Osteoporos Int (2008) 19:607 613 DOI 10.1007/s00198-007-0508-8 ORIGINAL ARTICLE Wrist fracture as a predictor of future fractures in younger versus older postmenopausal women: results from the National Osteoporosis Risk Assessment (NORA) E. Barrett-Connor & S. G. Sajjan & E. S. Siris & P. D. Miller & Y.-T. Chen & L. E. Markson Received: 18 April 2007 / Accepted: 26 September 2007 / Published online: 6 December 2007 # International Osteoporosis Foundation and National Osteoporosis Foundation 2007 Abstract Summary The short-term association between wrist-fracture history and future fracture has not been simultaneously compared between younger and older postmenopausal women. This 3-year follow-up study of 158,940 women showed a similar future fracture risk in younger and older women with wrist-fracture history. Introduction We examined the association between prior wrist fracture and future osteoporosis-related fractures within 3 years in younger and older postmenopausal women. Methods In the National Osteoporosis Risk Assessment (NORA) study, 158,940 postmenopausal women, aged 50-98 (median 63) years, provided information on fracture history since age 45, and responded to follow-up surveys 1 or 3 years later when new fractures were queried. Cox regression models were used to obtain relative risk (RR) and 95% confidence interval (CI) estimates. E. Barrett-Connor (*) Department of Family and Preventive Medicine, University of California, San Diego, Stein Clinical Research Building, Room 349, 9500 Gilman Drive, Mail Code: 0607, La Jolla, CA 92093-0607, USA e-mail: ebarrettconnor@ucsd.edu S. G. Sajjan : Y.-T. Chen : L. E. Markson Outcomes Research & Management, Merck & Co., Inc., West Point, PA, USA E. S. Siris Columbia University College of Physicians and Surgeons, New York, NY, USA P. D. Miller Colorado Center for Bone Research, Lakewood, CO, USA Results Of the 158,940 participants, 8,665 reported a history of wrist fracture at baseline; 4,316 women reported at least one new fracture within three years. The RR for any subsequent clinical fracture, adjusted for covariates and baseline BMD T-score, was 2.4 (2.0, 2.9) for younger and 2.1 (1.9, 2.3) for older women. A prior wrist fracture increased the risk of a future wrist fracture about 3-fold and doubled the risk of any osteoporotic fracture. Conclusions Prior wrist fracture strongly predicts threeyear risk of any future osteoporotic fracture for older and younger postmenopausal women, independent of baseline BMD and common osteoporosis risk factors. More consideration should be given to evaluating and managing osteoporosis in younger and older women with a history of wrist fracture, independent of their BMD. Keywords Bone mineral density. Osteoporosis. Osteoporotic fracture. Postmenopausal women. Wrist fracture Introduction The major consequence of osteoporosis is fragility fractures. It is well known that prior fracture is associated with a significantly increased risk of future fracture [1 3]. Others have reported that forearm fracture is associated with about a two-fold increase in future fracture risk [4, 5]. This consistent association between fracture history and increased risk of future fracture risk led to the inclusion of prior fracture as a risk factor in the National Osteoporosis Foundation (NOF) treatment guidelines [6]; however, guidelines currently include only hip or spine fractures. The association between wrist fracture and increased future fracture risk over 10 years has been shown in DO00508; No of Pages
608 Osteoporos Int (2008) 19:607 613 prospective studies of postmenopausal women [5, 7 9]. The impact of wrist fracture on short-term fracture risk has not been reported. Also, no published studies have had a large enough cohort to allow simultaneous comparison of wrist fracture risk among younger versus older postmenopausal women. The goal of the present study was to assess the association between a history of wrist fracture after age 45 with the three-year risk of osteoporosis-related fracture (hip, spine, rib, wrist, or forearm) in women less than 65 years of age compared to women 65 years and older. Methods Study population The National Osteoporosis Risk Assessment (NORA) is a longitudinal observational study of osteoporosis among 200,160 postmenopausal women in the United States. Details of the study design have been published [7, 8]. Briefly, postmenopausal women who were at least 50 years of age, who had not previously been diagnosed with osteoporosis, and who had not had BMD testing within the preceding 12 months were eligible to participate. Women were recruited through the offices of 4236 primary care physicians in 34 states. Each office provided randomly chosen names of up to 300 eligible women who received letters of invitation from their physicians; about 30% of those invited (N=40-100) agreed to enroll and gave written informed consent. The study documents were approved by Essex, a national institutional review board. Bone mineral density measurement At baseline each NORA participant had BMD measured at a single peripheral site: heel measurement with single x-ray absorptiometry (SXA) (Osteoanalyzer, Norland Medical Systems Inc., White Plains, NY) and forearm using peripheral DXA (pdexa, Norland Medical Systems Inc., White Plains, NY); finger using peripheral DXA (AccuDEXA, Schick Technologies Inc., Long Island City, NY); or heel using ultrasound (Sahara, Hologic Inc., Bedford, MA). BMD testing was performed in the physician s office by NORA field radiology technicians certified by the International Society for Clinical Densitometry. All instruments were calibrated daily and before use at each office site using the manufacturer s internal standard. Quality assurance of BMD tests was maintained by staff at Synarc in Portland, Oregon [9]. T-scores, defined as the number of standard deviations from the young adult mean, were calculated from manufacturers Caucasian reference databases. In NORA, the association of BMD T-score per SD decrease with risk of fracture has been shown to be comparable whether calculated from the manufacturer s young adult reference database or from the mean BMD of the NORA cohort [10]. NORA T-scores based on Caucasian reference databases also predicted fracture risk in ethnic minorities [11]. Baseline risk factors At enrollment, each participant completed mailed questionnaires that provided information on published risk factors for low BMD and fractures, including age, ethnicity, education, self-rated health status, weight, height, personal and family history of fracture, medication use (corticosteroid, estrogen), lifestyle (cigarette smoking, alcohol use, regular exercise), and age at menopause. Body mass index (BMI) was calculated from self-reported height and weight. History of wrist fracture was queried by asking, Since the age of 45, have you broken any of the following bones: hip, rib, wrist, spine (backbone)?" Each fracture history question had a yes/no answer. No information was obtained about the date of previous fractures or the circumstances in which they occurred. Longitudinal follow-up Participants were contacted for two follow-up surveys. The first survey was conducted approximately 12 months after enrollment (median: 13.5 months; 99% range: 10.4 26.1 months). Each participant received a questionnaire that asked about new fractures, general health status, current osteoporosis medication use, new BMD measurements, and history of falls. When responders to the first follow-up survey (N=164,006; 81.9%) were compared with non-responders, at baseline non-responders were older, in poorer reported health, less educated, less likely to be Caucasian, less likely to have used estrogen, less likely to exercise regularly, and more likely to be current smokers. The second follow-up survey was conducted approximately 3 years after baseline (median: 38.9 months; 99% range: 33.2 46.6 months). Of the 124,988 women who were mailed the three-year follow-up questionnaire, 100,697 (80.6%) responded, 94,521 of whom had also responded to the first follow-up survey. On average, the second follow-up survey non-responders, compared to responders, were older, less educated, in poorer general health, less likely to be Caucasian, less likely to have used estrogen, and more likely to be current smokers, to have lower BMD, and to have reported a fracture at the first follow-up survey. Overall, 170,086 women responded to either the first or second follow-up surveys. Incident osteoporosis-related fractures were queried, "Since joining NORA, have you broken or fractured any bone? Please tell us which bone(s) you broke, e.g., wrist (yes/no), forearm (yes/no). Reported
Osteoporos Int (2008) 19:607 613 609 fractures since baseline at the hip, wrist, forearm, spine, or rib are consistent with other reports [12 15]. Study population This analysis included women who answered the baseline history of wrist fracture after age 45 question, and who responded to the first or second follow-up survey or, in most cases, to both surveys. Women who reported a baseline history of fractures at other sites hip, rib, and spine (N=7,416) were not included in the analysis, nor were women who reported multiple fractures including wrist fractures (N=1,853), because they were at greater risk for future fracture than women who reported only a wrist fracture. Thus, in this paper the history of wrist fracture refers to women whose only reported fracture at baseline was a wrist fracture. The comparison group was women who reported no fracture at any site, because the purpose of this study was to determine the magnitude of future fracture risk in patients who present with a wrist fracture in the absence of other fractures. Figure 1 displays a flowchart of the study sample of women included in these analyses. Statistical analysis The chi-square and t-test statistics were used to compare characteristics between groups. The association between wrist-fracture history and future fracture by fracture site was stratified by age (<65 vs. 65 years). The absolute risk of future fracture was calculated as new fracture rate per Patients responded to baseline, year 1 and/or year 3 Survey (N=170,086) 1,000 eligible women-years. The excess or attributable risk, that is, how many fractures were attributable to history of wrist fracture, was calculated as the absolute risk difference between wrist-fracture-history and the no-fracture-history groups. Due to the large sample size, very small but statistically significant differences were observed; to avoid over-interpretation of associations that are not clinically meaningful, statistical significance reported in this paper was based on p values <0.0001. Multivariate analysis was performed to assess the association between baseline wrist-fracture history and future fractures by type of fracture and stratified by age. Cox models were developed for all future osteoporotic fractures combined and for each new fracture site separately. We did not combine wrist with forearm fractures. Relative risk (RR; 95% confidence interval) adjusted for covariates was estimated from Cox models using the time from baseline to the time of first new fracture. If no fractures occurred during follow-up, time from baseline to the last survey response (first or second follow-up survey) was used as the censoring time. Fifteen potential covariates were considered (age, BMD T-score, ethnicity, education, self-rated health status, maternal or family history of fracture, estrogen use, weight, current corticosteroid use, smoking status, menopause age before 40 years, thyroid medication use, BMI, exercise, and alcohol use). The final set of covariates was identified from the all future osteoporotic fractures model with stepwise selection criteria set at P<0.05. All variables but thyroid medication use, BMI, exercise, and alcohol were included in the final models for any future fracture and all other site specific fractures. An interaction term between age group and wristfracture history was used to determine whether the association varied by age group. Results are presented with and without the adjustment for BMD T-score to demonstrate the impact of BMD. Hip, spine, rib, arm or multiple Fxs (N=9,269) Prior Fracture Since Age 45 at baseline* (N=168,209) Wrist Fracture Hx (N=8,665) No Fracture Hx (N=150,275) Excluded Study Sample (N=158,940) *Since the age 45, have you broken any of the following bones? A) Hip B) Rib C) Wrist D) Spine (backbone) Fig. 1 Flowchart showing the selection of study sample of women with either a history of wrist fracture or history of no fracture since age 45 Results Of the 158,940 women who met the inclusion criteria for these analyses, 8,665 (5.5%) reported a history of wrist fracture since age 45 at baseline. Their median age was 63 years (range 50 98 years); 90.8% were Caucasian. Table 1 shows the baseline characteristics of women with and without wrist-fracture history stratified by age <65 and 65 years old. For both younger and older women, women with a wrist-fracture history had a lower BMD T-score, were more likely to have a family history of fracture, and were less likely be current users of estrogen, compared to women with no history of fracture. In addition, younger women with a wrist-fracture history were more likely to report a poor health status, and older women with a wrist-
610 Osteoporos Int (2008) 19:607 613 Table 1 Characteristics of women with history of wrist fracture vs. no fracture history * % are column percentages and may not add up to 100% due to missing/unknown a p value <0.0001 Characteristic <65 Years (N=87,899) 65 Years (N=71,041) Wrist fracture Hx* (N=2,609) No fracture Hx* (N=85,290) Wrist fracture Hx* (N=6,056) No fracture Hx* (N=64,985) Age: Mean (SD) yrs 58.6 (3.9) a 57.0 (4.2) 73.6 (5.8) a 72.0 (5.4) T-score: Mean (SD) -1.00 (1.0) a -0.4 (1.0) -1.7 (1.0) a -1.2 (1.1) Race (%) Caucasian 92 91 93 a 91 Education (%) College or more 45 48 36 35 Health status (%) Excellent/ V. Good 45 a 50 42 41 Good 37 36 41 42 Fair/Poor 17 14 16 16 Estrogen use (%) Never 29 a 25 47 a 41 Former 19 15 23 21 Current 50 59 26 35 Cigarette smoking (%) Never 49 50 59 a 57 Former 37 36 33 34 Current 13 13 6 7 Family fracture history (%) 29 a 23 27 a 21 Thyroid medication Use (%) 17 a 16 21 20 Alcohol use, drinks/wk (%) None 71 70 72 74 <7 19 20 16 15 7+ 8 8 8 8 Menopause age <40 years 22 20 13 13 Regular exercise 0-2 (vs. 3+ ) 47 46 42 a 44 times/wk (%) Weight <127 lbs (%) 13 12 20 a 17 fracture history were more likely to have low body weight. Also, there was a trend of more younger women (age <65) reporting a history of wrist facture with greater BMI level and normal BMD T-scores, whereas the trend was not seen in their older counterparts or in those with low BMD T- scores (data not shown). Tables 2 and 3 show the number of women with future fractures and future fracture rates per 1,000 women-years by wrist-fracture history and by age groups; data are shown by type of fracture and wrist-fracture history. Approximately 36 months after baseline, a total of 4,316 (1726 younger; 2590 older) women reported 4,708 new fractures. Future fracture rates were significantly higher among women with a wrist-fracture history compared to women with no fracture history; this was true regardless of age or new fracture location. The three-year excess risks of future fracture per 1,000 women-years associated with a prior wrist fracture for any new fracture in the younger and older Table 2 Number and rates (95% confidence intervals) of future fracture per 1,000 women-years by baseline history of wrist fracture (<65 years) Fracture site All women <65 years Wrist fracture Hx (N=2,609) No fracture Hx (N=85,290) Fracture rate N Fracture rate N Fracture rate Any fracture 8.2 (7.8, 8.6) 146 24.2 (20.4, 27.9) 1,580 7.7 (7.3, 8.1) Wrist 3.4 (3.2, 3.7) 90 14.7 (11.8, 17.7) 632 3.1 (2.8, 3.3) Rib 2.5 (2.3, 2.7) 30 4.9 (3.1, 6.6) 497 2.4 (2.2, 2.6) Hip 0.8 (0.7, 0.9) 11 1.8 (0.7, 2.8) 157 0.8 (0.6, 0.9) Spine 1.1 (1.0, 1.3) 9 1.5 (0.5, 2.4) 230 1.1 (1.0, 1.3) Forearm 1.0 (0.9, 1.1) 17 2.7 (1.4, 4.0) 196 0.9 (0.8, 1.1)
Osteoporos Int (2008) 19:607 613 611 Table 3 Number and rates (95% confidence intervals) of future fracture per 1,000 women-years by baseline history of wrist fracture ( 65 years) Fracture site All women 65 years Wrist fracture Hx (N=6,056) No fracture Hx (N=64,985) Fracture rate N Fracture rate N Fracture rate Any fracture 15.8 (15.2, 16.4) 493 36.5 (33.5, 39.6) 2,097 13.9 (13.4, 14.5) Wrist 6.1 (5.7, 6.4) 266 19.3 (17.1, 21.6) 738 4.9 (4.5, 5.2) Rib 3.7 (3.4, 4.0) 82 5.9 (4.6, 7.1) 538 3.5 (3.2, 3.8) Hip 3.2 (2.9, 3.5) 93 6.7 (5.3, 8.0) 436 2.9 (2.6, 3.1) Spine 2.6 (2.3, 2.8) 50 3.6 (2.6, 4.5) 380 2.5 (2.2, 2.7) Forearm 1.5 (1.3, 1.7) 54 3.9 (2.8, 4.9) 202 1.3 (1.1, 1.5) women, respectively, were 16.5 and 22.6. This translates to a new fracture rate ratio (FRR) approximately 3.1 and 2.6 times higher in women with a wrist-fracture history compared to those with no fracture history among younger and older women respectively. The fracture rate ratio estimates for future wrist fracture were 4.7 and 3.9 among younger and older women, respectively. Table 4 shows results from multivariate analyses. Wristfracture history was associated with a more than a 2-fold increased risk of any future osteoporotic fracture, with similar future fracture rate ratios and adjusted relative risks (Wrist-fracture history vs. No fracture history) in younger (FRR, 3.2; 95% CI, 2.7-3.7; RR, 2.9; 95% CI, 2.4, 3.4) and older women (FRR, 2.6; 95% CI, 2.4-2.9; RR, 2.5; 95% CI, 2.2, 2.7). After adjusting for the covariates including BMD T-score, as shown, the estimates of relative risk for future fracture were only slightly attenuated. The risk for future fracture did not differ significantly by age group, as Table 4 Adjusted relative risk (RR) and 95% confidence interval (CI) for future fractures in women with wrist-fracture history compared to women with no fracture history by age and fracture type Adjusted relative risk (95% CI)* Covariates Covariates + BMD T-score Any future fracture Overall 2.4 (2.2, 2.7) 2.1 (1.9, 2.3) Age <65 2.9 (2.4, 3.4) 2.4 (2.0, 2.9) Age 65 2.5 (2.2, 2.7) 2.1 (1.9, 2.3) Future wrist fracture Overall 3.8 (3.3, 4.3) 3.1 (2.8, 3.5) Age <65 4.4 (3.5, 5.5) 3.5 (2.8, 4.4) Age 65 3.7 (3.2, 4.3) 3.0 (2.6, 3.5) Future hip fracture Overall 1.9 (1.5, 2.3) 1.6 (1.3, 2.0) Age <65 2.1 (1.1, 3.8) 1.7 (0.9, 3.1) Age 65 2.1 (1.7, 2.7) 1.7 (1.4, 2.2) * Based on Cox regression models adjusting for age (in models for type of future fracture), ethnicity, education, self-rated health status, maternal or family history of fracture, estrogen use, weight, current corticosteroid use, menopause age <40 years, and smoking status. indicated by the non-significant interaction term of age group and prior wrist fracture (P>0.10). Results were similar for risk of future wrist fracture (Table 4). Only the risk of future hip fracture differed across age groups when adjusted for BMD T-score (Table 4). Regardless of age, women with a wrist-fracture history were at increased risk for future fracture at each fracture site hip (RR=1.9; 95% CI=1.5, 2.3); rib (RR=1.6; 95% CI=1.3, 2.0); spine (RR= 1.3; 95% CI=1.0, 1.6); wrist (RR=3.8; 95% CI=3.3, 4.3), and forearm (RR=2.7; 95% CI=2.1, 3.5). These RRs decreased only slightly after including BMD T-score and all other covariates in the model, and the independent associations of wrist-fracture history remained significant for all clinical fracture sites except the spine. Discussion Wrist fractures occur earlier than other osteoporotic fractures and are the most common clinical low trauma fracture in women before the age of 65 in the United States [16]. In Australian women, wrist fractures remain the most common clinical fracture even in old age [17]. The present NORA study is the first to show that both older and younger postmenopausal women with a positive history of wrist fracture after age 45 are at increased risk for a future clinical fracture and recurrent wrist fracture independent of BMD and other common risk factors for fracture. Indeed, the magnitude of increase in both absolute and relative risk for future fractures was similar for younger and older postmenopausal women who had experienced a wrist fracture after age 45, as shown in Tables 2, 3, and4. Similar findings were observed for the risk of future wrist fracture (Table 4). The NORA data support the concept that even relatively young postmenopausal women with prior fractures should be evaluated and managed to prevent future fractures, even in the absence of a low BMD T-score [18]. The increased risk of future fracture observed in the current study was only slightly attenuated when BMD T-score was
612 Osteoporos Int (2008) 19:607 613 added to the models. These NORA results are similar to results from a meta-analysis reported by Kanis and colleagues [2] in which low BMD explained only the minority of the risk for any fracture in persons with a prior fracture. This meta-analysis was based on a review of 11 published studies including 15,259 men and 44,902 women. The authors did not report wrist fracture risk separately. The Study of Osteoporotic Fractures (SOF) was restricted to 9,704 women who were aged at least 65 years at baseline; the age-adjusted 10-year incident hip fracture rate was 43 percent higher in women who had reported a prior wrist fracture since age 50 compared to women without a prior wrist-fracture history. In contrast to the present study, the SOF investigators reported that this association was no longer statistically significant after adjusting for hip BMD [9], a BMD site not measured in NORA. In the prospective Dubbo study, femoral neck BMD was the primary risk factor for wrist fracture; and BMD was not measured at the wrist [19]. It is unclear why a prior wrist fracture conveys a high risk for future fractures at other skeletal sites, in addition to the wrist, independent of the baseline BMD. It is likely that a fragility fracture at one site reflects systemic skeletal fragility, plausibly a sign of poor bone quality. The exact mechanism of systemic skeletal fragility beyond areal bone density needs to be studied. In the present study we asked only about fractures after age 45. In another study comparing women who had a prior wrist fracture before or after age 70 years, the relative risk of new wrist fracture was only slightly greater for women with wrist-fracture history before age 70 years compared to women with a wrist fracture after 70 years of age (2.8 versus 2.1); statistical significance between age groups was not reported [5]. In another study in which the prior fractures occurred between ages 20 34 years, the relative risk for any other new fracture was 2.4 [20], similar to the risks reported here. In two other studies, the relative risk for new fracture was somewhat lower, 1.5 to 2.1, when the prior wrist fracture occurred after age 45. [7, 8] Taken together, these studies do not suggest that the age of first wrist fracture adds importantly subsequent fracture risk. Older women are more likely to fall than younger women, but younger women may be performing vigorous physical activity leading to falls, and may have better reflexes, leading to a fall saved by the extended arm and sparing the hip." We did not include fall data in our analysis. However, the observation that fracture risk was similar between the younger as opposed to the older women raises the question of fall-independent factors that increase fracture risk. Some potential limitations of the NORA study should be considered. Selection bias for healthy participants may exist because participation was voluntary, and eligibility was dependent on having a private physician and no known osteoporosis. Further those who completed both surveys were healthier than those who did not. Ninety-three percent of NORA participants were Caucasian, and nearly 40% had a college education, so NORA is not a representative sample of the postmenopausal women in the US. However, given the large sample from 34 states, and the strong similarity of NORA results with those from the Kanis [2] meta-analysis of eight studies of European women, NORA results are likely to be generalizable to Caucasian women of northern European ancestry. A third limitation is that risk factors and fractures were self-reported. In NORA, neither the history of wrist fracture nor the incident fractures were validated by radiographs or record review. However, others have reported a low false positive rate for wrist fractures. In a community-based study of older adults in Australia, only 4.8% of self-reported hip and 2.2% of self-reported wrist fractures could not be confirmed by review of radiology reports [21]. In a fracture validation study from the Study of Osteoporotic Fractures, only 8% of women aged 65 or older reported a history of wrist fracture that could not be confirmed by radiographs [22]. In the Women s Health Initiative, 81% of reported forearm/wrist fractures were validated [23]. Diagnostic detection bias is possible for a second wrist fracture, however, if a woman with a prior fracture more often seeks medical care for a subsequent similar episode. The size of this problem is likely to be small for two reasons: first, a wrist fracture predicted fractures at other sites almost as well as a new wrist fracture; and second, all NORA women had a private physician, making it less likely that many wrist fractures were undiagnosed. Finally, the independence of association of prior wrist fracture from BMD in predicting future fracture in NORA was based on peripheral bone density. It is possible that BMD at the hip or spine would have explained more of the future fracture risk. In summary, the absolute and excess risk for future fracture was only slightly higher in older women than in younger women, and the relative risk for new fracture among women with prior wrist fracture was consistent in younger and older women. The increased risk was largely independent of multiple covariates and peripheral BMD. Given that wrist fracture is a strong predictor of a new fracture at the wrist as well as distant skeletal sites within three years in younger and older postmenopausal women, all postmenopausal women, regardless of age, who present with a wrist fracture should be appropriately managed for osteoporosis to reduce their future fracture risk. Acknowledgments National Osteoporosis Risk Assessment (NORA) was funded and managed by Merck & Co., Inc. in collaboration with the International Society for Clinical Densitometry. Disclaimer/Conflict of interest statement Drs. Barrett-Connor, Siris, and Miller receive consulting fees from Merck & Co., Inc. to compensate them for time spent working on the NORA project; Drs. Sajjan, Chen, and Markson are paid employees of Merck & Co., Inc.
Osteoporos Int (2008) 19:607 613 613 References 1. Klotzbuecher CM RP, Landsman PB, Abbott TA III, Berger M (2000) Patients with prior fractures have an increased risk of future fractures: A summary of the literature and statistical synthesis. J Bone Miner Res 15:721 739 2. 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 35:375 382 3. Porthouse J, Birks YF, Torgerson DJ, Cockayne S, Puffer S, Watt I (2004) Risk factors for fracture in a UK population: a prospective cohort study. Qjm 97:569 574 4. Mallmin H, Ljunghall S, Persson I, Naessen T, Krusemo UB, Bergstrom R (1993) Fracture of the distal forearm as a forecaster of subsequent hip fracture: a population-based cohort study with 24 years of follow-up. Calcif Tissue Int 52:269 272 5. Cuddihy MT, Gabriel SE, Crowson CS, O Fallon WM, Melton LJ 3rd (1999) Forearm fractures as predictors of subsequent osteoporotic fractures. Osteoporos Int 9:469 475 6. National OF (1998) Osteoporosis: review of the evidence for prevention, diagnosis and treatment and cost-effectiveness analysis. Osteoporos Int 8:7 80 7. Siris ES, Miller PD, Barrett-Connor E, Faulkner KG, Wehren LE, Abbott TA, Berger ML, Santora AC, Sherwood LM (2001) Identification and fracture outcomes of undiagnosed low bone mineral density in postmenopausal women: results from the National Osteoporosis Risk Assessment. Jama 286:2815 2822 8. Siris E, Miller P, Barrett-Connor E, Abbott T, Sherwood L, Berger M (1998) Design of NORA, the National Osteoporosis Risk Assessment Program: a longitudinal US registry of postmenopausal women. Osteoporos Int 8(Suppl 1):S62 S69 9. Fasano G, Gaither K, Siris E, Miller P, Barrett-Connor E, Berger M (1998) The impact of quality assurance on bone densitometry in the National Osteoporosis Risk Assessment. Bone 23:S590 10. Miller PD, Siris ES, Barrett-Connor E, Faulkner KG, Wehren LE, Abbott TA, Chen YT, Berger ML, Santora AC, Sherwood LM (2002) Prediction of fracture risk in postmenopausal white women with peripheral bone densitometry: evidence from the National Osteoporosis Risk Assessment. J Bone Miner Res 17:2222 2230 11. Barrett-Connor E, Siris ES, Wehren LE, Miller PD, Abbott TA, Berger ML, Santora AC, Sherwood LM (2005) Osteoporosis and fracture risk in women of different ethnic groups. J Bone Miner Res 20:185 194 12. Seeley DG, Browner WS, Nevitt MC, Genant HK, Scott JC, Cummings SR (1991) Which fractures are associated with low appendicular bone mass in elderly women? The study of Osteoporotic Fractures Research Group. Ann Intern Med 115:837 842 13. Dias JJ, Wray CC, Jones JM (1987) Osteoporosis and Colles fractures in the elderly. J Hand Surg [Br] 12:57 59 14. Earnshaw SA, Cawte SA, Worley A, Hosking DJ (1998) Colles fracture of the wrist as an indicator of underlying osteoporosis in postmenopausal women: a prospective study of bone mineral density and bone turnover rate. Osteoporos Int 8:53 60 15. Horowitz M, Wishart JM, Bochner M, Need AG, Chatterton BE, Nordin BE (1988) Mineral density of bone in the forearm in premenopausal women with fractured wrists. Bmj 297:1314 1315 16. Black DM, Cooper C (2000) Epidemiology of fractures and assessment of fracture risk. Clin Lab Med 20:439 453 17. Eisman J, Clapham S, Kehoe L (2004) Osteoporosis prevalence and levels of treatment in primary care: the Australian Bone Care Study. J Bone Miner Res 19:1969 1975 18. Gunter MJ, Beaton SJ, Brenneman SK, Chen YT, Abbott TA 3rd, Gleeson JM (2003) Management of osteoporosis in women aged 50 and older with osteoporosis-related fractures in a managed care population. Dis Manag 6:83 91 19. Nguyen TV, Center JR, Sambrook PN, Eisman JA (2001) Risk factors for proximal humerus, forearm, and wrist fractures in elderly men and women: the Dubbo Osteoporosis Epidemiology Study. Am J Epidemiol 153:587 595 20. Honkanen R, Tuppurainen M, Kroger H, Alhava E, Puntila E (1997) Associations of early premenopausal fractures with subsequent fractures vary by sites and mechanisms of fractures. Calcif Tissue Int 60:327 331 21. Ivers RQ, Cumming RG, Mitchell P, Peduto AJ (2002) The accuracy of self-reported fractures in older people. J Clin Epidemiol 55:452 457 22. Nevitt MC, Cummings SR, Browner WS, Seeley DG, Cauley JA, Vogt TM, Black DM (1992) The accuracy of self-report of fractures in elderly women: evidence from a prospective study. Am J Epidemiol 135:490 499 23. Chen Z, Kooperberg C, Pettinger MB, Bassford T, Cauley JA, LaCroix AZ, Lewis CE, Kipersztok S, Borne C, Jackson RD (2004) Validity of self-report for fractures among a multiethnic cohort of postmenopausal women: results from the Women s Health Initiative observational study and clinical trials. Menopause 11:264 274