Bone loss and the risk of non-vertebral fractures in women and men: the Tromsø study

Osteoporos Int (2010) 21:1503 1511 DOI 10.1007/s00198-009-1102-z ORIGINAL ARTICLE Bone loss and the risk of non-vertebral fractures in women and men: the Tromsø study L. A. Ahmed & N. Emaus & G. K. Berntsen & Å. Bjørnerem & V. Fønnebø & L. Jørgensen & H. Schirmer & J. Størmer & R. M. Joakimsen Received: 7 May 2009 /Accepted: 8 October 2009 /Published online: 21 November 2009 # International Osteoporosis Foundation and National Osteoporosis Foundation 2009 Abstract Summary We assessed the association between the rate of forearm bone loss and non-vertebral fracture. Bone loss at the distal forearm predicted fractures, independently of baseline BMD, but not independently of follow-up BMD in women. The BMD level where an individual ends up is the significant predictor of fracture risk. Introduction Bone loss may predict fracture risk independently of baseline BMD. The influence of follow-up BMD on this prediction is unknown. The aim of this study was to Electronic supplementary material The online version of this article (doi:10.1007/s00198-009-1102-z) contains supplementary material, which is available to authorized users. L. A. Ahmed (*) : N. Emaus : Å. Bjørnerem : V. Fønnebø : L. Jørgensen Institute of community medicine, University of Tromsø, 9037 e-mail: luai.ahmed@uit.no N. Emaus SKDE, University Hospital of North Norway, G. K. Berntsen NST, University Hospital of North Norway, L. Jørgensen Department of Health and Care Sciences, University of Tromsø, L. Jørgensen: H. Schirmer : J. Størmer : R. M. Joakimsen University Hospital of North Norway, R. M. Joakimsen Institute of Clinical Medicine, University of Tromsø, assess the association between bone loss and fracture risk in both sexes in a prospective population-based study. Methods We included 1,208 postmenopausal women (50 to 74 years), and 1,336 men (55 to 74 years) from the Tromsø Study, who had repeated distal and ultra-distal forearm BMD measurements. Non-vertebral fractures were registered from 2001 to 2005. Results A total of 100 women and 46 men sustained fractures during the follow-up time. Independent of baseline BMD, the RR associated with distal site bone loss of 1 SD %/year was 1.23 (1.01-1.50) for low-trauma fractures (excluding hand, foot, skull & high-trauma) and 1.32 (1.07-1.62) for osteoporotic fractures (hip, wrist and shoulder). However, bone loss did not predict fracture after adjusting for follow-up BMD. The BMD level where an individual ends up became the significant predictor of fracture risk and not the rate of bone loss. Follow-up BMD at ultra-distal site was associated with low-trauma fractures in both sexes. While ultra-distal site BMD changes were not associated with fracture risk in both sexes. Conclusion Bone loss at the distal forearm predicted nonvertebral fractures, independently of baseline BMD, but not independently of follow-up BMD, in women. The BMD level where an individual ends up is the significant predictor of fracture risk and not the rate of bone loss. Keywords Bone loss. Bone mineral density. Fracture risk. Non-vertebral fractures. Osteoporotic fractures. Postmenopausal women Abbreviations BMD Bone mineral density RR Relative risk SD Standard deviation

1504 Osteoporos Int (2010) 21:1503 1511 Introduction Osteoporotic fractures constitute a major health problem with substantial morbidity and costs [1, 2]. Osteoporosis is described as a systemic disease characterized by low bone mass and micro-architectural deterioration of bone tissue, with a consequent increase in bone fragility and susceptibility to fracture [3]. As there are no other satisfactory clinical tools available to assess bone quality, the diagnosis of osteoporosis centers on the assessment of skeletal mass, by measurements of bone mineral density (BMD) [4]. BMD is a function of the amount of bone gained during growth and the amount of bone lost during aging [5, 6]. Although BMD is a strong predictor, most fractures occur in individuals without low BMD, therefore, other tools to identify women and men with increased fracture risk independent of BMD levels are needed [7]. Several studies have demonstrated an association between rate of bone turnover and osteoporotic fractures independently of BMD levels in women at the time of menopause as well as in elderly women [8 13]. On the other hand, with studies suggesting that BMD change plays a stronger part [14 17] and others suggesting BMD change to be an independent but weak predictor of future fracture risk [18, 19], the role of bone loss as a predictor of fracture risk still remains unclear. The aim of this study was to assess the association and its magnitude, between the rate of bone loss and the risk of the first any non-vertebral, low-trauma non-vertebral (excluding hand, foot, skull, and high-trauma), and osteoporotic (hip, wrist, and proximal humerus) fracture, independent of BMD level, in both women and men in a prospective population-based study. Material and method Study population The Tromsø study is a longitudinal population-based multipurpose study focusing on lifestyle-related diseases. The first Tromsø study (Tromsø I) took place in 1974 and the fifth survey in 2001 (Tromsø V). In 1994 (Tromsø IV), all women aged 50 to 74 years and men aged 55 to 74 years together with random samples (of 5 10%) of younger and older age groups were invited to an extended examination that included bone densitometry. The attendance rates were 79% and 77% in women and men, respectively, and BMD was measured on a total of 7,948 persons (4,558 women and 3,390 men) [20]. In 2001, 7,386 persons still alive and living in Tromsø were invited for a re-examination. Bone densitometry was performed on 5,771 persons (3,427 women and 2,344 men), with attendance rate of 79% and 77% in women and men, respectively. The follow-up examination included 59% and 53% of the originally invited women and men from 1994. The average interval between the two forearm BMD measurements was 6.5 years (SD, 0.6 years). Informed consent was obtained, and the Regional Committee of Research Ethics and the Norwegian Data Inspectorate approved the study. Questionnaires In each survey, the participants filled in two self-administered questionnaires. The questionnaires covered, among others, family history of osteoporosis and fractures, physical activity, smoking habits, self-rated health status, history of previous fractures, menopausal status, and use of vitamin D, calcium supplements, hormone replacement therapy, and osteoporosis medications. BMD measurements Bone densitometry was performed in both surveys at the forearm, distal, and ultra-distal sites, with two single X-ray absorptiometric devices (DTX-100; Osteometer MediTech, Inc., Hawthorne, California). The distal site includes both the radius and ulna from 8-mm point (point where the ulna and radius are separated by 8 mm) and 24 mm proximally, containing mainly cortical bone [21]. The ultra-distal site includes only the radius, and stretches from the 8-mm point up to the radial endplate, containing mainly trabecular bone [21]. The non-dominant arm was measured except when it was ineligible because of wounds, plaster casts, and so on. The same rigorous study protocol was used in both surveys, and quality control with respect to densitometer precision and correction of artifacts are reported [22 24]. Other measurements Height and weight was measured to the nearest centimeter/ half kilogram with attendants wearing light clothing and no shoes. Body mass index (BMI) was calculated as weight in kilograms divided by the square of the height in meters. Fracture registration The fracture registry is based on the radiographic archives at the University Hospital in Tromsø. The nearest alternative radiographic service or fracture treatment facility is located 250 km from Tromsø. The only fractures that would be missed are fractures occurring while inhabitants were traveling, and no control radiographic examination was done after returning home, in addition to fractures not radiographically examined. Similar registration for participants in the second and third Tromsø surveys was

Osteoporos Int (2010) 21:1503 1511 1505 described earlier [25]. The computerized records in the radiographic archives of the University Hospital contain the national personal identification number, time of investigation, fracture codes, and descriptions. All radiographic examinations coded abnormal on participants in the fourth survey were reviewed to ascertain the fracture code, to identify exact fracture type and anatomical location, to distinguish consecutive fracture cases from one another, and to capture fractures that had not been coded correctly as fractures. In addition, the discharge records were checked with respect to hip fractures. The fracture registry covered the period from the 1 st of January 1994 to the 12th of February 2005, with respect to all non-vertebral fractures. For our target population, the follow-up time was assigned from the date of the repeated BMD measurement at the fifth survey (in 2001 2002) for each participant, to date of first fracture, migration, death, or to the 12th of February 2005. Data preparation and statistical analysis After exclusion of invalid scans and with 179 persons withdrawing the informed consent at a later stage in the study, 3,263 women and 2,271 men with repeated BMD measurements at both the distal and ultra-distal forearm sites remained. Only postmenopausal women aged 50 74 years who were not reporting any use of hormone replacement therapy during the follow-up period, and men aged 55 to 74 years in 1994 1995 were included in the analyses. All women and men who reported use of osteoporosis medications were excluded. One man was excluded for having a pathological fracture. In order to assess the predicting effect of BMD changes on fracture risk, it is necessary to start the follow-up of the outcome after complete determination of the exposure [18, 19]. Therefore, we excluded all women and men who suffered fractures between the two BMD measurements. Furthermore, to assess the risk of first fracture, all participants who reported history of previous fracture, before the first BMD measurements, were excluded. Analyses including fractures between the BMD measurements controlled for previous fractures (before baseline) did not change the conclusion (data not shown). A total of 1,208 women and 1,336 men with repeated measurements at the distal and ultra-distal sites were included in this study (Fig. 1). Change in BMD in grams per square centimeters was estimated by calculating the difference between measurements in Tromsø IV (baseline BMD) and Tromsø V (followup BMD). This total estimate was divided by the length of each participant's follow-up time to calculate the annual changes. The annual BMD changes are expressed as percent per year and as grams per square centimeters per year. A t test was used to determine the differences between fractured and not fractured subjects for continuous variables and χ 2 test, for categorical variables. The relative risks (RR) [with 95% confidence interval (CI)] of fracture were calculated using the Cox's proportional hazards regression model. Covariates measured at Tromsø V including BMI, family history of osteoporosis or fractures, physical inactivity, smoking, self-reported poor health, and the use of vitamin D and calcium supplements were left out of the final models, as they did not significantly affect the relationships between BMD changes and fracture risk. The final gender-specific models were adjusted for age and separately for baseline and follow-up BMD. BMD change was ranked in tertiles, and linear trends of fracture risk were assessed. In addition, RR of fracture was calculated for 1-year increase of follow-up age, 1 SD% decrease in baseline and follow-up BMD, and 1 SD% decrease in annual percentage rate of BMD at both distal and ultra-distal sites. RR were calculated for all nonvertebral, low-trauma non-vertebral (excluding hand, foot, skull, and high-trauma fractures), and for osteoporotic (hip, distal forearm, and proximal humerus) fractures. Analyses were performed in the SAS statistical package, v9 (SAS Institute Inc., Cary, NC, USA). Results Tromsø IV (1994): 7948 persons (aged 25 to 84 years) had forearm BMD measurements Persons with repeated distal and ultra-distal 2271 BMD measurements 3263 Persons excluded from analyses*: 490 Age < 55 (men) Age < 50 (women) 335 10 Age > 74 15 135 Non-vertebral fracture between 1994-2001 427 293 Self-reported fractures before 1994 375 1 Pathological fracture 0 - Premenopausal 983 0 Use of hormone replacement therapy 538 60 Use of osteoporosis medication 284 3 Invalid repeated BMD measurements 55 * The numbers are overlapping, not additive. 1,337 men 46 Tromsø V (2001): 5771 persons had repeated forearm BMD measurements 179 persons excluded (did not consent to medical research) Final numbers of persons included in the analyses Final numbers of fractures after Tromsø V 100 1,208 women Fig. 1 Flow chart presenting numbers of persons with forearm bone densitometry measurements in the fourth and fifth Tromsø surveys and the final numbers of men and women included in this study For women, the mean baseline age was 60 years; BMD at the distal site was 0.42 g/cm 2, and the mean annual BMD change at distal site was 0.95% per year. For men, the corresponding values were 62.7 years, 0.54 g/cm 2, and 0.52% per year, respectively. During follow-up period (mean, 3.3 years; SD, 0.7 years), 100 women and 46 men sustained non-vertebral fractures. had 11 hip, 40 wrist, and 18 proximal humerus fractures, whereas, men

1506 Osteoporos Int (2010) 21:1503 1511 Table 1 Baseline characteristics of men (aged 55 74 years) and postmenopausal women (aged 50 74 years) with and without non-vertebral fractures a Data are mean ± SD b Significant age-adjusted difference from those without fractures (p<0.05) No fracture Any fracture No fracture Any fracture Number 1,290 46 1,108 100 Age (years) by 31 Dec. 1994 a 62.6±5.3 64.0±6.3 59.9±6.5 61.4±6.7 b Body mass index (BMI) a 26.1±3.3 25.9±3.2 26.3±4.4 26.2±4.6 Physical inactive (%) 56.0 45.6 70.2 71.0 Smokers (%) 28.2 28.3 30.7 27.0 Poor health (%) 40.8 32.6 44.4 46.0 Using vitamin D supplements (%) 4.8 4.3 8.2 10.0 Using calcium supplements (%) 16.4 17.4 24.9 28.0 Distal forearm site Baseline BMD (g/cm 2 ) a 0.54±0.06 0.53±0.06 0.42±0.06 0.40±0.06 b Rate of annual change (%/year) a 0.52±0.59 0.54±0.71 0.93±0.92 1.13±0.95 b Absolute annual change (mg/year) a 2.74±3.03 2.76±3.91 3.79±3.67 4.43±3.72 Ultra-distal forearm site Baseline BMD (g/cm 2 ) a 0.44±0.07 0.43±0.06 0.32±0.06 0.29±0.06 b Rate of annual change (%/year) a 0.42±0.80 0.50±1.00 0.90±1.18 1.00±1.20 Absolute annual change (mg/year) a 1.85±3.41 1.96±4.03 2.90±3.64 2.92±3.58 had ten hip, five wrist, and one proximal humerus fracture. Baseline characteristics of women and men by fracture status are presented in Table 1. with fractures were older and had lower BMD levels at baseline. No significant differences were detected in men. Distal forearm BMD Independent of baseline BMD, higher levels of bone loss were associated with increased fracture risks in postmenopausal women. in the highest bone-loss tertile showed an increased risk of non-vertebral fractures (RR, 1.77; 95% CI 1.05 2.98), low-trauma non-vertebral fractures (RR 2.04; 95% CI 1.14 3.65), and osteoporotic fractures (RR 2.64; 95% CI 1.28 5.46) compared with those with the lowest bone loss (Fig. 2). The p values for the trends were 0.036, 0.015, and 0.012 for non-vertebral, low-trauma non-vertebral, and osteoporotic fractures, respectively. For osteoporotic fractures, women in the middle tertile also showed an increased risk (RR 2.60; 95% CI 1.26 5.40) in comparison to those in the lowest bone loss tertile. in the highest bone loss tertile showed no significant reduction in fractures risk compared with those with the lowest bone loss. Fig. 2 Relative risk of fractures in men and women by tertiles of annual BMD change (%/year) at the distal forearm site Relarive risk of fracture 6 5 4 3 2 1 * (p for trend <0.05) * * * 0 Lowest loss Middle Heighest loss Lowest loss Middle Heighest loss Tertiles of BMD change (%/yr) All non-vertebral fractures Low-trauma non-vertebral fractures Osteoporotic fractures

Osteoporos Int (2010) 21:1503 1511 1507 Table 2 Relative risk of fractures according to annual distal forearm BMD percentage changes in men and women All models were adjusted for follow-up age a Excluding hand, foot, skull, and high-trauma fractures b Hip, wrist, proximal humerus fractures 55 74years 50 74years Number of persons 1,336 1,208 All non-vertebral fractures Number of fractures 46 100 1: Baseline BMD 1.19 (0.89 1.58) 1.33 (1.07 1.65) 2: Follow-up BMD 1.16 (0.87 1.55) 1.39 (1.12 1.73) 3: BMD change (1SD%/year) 0.97 (0.72 1.30) 1.20 (1.00 1.44) Baseline BMD 1.21 (0.90 1.62) 1.31 (1.05 1.63) BMD change (1SD%/year) 0.93 (0.69 1.26) 1.17 (0.97 1.40) Follow-up BMD 1.24 (0.89 1.73) 1.35 (1.05 1.73) BMD change, unit: (1SD%/year) 0.87 (0.62 1.22) 1.05 (0.85 1.29) Low-trauma non-vertebral fractures a Number of fractures 23 81 1: Baseline BMD 1.55 (1.06 2.26) 1.31 (1.03 1.66) 2: Follow-up BMD 1.47 (0.99 2.18) 1.41 (1.11 1.79) 3: BMD change (1SD%/year) 0.87 (0.56 1.36) 1.26 (1.04 1.53) Baseline BMD 1.62 (1.11 2.36) 1.27 (1.00 1.62) BMD change (1SD%/year) 0.79 (0.51 1.23) 1.23 (1.01 1.50) Follow-up BMD 1.74 (1.13 2.67) 1.32 (1.00 1.75) BMD change (1SD%/year) 0.67 (0.42 1.09) 1.11 (0.89 1.40) Osteoporotic fractures b Number of fractures 16 66 1: Baseline BMD 1.18 (0.74 1.88) 1.29 (0.99 1.67) 2: Follow-up BMD 1.10 (0.68 1.76) 1.44 (1.10 1.87) 3: BMD change (1SD%/year) 0.77 (0.45 1.31) 1.35 (1.10 1.65) Baseline BMD 1.25 (0.79 2.00) 1.24 (0.95 1.62) BMD change (1SD%/year) 0.73 (0.43 1.25) 1.32 (1.07 1.62) Follow-up BMD 1.31 (0.77 2.21) 1.28 (0.94 1.74) BMD change (1SD%/year) 0.68 (0.38 1.21) 1.21 (0.95 1.54) In women but not in men, a reduction by 1% SD per year of BMD was associated with low-trauma non-vertebral fractures independent of baseline BMD (RR 1.23; 95% CI 1.01 1.50). The association was strengthened further when restricting the analysis to osteoporotic fractures. Each SD percent loss of BMD was associated with 32% (95% CI 1.07 1.62) increased risk of osteoporotic fracture in women (Table 2). However, bone loss did not predict fractures after adjusting for follow-up BMD. In contrast, each 1 ± SD lower follow-up BMD predicted low-trauma fractures independently of BMD change in women (RR 1.32; 95% CI 1.00 1.75) and in men (RR 1.74; 95% CI 1.13 2.67). Ultra-distal forearm BMD There were no significant differences in fracture risk between tertiles of bone loss in women and men (Fig. 3). BMD change was not associated with fracture risk, neither before nor after adjusting for baseline or follow-up BMD in both sexes (Table 3). However, follow-up BMD was associated with

1508 Osteoporos Int (2010) 21:1503 1511 Fig. 3 Relative risk of fractures in men and women by tertiles of annual BMD change (%/year) at the ultra-distal forearm site Relative risk of fracture 6 5 4 3 2 1 0 Lowest loss Middle Heighest loss Lowest loss Middle Heighest loss Tertiles of BMD change (%/yr) All non-vertebral fractures Low-trauma non-vertebral fractures Osteoporotic fractures low-trauma fractures in women (RR 1.52; 95% CI 1.17 1.98) and in men (RR 1.67; 95% CI 1.05 2.66) and osteoporotic fractures in women independently of BMD change. Discussion Bone loss at the distal forearm predicted non-vertebral fractures independently of baseline BMD, but not independently of follow-up BMD, in women. Our results suggested that the BMD level where an individual ends up is the most strongly predictive of fracture risk and not the rate of bone loss. Implications Previous studies have shown substantial variation of loss rates in both sexes at any age [26 31]. lose approximately 1% BMD annually after the menopausal transition phase [26, 27, 32 37], whereas older men lose annually approximately 0.8% [26, 33, 34, 36]. Rapid bone loss may result in destruction of bone trabeculae and change the orientation and connectivity of the trabecular network leading to bone fragility which partly may be independent of bone mass [29, 35]. Prospective studies suggest that the rate of forearm bone loss is associated with an increased risk of fracture independently of baseline BMD and age in postmenopausal women [17] and that rate of femoral bone loss is a predictor of fracture risk in elderly women [16]. Our results indicated that distal forearm bone loss is a predictor of incident non-vertebral fracture, independent of baseline BMD, in postmenopausal women and can be compared with data from other prospective studies [14 19], only one of them including men [19]. Riis et al. reported that baseline forearm bone mass and rate of bone loss predisposed to the same extent to fractures [14], whereas, Gnudi et al. reported that rate of bone loss at the proximal but not distal radius was significantly associated with fracture risk in women over the age of 65 years [15]. Nevertheless, in the OFELY study, women with incident fractures had 38 53% higher rates of bone loss prior to fracture compared with those without fracture, and bone loss in the highest tertiles at the distal, mid-, and ultra-distal radius was significantly associated with increased risk of all fractures [17]. Moreover, the rate of bone loss at the hip was an independent fracture risk predictor in older postmenopausal women [16, 18]. The Canadian Multicentre Study reported that, in women but not in men, models that included BMD change instead of baseline BMD were more likely to better estimate risk of hip and forearm fractures [19]. They found that a decrease of 0.01 g/cm 2 per year in the greater trochanter increased hip fracture risk by 2.39-fold, and a decrease of 0.03 g/cm 2 per year led to a 13.65-fold increased risk in women not using antiresorptive agents [19]. In men, BMD changes at the total hip of 0.01 g/cm 2 per year was associated with a 1.60-fold increase in any type of fracture, and a decrease of 0.03 g/cm 2 per year was associated with a 3.4-fold increase in fracture risk [19]. Our data did not show that forearm bone loss is predictive of incident fractures in men, which may be due to a lack of statistical power; therefore, larger studies are required to address this question in men. Although not significantly different, men without fracture were more physically inactive and reported worse health compared with men with fracture. Therefore, one can speculate that healthy men live an active lifestyle with increased risks of falling and

Osteoporos Int (2010) 21:1503 1511 1509 Table 3 Relative risk of fractures according to annual ultradistal forearm BMD percentage changes in men and women All models were adjusted for follow-up age a Excluding hand, foot, skull, and high-trauma fractures b Hip, wrist, and proximal humerus fractures 55 74years 50 74years Number of persons 1,336 1,208 All non-vertebral fractures Number of fractures 46 100 1: Baseline BMD 1.14 (0.85 1.54) 1.47 (1.18 1.84) 2: Follow-up BMD 1.14 (0.85 1.54) 1.48 (1.19 1.85) 3: BMD change (1SD%/year) 1.08 (0.81 1.43) 1.11 (0.90 1.36) Baseline BMD 1.14 (0.85 1.53) 1.48 (1.18 1.85) BMD change (1SD%/year) 1.06 (0.80 1.41) 1.11 (0.91 1.35) Follow-up BMD 1.13 (0.82 1.56) 1.50 (1.19 1.90) BMD change (1SD%/year) 1.03 (0.76 1.39) 0.97 (0.79 1.18) Low-trauma non-vertebral fractures a Number of fractures 23 81 : 1: Baseline BMD 1.61 (1.05 2.47) 1.48 (1.16 1.90) 2: Follow-up BMD 1.56 (1.02 2.40) 1.54 (1.21 1.97) 3: BMD change (1SD%/year) 1.03 (0.69 1.54) 1.20 (0.96 1.50) Baseline BMD 1.61 (1.05 2.47) 1.49 (1.16 1.91) BMD change (1SD%/year) 1.00 (0.68 1.46) 1.20 (0.96 1.49) Follow-up BMD 1.67(1.05 2.66) 1.52 (1.17 1.98) BMD change (1SD%/year) 0.86 (0.57 1.29) 1.04 (0.82 1.30) Osteoporotic fractures b Number of fractures 16 66 1: Baseline BMD 1.14 (0.69 1.86) 1.61 (1.22 2.13) 2: Follow-up BMD 1.15 (0.70 1.89) 1.66 (1.26 2.18) 3: BMD change (1SD%/year) 1.07 (0.68 1.69) 1.16 (0.91 1.49) Baseline BMD 1.13 (0.69 1.86) 1.62 (1.23 2.14) BMD change (1SD%/year) 1.07 (0.67 1.66) 1.16 (0.91 1.47) Follow-up BMD 1.14 (0.67 1.97) 1.67 (1.25 2.24) BMD change (1SD%/year) 1.02 (0.62 1.67) 0.98 (0.76 1.25) sustaining high impact fractures in comparison to less healthy men. Although there are controversies whether BMD predicts fracture risk equally in the two sexes, BMD is a predictor of fracture risk in women and men [38 40]. However, the larger skeleton achieved during growth produces stronger bones in men than in women, bones that tolerate larger absolute load [41]. Trabecular bone lost during aging mainly results in thinning of trabeculae in men and in loss of connectivity in women [42]. In addition, periosteal apposition increases the cross-sectional area of the bone adding more bone to the outer perimeter of the bone in men, and structural failure occurs less in men than in women because the bone strength is better maintained in men than in women [41, 42]. Although men lose less bone mass, they might even tolerate bone loss better than women when it comes to fracture risk. Overall, our findings are consistent with findings of previous studies that BMD loss is a predictor of fracture risk independent baseline BMD in women. However, none

1510 Osteoporos Int (2010) 21:1503 1511 of these studies adjusted for follow-up BMD. Our results suggested that the BMD level where an individual ends up (follow-up BMD) is the significant predictor of fracture risk and not the rate of bone loss. There was a significant correlation between rate of bone loss and follow-up BMD (r=0.54). This could indicate that increased rate of bone loss results in low follow-up BMD and subsequently higher fracture risk. On the other hand, stratifying the analyses by levels of follow-up BMD did not show any significant association between bone loss and fracture risk at any level of follow-up BMD; however, this cannot confirm the true association due to lack of statistical power. Therefore, more research is needed to clarify this point. Bias considerations The strength of this study is its population-based design, a high attendance rate in both sexes, a comprehensive quality control of densitometer performance, and a long period between BMD measurements which gives the possibility of obtaining reliable estimates of BMD changes and a well validated fracture registration. Notwithstanding a high response rate among males and females older than 50 years, participants attending both studies were younger, taller, had lower BMI (women), and better self-perceived health. They also smoked less and had higher BMD at both forearm sites [26]. Participants lost for follow-up may therefore be less healthy or had a less healthy lifestyle than those who participated in both studies. Nevertheless, loss of participants with possible higher bone loss rates might lead to an underestimation of effect in our study. Our follow-up data includes only the two forearm sites. However, measurements at any site predict any osteoporosis fracture equally well with a gradient of risk of approximately 1.5 per SD decrease in BMD [43]. Since we are interested in the overall risk, BMD changes at the distal forearm sites remain valid [44]. In this study, we excluded all women and men who suffered fractures between the two BMD measurements, in addition to all those reported history of previous fractures before the first BMD measurement. This was necessary in order to assess the effect the two BMD measurements have in predicting the same fracture outcomes (only incident fractures after Tromsø V) and also to exclude the bias caused by a potential accelerated bone loss induced by fractures between the BMD measurements [17 19]. Nevertheless, analyses including fractures between the BMD measurements controlled for previous fractures (data not shown) did not change the conclusion and therefore support including only incident fractures in the analyses. Conclusion The annual rate of bone loss at the distal forearm predicted non-vertebral fractures independently of baseline BMD, but not independently of follow-up BMD, in women. The BMD level, where an individual ends up, is the most strongly predictive of fracture risk in both sexes. Further research is needed to determine whether the BMD change or the end-up BMD level mostly predicts the fracture risk. Acknowledgment This work was supported by grant from the Centre for Research in the Elderly in Tromsø, the Northern Norway Regional Health Authority. Conflicts of interest References None. 1. Cummings SR, Melton LJ (2002) Epidemiology and outcomes of osteoporotic fractures. Lancet 359:1761 1767 [see comment] 2. Melton LJ 3rd (2003) Adverse outcomes of osteoporotic fractures in the general population. J Bone Miner Res 18:1139 1141 3. NIH Consensus Panel (1993) Consensus development conference: diagnosis, prophylaxis, and treatment of osteoporosis. Am J Med 9646 650 4. Kanis JA (2002) Diagnosis of osteoporosis and assessment of fracture risk. Lancet 359:1929 1936 5. Bass S, Delmas PD, Pearce G, Hendrich E, Tabensky A, Seeman E (1999) The differing tempo of growth in bone size, mass, and density in girls is region-specific. J Clin Invest 10795 804 [see comment] 6. Hough S (1998) Fast and slow bone losers. Relevance to the management of osteoporosis. Drugs & Aging 12(Suppl 1):1 7 7. Kanis JA, Borgstrom F, De Laet C, Johansson H, Johnell O, Jonsson B, Oden A, Zethraeus N, Pfleger B, Khaltaev N (2005) Assessment of fracture risk. Osteoporos Int 16:581 589 8. Garnero P, Sornay-Rendu E, Chapuy MC, Delmas PD (1996) Increased bone turnover in late postmenopausal women is a major determinant of osteoporosis. J Bone Miner Res 11:337 349 9. Garnero P, Hausherr E, Chapuy MC, Marcelli C, Grandjean H, Muller C, Cormier C, Breart G, Meunier PJ, Delmas PD (1996) Markers of bone resorption predict hip fracture in elderly women: the EPIDOS Prospective Study. J Bone Miner Res 11:1531 1538 10. Garnero P, Dargent MP, Hans D, Schott AM, Breart G, Meunier PJ, Delmas PD (1998) Do markers of bone resorption add to bone mineral density and ultrasonographic heel measurement for the prediction of hip fracture in elderly women? The EPIDOS Prospective Study. Osteoporos Int 8:563 569 11. Garnero P, Sornay-Rendu E, Claustrat B, Delmas PD (2000) Biochemical markers of bone turnover, endogenous hormones and the risk of fractures in postmenopausal women: the OFELY study. J Bone Miner Res 11526 1536 12. Johnell O, Oden A, De Laet C, Garnero P, Delmas PD, Kanis JA (2002) Biochemical indices of bone turnover and the assessment of fracture probability. Osteoporos Int 13:523 526 13. Vergnaud P, Garnero P, Meunier PJ, Breart G, Kamihagi K, Delmas PD (1997) Undercarboxylated osteocalcin measured with a specific immunoassay predicts hip fracture in elderly women: the EPIDOS study. JClin EndocrinolMetab 82:719 724 14. Riis BJ, Hansen MA, Jensen AM, Overgaard K, Christiansen C (1996) Low bone mass and fast rate of bone loss at menopause:

Osteoporos Int (2010) 21:1503 1511 1511 equal risk factors for future fracture: a 15-year follow- up study. Bone 19:9 12 15. Gnudi S, Malavolta N, Lisi L, Ripamonti C (2001) Bone mineral density and bone loss measured at the radius to predict the risk of nonspinal osteoporotic fracture. J Bone Miner Res 16:1130 1135 16. Nguyen TV, Center JR, Eisman JA (2005) Femoral neck bone loss predicts fracture risk independent of baseline BMD. J Bone Miner Res 20:1195 1201 17. Sornay-Rendu E, Munoz F, Duboeuf F, Delmas PD (2005) Rate of forearm bone loss is associated with an increased risk of fracture independently of bone mass in postmenopausal women: the OFELY study. J Bone Miner Res 20:1929 1935 18. Hillier TA, Stone KL, Bauer DC, Rizzo JH, Pedula KL, Cauley JA, Ensrud KE, Hochberg MC, Cummings SR (2007) Evaluating the value of repeat bone mineral density measurement and prediction of fractures in older women: the study of osteoporotic fractures. Arch Intern Med 167:155 160 19. Berger C, Langsetmo L, Joseph L, Hanley DA, Davison KS, Josse RG, Prior JC, Kreiger N, Tenenhouse A, Goltzman D (2009) Association between change in BMD and fragility fracture in women and men. J Bone Miner Res 2361 20. Berntsen GK, Fonnebo V, Tollan A, Sogaard AJ, Magnus JH (2001) Forearm bone mineral density by age in 7,620 men and women: the Tromso study, a population-based study. Am J Epidemiol 153:465 473 21. Schlenker RA, VonSeggen WW (1976) The distribution of cortical and trabecular bone mass along the lengths of the radius and ulna and the implications for in vivo bone mass measurements. Calcif Tissue Res 20:41 52 22. Berntsen GK, Tollan A, Magnus JH, Sogaard AJ, Ringberg T, Fonnebo V (1999) The Tromsø Study: artifacts in forearm bone densitometry prevalence and effects. Osteoporos Int 10:425 432 23. Berntsen GK, Fonnebo V, Tollan A, Sogaard AJ, Joakimsen RM, Magnus JH (2000) The Tromsø Study: determinants of precision in bone densitometry. J Clin Epidemiol 53:1104 1112 24. Emaus N, Berntsen GK, Joakimsen R, Fonnebo V (2005) Bone mineral density measures in longitudinal studies: the choice of phantom is crucial for quality assessment. The Tromso study, a population-based study. Osteoporos Int 16:1597 1603 25. Joakimsen RM, Fonnebo V, Sogaard AJ, Tollan A, Stormer J, Magnus JH (2001) The Tromso study: registration of fractures, how good are self-reports, a computerized radiographic register and a discharge register? Osteoporos Int 12:1001 1005 26. Emaus N, Berntsen GK, Joakimsen R, Fonnebo V (2006) Longitudinal changes in forearm bone mineral density in women and men aged 45 84 years: the Tromso Study, a population-based study. Am J Epidemiol 163:441 449 27. Ensrud KE, Palermo L, Black DM, Cauley J, Jergas M, Orwoll ES, Nevitt MC, Fox KM, Cummings SR (1995) Hip and calcaneal bone loss increase with advancing age: longitudinal results from the study of osteoporotic fractures. J Bone Miner Res 10:1778 1787 28. Nguyen TV, Sambrook PN, Eisman JA (1998) Bone loss, physical activity, and weight change in elderly women: the Dubbo Osteoporosis Epidemiology Study. J Bone Miner Res 13:1458 1467 29. Riis BJ (1995) The role of bone loss. Am J Med 98:S29 S32 30. Sowers M, Clark K, Wallace R, Jannausch M, Lemke J (1991) Prospective study of radial bone mineral density in a geographically defined population of postmenopausal Caucasian women. Calcif Tissue Int 48:232 239 31. Cauley JA, Lui L-Y, Barnes D, Ensrud KE, Zmuda JM, Hillier TA, Hochberg MC, Schwartz AV, Yaffe K, Cummings SR, Newman AB (2009) Successful skeletal aging: a marker of low fracture risk and longevity. The study of osteoporotic fractures (SOF). J Bone Miner Res 2134 143 32. Bainbridge KE, Sowers MF, Crutchfield M, Lin X, Jannausch M, Harlow SD (2002) Natural history of bone loss over 6 years among premenopausal and early postmenopausal women. Am J Epidemiol 156:410 417 33. Burger H, de-laet CE, van-daele PL, Weel AE, Witteman JC, Hofman A, Pols HA (1998) Risk factors for increased bone loss in an elderly population: the Rotterdam Study. Am J Epidemiol 147:871 879 34. Hannan MT, Felson DT, Dawson-Hughes B, Tucker KL, Cupples LA, Wilson PW, Kiel DP (2000) Risk factors for longitudinal bone loss in elderly men and women: the Framingham Osteoporosis Study. J Bone Miner Res 1710 720 35. Hui SL, Slemenda CW, Johnston CC Jr (1990) The contribution of bone loss to postmenopausal osteoporosis. Osteoporos Int 1:30 34 36. Jones G, Nguyen T, Sambrook P, Kelly PJ, Eisman JA (1994) Progressive loss of bone in the femoral neck in elderly people: longitudinal findings from the Dubbo osteoporosis epidemiology study. BMJ 309:691 695 37. Melton LJ, Atkinson EJ, O'Connor MK, O'Fallon WM, Riggs BL (2000) Determinants of bone loss from the femoral neck in women of different ages. J Bone Miner Res 124 31 38. Kanis JA, Johnell O, Oden A, De Laet C, Mellstrom D (2004) Epidemiology of osteoporosis and fracture in men. Calcif Tissue Int 790 99 39. Melton LJ 3rd, Orwoll ES, Wasnich RD (2001) Does bone density predict fractures comparably in men and women? Osteoporos Int 12:707 709 40. Orwoll E (2000) Assessing bone density in men. J Bone Miner Res 11867 1870 41. Seeman E (2003) The structural and biomechanical basis of the gain and loss of bone strength in women and men. Endocrinol Metab Clin North Am 32:25 38 42. Seeman E (2008) Bone quality: the material and structural basis of bone strength. J Bone Miner Metab 26:1 8 43. Marshall D, Johnell O, Wedel H (1996) Meta-analysis of how well measures of bone mineral density predict occurrence of osteoporotic fractures. BMJ 312:1254 1259 44. Stone KL, Seeley DG, Lui LY, Cauley JA, Ensrud K, Browner WS, Nevitt MC, Cummings SR, Osteoporotic Fractures Research Group (2003) BMD at multiple sites and risk of fracture of multiple types: long-term results from the study of osteoporotic fractures. J Bone Miner Res 18:1947 1954