Osteoporosis International

, Volume 17, Issue 3, pp 471–477 | Cite as

The long-term predictive value of bone mineral density measurements for fracture risk is independent of the site of measurement and the age at diagnosis: results from the Prospective Epidemiological Risk Factors study

  • Yu Z. Bagger
  • László B. Tankó
  • Peter Alexandersen
  • Henrik B. Hansen
  • Gerong Qin
  • Claus Christiansen
Original Article


Although the utility of bone mass measurements has been the subjects of extensive investigations, the number of studies comparing the predictive value of bone mass measurement at different skeletal sites in the same cohort with a large number of clinically verified endpoints is limited. Furthermore, scant information is available on how age at the time of diagnosis influence the risk of future fractures posed by low bone mineral density (BMD). We have followed 5,564 Danish postmenopausal women for a mean period of 7.3 years. Bone mineral content (BMC) at the forearm and BMD at the spine and hip were assessed at baseline. Vertebral fractures were assessed on digitalized images of lateral X-rays of the thoracic and lumbar spine, whereas non-vertebral fractures were self-reported. At follow-up, 17.6% of the women revealed an incidental vertebral fracture and 14.2% reported a new non-vertebral fracture. The absolute risk per 1,000 person-years of osteoporotic fracture increased significantly with decreasing bone mass at all three skeletal sites (P<0.001). Osteoporotic BMD (T-score ≤−2.5) had similar predictive values of fractures regardless of the skeletal site of measurement. Furthermore, the absolute risk of osteoporotic fractures increased significantly with increasing age at the same level of bone mass. Interestingly, the relative risk (RR) of vertebral fracture accompanying 1 SD decrease in spine BMD was similar across different age groups: <55 years (RR:2.1, 95% CI 1.3–3.3), 55–64 years (RR:2.3, 95%CI 1.7–3.2), 65–74 years (RR:2.0; 95%CI 1.5–2.6). Furthermore, women with any prior osteoporotic fracture had a 2.4-fold (95% CI 2.01–2.75, P<0.001) increased risk of a new vertebral fracture. Both age and prior fracture are strong predictors of future fractures. The long-term predictive value of bone mass measurement is independent of the site of measurement and the age at diagnosis.


Age Baseline bone mass measurement Different skeletal sites Prior fracture Subsequent fracture 


Osteoporosis and its consequences, vertebral and hip fractures, are increasing health concerns of industrialized countries. Increasing longevity of the elderly in these societies is among the main factors contributing to this trend. Osteoporosis is not only a disease of bones, it is also a common cause of morbidity and mortality and, thereby, frequently leads to prolonged hospitalization, demanding a considerable amount of financial resources from health care systems. In Europe it is estimated that 179,000 men and 611,000 women will suffer a hip fracture in each upcoming year, and related costs will provisionally reach 25 billion euros [1, 2, 3]. Effective counteraction of these unfavorable trends requires effective identification of high-risk individuals, which is a sine qua non of effective primary prevention.

Bone mass measurement by densitometry is an important tool for diagnosing osteoporosis [4, 5, 6, 7]. Although, recently, there has been much attention with regard to bone quality and its implications, low bone mineral density (BMD) remains among the most important determinants of osteoporotic fracture risk [8, 9, 10, 11, 12, 13, 14, 15, 16, 17]. It is known that the absolute risk of fractures at the same level of bone mass increases with increasing age. However, only limited data are available on whether the predictive value of BMD differs depending on the age of the subject at the time of assessment.

In the present study, we assessed the predictive value of bone mineral content (BMC)/BMD measurements at different skeletal sites and prior fractures for future fractures, and further investigated the influence of age on the predictive value. These questions were addressed in a large population-based cohort of 5,564 postmenopausal Danish women with a mean age of 63.7 years at baseline (range 45–70 years), who were followed for an average period of 7.3 years.

Subjects and methods

Study population

The Prospective Epidemiological Risk Factors (PERF) study follows-up on a large population-based cohort initiated to obtain further insight into the epidemiology and pathogenesis of menopause-related diseases, with special focus on osteoporosis. Detailed descriptions of the cohort and the study parameters assessed have been discussed elsewhere [18]. In brief, the original study population of 8,502 postmenopausal women, 45–70 years old (a mean age of 63.7±8.1 years), was recruited between 1977 and 1997 by large questionnaire surveys that invited women to participate in screening for various placebo-controlled clinical trials and epidemiological studies performed at the Center for Clinical and Basic Research. In 2000 to 2001, individuals were recalled for a follow-up examination. The follow-up information was obtained on 6,573 women (77.3%); 5,847 women of the original population participated in the PERF study and underwent re-examinations. Seven hundred and twenty-four women (8.5%) died during the overall follow-up period. Approximately 18.7% of participates received trial medications, including hormone replacement therapy (HRT) or bisphosphonates, for a period of 1–3 years during the original studies.

All women gave their written informed consent to inclusion in the PERF study, and the study was carried out in compliance with the Declaration of Helsinki II and the European Standards for Good Clinical Practice. The study protocol was approved by the local ethics committee.

Demographic characteristics, risk factors, and clinical events

At both baseline and follow-up, data on body weight and height, body mass index (BMI), age at menarche and menopause, number of births, education (primary/secondary/university), smoking habits (never/previous/current), regular alcohol consumption, diet, medical history, family history, and current or previous long-term use of drugs [e.g., HRT, bisphosphonates, selective estrogen receptor modulator (SERMs)] were gathered during personal interviews using a preformed questionnaire.

Bone mass measurements

During the period of 1977 to 1990, BMC of the forearm was measured by single-photon absorptiometry (SPA) using a bone mineral analyzer 1100 scanner (Mølsgaard Medical, Hørsholm, Denmark) or a DT-100 scanner (Osteometer MediTech A/S, Rødovre, Denmark). From 1992, BMC or BMD of the forearm was assessed by single-energy X-ray absorptiometry using a DTX-100/DTX-200 scanner.

In 1983–1984, BMD of the lumbar spine (L1–L4) was measured by dual photon absorptiometry (DPA) using a Lunar Radiation Corporation DP3 scanner (Lunar Corporation, Madison, Wis., USA). Since 1988, BMD of the lumbar spine and hip has been measured by dual-energy X-ray absorptiometry (DXA) using a QDR 1000 or a QDR-2000 scanner (Hologic, Waltham, Mass., USA).

Upon the introduction of any new scanner, cross-calibration of the scanners was performed by measuring BMD in 20 postmenopausal subjects aged 50–65 years by both the new and the old scanner. The correlation coefficients between the BMA-1100 and DT-100 scanners and the DT-100 and DTX 100/200 scanners were r=0.99 and r=0.95, respectively [19]. Similar coefficients were obtained when DP3 was correlated with the QDR 1000 scanner and the QDR1000 scanner with the QDR 2000 scanner (r=0.94, r=0.99, respectively). BMC/BMD values were then always corrected by the respective correction factors. Furthermore, the same BMD phantom and daily calibration regimen have been used ever since 1977 for the comparison of results at different time points.

A total of 5,564 women who had bone mass measurements performed at baseline and information on fractures were included in the present analysis. Among them, 5,409 women had a valid spine BMD at baseline, 2,738 women had total hip BMD, and 2,354 had forearm BMC. For the comparison of predictive abilities of central and peripheral bone mass measurements, a subgroup of 1,804 women, who underwent bone mass measurements at all three sites, was included in this analysis.

Fracture diagnosis

Lateral X-rays of the thoracic and lumbar spine were obtained at baseline and at the follow-up visit using standard X-ray equipment. Vertebral deformities from T4 to L4 were assessed by digital measurements of morphologic changes using the Image Pro Image Analyzer software (version 4.5 for Windows, Media Cybernetics, Silver Spring, Md., USA). The ratio of the anterior and posterior heights of each vertebral body was determined digitally, and a difference of more than 20% between the anterior and posterior edges was considered as a radiographic vertebral fracture. A new fracture or progression of previously existing vertebral fractures was determined by comparison of current status with previous X-rays. All vertebral fractures were based on radiographic diagnosis. At baseline and at the follow-up examination, information on prevalent non-vertebral fractures (wrist, hip, humeral fracture, rib, ankle, and foot) was collected and later verified by X-rays or hospital discharge summaries. Fractures caused by traffic accidents were excluded from the analyses.

Statistical analysis

The statistical analysis was carried out with SPSS data analysis software (version 11.01, SPSS, Chicago, Ill., USA). Demographic characteristics in women with or without incidental fractures were assessed by Student’s t-test for continuous variables or X2 tests for categorical variables. A general linear model was used to provide age-adjusted BMD/BMC at baseline. The study population was stratified into categories of normal BMD (T-score >−1), osteopenic BMD (T-score between ≤−1 and >−2.5) and osteoporotic BMD (T-score ≤−2.5). T-scores were calculated from the young adult normal white reference database as reported by the different manufacturers and then pooled. The reference ranges of BMD and their standard deviations (SD) from Hologic were 1.05±0.10 g/cm2 for the spine L1–L4, and 0.975±0.100 g/cm2 for the total hip, respectively. The reference range of forearm BMC and SD was 3.330±0.470 g. Furthermore, Z scores were calculated from the reference database of the scanner for each age group. Fracture incidence rates were expressed as numbers of fractures per 1,000 person-years, and the 95% confidence intervals (CIs) were calculated from the exact Poisson formula. We further estimated the relative risk of fractures associated with baseline BMD or prevalent fractures and the predictive ability of the spine BMD in women of different ages, using a logistic regression. Potential confounding variables were selected on the basis of significance in the bivariate analysis. All statistical tests were two-tailed, and differences were considered significant if the P value was less than 0.05.


Characteristics of the study population

The mean age of this population at baseline was 63.7±8.1 years, with a distribution across the different age groups as follows: >55 years, 18.6% (n=1,037); 55–64 years, 32.0% (n=1,781); 65–74 years, 42.4% (n=2,359); ≥75 years, 7.0% (n=387).The mean follow-up period was 7.3±3.9 years.

Table 1 summarizes the demographic characteristics of women with or without incidental fractures. Women with a fracture were, on average, 2 years older (P<0.001). Other characteristics addressed showed significant differences, such as a higher frequency of ever smoking, an earlier entry into menopause, and higher percentages of prior fractures and family history of osteoporosis. Furthermore, women with manifest osteoporosis at follow-up more frequently reported ongoing treatment for osteoporosis (15.9% vs 7.3%, P<0.001 ). As expected, there was a significantly greater reduction in the height of women with at least one vertebral fracture (−2.2±9.2 cm) than in those without (−1.4±7.6 cm, P=0.003). A comparison of age-adjusted BMD/BMC at the different skeletal sites shows that women with fractures had significantly lower values at baseline. However, there were no differences between women with or without fractures regarding age at menarche, BMI, past use of HRT, and previously receiving any treatment for prevention of osteoporosis in clinical trials.
Table 1

Demographic characteristics in women with or without new fractures


Women without a fracture (n=3,973)

Women with a fracture (n=1,591)


Age at baseline (years)




Age at menarche (years)




Age at menopause (years)




BMI at baseline (kg/m2)




Spine BMD (g/cm2) at baselinea




Total hip BMD (g/cm2) at baselinea




Forearm BMC (g) at baselinea




Prior fracture (%)




Ever smoking (%)




Family history of osteoporosis (%)




Past use of HRT(>1 year, %)




Receiving treatment in clinical trials (>1 year, %)




Current treatment for osteoporosis (%)




Values are means±SDs

aAge-adjusted BMD/BMC expressed as mean BMD/BMC±SEM

At follow-up, 17.6% of the women had a new radiologically proven vertebral fracture and 14.2% reported occurrence of an incidental non-vertebral fracture. Among non-vertebral fractures, a wrist fracture had been experienced by 38.1% of women. It is to be emphasized that the number of hip fractures was low in this study population, and only 1.9% of this population reported a hip fracture, and 1% of them had experienced a new hip fracture during the follow-up period (e.g., the incidence rate was 1.3 per 1,000 person-years). If all kinds of fractures were taken into account, the incidence of all fractures in this population was 28.6%.

Baseline bone mass and fracture

Women experiencing a fracture had significantly lower baseline BMD/BMC at the different skeletal sites independently of age (Table 1). This population was stratified according to their baseline BMD/BMC into categories of normal, osteopenic and osteoporotic BMD/BMC in terms of the WHO criteria. As indicated in Table 2, the incident rates of the different fractures per 1,000 person-years increased significantly with decreasing BMD/BMC across the three categories (P<0.001). The absolute risks of fractures in the same category was independent of the site of measurement.
Table 2

Incident rate per 1,000 person-years of future osteoporotic fractures in women, stratified by WHO criteria based on the different bone mass measurements


No. of women with a fracture/incident rate for an incidental fracture per 1,000 person-years (95% confidence interval)



Vertebral fracture

Non-vertebral fracture

Any fracture



No. of events (no. at risk)

Incident rate (95%CI)

No. of events (no. at risk)

Incident rate (95%CI)

No. of events (no. at risk)

Incident rate (95%CI)

Spine BMD (follow-up =6.9 years)

Normal BMD

251 (1,753)

19.4 (16.9–22.1)

230 (1774)

17.6 (15.3–20.3)

437 (1567)

35.5 (31.9–39.4)


373 (1,891)

26.0 (23.2–29.1)

332 (1932)

22.9 (20.4–25.8)

642 (1640)

47.4 (43.3–51.9)


323 (818)

47.8 (42.0–54.3)

206 (935)

28.8 (24.7–33.5)

458 (683)

72.8 (64.8–81.7)

Total hip BMD (follow-up =7.9 years)

Normal BMD

96 (897)

12.9 (10.3–15.9)

99 (894)

13.3 (10.7–16.4)

182 (811)

25.5 (22.0–29.5)


224 (1,194)

21.7 (18.8–25.0)

238 (1180)

23.2 (20.1–26.7)

423 (995)

44.4 (39.6–49.7)


92 (235)

41.4 (32.4–52.8)

68 (259)

29.4 (22.2–38.4)

143 (184)

70.9 (57.2–87.8)

Forearm BMC (follow-up =9.9 years)

Normal BMD

149 (1,120)

12.6 (10.6–15.0)

165 (1104)

14.0 (11.9–16.6)

289 (980)

26.0 (22.7–29.6)


159 (758)

19.2 (16.1–22.8)

189 (728)

23.2 (19.7–27.2)

312 (605)

41.4 (36.1–47.4)


48 (120)

33.7 (23.8–47.2)

36 (132)

24.2 (16.4–35.2)

73 (95)

56.1 (41.5–75.9)

In order to compare predictive abilities of central and peripheral bone mass measurements, we studied a subgroup of 1,804 women who had undergone bone mass measurements at all three sites (Table 3). Women with osteoporotic bone mass had a 2–3-times higher relative risk of fractures than those women with normal bone mass, even after adjustment for age, prior fractures, previous HRT or other treatments for osteoporosis. Bone mass measurements assessed at the spine, hip or forearm had comparable predictive value for predicting the different types of fractures (Table 3).
Table 3

The predictive value of baseline bone mass measured at the different skeletal sites in a subpopulation (n=1,804)



Percentage of women with a fracture/relative risk (RR, 95% confidence interval) for an incidental fracture



Vertebral fracture

Non-vertebral fracture

Any fracture



Percentage of women

Adjusted RRa

Percentage of women

Adjusted RRa

Percentage of women

Adjusted RRa

Spine BMD

Normal BMD









1.3 (0.9–1.9)


1.5 (1.1–2.0)


1.5 (1.1–1.9)



3.1 (2.1–4.6)


2.2 (1.5–3.2)


3.0 (2.2–4.1)

Total hip BMD

Normal BMD









1.4 (1.0 –1.9)


1.6 (1.2–2.2)


1.7 (1.3–2.1)



2.9 (1.8–4.6)


2.3 (1.4–3.5)


2.9 (1.9–4.3)

Forearm BMC

Normal BMC









1.3 (0.9–1.7)


1.9 (1.5–2.6)


1.7 (1.3–2.1)



2.2 (1.3–3.7)


2.2 (1.3–3.7)


2.4 (1.6–3.7)

a Based on logistic regression model adjusted for age, prior fractures, previous HRT and other treatment for osteoporosis

Age and the risk of a new fracture

The study population was stratified into four age categories as follows: age <55 years; age between 55 years and 64 years; age between 65 years and 74 years; age ≥75 years. Figure 1 illustrates the incident rate of the different fractures per 1,000 person-years in 10-year age categories. The incident rates of fractures increased exponentially with advancing age. The vertebral fracture rate per 1,000 person-years was 42.4 (95%CI 19.7–25.5) in women older than 65 years of age, and 100.5 (95% CI 84.1–118.9) in women aged over 75 years. The incident hip fracture rate was 2.6 (95% CI 1.8–3.6) in women over 65 years, and 4.2 (95% CI 1.6–9.2) in women over 75 years. However, the incident rate of wrist fracture was relatively constant between the ages of early 60s to late 70s (from 8.4 to 10.7, Fig. 1). Furthermore, the new incidence rate of any osteoporotic fractures per 1,000 person-years increased dramatically across age groups (the rate changes from 24.8 per 1,000 person-years to 133.6 per 1,000 person-years, Fig. 1).
Fig. 1

Incident rate of osteoporotic fractures per 1,000 person-years in the 10-year age categories

Figure 2 shows the absolute risk of incidental fractures per 1,000 person-years in the different Z scores of baseline spine BMD by the various age groups. As shown in the figure, the rate of fracture increased significantly with increasing age. For a given level of BMD, the absolute risk of fractures, especially vertebral fracture risk, is much greater for older than for younger women. For example, a difference in age of 10 years between 55–64 years and 65–74 years at a Z score of 0.1 to >-1 confers a difference in vertebral fracture risk of 55.8 per 1,000 person-years (Fig. 2, left panel), indicating that age appears to have a stronger effect on fracture rate. However, for women aged over 75 years, the non-vertebral fracture rate was lower than those for women 10 years younger (Fig. 2, middle panel).
Fig. 2

Incident rate of osteoporotic fractures per 1,000 person-years in the different Z scores of baseline spine BMD by the various age groups

We further studied whether the predictive value of BMD for fractures differs depending on the age of the subject at the time of assessment. Table 4 shows the relative risk of fractures for a 1 SD decrease in Z score of baseline spine BMD by 10-year age group. The relative risk of vertebral, non-vertebral, or any osteoporotic fractures by a decrease of 1 SD of spine BMD was comparable for women aged from their early 50s to 75 years, even after adjustment for prior fractures, previous HRT or any treatment for osteoporosis (Table 4). However, women aged over 75 years had slightly higher risk of vertebral or any fractures than those women younger than 75 years; in contrast, the relative risk of non-vertebral fracture was lower (Table 4).
Table 4

Relative risk (RR) of vertebral fracture for a 1 SD decrease in Z score of spine BMD at baseline by each age group

Age group

RRa (95% confidence interval)


Vertebral fracture

Non-vertebral fracture

All fractures

<55 years

2.1 (1.3–3.3)

1.4 (0.9–2.1)

1.7 (1.2–2.4)

55–64 years

2.3 (1.7–3.2)

1.6 (1.2–2.2)

2.0 (1.5–2.6)

65–74 years

2.0 (1.5–2.6)

1.7 (1.3–2.4)

2.0 (1.6–2.6)

>75 years

3.5 (1.7–7.4)

0.9 (0.3–2.9)

2.8 (1.3–5.8)

aBased on logistic regression model adjusted for prior fractures, past use of HRT or other treatment for osteoporosis, and follow-up period (reference group Z score >−1.0)

Prior fracture and the risk of a new fracture

At baseline, 4.1% of the women had at least one prevalent vertebral fracture, whereas 15.3% reported at least one prior non-vertebral fracture. Women with any prior fractures (totally, 18.9%) at baseline had a significantly higher risk for a new vertebral or non-vertebral fracture. Women with a prior fracture had a 2.3-times (95% CI 2.01–2.72, P<0.001) higher risk for a future vertebral fracture than did those without a prior fracture. After adjustment for age and baseline spine BMD, the relative risk remained significant (RR 1.7, 95% CI 1.5–2.1, P<0.001). The relative risk for all osteoporotic fractures combined was 1.8-times higher (95% CI 1.56–2.08, P<0.001) without any adjustment, and 1.5-times higher (95%CI 1.2–1.6, P<0.001) after adjustment for age and BMD. Furthermore, women with a prior vertebral fracture had a 4.3-times (95% CI, 3.2–5.6, P<0.001) higher risk for a subsequent vertebral fracture, even after adjustment for age and baseline BMD.


This study has investigated prospective associations between baseline bone mass measured at the different skeletal sites, prior fractures, and the subsequent risk of vertebral and non-vertebral fractures, in a large population-based cohort including 5,564 postmenopausal women. The main findings were as follows: (1) the baseline BMD/BMC of the spine, hip, or forearm was a significant predictor of future osteoporotic fractures with virtually similar predictive values; (2) the fracture rates, especially vertebral or hip fractures, increased exponentially with advancing age; (3) prior fracture was a strong and independent risk factor for subsequent fractures and, in particular, vertebral fractures; (4) the long-term predictive value of bone mass measurements for fracture risk is independent of age at diagnosis.

It is well-established that the fracture risk in postmenopausal women increases with decreasing BMD [8, 9]. Some studies have indicated that a given site of bone mass measurement best predicts fracture risk for that specific site [11, 20]. Until now most studies have focused on a single skeletal bone mass measurement (calcaneus, proximal radius, distal radius, or femoral neck) for prediction of fractures [8, 12, 13, 14, 15, 16]. Data on comparison of the different skeletal sites in the same cohort are limited [9, 10, 11]. In the present study, we have been able to demonstrate relationships between the absolute and relative risk of the different skeletal fractures and bone mass measurements assessed at the spine, hip and forearm in a large population. Our findings indicated that bone mass assessed at the spine, hip or forearm has virtually the same predictive ability for vertebral, non-vertebral, and all fractures (Tables 2 and 3). Furthermore, although previous studies suggested that the predictive value of BMD measurements is site specific (e.g., hip BMD is predictive for hip fracture only) [11, 20], our study indicated that BMD at the spine or hip was equally good for predicting vertebral or non-vertebral fractures (Table 3).

Our data further emphasize the importance of a prior fracture in determining susceptibility to further fractures [15, 21, 22]. In line with these studies, the risk of incident vertebral fracture was increased by a two-fold factor in those with any prior fracture, and four-fold in those with a baseline vertebral fracture. Adjustment for age and BMD did not alter the relative risk, indicating that a prior fracture is an independent predictor of subsequent fractures. Since many fractures and vertebral deformity are asymptomatic and may be neglected by the patient, it remains an ongoing challenge to find these patients for timely initiation of secondary prevention.

Age is also an important factor for osteoporosis. In the present study we have shown that the vertebral fracture rate, but also the hip fracture rate, increases exponentially with increasing age (Fig. 1), suggesting that the fracture risk is strongly modulated by age per se. In contrast, the new incidence of wrist fracture was relatively constant between the ages of 60 years to late 70s. Furthermore, our study has confirmed previous observations [8] demonstrating that the absolute risk for fracture increases with increasing age at the same level of BMD. However, it has been less studied whether the predictive ability of BMD for fractures differs with age. Most published studies have focused on the elderly population aged ≥65 years. Very recently, the National Osteoporosis Risk Assessment (NORA) study has demonstrated for the first time that low BMD in younger postmenopausal women, 50–64 years of age, showed a 1-year relative risk of fracture similar to that found in women aged 65 years and older [23]. In accordance with the NORA study, our study also included 50% of women aged <65 years (>55 years, 18.6% and 55–64 years, 32.0%). We found similar results to those of the NORA study, showing that the relative risk of future fractures for a decrease of 1 SD of BMD was similar for women in 10-year age intervals between <55 years and 74 years but not for those aged over 75 years (Table 4). These findings can confirm to physicians that the predictive value of BMD for future fracture risk has a comparable relevance in both younger and older postmenopausal women.

Bone mass measurement combined with other well-established risk factors, such as smoking, age at menopause, prior fractures and family history of osteoporosis, will likely enable a more accurate identification of subjects at increased risk for fractures and facilitate adequate decision making for the subsequent clinical management of the condition of postmenopausal women with low bone mass. Thus, the number of patients needed to treat (NNT) with an intervention could be more precise, and the cost-effectiveness of the intervention could increase. However, no treatments are presently given for a lifetime due to side effects of continued treatment. Recently, some studies have shown that the effects of antiresorptive interventions (HRT or bisphosphonates) of 2–6 years appear to persist when treatment is stopped [24, 25, 26, 27]. Furthermore, the risk of osteoporotic fractures was also significantly reduced by 52% in women who received 2–3 years’ treatment with HRT [27]. These findings have important socioeconomic implications for osteoporosis management.

Our study has some limitations to be pointed out. The incidence of hip fractures in the present cohort was 1.3 per 1,000 person-years, which is considerably lower than that recently reported by Vestergaard et al. who used data from the Central Patient Registry in Denmark (8.2 per 1,000 person-years) [28]. The lower incidence rate is likely attributable to selection bias in our population that consisted of voluntary subjects recruited via questionnaire surveys. Thus, we cannot exclude the possibility that subjects with limited mobility following a hip fracture chose not to attend the clinic for re-examination. The likelihood of selection bias is also corroborated by the fact that non-responders (22.7%) were older than responders (baseline age 65.5 vs 63.7 years, P<0.001). Furthermore, a number of methodological issues should also be considered when one is interpreting the results. Because of rapid advances in the technology of bone measurement, bone mass has been measured by different bone densitometers over the past decades. With strictly standardized quality control procedures since 1977, we made every effort to make measurements comparable. In addition, scan quality has also been reviewed and cross-calibrated by a central facility over time.

In conclusion, this study further advocates the utility of bone mass measurement, which seems to offer effective fracture prediction independent of the site of measurement and age of the patient. Therefore, women with well-established risk factors for osteoporosis should be identified at an earlier stage so that adequate measures of diagnosis and therapy can be initiated in a timely manner. The need for the achievement of timely action is also emphasized by the high risk of further fractures in those who already have manifest disease (prior fracture), which, with currently available medications, can only partly be counteracted (~50%).


  1. 1.
    Seeman E, Hopper JL, Bach LA, Cooper ME, Parkinson E, McKay J, et al (1989) Reduced bone mass in daughters of women with osteoporosis. N Engl J Med 320:554–558PubMedCrossRefGoogle Scholar
  2. 2.
    Black DM, Cooper C (2000) Epidemiology of fractures and assessment of fracture risk. Clin Lab Med 20:439–453PubMedGoogle Scholar
  3. 3.
    Compston J (2004) Action plan for the prevention of osteoporotic fractures in the European Community. Osteoporos Int 15:259–262PubMedCrossRefGoogle Scholar
  4. 4.
    Kanis JA, Melton LJ III, Christiansen C, Johnston CC, Khaltaev N (1994) The diagnosis of osteoporosis. J Bone Miner Res 9:1137–1141PubMedCrossRefGoogle Scholar
  5. 5.
    Kanis JA, Delmas P, Burckhardt P, Cooper C, Torgerson D (1997) Guidelines for diagnosis and management of osteoporosis. The European foundation for osteoporosis and bone disease. Osteoporos Int 7:390–406PubMedCrossRefGoogle Scholar
  6. 6.
    Kanis JA (2002) Diagnosis of osteoporosis and assessment of fracture risk. Lancet 359:1929–1936CrossRefPubMedGoogle Scholar
  7. 7.
    World Health Organization (1998) Guidelines for preclinical evaluation and clinical trials in osteoporosis. WHOGoogle Scholar
  8. 8.
    Hui SL, Slemenda CW, Johnston CC Jr (1989) Baseline measurement of bone mass predicts fracture in white women. Ann Intern Med 111:355–361PubMedGoogle Scholar
  9. 9.
    Wasnich RD, Ross PD, Davis JW, Vogel JM (1989) A comparison of single and multi-site BMC measurements for assessment of spine fracture probability. J Nucl Med 30:1166–1171PubMedGoogle Scholar
  10. 10.
    Melton LJ III, Atkinson EJ, O’Fallon WM, Wahner HW, Riggs BL (1993) Long-term fracture prediction by bone mineral assessed at different skeletal sites. J Bone Miner Res 8:1227–1233PubMedGoogle Scholar
  11. 11.
    Cummings SR, Black DM, Nevitt MC, Browner W, Cauley J, Ensrud K, et al (1993) Bone density at various sites for prediction of hip fractures. The study of osteoporotic fractures research group. Lancet 341:72–75CrossRefPubMedGoogle Scholar
  12. 12.
    Gardsell P, Johnell O, Nilsson BE, Gullberg B (1993) Predicting various fragility fractures in women by forearm bone densitometry: a follow-up study. Calcif Tissue Int 52:348–353PubMedCrossRefGoogle Scholar
  13. 13.
    Nguyen T, Sambrook P, Kelly P, Jones G, Lord S, Freund J, et al (1993) Prediction of osteoporotic fractures by postural instability and bone density. BMJ 307:1111–1115PubMedCrossRefGoogle Scholar
  14. 14.
    O’Neill TW, Lunt M, Silman AJ, Felsenberg D, Benevolenskaya LI, Bhalla AK, et al (2002) Relationship between bone density and incident vertebral fracture in men and women. J Bone Miner Res 17:2214–2221PubMedCrossRefGoogle Scholar
  15. 15.
    Albrand G, Munoz F, Sornay–Rendu E, DuBoeuf F, Delmas PD (2003) Independent predictors of all osteoporosis-related fractures in healthy postmenopausal women: the OFELY study. Bone 32:78–85CrossRefPubMedGoogle Scholar
  16. 16.
    Schuit SC, van der KM, Weel AE, De Laet CE, Burger H, Seeman E, et al (2004) Fracture incidence and association with bone mineral density in elderly men and women: the Rotterdam study. Bone 34:195–202CrossRefPubMedGoogle Scholar
  17. 17.
    Kanis JA, Johnell O, Oden A, Dawson A, De Laet C, Jonsson B (2001) Ten year probabilities of osteoporotic fractures according to BMD and diagnostic thresholds. Osteoporos Int 12:989–995CrossRefPubMedGoogle Scholar
  18. 18.
    Bagger YZ, Riis BJ, Alexandersen P, Tankó LB, Christiansen C (2001) Risk factor for development of osteoporosis and cardiovascular disease in postmenopausal Danish women: The PERF study. J Bone Miner Res 16:396Google Scholar
  19. 19.
    Borg J, Mollgaard A, Riis BJ (1995) Single X-ray absorptiometry: performance characteristics and comparison with single photon absorptiometry. Osteoporos Int 5:377–381CrossRefPubMedGoogle Scholar
  20. 20.
    Marshall D, Johnell O, Wedel H (1996) Meta-analysis of how well measures of bone mineral density predict occurrence of osteoporotic fractures. BMJ 312:1254–1259PubMedGoogle Scholar
  21. 21.
    Klotzbuecher CM, Ross PD, 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 15:721–739PubMedCrossRefGoogle Scholar
  22. 22.
    Kanis JA, Johnell O, De Laet C, Johansson H, Oden A, Delmas P, et al (2004) A meta-analysis of previous fracture and subsequent fracture risk. Bone 35:375–382CrossRefPubMedGoogle Scholar
  23. 23.
    Siris ES, Brenneman SK, Miller PD, Barrett–Connor E, Chen YT, Sherwood LM, et al (2004) Predictive value of low BMD for 1-year fracture outcomes is similar for postmenopausal women ages 50–64 and 65 and older: results from the National Osteoporosis Risk Assessment (NORA). J Bone Miner Res 19:1215–1220PubMedCrossRefGoogle Scholar
  24. 24.
    Bagger YZ, Tankó LB, Alexandersen P, Ravn P, Christiansen C (2003) Alendronate has a residual effect on bone mass in postmenopausal Danish women up to 7 years after treatment withdrawal. Bone 33:301–307CrossRefPubMedGoogle Scholar
  25. 25.
    Stock JL, Bell NH, Chesnut CH III, Ensrud KE, Genant HK, Harris ST, et al (1997) Increments in bone mineral density of the lumbar spine and hip and suppression of bone turnover are maintained after discontinuation of alendronate in postmenopausal women. Am J Med 103:291–297CrossRefPubMedGoogle Scholar
  26. 26.
    Orr–Walker B, Wattie DJ, Evans MC, Reid IR (1997) Effects of prolonged bisphosphonate therapy and its discontinuation on bone mineral density in post-menopausal osteoporosis. Clin Endocrinol (Oxf) 46:87–92CrossRefGoogle Scholar
  27. 27.
    Bagger YZ, Tanko LB, Alexandersen P, Hansen HB, Mollgaard A, Ravn P et al (2004) Two to three years of hormone replacement treatment in healthy women have long-term preventive effects on bone mass and osteoporotic fractures: the PERF study. Bone 34:728–735CrossRefPubMedGoogle Scholar
  28. 28.
    Vestergaard P, Rejnmark L, Mosekilde L (2005) Osteoporosis is markedly underdiagnosed: a nationwide study from Denmark. Osteoporos Int 16:134–141CrossRefPubMedGoogle Scholar

Copyright information

© International Osteoporosis Foundation and National Osteoporosis Foundation 2005

Authors and Affiliations

  • Yu Z. Bagger
    • 1
  • László B. Tankó
    • 1
  • Peter Alexandersen
    • 1
  • Henrik B. Hansen
    • 1
  • Gerong Qin
    • 1
  • Claus Christiansen
    • 1
  1. 1.Center for Clinical and Basic Research A/SBallerupDenmark

Personalised recommendations