| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Bone and Mineral Research Program (N.D.N., J.A.E., J.R.C., T.V.N.), Garvan Institute of Medical Research, and the Department of Endocrinology (J.A.E., J.R.C.), St. Vincent’s Hospital, and the Faculty of Medicine (J.A.E., T.V.N.), The University of New South Wales, Sydney, New South Wales 2010, Australia
Address all correspondence and requests for reprints to: Tuan V. Nguyen, Bone and Mineral Research Program, Garvan Institute of Medical Research, St. Vincent’s Hospital, 384 Victoria Street, Darlinghurst, Sydney, New South Wales 2010, Australia. E-mail: t.nguyen{at}garvan.org.au.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Design: This study was part of the ongoing Dubbo Osteoporosis Epidemiology Study, which was designed as a prospective population-based cohort investigation.
Participants: At baseline, 924 women and 723 men aged 60+ yr did not have osteoporosis [bone mineral density (BMD) T-scores > –2.5]. The individuals have been followed for up to 15 yr.
Main Outcome Measures: Atraumatic fractures were prospectively identified through radiologists’ reports.
Risk Factors: At baseline, femoral neck BMD (FNBMD) was measured by dual energy x-ray absorptiometry (DXA); history of fall, postural stability, and quadriceps strength was obtained.
Results: During the follow-up period, among the nonosteoporotic group, 221 women and 105 men had sustained a fracture, accounting for 55 and 74% of total fractures in the entire Dubbo Osteoporosis Epidemiology Study sample, respectively. The following factors were independent risk factors for any fracture: in women, age per SD (hazard ratio, 1.2; 95% CI, 1.0–1.3), postural sway per SD (1.1, 1.0–1.2), FNBMD per SD (1.6, 1.3–1.9), fall in the previous 12 months (2.1, 1.6–2.7), and prior fracture (1.8, 1.2–2.7); in men, age (1.4, 1.1–1.6), postural sway (1.2, 1.0–1.3), FNBMD (1.2, 1.0–1.5), and fall in the previous 12 months (1.9, 1.2–3.0). Exposure to at least one of the risk factors could account for 49% (women) and 39% (men) of any fractures in this population.
Conclusion: In nonosteoporotic elderly, the combination of low BMD, advancing age, fall during the last 12 months, and prior fracture could identify a subgroup of individuals with high risk of fracture.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
–2.5) should be treated to reduce their risk of fracture (4, 5). Even in these high-risk individuals, treatment can reduce the risk by around 50% (6). In recent years, it has been realized that approximately half of all fracture cases occur in individuals without osteoporosis (7). If the aim is to reduce fracture burden in the community, some individuals in this group should also be considered for intervention. However, it is not known which factors are associated with fracture risk in this group. The aim of this study was to estimate the incidence fracture and to assess the association between risk factors and fracture risk in men and women with osteopenia and "normal" BMD.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Data used in this analysis were derived from the ongoing Dubbo Osteoporosis Epidemiology Study (DOES) for which details of protocol and study design have been previously described (8, 9). The sampling frame for DOES is the city of Dubbo, New South Wales (Australia), a locality of approximately 32,000 people, 98.6% white, of which 1581 men and 2095 women were aged 60 yr or older in 1989. Dubbo had been selected for the study because its age and gender distribution of the population closely resemble the Australian population and its relative isolation in terms of medical care allows virtually complete ascertainment of all fractures. In total, the rate of participation was higher in women (69%) than in men (58%). Among those who agreed to participate in the study, approximately 90% also agreed to have BMD measured. This study was approved by the St. Vincent’s Campus Research Ethics Committee and informed written consent was obtained from each participant.
Assessment of risk factors
In addition to anthropometric measurements, each individual was interviewed by a nurse coordinator who administered a structured questionnaire to obtain information on lifestyle and dietary calcium intake. Body weight (kilograms) with light clothing and without shoes was obtained (to the nearest 0.1 kg) on an electronic scale. Standing height (centimeters) without shoes was measured (to the nearest 0.1 cm) by a wall-mounted stadiometer. Lifestyle and clinical data (8, 9, 10), any history of falls in the preceding 12 months, and any history of fractures in the past 5 yr were obtained. Information on current and past smoking habit was obtained from each person that was quantified as the number of pack-years consumed in each 10-yr interval age group.
Dietary calcium intake was assessed based on a food frequency questionnaire as described elsewhere (11). The questionnaire instructed subjects to answer the frequency of consumption of individual food items such as milk and dairy products, eggs, fish, cereals, fruits and vegetables, and others. The total amount of dietary calcium intake then was estimated for each individual nutrient by using a standardized calcium content value.
Physical activity was assessed at baseline as the average number of hours per day spent in each of five levels of activity based on the Framingham Massachusetts Heart Study (12). The five activities were: basal activity (weighting factor 1.0, e.g. sleeping, lying down), sedentary (weighting factor 1.1, e.g. sitting, standing), light (weighting factor 1.5, e.g. casual walking), moderate (weighting factor 2.4, e.g. gardening, carpentry), or heavy (weighting factor 5, e.g. lifting, digging). The weighting or intensity factor was used based on the approximate oxygen consumption needed for each level of activity (12). The products of hours and weighting factor for all activities were then summed to yield an index of total physical activity per week. High total physical activity index corresponds to lifestyles that are physically active, and low levels correspond to habitual inactivity.
Various risk factors for falling were tested on each subject at baseline. Quadriceps strength (maximum isometric contraction) was measured in the sitting position in the subject’s dominant (stronger) leg with a horizontal spring gauge calibrated up to 50 kg force. This method has a reliability coefficient of 0.92 (13). Body sway was measured as displacement of the body at the level of the waist in 30-sec periods. The area (square millimeters) encompassing all movements, forward and backward, left and right, was used as the sway. Four test conditions were used comprising: eyes open and closed on firm surface (wooden floor) and eyes open and closed on compliant surface (high density foam 15-cm high). Full descriptions of these assessments, their test and retest reliability scores, and confidence intervals (CIs) have been given elsewhere (13). The highest value of area of sway obtained was used in the analysis.
BMD (grams per square centimeter) was measured at the lumbar spine and femoral neck by dual energy x-ray absorptiometry using a LUNAR DPX densitometer (GE-LUNAR, Madison, WI). The radiation dose with this method is less than 0.1 µGy. The coefficient of variation of BMD in our institution in normal subjects was 1.5% and 2% at the proximal femur and lumbar spine, respectively (14). Based on the actual values of femoral neck BMD (FNBMD) obtained, each subject was classified as "osteoporotic" with a BMD being 2.5 SD or more below the young normal level or "osteopenic" with a BMD between 2.5 to 1.1 SD below the young normal level or as "normal." The "young normal" BMD was obtained from a sample of 52 Australian men and women aged between 20 to 32 yr. These values are identical to those of LUNAR white database. In this study, individuals with BMD T-scores more than –2.5 were included in the analysis.
Fracture ascertainment
Fractures occurring during the study period were identified for residents of the Dubbo local government area through radiologists’ reports from the two centers providing x-ray services as previously described (8, 9). Fractures were only included if the report of fracture was definite and, on interview, had occurred with minimal trauma (fall from standing height or less). Fractures clearly resulting from major trauma (such as motor vehicle accidents) or from underlying diseases (such as cancer or bone-related diseases) were excluded from the analysis. Any fractures more than 3 months before study entry were not considered in the analysis. Fracture was analyzed as any osteoporotic fracture or as subgroups of hip; symptomatic vertebral; wrist and forearm, including Colles’, Smith’s, and metacarpal fractures; and major fracture, including any osteoporotic fracture, except fracture of the skull, face, fingers, toes, patella, distal tibia ankle, distal tibia, malleus, less than two ribs, and other nonspecified fractures.
Statistical analysis
Incidence of fractures was calculated as fracture cases per 1000 person-yr. Cox’s proportional hazards regression model was used to estimate relative risk and 95% CI for each SD change or in specified groups compared with the reference group with categorized risk factors. The outcomes in this model were fracture incidence and the time to fracture from baseline BMD measurement. Stepwise and backward algorithms were used to search for a model with maximum discriminatory power. The significance of parameter estimates derived from the Cox’s proportional hazards model (15) was tested with the likelihood ratio statistic (16). The assumption of proportional hazards for the levels of each risk factor was tested by evaluating the linearity of plots of log(–log(S[ti])j), where S(ti)j describes the jth survival time for the ith level (i = 1, 2) for each risk factor. The contribution of risk factors to hip fracture risk was further evaluated in a descriptive analysis in which the measurement of each non-BMD risk factor was dichotomized into two categories (presence or absence). A risk score, or more specifically the number of risk factors, was then derived as the sum of all individual risk factors for each individual. Incidence and relative risk of hip fracture were then computed for each risk score independent of or in combination with BMD values.
To estimate the proportion of fracture that may be hypothetically reduced by the elimination of the risk factors, population-attributable risk fraction (PARF) was calculated. The PARF is a function of two parameters: the prevalence of a risk factor and the relative risk associated with the risk factor. To estimate the prevalence, each of the independent risk factors (obtained from the first stage of analysis) was dichotomized into high-risk and low-risk groups. The relative risk associated with risk of specific fracture types for the high-risk category was then estimated from the multiple logistic model. The statistical estimation of PARF was based on the "sequential attributable fractions" (17). Briefly, for each threshold criterion used to define high-risk individuals, the expected probability of fracture was calculated from the estimated coefficients of the multivariate logistic analysis model. The expected probability was then compared with the observed probability, and components of attributable fraction were subsequently estimated for each possible combination of risk factors.
The PARF can sum to more than one because some individuals with more than one risk factor can have fracture prevented in more than one way and the prevented cases of these individuals could be counted more than once (18). It has been stated that except under special circumstances, PARF estimated the usual way should not be added together (19, 20). To avoid this problem, therefore, we estimated the PARF limits using methods accordingly (18). All database management and statistical analyses were performed with the SAS Statistical Analysis System version 9.1.3 (21).
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
).
|
The overall incidence of fracture (per 1000 person-yr) was 23.1 (95% CI, 15.4–34.6) in women and 14.4 (8.7–24.0) in men. In both sexes, the most common fractures were observed at the vertebrae (27.3% in women and 34.1% in men) followed by the wrist and forearm (25.5% in women and 4.2% in men) and the hip (8.6% in women and 14.4% in men). As expected, there was an inverse relationship between the incidence of fracture and BMD levels (Fig. 2). In both sexes, FNBMD in fracture subjects was approximately 0.4 SD lower than in nonfracture subjects and this difference was more pronounced in those with hip fracture (0.7 SD).
|
).
|
).
|
When all risk factors were simultaneously considered in multivariable Cox’s proportional hazards model, the following factors were found to be independent predictors of any fracture risk in both sexes: in women, per SD difference, age (hazard ratio, 1.2; 1.0–1.3), postural instability (1.1, 1.0–1.2), history of fall (2.1, 1.6–2.8), and prior fracture (1.7, 1.2–2.7), and in men, age (1.4, 1.1–1.6), postural instability (1.2, 1.1–1.3), and history of fall (1.9, 1.2–3.0) (Table 3).
|
Further analysis by individual fracture site in both sexes indicated that the effect of age, baseline BMD, and prior fracture on the fracture risk were more pronounced on hip, vertebral, and wrist/forearm fractures. In men, however, there was no significant association between baseline BMD and fracture risk at hip and wrist/forearm. Fall was independently associated with any fracture risk in both sexes (Table 4).
|
To estimate the PARF, each independent continuous risk factor was dichotomized (age 
70 yr vs. age < 70 yr and osteopenic BMD vs. normal BMD), and hazard ratio of fracture was estimated for each and combined factors. Results of the analysis of PARF (Table 5) suggested that in both sexes, the attributable risk fraction generally increased as a function of the cumulative presence of risk factors and the incidence of fracture of the corresponding risk factors. In this sample, 68.5% in women and 46.2% in men had osteopenia (–2.5 < BMD T-score <–1.0), 39.9% women and 45.5% men were 70 yr of age or older, 25.0% women and 17.3% men had a history of fall in the last 12 months, and 10.2% women and 8.4% men had fractures in the last 5 yr. Approximately 82% of women and 73% of men had at least one risk factor, 38% women and 47% men had at least two risk factors, 13% women and 8% men had at least three risk factors, and 1% had four risk factors in both sexes.
|
70 yr, osteopenic BMD, history of fall and fracture) were approximately 2, 15, and 9% for any fracture, hip fracture, and vertebral fracture, respectively. However, in women, the major contribution to PARFs was osteopenia. Indeed, in women, osteopenia and fall were the two most important risk factors for any fracture and the major after adjustment for the other factors, whereas osteopenia and prior fracture played the most important role for the fracture risk at hip and vertebral. In this study sample, if osteopenia were eliminated and all other risk factors were unchanged, 32 to 40% of any fracture cases in women and 13 to 20% in men would be prevented; the figures were less pronounced for the elimination of fall (11–17% in women and 6–10% in men), age (4–7% in women and 19–26% in men), and prior fracture (2–4% for both sexes) risk factors. |
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
As expected, the incidence rates of fracture among nonosteoporotic women and men were lower than their counterparts in the general population and particularly lower than those with osteoporotic BMD. Although the incidence rate of any osteoporotic fracture in women was 1.6-fold higher than that in men, the incidence rate of hip fracture was the same in nonosteoporotic women and men (2.2 per 1000 person-yr). This is lower than a previous report of the incidence among nonosteoporotic women was 4.1 person-per 1000 yr (29). The discrepancy for this could be the result of the age difference. The average age among fractured women in this study was 70 yr vs. 77 yr in the other report, and it is recognized that incidence of fracture increases with advancing age independently of BMD. In the present study, the incidence of wrist and forearm fracture was much higher in women compared with that in men (relative risk = 10.5; 95% CI, 4.3–26.1).
The magnitude of the association between BMD and fracture in nonosteoporotic women was comparable to that in the entire sample and in general white women population (3) in which each SD decrease in BMD is associated with approximately 50% increase in fracture risk. The present study also showed that, in both sexes, in predicting hip fracture risk, FNBMD had better predictive value than lumbar spine BMD. This finding is consistent with previous Dubbo and meta-analysis data (3, 30). However, both measurements at the femoral neck and lumbar spine had similar predictive values for vertebral and other fractures.
These data suggest the concept that fracture prevention should be expanded beyond BMD to include fall prevention as a primary component, because at least 12% of fracture cases in women could be attributed to falls. However, this figure is less in men (6%), which might be the result of a lower prevalence of prior fall among men without osteoporosis (17 vs. 25% in women). The association between falls and fracture risk may be mediated by vitamin D deficiency and high PTH. Muscle weakness, a predictor of fall and fracture, has been shown to be associated with long-term vitamin D deficiency (31, 32). Moreover, hyperparathyroidism is associated with motor neuron impairment (33). The present study did not measure serum calcium, vitamin D, and PTH levels for all participants; therefore, it was not possible to analyze the association between these hormones and fracture risk. Nevertheless, a previous analysis of data from a subsample of men in DOES found that fracture risk was significantly associated with reduced 25(OH)D but not with PTH (34), and none of the men in the study sample had vitamin D deficiency status.
The finding that postural sway was an independent predictor of any fracture is consistent with previous observation in the general population (8). The association between postural sway and quadriceps weakness and fracture is presumably mediated by falls, because increased body sway as a measure of postural instability (35, 36) is a predictor of falls (37) along with impaired tactile sensitivity, joint position sense, and reaction time alone (13).
The present results also confirmed that prior fracture was a predictor of subsequent fracture (38, 39). It is noted that the magnitude of the prior fracture–fracture risk relationship in men was comparable to that in women. It seems prior fracture had a "stringent" effect on subsequent fracture in men than in women. The magnitude of association between prior fracture and subsequent fracture were more pronounced at specific sites such as hip, vertebrae, and wrist fracture. This is because incident hip and vertebral fractures in this study were the secondary fractures rather than primarily incident events. The relation between prior fracture and hip fracture among nonosteoporotic women in the present study is consistent with that of previous study (40). However, only 4% (women) and 7% (men) of incident hip fractures in nonosteoporotic subjects were attributable to prior fracture. The corresponding rates were more pronounced in vertebral (14%) and wrist/forearm (16%) fractures in women compared with 6% in vertebral in men.
In the general population, the incidence of fracture increases with advancing age (8, 41, 42, 43). Although the present study showed that the effect of age on fracture risk in women without osteoporosis was independent of BMD, the magnitude of effect was modest. Only 4 to 7% of fracture cases in nonosteoporotic women could be attributed to advancing age per se. This is because among those who fractured, approximately half were aged 70+ yr, and in this group, the magnitude of the association between advancing age and fracture was modest (hazard ratio = 1.3). However, the effect of age on fracture risk was greater in men with a PARF at least 20% of fracture cases.
The existence of the other independent risk factors for fracture in women and men without osteoporosis cannot be ruled out, because approximately half of women and two thirds of men who experienced fracture were not "explained" by the combination of osteopenia, advancing age, fall, and prior fracture. However, additional risk factors would have to be sufficiently prevalent and strongly related to fracture to make a substantial contribution to the fracture incidence in the general population. In the present study, lifestyle factors such as calcium intake, physical activity, and smoking status were not found to have any association with fracture risk among nonosteoporotic women and men.
This prospective study overcomes several biases inherent in cross-sectional or case-control studies of the association between BMD and other risk factors and fractures. The study was based on a relatively large sample size with long duration of follow up. Furthermore, participants in this study were essentially volunteers and were generally healthier than those who did not participate (44). Therefore, the magnitude of association between risk factors and fracture risk could have been underestimated.
The findings from the present study have important implications in clinical practice. Currently, treatment is recommended for individuals with low BMD (i.e. BMD T-scores –2.5) to reduce their risk of fracture (4, 5). BMD T-scores have also been used as criterion to recruit participants for intervention trials. The findings from this study confirm that the BMD-fracture risk relationship is a continuous one. That means there is no arbitrary cutoff value that can better discriminate fracture subjects from nonfracture counterparts. Therefore, the criteria for treatment recommendation and for recruitment of trials could perhaps better be based on individual absolute risk rather than any cutoff BMD value.
In conclusion, postural instability, quadriceps weakness, and history of fall or prior fracture were significant predictors of subsequent osteoporotic fractures independent of baseline BMD and age in women and men without osteoporosis. The combination of those predictors could identify a subgroup of women and men who were not osteoporotic but at high risk of fractures. Selecting patients for intervention to reduce the community burden of fractures, therefore, should be not only based on BMD values and interventions should arguably be focused on the contributory risk factors.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Disclosure Statement: The authors have nothing to disclose.
First Published Online December 12, 2006
Abbreviations: BMD, Bone mineral density; BMI, body mass index; CI, confidence interval; DOES, Dubbo Osteoporosis Epidemiology Study; FNBMD, femoral neck BMD; PARF, population-attributable risk fraction.
Received July 11, 2006.
Accepted December 1, 2006.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  S. Parikh, H. Mogun, J. Avorn, and D. H. Solomon Osteoporosis Medication Use in Nursing Home Patients With Fractures in 1 US State Arch Intern Med, May 26, 2008; 168(10): 1111 - 1115. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
