
compared with naproxen or ibuprofen therapy, but reported to the Food and Drug Administration,
only among patients who were not taking aspirin. demonstrated an increase in cardiovascular events
In contrast, to our knowledge, neither phar- among patients receiving celecoxib.33
macoepidemiologic studies nor randomized, con- Although we found that patients with an in-
trolled trials have reported clear evidence of an creased cardiovascular risk at baseline appeared to
increased cardiovascular risk associated with cele- have a higher absolute rate of events than those with
coxib. The failure of pharmacoepidemiologic stud- no increase in cardiovascular risk at baseline, for-
ies to show an increased risk may be due in part to mal statistical tests of interaction showed no dif-
the lower doses and shorter duration of use in these ferential effect of celecoxib with respect to baseline
studies than in clinical trials and in part to the po- cardiovascular risk. One prespecified subgroup in-
tential for selection bias in nonrandomized studies. cluded users of cardioprotective aspirin at base-
Nevertheless, the Celecoxib in Long-term Arthritis line. Although the overall absolute risk appeared to
Safety Study (CLASS),2 which used the same dose be higher among such patients, analysis of the data
of celecoxib (400 mg twice daily) that was given to on aspirin users in this study shows that they had a
one group in the APC study and compared celecoxib higher frequency of cardiovascular risk factors at
with two nonselective NSAIDs, did not show an in- baseline than did nonusers.
creased rate of cardiovascular events.2 CLASS dif- The cardiovascular findings with regard to cele-
fered from the VIGOR study in several important coxib use in the APC study are consistent with those
ways. A short-term study not designed for system- identified for rofecoxib use in the APPROVe trial. In
atic and formal assessment of cardiovascular contrast, preliminary analyses from the PreSAP trial,
events, CLASS enrolled relatively low-risk patients which involved a daily dose of 400 mg of celecoxib,
and allowed the use of aspirin for cardiovascular showed no apparent increase in cardiovascular risk.
protection. In addition, FitzGerald has suggested The differences in the dosing regimens between
that CLASS did not completely refute evidence of these two trials — twice daily in the APC study, as

n engl j med 352;11 www.nejm.org march 17, 2005 1077

Downloaded from www.nejm.org on March 18, 2005 . This article is being provided free of charge for use in Philippines.
Copyright © 2005 Massachusetts Medical Society. All rights reserved.
The new england journal of medicine

Table 4. Incidence of Death from Cardiovascular Causes, Myocardial Infarction, Stroke, or Heart Failure According to
Baseline Characteristics.

No. of Both Celecoxib Hazard Ratio P Value for


Subgroup Patients Placebo Groups (95% CI)* Interaction

no./total no. (%)


Age 0.61
<60 yr 1078 4/372 (1.1) 14/706 (2.0) 1.8 (0.6–5.6)
≥60 yr 957 3/307 (1.0) 25/650 (3.8) 4.0 (1.2–13.2)
Sex 0.55
Female 648 2/206 (1.0) 8/442 (1.8) 1.9 (0.4–8.7)
Male 1387 5/473 (1.1) 31/914 (3.4) 3.2 (1.3–8.3)
Baseline cardiovascular risk factors 0.44
Yes 963 4/321 (1.2) 28/642 (4.4) 3.5 (1.2–10.1)
No 1072 3/358 (0.8) 11/714 (1.5) 1.8 (0.5–6.6)
Diabetes 0.86
Yes 191 1/61 (1.6) 5/130 (3.8) 2.3 (0.3–19.9)
No 1843 6/617 (1.0) 34/1226 (2.8) 2.9 (1.2–6.8)
Aspirin use 0.63
Yes 614 2/213 (0.9) 14/401 (3.5) 3.8 (0.9–16.6)
No 1421 5/466 (1.1) 25/955 (2.6) 2.4 (0.9–6.4)
Use of lipid-lowering drug 0.79
Yes 563 3/184 (1.6) 15/379 (4.0) 2.4 (0.7–8.4)
No 1472 4/495 (0.8) 24/977 (2.5) 3.1 (1.1–8.8)

* CI denotes confidence interval.

compared with once daily in the PreSAP study — has substantial implications for public health,11,34
support the hypothesis that sustained inhibition of patient education,35 and drug regulation.36,37 Given
prostacyclin may contribute to the increase in car- the experience with COX-2 inhibitors, we support
diovascular risk. Other potential differences in the the call for regulatory agencies to consider request-
trials, including geographic differences, differences ing a formal evaluation of long-term cardiovascu-
in the patient population, and differences in use of lar outcomes of any new drug with a mechanism of
concomitant medications, may have contributed to action that could augment the risk of cardiac and
the disparity in the preliminary findings. vascular events, especially if many patients who are
The increased cardiovascular risk in the APC trial likely to use the new agent are prone to cardiovas-
was based on a small number of events in a trial that cular disease.25 This category may include nonse-
was not designed or statistically powered to evalu- lective NSAIDs (other than aspirin), as discussed
ate cardiovascular risk. Although we believe we have earlier. More broadly, this experience underscores
identified all adverse cardiovascular events, we can- both the need for long-term, placebo-controlled tri-
not rule out the possibility that some events re- als to assess safety as well as efficacy and the need
mained unreported. Our results must therefore be to improve methods for assessing potential adverse
interpreted with caution. cardiovascular outcomes in studies with noncar-
Still, in the context of the results of the other tri- diovascular primary end points.
als reviewed involving agents in the same class, In summary, a blinded review of cardiovascular
these data suggest that there may be a real increase events in a large, randomized, controlled study of
in cardiovascular risk associated with the use of cele- two doses of celecoxib for the prevention of colorec-
coxib in particular and the class of selective COX-2 tal adenomas showed a dose-related risk of such
inhibitors in general. If correct, this interpretation events, including death from cardiovascular causes,

1078 n engl j med 352;11 www.nejm.org march 17 , 2005

Downloaded from www.nejm.org on March 18, 2005 . This article is being provided free of charge for use in Philippines.
Copyright © 2005 Massachusetts Medical Society. All rights reserved.
celecoxib and cardiovascular risk

myocardial infarction, stroke, and heart failure. In ib in preventing colorectal neoplasia and in reliev-
light of other recent reports of the adverse cardio- ing pain.
vascular effects of other agents in this class, these The APC was sponsored by the National Cancer Institute and co-
data provide further evidence that long-term use of sponsored by Pfizer. This cardiovascular review was funded solely by
the National Cancer Institute.
COX-2 inhibitors may increase the risk of serious Drs. McMurray, Pfeffer, and Zauber report having received con-
cardiovascular events. These risks will need to be sulting fees from Pfizer. Drs. Solomon, McMurray, and Pfeffer re-
weighed against any potential benefits of celecox- port having received lecture fees from Pfizer. Dr. Wittes reports hav-
ing received consulting fees from Merck within the past two years.

appendix
The following persons participated in the APC Study: Steering Committee: M.M. Bertagnolli, E. Hawk, C. Eagle; Statistical Team: A. Zaub-
er, K.M. Kim, D. Corle, R. Rosenstein, J. Tang, T. Hess, A. Wilton; Medical Monitors: W. Anderson, L. Doody; Central Pathology Review: M.
Redston; Project Directors: M. Woloj, D. Bagheri, A. Crawford, M. Schietrum, V. Ladouceur; Data and Safety Monitoring Board: S. Rosen
(chair), L. Friedman, R. Makuch, R. Phillips, P. Taylor; Principal Investigators: United States: S. Auerbach (California Professional Re-
search, Newport Beach), C.F. Barish (Wake Research Associates, Raleigh, N.C.), T. Barringer (Carolinas Medical Center, Charlotte, N.C.),
R.W. Bennetts (Northwest Gastroenterology Clinic, Portland, Oreg.), M. Blitstein (Associates in Gastroenterology and Liver Disease, Lake
Forest, Ill.), J. Bruggen (Wake Forest University Baptist Medical Center, Winston-Salem, N.C.), P. Carricaburu (Veterans Affairs Hospital,
Sheridan, Wyo.), D. Chung (Massachusetts General Hospital, Boston), F. Colizzo (Pentucket Medical Associates, Haverhill, Mass.), R. Cur-
tis (Newton–Wellesley Hospital, Newton, Mass.), T. Dewar (Harris Methodist Hospital Fort Worth, Ft. Worth, Tex.), R. DuBois (Vanderbilt
University Medical Center, Nashville), T. Feinstat (Gastroenterology Consultants of Sacramento, Roseville, Calif.), T.R. Foley (Regional
Gastroenterology Associates of Lancaster, Lancaster, Pa.), D. Gabbaizadeh (Huntington Research Group, Huntington Station, N.Y.), J.
Geenen (Wisconsin Center for Advanced Research, Milwaukee), F. Giardiello (Johns Hopkins Hospital, Baltimore), A. Goetsch (nTouch Re-
search, Huntsville, Ala.), M. Goldberg (Regional Gastroenterology Associates of Lancaster, Evanston, Ill.), J.L. Goldstein (University of Illi-
nois at Chicago, Chicago), W. Harlan, III (Asheville Gastroenterology Associates, Asheville, N.C.), R. Hogan (Gastrointestinal Associates,
Jackson, Miss.), M. Kamionkowski (Gastroenterology Associates of Cleveland, Mayfield Heights, Ohio), M. Kelfer (Fallon Clinic, West
Boylston, Mass.), B. Kerzner (Health Trends Research, Baltimore), K. Kim (University of Chicago Medical Center, Chicago), I. Klimberg
(Gastroenterology Associates of Ocala, Ocala, Fla.), G. Koval (West Hills Gastroenterology Associates, Portland, Oreg.), C. Krone (Ad-
vanced Clinical Therapeutics, Tucson, Ariz.), S. Krumholz (Waterside Clinical Research, West Palm Beach, Fla.), M.W. Layton (South Puget
Sound Clinical Research Center, Olympia, Wash.), C. Lightdale (Columbia-Presbyterian Medical Center, New York), P.J. Limburg (Mayo
Clinic, Rochester, Minn.), C. Lind (Vanderbilt University Medical Center, Nashville), D. Lipkis (Institute for Health Care Assessment, San
Diego, Calif.), M. Lloyd (Idaho Gastroenterology, Meridian), D. Maccini (Spokane Digestive Disease Center, Spokane, Wash.), F. MacMilan,
Sr. (Pentucket Medical Associates, Haverhill, Mass.), R. Madoff (University of Minnesota, Minneapolis), A. Malik (Advanced Clinical Re-
search, North Providence, R.I.), A. Markowitz (Memorial Sloan-Kettering Cancer Center, New York), R. Marks (Alabama Digestive Re-
search Center, Alabaster), C.J. McDougall (Manhattan Associates, New York), P. Miner (Oklahoma Foundation for Digestive Research,
Oklahoma City), M. Murphy (Southeastern Digestive and Liver Disease Institute, Savannah, Ga.), A. Namias (Gastrointestinal Physicians,
Salem, Mass.), N. Nickl (University of Kentucky Medical Center, Lexington), M. Pochapin (Jay Monahan Center for Gastrointestinal Health,
New York), R.E. Pruitt (Nashville Medical Research Institute, Nashville), J. Puolos (Cumberland Research Associates, Fayetteville, N.C.),
D.S. Riff (AGMG Clinical Research, Anaheim, Calif.), R. Roman (South Denver Gastroenterology, Englewood, Colo.), L. Rubin (New Jersey
Physicians, Passaic), D. Ruff (Healthcare Discoveries, San Antonio, Tex.), M. Safdi (Consultants for Clinical Research, Cincinnati), J. Saltz-
man (Brigham and Women’s Hospital, Boston), B. Salzberg (Atlanta Gastroenterology Associates, Atlanta), J.A. Sattler (Western Clinical
Research, Torrance, Calif.), P. Schleinitz (Americas Doctors Research, Medford, Oreg.), J. Schwartz (Northwest Gastroenterologists, Ar-
lington Heights, Ill.), M. Schwartz (Jupiter Research Association, Jupiter, Fla.), M. Silpa (Gastroenterology Associates of The East Bay Med-
ical Group, Berkeley, Calif.), D. Silvers (Drug Research Services, Metairie, La.), D. Smoot (Howard University Cancer Center, Washington,
D.C.), S. Sontag (Veterans Affairs Medical Center, Hines, Ill.), R.J. Sorrell (Gastroenterology Specialties, Lincoln, Nebr.), D. Stanton (Com-
munity Clinical Trials, Orange, Calif.), J. Sturgeon (Americas Doctors Research, Shawnee Mission, Kans.), J.P. Tracey (Hawthorne Medical
Associates, North Dartmouth, Mass.), T. Werth (Charlotte Gastroenterology and Hepatology, Charlotte, N.C.), C.M. Wilcox (University of
Alabama at Birmingham, Birmingham), R. Wohlman (Northwest Gastroenterology Associates, Bellevue, Wash.), S. Woods (Gastroenter-
ology Associates of Fairfield County, Bridgeport, Conn.); United Kingdom: J. Burn (South Cleveland Hospital, Middlesbrough); Australia:
H. Ee (Sir Charles Gairdner Hospital, Nedlands, W.A.), M. Korman (Monash Medical Centre, Clayton, Victoria), A. Lee (Concord Repatria-
tion and General Hospital, Concord, N.S.W.), B. Leggett (Royal Brisbane Hospital, Herston, Queensland), F. Macrae (Royal Melbourne
Hospital, Melbourne, Victoria), L. Mollison (Freemantle Hospital, Freemantle, W.A.), N. Yeomans (Western Hospital, Footscray, Victo-
ria), G. Young (Flinders Medical Centre, Bedford, S.A.); Canada: G. Aumais (Hospital Maisonneuve-Rosemont, Montreal), R. Bailey (Hys
Medical Centre, Edmonton, Alta.), C. Bernstein (Winnipeg Health Sciences Centre, Winnipeg, Man.), L. Cohen (Sunnybrook and Wom-
en’s Hospital, Toronto), C. Dallaire, R. Dube (Centre Hospitalier Universitaire de Quebec, Quebec, Que.), D. Morgan (McMaster University,
Hamilton, Ont.), T. Sylwestrowicz (St. Paul’s Hospital, Saskatoon, Sask.), G. Van Rosendaal (University of Calgary Health Sciences Centre,
Calgary, Alta.), S.J. Van Zantan (Queen Elizabeth II Health Sciences Centre, Halifax, N.S.).

references
1. Marnett LJ, Rowlinson SW, Goodwin osteoarthritis and rheumatoid arthritis: the 4. Marnett LJ, Kalgutkar AS. Cyclooxygen-
DC, Kalgutkar AS, Lanzo CA. Arachidonic CLASS study: a randomized controlled trial. ase 2 inhibitors: discovery, selectivity and
acid oxygenation by COX-1 and COX-2: JAMA 2000;284:1247-55. the future. Trends Pharmacol Sci 1999;20:
mechanisms of catalysis and inhibition. 3. Bombardier C, Laine L, Reicin A, et al. 465-9.
J Biol Chem 1999;274:22903-6. Comparison of upper gastrointestinal toxic- 5. Steinbach G, Lynch PM, Phillips RKS, et
2. Silverstein FE, Faich G, Goldstein JL, et ity of rofecoxib and naproxen in patients al. The effect of celecoxib, a cyclooxygenase-
al. Gastrointestinal toxicity with celecoxib vs with rheumatoid arthritis. N Engl J Med 2 inhibitor, in familial adenomatous poly-
nonsteroidal anti-inflammatory drugs for 2000;343:1520-8. posis. N Engl J Med 2000;342:1946-52.

n engl j med 352;11 www.nejm.org march 17, 2005 1079

Downloaded from www.nejm.org on March 18, 2005 . This article is being provided free of charge for use in Philippines.
Copyright © 2005 Massachusetts Medical Society. All rights reserved.
celecoxib and cardiovascular risk

6. Giardiello FM, Yang VW, Hylind LM, et Parecoxib, valdecoxib, and cardiovascular 26. Bresalier RS, Sandler RS, Quan H, et al.
al. Primary chemoprevention of familial ad- risk. Circulation 2005;111:249. Cardiovascular events associated with rofe-
enomatous polyposis with sulindac. N Engl 18. Juni P, Nartey L, Reichenbach S, Sterchi coxib in a colorectal adenoma chemopreven-
J Med 2002;346:1054-9. R, Dieppe PA, Egger M. Risk of cardiovascu- tion trial. N Engl J Med 2005;352:1092-102.
7. Baron JA. Epidemiology of non-steroi- lar events and rofecoxib: cumulative meta- 27. Mukherjee D, Nissen SE, Topol EJ. Risk
dal anti-inflammatory drugs and cancer. analysis. Lancet 2004;364:2021-9. of cardiovascular events associated with se-
Prog Exp Tumor Res 2003;37:1-24. 19. Crofford LJ, Lipsky PE, Brooks P, lective COX-2 inhibitors. JAMA 2001;286:
8. Hawk ET, Viner J, Richmond E, Umar A. Abramson SB, Simon LS, van de Putte LB. 954-9.
Non-steroidal anti-inflammatory drugs Basic biology and clinical application of spe- 28. Weir MR, Sperling RS, Reicin A, Gertz
(NSAIDs) for colorectal cancer prevention. cific cyclooxygenase-2 inhibitors. Arthritis BJ. Selective COX-2 inhibition and cardio-
Cancer Chemother Biol Response Modif Rheum 2000;43:4-13. vascular effects: a review of the rofecoxib de-
2003;21:759-89. 20. FitzGerald GA. COX-2 and beyond: ap- velopment program. Am Heart J 2003;146:
9. FitzGerald GA, Patrono C. The coxibs, proaches to prostaglandin inhibition in hu- 591-604.
selective inhibitors of cyclooxygenase-2. man disease. Nat Rev Drug Discov 2003;2: 29. Konstam MA, Weir MR, Reicin A, et al.
N Engl J Med 2001;345:433-42. 879-90. Cardiovascular thrombotic events in con-
10. FitzGerald GA. Coxibs and cardiovascu- 21. McAdam BF, Catella-Lawson F, Mardini trolled, clinical trials of rofecoxib. Circula-
lar disease. N Engl J Med 2004;351:1709-11. IA, Kapoor S, Lawson JA, FitzGerald GA. tion 2001;104:2280-8.
11. Topol EJ, Falk GW. A coxib a day won’t Systemic biosynthesis of prostacyclin by cy- 30. Topol EJ. Rofecoxib, Merck, and the
keep the doctor away. Lancet 2004;364:639- clooxygenase (COX)-2: the human pharma- FDA. N Engl J Med 2004;351:2877-8.
40. cology of a selective inhibitor of COX-2. Proc 31. Ott E, Nussmeier NA, Duke PC, et al. Ef-
12. White WB, Faich G, Borer JS, Makuch Natl Acad Sci U S A 1999;96:272-7. [Erra- ficacy and safety of the cyclooxygenase 2 in-
RW. Cardiovascular thrombotic events in ar- tum, Proc Natl Acad Sci U S A 1999;96: hibitors parecoxib and valdecoxib in pa-
thritis trials of the cyclooxygenase-2 inhibi- 5890.] tients undergoing coronary artery bypass
tor celecoxib. Am J Cardiol 2003;92:411-8. 22. Brock TG, McNish RW, Peters-Golden surgery. J Thorac Cardiovasc Surg 2003;125:
13. Mamdani M, Juurlink DN, Lee DS, et al. M. Arachidonic acid is preferentially metab- 1481-92.
Cyclo-oxygenase-2 inhibitors versus non- olized by cyclooxygenase-2 to prostacyclin 32. Nussmeier NA, Whelton AA, Brown
selective non-steroidal anti-inflammatory and prostaglandin E2. J Biol Chem 1999; MT, et al. Complications of the COX-2 inhib-
drugs and congestive heart failure outcomes 274:11660-6. itors parecoxib and valdecoxib after cardiac
in elderly patients: a population-based co- 23. Chenevard R, Hurlimann D, Bechir M, surgery. N Engl J Med 2005;352:1081-91.
hort study. Lancet 2004;363:1751-6. et al. Selective COX-2 inhibition improves 33. Celecoxib. PhRMA IQ5-97-02-001. (Ac-
14. Solomon DH, Schneeweiss S, Glynn RJ, endothelial function in coronary artery dis- cessed February 18, 2005, at http://www.
et al. Relationship between selective cyclo- ease. Circulation 2003;107:405-9. clinicalstudyresults.org/documents/
oxygenase-2 inhibitors and acute myocar- 24. Farkouh ME, Kirshner H, Harrington company-study_76_0.pdf.)
dial infarction in older adults. Circulation RA, et al. Comparison of lumiracoxib with 34. Topol EJ. Failing the public health —
2004;109:2068-73. naproxen and ibuprofen in the Therapeutic rofecoxib, Merck, and the FDA. N Engl J Med
15. Ray WA, Stein CM, Daugherty JR, Hall Arthritis Research and Gastrointestinal 2004;351:1707-9.
K, Arbogast PG, Griffin MR. COX-2 selective Event Trial (TARGET), cardiovascular out- 35. Dieppe PA, Ebrahim S, Martin RM, Juni
non-steroidal anti-inflammatory drugs and comes: randomised controlled trial. Lancet P. Lessons from the withdrawal of rofecox-
risk of serious coronary heart disease. Lan- 2004;364:675-84. ib. BMJ 2004;329:867-8.
cet 2002;360:1071-3. 25. Zhao SZ, Burke TA, Whelton A, von All- 36. Josefson D. FDA warns Merck over its
16. Mamdani M, Rochon P, Juurlink DN, et men H, Henderson SC. Comparison of the promotion of rofecoxib. BMJ 2001;323:767.
al. Effect of selective cyclooxygenase 2 in- baseline cardiovascular risk profile among 37. Mukherjee D, Topol EJ. Pharmaceutical
hibitors and naproxen on short-term risk of hypertensive patients prescribed COX-2- advertising versus research spending: are
acute myocardial infarction in the elderly. specific inhibitors or nonspecific NSAIDs: profits more important than patients? Am
Arch Intern Med 2003;163:481-6. data from real-life practice. Am J Manag Heart J 2003;146:563-4.
17. Furberg CD, Psaty BM, FitzGerald GA. Care 2002;8:Suppl:S392-S400. Copyright © 2005 Massachusetts Medical Society.

1080 n engl j med 352;11 www.nejm.org march 17 , 2005

Downloaded from www.nejm.org on March 18, 2005 . This article is being provided free of charge for use in Philippines.
Copyright © 2005 Massachusetts Medical Society. All rights reserved.