0021-972X/05/$15.00/0 Printed in U.S.A.

The Journal of Clinical Endocrinology & Metabolism 90(3):15831587 Copyright 2005 by The Endocrine Society doi: 10.1210/jc.2004-0826

Teriparatide Reduces the Fracture Risk Associated with Increasing Number and Severity of Osteoporotic Fractures
J. Christopher Gallagher, Harry K. Genant, Gerald G. Crans, Socorro Juan Vargas, and John H. Krege
Bone Metabolism Section, Creighton University Medical Center (J.C.G.), Omaha, Nebraska 68131; Osteoporosis and Arthritis Research Group, University of California (H.K.G.), San Francisco, California 94143; Synarc, Inc. (H.K.G.), San Francisco, California 94143; Lilly Research Laboratories, Eli Lilly & Company (G.G.C., J.H.K.), Indianapolis, Indiana 46285; and Department of Endocrinology, William W. Backus Hospital (S.J.V.), Norwich, Connecticut 06360
The relationship between prior fractures and risk of new fractures was evaluated in 931 postmenopausal women with prevalent vertebral fractures randomized to daily placebo or teriparatide (20 g) in the Fracture Prevention Trial. The median observation time was 21 months. Among placebo patients with one, two, or three or more prevalent vertebral fractures, 7%, 16%, and 23%, respectively, developed vertebral fractures (by Cochran-Armitage trend test, P < 0.001), and 3%, 9%, and 17% developed moderate or severe vertebral fractures (P < 0.001). Among placebo patients with mild, moderate, or severe prevalent vertebral fractures, 10%, 13%, and 28%, respectively, developed vertebral fractures (P < 0.001), and 4%, 8%, and 23% developed moderate or severe vertebral fractures (P < 0.001). Among placebo patients with zero, one, or two or more prior nonvertebral fragility fractures, 4%, 8%, and 18%, respectively, developed nonvertebral fragility fractures (P < 0.001). In the teriparatide-treated group, there was no significant increase in vertebral or nonvertebral fracture risk in these subgroups. In summary, the number and severity of prevalent vertebral fractures independently predicted the risk for new vertebral fractures, and the number of prior nonvertebral fractures predicted the risk for new nonvertebral fractures in placebo patients. However, in teriparatidetreated patients, the increased fracture risk associated with prior number and severity of fracture was not observed. (J Clin Endocrinol Metab 90: 15831587, 2005)

HE PRESENCE OF prior osteoporotic fractures is predictive of future fracture risk (13). Patients with increasing number and severity of prevalent vertebral fractures have increasing risk for future vertebral fracture (4, 5). Thus, patients with a history of skeletal pathology, as reflected by previous fractures, are at increased risk for multiple and severe vertebral fractures and for nonvertebral fractures. The majority of current agents for the prevention or treatment of osteoporosis, such as bisphosphonates, calcitonin, estrogens, and selective estrogen receptor modulators (SERMs), reduce fracture risk by reducing bone resorption and preserving bone mass. Teriparatide (rDNA origin) injection [recombinant human PTH-(134)] is a bone formation agent that has been shown to increase bone mass in men and women (6). In the Fracture Prevention Trial, an international, multicenter, randomized, placebo-controlled trial of 1637 postmenopausal women with osteoporosis, treatment with teriparatide (20 g) once daily reduced the risk of new vertebral fractures by 65% and the risk of new moderate or severe vertebral fractures by 90%. The risk of new nonvertebral fragility fractures was reduced by 53%, and the risk of all new nonvertebral fractures, regardless of trauma, was reduced by 35% (6). An analysis of data from this trial was conducted to evaluate the risk for new vertebral fractures
First Published Online December 21, 2004 Abbreviations: BMD, Bone mineral density; SERM, selective estrogen receptor modulator. JCEM is published monthly by The Endocrine Society (http://www. endo-society.org), the foremost professional society serving the endocrine community.

associated with increasing number and severity of prevalent vertebral fractures among women who received placebo or teriparatide (20 g). The risk for new nonvertebral fractures associated with increasing number of prior nonvertebral fragility fractures was also assessed.
Patients and Methods
The analyses include data from women who participated in the Fracture Prevention Trial, a prospective, randomized, double-blind study of teriparatide vs. placebo in postmenopausal women (6). A brief description of the participants and study methods follows.

Study participants and treatment regimens

Patients were eligible for inclusion in the trial if they were ambulatory women at least 5 yr past menopause who had at least one moderate or two mild atraumatic vertebral fractures. For women who had fewer than two moderate vertebral fractures, an additional inclusion criterion included bone mineral density (BMD) of the lumbar spine or proximal femur at least 1 sd below the mean value in healthy young (20 35 yr old) white women. Women with diseases related to bone or calcium metabolism, urolithiasis within the preceding 2 yr, alcohol or drug abuse, impaired hepatic function, or a serum creatinine concentration of 177 m or higher were excluded from the study. Women who had taken drugs affecting bone metabolism within the past 224 months, depending on the drug, were also excluded. The study was approved by the ethics committee at each participating center; all participants gave written informed consent. Women in the Fracture Prevention Trial were randomized to placebo (n 544), 20 g teriparatide (n 541), or 40 g teriparatide (n 552) by daily self-administered sc injection. Participants received daily supplements of calcium (1000 mg) and vitamin D (400 1200 IU). Nonvertebral fractures were assessed at each follow-up visit, scheduled 1, 3, 6, 9, 12, 15, 18, and 21 months after baseline and at the end of the study. Lumbar spine BMD was measured in all participants at baseline, 12 and

1583

1584

J Clin Endocrinol Metab, March 2005, 90(3):15831587

Gallagher et al. Fracture Risk Reduced by Teriparatide

18 months, and the end of the study by dual energy x-ray absorptiometry, as previously described (6). Vertebral fractures were determined from lateral thoracic and lumbar spine radiographs that were obtained at baseline and the end of the study (6). The study was terminated after a median 19-month treatment, because a long-term teriparatide carcinogenicity study in rats revealed the occurrence of skeletal proliferative lesions, including osteosarcoma. These findings were later determined to be unlikely to have significant predictive ability in humans (79), and a dose of 20 g teriparatide was subsequently approved by various regulatory agencies for the treatment of osteoporosis. The present analysis included 931 postmenopausal women who received either placebo (n 464) or the approved clinical dose of 20 g teriparatide (n 467) in the Fracture Prevention Trial and had at least one vertebral fracture at baseline. Vertebral fracture analyses were based on 801 women (placebo, n 398; 20 g teriparatide, n 403) who had postbaseline radiographs that were adequate for evaluation. The protocol required that patients have one or more prevalent vertebral fractures, as assessed by investigative site. However, some placebo (n 50) and teriparatide (n 41) patients were subsequently judged by the central reader not to have prevalent vertebral fractures. Because the vertebral fracture status of these patients is uncertain and the analyses depend on this variable, these data were not included. Also, there were some placebo (n 96) and teriparatide (n 97) patients who lacked a baseline and/or postbaseline radiograph. Hence, their vertebral fracture results were truly missing. We believe that these values were missing at random, because it is unlikely that the unknown results for these subjects depended on any missing data. Furthermore, no statistically significant differences in baseline characteristics were observed between placebo and teriparatide subjects with missing radiographs or those included in the analyses, supporting the belief that the missing data did not impact the results. Results for patients randomized to 40 g teriparatide were similar to those for patients given 20 g teriparatide.

fractures. New nonvertebral fractures were confirmed by review of radiographs or radiological reports.

Statistical analyses
To determine the relationship between prevalent vertebral fracture grade and incident vertebral fractures in both the placebo and treated groups, the Cochran-Armitage trend test, which tests for trends in binomial proportions across levels of either a single factor or a covariate, was performed (11). This statistical test is sensitive to linearity between response and covariates and detects trends that elude less specific statistical methods. This analysis was also performed using the number of prevalent vertebral fractures as the covariate. When a significant trend was found in one group, but not the other, an additional analysis was performed to compare the magnitude of the slope differential (where the slope is calculated from the linear regression between fracture risk and baseline risk factor) by constructing 95% confidence intervals for the difference in slopes between the placebo and teriparatide subjects using the bootstrap technique (12). The above analysis was also used for the new nonvertebral fracture and new nonvertebral fragility fracture end points, using number of prior nonvertebral fragility fractures as the covariate. We also investigated the relationship between the number and severity of prevalent vertebral fractures to determine whether these variables were independent risk factors. This was done using a logistic regression analysis with placebo-treated patients, modeling vertebral fracture risk as a function of the two variables and their interaction. The power of the study to detect a placebo-like risk trend in the teriparatide group was calculated by applying the overall teriparatide vertebral fracture risk reduction (for both any new vertebral fracture and any new moderate or severe vertebral fracture) to the observed placebo proportions to obtain the placebo-like risk trend for the teriparatide group. This was done across prevalent vertebral fracture number and severity risk categories. Simulations were performed based on these proportions and demonstrated that the power of the study to detect placebo-like trends in the teriparatide group was at least 70% for all analyses, indicating that the study was adequately powered to detect a pattern of risk in the teriparatide group similar to that observed in the placebo group.

Assessment of number and severity of vertebral fractures

The primary study end point was the proportion of women developing one or more new vertebral fractures. Evaluation of radiographs of the thoracic and lumbar spine of each participant at baseline and at study end point was performed at a central location (Osteoporosis and Arthritis Research Group, University of California, San Francisco, CA) by radiologists who were unaware of patient treatment assignment but were not blinded to radiograph temporal sequence. Vertebrae were graded individually, using the semiquantitative criteria of Genant et al. (10), as normal (semiquantitative grade 0) or with mild, moderate, or severe deformity (semiquantitative grades 1, 2, and 3, respectively). Briefly, a mild fracture was defined visually as an approximately 20 25% reduction in anterior, middle, or posterior vertebral height. Moderate and severe vertebral fractures were defined as an approximately 25 40% reduction and greater than an approximately 40% reduction in vertebral height, respectively. Vertebrae exhibiting abnormalities such as scoliosis, fusion, or other anomalies were excluded from the analyses. A new vertebral fracture was reported if a previously normal vertebra became deformed. Two variables, derived from the baseline vertebral fracture assessment, were used in the subsequent analyses: the number of fractures (one, two, or three or more) and the maximum severity grade (mild, moderate, or severe),

Results

Of 1085 women who received either 20 g teriparatide or placebo in the Fracture Prevention Trial, baseline radiographs were available for 931 women with prevalent vertebral fractures. Although an existing vertebral fracture was an inclusion criteria, after evaluation of the radiographs by the central reader, it was determined that 91 women (50 in the placebo group and 41 in the teriparatide group) did not have a vertebral fracture. No significant differences existed between women in the placebo group (n 464) and the teriparatide group (n 467) in age, years past menopause, lumbar spine and femoral neck T-scores, number and severity of prevalent vertebral fractures, and number of prior nonvertebral fractures (Table 1).
TABLE 1. Baseline characteristics of women with confirmed prevalent vertebral fractures
Characteristic Placebo (n 464) Teriparatide (n 467)

Assessment of nonvertebral fractures

Patients were queried regarding the occurrence of nonvertebral fractures at each follow-up visit. A nonvertebral fracture was defined as a fracture at any of the following nonvertebral sites: clavicle, scapula, ribs, sacrum, humerus, forearm, wrist, pelvis, femur, patella, tibia, fibula, ankle, calcaneus, tarsus, or metatarsal. Fractures of the skull, face, metacarpals, fingers, and toes were excluded from analysis. Nonvertebral fractures were classified as fragility fractures (the protocol-specified nonvertebral fracture end point) if, in the opinion of the local investigator, the associated trauma would not have resulted in the fracture of a normal bone. Therefore, pathological fractures and fractures resulting from automobile accidents or from falls from greater than standing height were included as nonvertebral fractures, but not as fragility

Age (yr) Years past menopause Lumbar spine BMD T-score Femoral neck BMD T-score No. of vertebral fractures Maximum vertebral fracture grade No. of prior nonvertebral fragility fractures Data are presented as the mean

69 7 21 9 2.5 1.6 2.4 0.8 2.6 1.7 1.9 0.7 0.3 0.8
SD.

70 7 21 9 2.5 1.5 2.4 0.8 2.5 1.7 1.9 0.7 0.4 0.9

Gallagher et al. Fracture Risk Reduced by Teriparatide

J Clin Endocrinol Metab, March 2005, 90(3):15831587

1585

Number of prevalent vertebral fractures and risk of incident vertebral fracture: placebo patients

In placebo patients with one, two, and three or more prevalent vertebral fractures, 6.8%, 15.7%, and 22.6%, respectively, developed new vertebral fractures, whereas 3.0%, 8.8%, and 17.1%, respectively, developed new moderate or severe vertebral fractures (Fig. 1). The Cochran-Armitage trend test revealed a statistically significant association between the number of prevalent vertebral fractures and both incident vertebral fracture risk (P 0.001) and incident moderate or severe vertebral fracture risk (P 0.001).
Baseline vertebral fracture grade and risk of incident vertebral fracture: placebo patients

In placebo patients with mild, moderate, and severe prevalent vertebral fractures, 9.6%, 12.9%, and 28.4%, respectively, developed new vertebral fractures, whereas 4.0%, 7.9%, and 23.2%, respectively, experienced new moderate or severe vertebral fractures (Fig. 2). The Cochran-Armitage trend test revealed a statistically significant association between the severity of prevalent vertebral fractures and both incident vertebral fracture risk (P 0.001) and incident moderate or severe vertebral fracture risk (P 0.001).
Number of prior nonvertebral fragility fractures and risk of incident nonvertebral fracture: placebo patients

FIG. 2. New vertebral fractures in the placebo and 20-g teriparatide groups by increasing severity of prevalent vertebral fracture grade. , Percentage of women with any new vertebral fracture; o, percentage of women with new moderate or severe fractures. The number of women in each baseline category is shown in parentheses below each bar.

In placebo patients with zero, one, and two or more prior nonvertebral fragility fractures, 8.2%, 9.8%, and 25.6%, respectively, developed new nonvertebral fractures, whereas 3.6%, 8.2%, and 18.0%, respectively, experienced new nonvertebral fragility fractures (Fig. 3). The Cochran-Armitage trend test revealed a statistically significant association be-

tween the number of prior nonvertebral fragility fractures and both incident nonvertebral fracture risk (P 0.002) and incident nonvertebral fragility fracture risk (P 0.001).
Number of prevalent vertebral fractures and risk of incident vertebral fracture: teriparatide-treated patients

New vertebral fractures in the 20-g teriparatide treatment group occurred in 3.4%, 5.8%, and 7.2% for patients with one, two, and three or more prevalent vertebral fractures, respectively. New moderate or severe fractures occurred in 0%, 1.9%, and 1.3% of patients with one, two, and three or more prevalent vertebral fractures, respectively (Fig. 1). In contrast with placebo, no statistically significant trend was detected for either outcome (P 0.15 and P 0.26, respectively) for the 20-g teriparatide treatment group. Compared with teriparatide, the placebo slopes were statistically greater in magnitude for both outcomes. The estimates (95% confidence interval) for the slope differences (placebo minus teriparatide) were 6.0% (1.4 10.7) and 6.3% (3.29.7) for any new vertebral fractures and new moderate or severe vertebral fractures, respectively. Compared with teriparatide patients, placebo patients had 6.0% and 6.3% increased risks of experiencing any new or new moderate or severe vertebral fracture, respectively, with each additional prevalent vertebral fracture.
Baseline vertebral fracture grade and risk of incident vertebral fracture: teriparatide-treated patients

FIG. 1. New vertebral fractures in the placebo and 20-g teriparatide groups by increasing number of prevalent vertebral fractures. , Percentage of women with any new vertebral fracture; o, percentage of women with new moderate or severe fractures. The number of women in each baseline category is shown in parentheses below each bar.

New vertebral fractures in the 20-g teriparatide treatment group occurred in 3.5%, 6.4%, and 5.8% of patients with mild, moderate, and severe baseline vertebral fracture grades, respectively. New moderate or severe vertebral fractures occurred in 1.8%, 0.5%, and 1.2% of patients with mild,

1586

J Clin Endocrinol Metab, March 2005, 90(3):15831587

Gallagher et al. Fracture Risk Reduced by Teriparatide

Additional analyses

Logistic regression modeling performed in the placebo group demonstrated that there were no significant interactions between the number and severity of vertebral fractures. This suggests that the number and severity of prevalent vertebral fractures acted as independent risk factors, and that the effect of each on subsequent vertebral fracture risk was additive. We repeated the logistic regression modeling including both lumbar spine and femoral neck BMD in the model. The main effect of spine BMD was statistically significant, but that of femoral neck BMD was not. There were no statistically significant first-order interactions, suggesting that the effects of baseline number and severity of vertebral fracture severity were consistent across the range of baseline BMD values.
Discussion

FIG. 3. New nonvertebral fractures in the placebo and 20-g teriparatide groups by increasing number of prior nonvertebral fractures. , Percentage of women with any new nonvertebral fracture; o, percentage of women with new nonvertebral fragility fractures. The number of women in each baseline category is shown in parentheses below each bar.

moderate, and severe grades, respectively (Fig. 2). In contrast with placebo, no statistically significant trend was detected for either outcome (P 0.43 and P 0.60, respectively). Compared with teriparatide, the placebo slopes were statistically larger in magnitude for both outcomes. The estimates (95% confidence interval) for the slope differences (placebo minus teriparatide) were 8.3 (2.4 14.3) and 9.9 (5.0 14.9) for any new vertebral fractures and new moderate or severe vertebral fractures, respectively.
Number of prior nonvertebral fragility fractures and risk of incident nonvertebral fractures: teriparatide-treated patients

New nonvertebral fractures in the 20-g teriparatide treatment group occurred in 6.3%, 2.0%, and 11.8% of patients with zero, one, and two or more prior nonvertebral fragility fractures, respectively. New nonvertebral fragility fractures occurred in 2.7%, 0%, and 3.9% of patients with zero, one, and two or more prior nonvertebral fragility fractures, respectively (Fig. 3). Again, in contrast with placebo, no statistically significant trend was detected for either outcome (P 0.35 and P 0.96, respectively) for the 20-g teriparatide treatment group. Compared with teriparatide, the placebo slopes were larger in magnitude for both outcomes. The estimates (95% confidence interval) for the slope differences (placebo minus teriparatide) were 6.0 (2.3 to 14.4) and 8.4 (2.0 15.2) for any new nonvertebral fracture and new nonvertebral fragility fractures, respectively. The reduction in total fractures in the teriparatide group was due to a reduction in fragility fractures and not to a reduction in nonfragility fractures, e.g. severe trauma. There was no statistically significant increased risk of nonvertebral fractures or nonvertebral fragility fractures in the teriparatide or placebo group with increasing number and severity of prevalent vertebral fractures.

This new analysis from the Fracture Prevention Trial (6) demonstrated that placebo-treated women with increasing number and severity of prevalent vertebral fractures were at increasing risk for both new vertebral fractures and new moderate or severe vertebral fractures. Patients in the placebo group with increasing numbers of prior nonvertebral fractures were at increasing risk for new nonvertebral fractures. These significant trends for increasing risk with increasing burden of prior fractures were not observed in women treated with teriparatide. The current results are consistent with previous findings. A recent review of more than 40 publications concluded that a history of prior fractures predicts future fracture events, and that the risk increases with the number of prior fractures (2). For example, Torgerson et al. (13) reported that in women aged 4751 yr, the number of prior fractures of wrist, arm, leg, ribs, clavicle, foot, and ankle predicted an increased future risk for nonvertebral fractures. During 2 yr of observation, women with one and more than one prior fracture had 2- and 6-fold increases, respectively, in the risk for nonvertebral fractures compared with women without a previous history of fracture. In clinical trials, Lindsay et al. (14) found that in subgroups of women with baseline zero, one, and two or more vertebral fractures, 1.9%, 4.6%, and 12.5% of women developed new vertebral fractures, respectively, during the first year of observation (14). In the MORE trial, subgroups of women with zero, one, two, and three or more prevalent vertebral fractures had 4.3%, 13.5%, 18%, and 36.6% probability, respectively, of developing a new vertebral fracture during 3 yr of observation (4). Not only does an increased number of prevalent vertebral fractures increase the fracture risk, but so does the severity of the vertebral fracture. In the MORE trial, women with no fracture and those with mild, moderate, or severe prevalent vertebral fractures had 4.3%, 10.5%, 23.6%, and 38.1% incident vertebral fracture risk, respectively (4). Bone strength has been described as a reflection of the integration of bone density and bone quality (15). Because the presence of vertebral fracture increases future vertebral fracture risk even after controlling for BMD (1), the presence of vertebral fractures might be an indicator of poor bone quality. Because teriparatide improves bone density (6) and bone

Gallagher et al. Fracture Risk Reduced by Teriparatide

J Clin Endocrinol Metab, March 2005, 90(3):15831587

1587

quality (16) and prevents the increasing risk of future fracture associated with a history of fractures, this therapy may reduce the expensive and disabling consequences of osteoporosis in women with a history of fracture. Vertebral fractures are quite common, with more than 30% of women aged 75 yr and 50% of women aged 85 yr and older having sustained a vertebral fracture (17). These results should encourage the increased use of spine imaging by healthcare providers to assess future vertebral fracture risk more accurately in individual patients. Nonvertebral fractures are also common, and a history of fragility fracture predicts future fractures, as shown in this and other studies (3). In conclusion, the incidence of new vertebral fractures markedly increased in placebo-treated women with greater number or increased severity of prevalent vertebral fractures. The risk of new nonvertebral fractures also increased in placebo-treated women with increasing number of prior nonvertebral fragility fractures. These trends for increased fracture risk with increasing prior fracture burden were not observed in women treated with teriparatide.
Acknowledgments
We thank Mary Ellen Perron for assistance with the graphics. Received May 7, 2004. Accepted December 1, 2004. Address all correspondence and requests for reprints to: Dr. John H. Krege, Lilly Research Laboratories, DC 6121, Eli Lilly & Co., Lilly Corporate Center, Indianapolis, Indiana 46285. E-mail: krege_john_ henry@lilly.com.

3.

4.

5.

6.

7.

8.

9.

10. 11. 12. 13.

14.

15.

References
1. Ross PD, Davis JW, Epstein RS, Wasnich RD 1991 Pre-existing fractures and bone mass predict vertebral fracture incidence in women. Ann Intern Med 114:919 923 2. Klotzbuecher CM, Ross PD, Landsman PB, Abbott III TA, Berger M 2000

16.

17.

Patients with prior fractures have an increased risk of future fractures: a summary of the literature and statistical synthesis. J Bone Miner Res 15:721739 Kanis JA, Johnell O, de Laet C, Johansson H, Oden A, Delmas P, Eisman J, Fujiwara S, Garnero P, Kroger H, McCloskey EV, Mellstrom D, Melton LJ, Pols H, Reeve J, Silman A, Tenenhouse A 2004 A meta-analysis of previous fracture and subsequent fracture risk. Bone 35:375382 Delmas PD, Genant HK, Crans GG, Stock JL, Wong M, Siris E, Adachi JD 2003 Severity of prevalent vertebral fractures and the risk of subsequent vertebral and nonvertebral fractures: results from the MORE trial. Bone 33:522 532 Watts NB, Josse RG, Hamdy RC, Hughes RA, Manhart MD, Barton I, Calligeros D, Felsenberg D 2003 Risedronate prevents new vertebral fractures in postmenopausal women at high risk. J Clin Endocrinol Metab 88:542549 Neer RM, Arnaud CD, Zanchetta JR, Prince R, Gaich GA, Reginster JY, Hodsman AB, Eriksen EF, Ish-Shalom S, Genant HK, Wang O, Mitlak BH 2001 Effect of parathyroid hormone (134) on fractures and bone mineral density in postmenopausal women with osteoporosis. N Engl J Med 344:1434 1441 Vahle JL, Sato M, Long GG, Young JK, Francis PC, Engelhardt JA, Westmore M, Ma L, Nold JB 2002 Skeletal changes in rats given daily subcutaneous injections of recombinant human parathyroid hormone(134) for 2 years and relevance to human safety. Toxicol Pathol 30:312321 Tashjian AH, Chabner BA 2002 Clinical safety of recombinant human parathyroid hormone (134) in the treatment of osteoporosis in postmenopausal women and men [commentary]. J Bone Miner Res 17:11511161 Vahle JL, Long GG, Sandusky G, Westmore M, Ma YL, Sato M 2004 Bone neoplasms in F344 rats given teriparatide [rhPTH(134)] are dependent on duration of treatment and dose. Toxicol Pathol 32:426 438 Genant HK, Wu CY, van Kuijk C, Nevitt MC 1993 Vertebral fracture assessment using a semiquantitative technique. J Bone Miner Res 8:11371148 Agresti A 1990 Categorical data analysis. New York: Wiley & Sons Efron B, Tibshirani RJ 1993 An introduction to the bootstrap. Monographs on statistics and applied probability. New York: Chapman, Hall; 57 Torgerson DJ, Campbell MK, Thomas RE, Reid DM 1996 Prediction of perimenopausal fractures by bone mineral density and other risk factors. J Bone Miner Res 11:293297 Lindsay R, Silverman SL, Cooper C, Hanley DA, Barton I, Broy SB, Licata A, Benhamou L, Geusens P, Flowers K, Stracke H, Seeman E 2001 Risk of new vertebral fracture in the year following a fracture. JAMA 285:320 323 NIH Consensus Development Panel 2001 Osteoporosis prevention, diagnosis, and therapy. JAMA 285:785795 Jiang Y, Zhao JJ, Mitlak BH, Wang O, Genant HK, Eriksen EF 2003 Teriparatide [recombinant human parathyroid hormone (134)] improves both cortical and cancellous bone structure. J Bone Miner Res 18:19321941 Melton III LJ, Lane AW, Cooper C, Eastell R, OFallon WM, Riggs BL 1993 Prevalence and incidence of vertebral deformities. Osteoporos Int 3:113119

JCEM is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.