Dr Eisman receives corporate appointments and serves as a consultant for NPS Pharmaceuticals, Novartis, Organon, Roche, and Servier. All other authors have no conflict of interest.
Asymptomatic Vertebral Deformity as a Major Risk Factor for Subsequent Fractures and Mortality: A Long-Term Prospective Study
Article first published online: 21 MAR 2005
Copyright © 2005 ASBMR
Journal of Bone and Mineral Research
Volume 20, Issue 8, pages 1349–1355, August 2005
How to Cite
Pongchaiyakul, C., Nguyen, N. D., Jones, G., Center, J. R., Eisman, J. A. and Nguyen, T. V. (2005), Asymptomatic Vertebral Deformity as a Major Risk Factor for Subsequent Fractures and Mortality: A Long-Term Prospective Study. J Bone Miner Res, 20: 1349–1355. doi: 10.1359/JBMR.050317
- Issue published online: 4 DEC 2009
- Article first published online: 21 MAR 2005
- Manuscript Accepted: 16 MAR 2005
- Manuscript Revised: 4 DEC 2004
- Manuscript Received: 13 SEP 2004
- vertebral deformity;
- vertebral fracture;
In elderly men and women, asymptomatic vertebral deformity was found to be associated with subsequent risk of symptomatic fractures, particularly vertebral fracture, and increased risk of mortality after a fracture.
Introduction: Vertebral deformity is associated with an increased risk of fracture and mortality. However, it is unclear whether the three events of vertebral deformity, fracture, and mortality are linked with each other and what role BMD plays in these linkages.
Materials and Methods: Vertebral deformity was determined from quantitative analysis of thoracolumbar spine X-rays in 300 randomly individuals (114 men and 186 women) ⩾60 years of age (as of mid-1989), who were randomly selected from the prospective Dubbo Osteoporosis Epidemiology Study. Incidence of atraumatic fractures and subsequent mortality were ascertained from 1989 to 2003. Cox's proportional hazards model was used to determine the association between asymptomatic vertebral deformities, osteoporotic fractures, and risk of mortality.
Results: The prevalence of asymptomatic vertebral deformity was 31% in men and 17% in women. During the follow-up period, subjects with vertebral deformity had a significantly higher risk of any fracture than those without vertebral deformity (44% versus 29%; hazards ratio [HR], 2.2; 95% CI, 1.4-3.7), particularly symptomatic vertebral fracture (relative risk [RR], 7.4; 95% CI, 3.2-17.0). Mortality rate was highest after a symptomatic fracture among those with vertebral deformity (HR, 9.0; 95% CI, 3.1-26.0). These associations were independent of age, sex, and BMD.
Conclusion: Vertebral deformity was a strong predictor of subsequent risk of fractures, particularly symptomatic vertebral fracture, and may modify fracture-associated mortality in both elderly men and women.
VERTEBRAL DEFORMITY, A hallmark of osteoporosis, affects at least 20% of the elderly population,(1–3) but only 30–40% of all these cases come to clinical attention.(4, 5) Vertebral deformity is associated with an increased risk of subsequent fractures, including vertebral(6–9) and hip fractures.(8–13) Recent studies have further suggested that vertebral deformity could be associated with increased risk of mortality in men(10) and women.(8, 10, 14, 15)
Fracture is associated with increased risk of mortality.(16, 17) Low BMD is a primary predictor of fracture risk(18, 19) and is suggested to be a risk factor for mortality as is fracture per se.(20) However, it is not known whether the three events (e.g., vertebral deformity, fracture, and mortality) are linked with each other or are independent of each other and of BMD. This study was designed to examine the interrelationship between asymptomatic vertebral deformity, subsequent symptomatic fractures, and fracture-associated mortality in relation to BMD in a group of elderly men and women.
MATERIALS AND METHODS
This study was part of the larger on-going Dubbo Osteoporosis Epidemiology Study (DOES); of which the study design and protocol have been described elsewhere.(21) Briefly, DOES is a longitudinal, population-based study of risk factors for fracture and mortality. The sampling frame is the city of Dubbo, New South Wales (Australia), a locality of ∼32,000 people, 98.6% white, of which 1581 men and 2095 women were ⩾60 years of age in 1989. It is relatively isolated in terms of medical care, allowing virtually complete ascertainment of all fractures.
This study consisted of 300 subjects (114 men and 186 women) who were randomly selected from a database of DOES participants. All subjects were ⩾60 years of age as of June 1989 and were free of any illness deemed likely to affect bone metabolism.
Radiographs were performed by a standardized procedure with a target-to-film distance of 105 cm. The presence or absence of vertebral deformity was assessed separately by one investigator and read in a masked fashion to BMD. Vertebral deformity was defined as a reduction of at least 3 SD from same-sex normals, according to Eastell et al.,(22) except that posterior deformities were defined by comparison with same-sex normals rather than by comparison with levels above and below, because this may lead to practical difficulties especially in subjects with multiple deformities. Normal values were derived from a subsample of 50 women (mean age, 67.9 years) and 30 men (mean age, 68.2) who were clinically assessed by two experienced clinicians to be free of deformity.
BMD at the lumbar spine and femoral neck was measured by DXA using Lunar DPX densitometer (GE Lunar, Madison, WI, USA). The CV of BMD at our institution for normal subjects is 1.5% for the lumbar spine and 1.3% for the femoral neck.(23) As part of the estimation of the risk of fracture caused by low BMD and vertebral deformity, individuals were classified as “osteoporotic” if the femoral neck BMD was ≥2.5 SD below the young normal level; otherwise, a “normal” classification was made.
Incidence of fractures and mortality
The primary outcomes in this study were the incidence of symptomatic fractures and of death. During the follow-up period (from mid-1989 to June 2003), the incidence of symptomatic atraumatic fractures was ascertained. Low-trauma symptomatic fractures were identified for residents of the Dubbo local government area through radiologists' reports from the only two centers providing X-ray services as previously described.(21) They were defined as fractures with minor or no trauma (e.g., fall from a standing height or less). The fractures were grouped within the following categories: (1) any type of fracture, (2) hip fracture, (3) vertebral fracture, (4) Colles' fracture, and (5) major upper or lower limb or multiple rib fractures but excluding hip fracture. Subsequent symptomatic vertebral fractures were confirmed by comparison with the baseline radiographs.
Death was also recorded during the study period. They were subgrouped into those after a symptomatic fracture event and those with no intervening fracture event.
The association between symptomatic fracture and mortality (outcome) and vertebral deformity (risk factor) was analyzed by the Cox's proportional hazards model, in which the hazards ratio (HR) and 95% CI of fracture and mortality associated with a vertebral deformity were estimated. Further analyses taking into account age, sex, BMD, and body weight were also performed. Because each HR is subjected to sampling variability (as represented by the CI), it was also of interest to estimate the probability that an association with an HR being >2. The HR of 2 was selected as a cut-point for defining “effect,” because it is equivalent to the probability of 0.67 that a vertebral deformity individual has a fracture first compared with a nonvertebral deformity individual. All analyses were conducted within the SAS software version 9.(24)
A total of 300 subjects (114 men and 186 women) were included in this analysis. The median length of follow-up was 10.2 ± 4 (SD) years. There was no significant difference in the follow-up duration between men (10.5 years) and women (10 years).
At baseline, men had a higher body weight, height, and body mass index (BMI) than women (p < 0.05). As expected, lumbar spine and femoral neck BMD were significantly higher in men than in women (p < 0.001) Using the 3 SD criteria, 35 men (30.7%) and 31 women (16.7%) had at least one vertebral deformity. The prevalence of deformity only increased moderately with age in both men and women, such that by the age of 75 years or above, 38.5% of men and 23.9% of women had a vertebral deformity. Seventeen men (14.9%) and 15 women (8.1%) had a single vertebral deformity, whereas 18 men (15.8%) and 16 women (8.6%) had two or more vertebral deformities. There was no significant difference in age, body weight, height, and BMI between subjects with and without vertebral deformity among either men or women, but femoral neck BMD in both men and women and lumbar spine BMD in men with baseline vertebral deformity were significantly lower than those without vertebral deformity (Table 1).
Incidence of symptomatic fractures
During the follow-up period, 26 men (22.8%) and 67 women (36%) sustained at least one symptomatic atraumatic fracture. Among those, 67 had one fracture and 26 had two or more fractures. The incidence of subsequent fractures among subjects with baseline vertebral deformity was 44% (n = 29/66), significantly higher than those without baseline vertebral deformity (27%; n = 64/234; p = 0.01), making the relative hazard of fracture at 2.2 (95% CI, 1.4–3.4). The probability that a vertebral deformity incurred an HR of >2 was estimated to be 66%. However, the probability that a vertebral deformity incurred an HR of >1 was 99% (data not shown). The difference was most pronounced for symptomatic vertebral fracture, with a 5-fold (95% CI, 1.3–20.0) increase in the relative hazard of fracture in men and almost an 11-fold (95% CI, 3.9–30.9) increase in women being observed. In both sexes, the effect of vertebral deformity on symptomatic vertebral fracture was virtually confirmatory, with the probability of an HR of >2 being 91% (for men) and 100% (for women). However, there was no significant effect of vertebral deformity on nonvertebral fractures (Table 2).
In the Cox's proportional hazards regression model taking into account the effect of age, sex, and BMD, the relative hazard of any subsequent fracture associated with a vertebral deformity was 2.3 (95% CI, 1.4–3.4). The strength of association was stronger for symptomatic vertebral fractures (HR, 6.5; 95% CI, 2.7–15.7). The association was more pronounced after adjusting for age, sex, and weight (instead of BMD), although the magnitude of relative risk did not significantly increase. In both analyses, the strength of associations for symptomatic vertebral fractures was stronger in women than in men (Table 3).
The effect of vertebral deformity on fracture risk was most pronounced in those with BMD T scores ≤ −2.5 SD (i.e., osteoporotic), among whom, the relative hazard for any fracture was 3.2 (95% CI, 1.2–8.7) and 6.7 (95% CI, 1.5–29.1) for vertebral fracture. Among those with osteopenia and normal BMD, the effect of vertebral deformity was significant only for the risk of symptomatic vertebral fracture (HR, 7.0; 95% CI, 2.3–18.1; Fig. 1).
During the follow-up period, 20 men and 25 women died; of those, 17 (7 men and 10 women) had had symptomatic fractures. The unadjusted relative hazard of death associated with a baseline vertebral deformity was 2.7-fold higher (95% CI, 1.5–4.9) than in those without a baseline vertebral deformity (Table 4), and this effect seemed to be more pronounced with time (Fig. 2). There was an 83% chance that the relative hazard [of mortality associated with a vertebral deformity] was >2 (Table 4). Moreover, the risk of death was highest in those with a vertebral deformity who had had a subsequent symptomatic fracture (HR, 9.0; 95% CI, 3.1–26.0). On the other hand, in those without a subsequent fracture, the risk of death was not significantly different between those with and without a vertebral deformity (HR, 1.0; 95% CI, 0.3–3.2). Adjustment for age and BMD did not significantly change the associations (Table 4).
The association between vertebral deformity and mortality was independent of femoral neck BMD. For example, individuals with vertebral deformity and an osteoporotic BMD (∼8% of the sample) had the highest risk of mortality compared with individuals with vertebral deformity only or low BMD only who, in turn, had higher risk of death than those without vertebral deformity and normal BMD (Fig. 3).
Vertebral deformity is one of the classic hallmarks of osteoporosis and is an important public health problem, because it is associated with increased morbidity (back pain and disability)(25–27) and mortality.(8, 10, 14, 15) In this study, baseline vertebral deformity was associated with an increased risk of subsequent symptomatic fractures, particularly vertebral fracture. Moreover, vertebral deformity individuals—regardless of age, sex, and BMD—also had a substantially higher risk of death, particularly after sustaining a fracture.
The strength of association between vertebral deformity and subsequent symptomatic fractures observed in this study is comparable with previous reports.(9, 10) That is, the presence of a vertebral deformity increased the risk of any subsequent fracture by between 2.0- and 2.8-fold. However, in contrast to previous studies,(10, 13) which found no significant association between vertebral deformity and future fractures in men, this study suggested that the relationship of vertebral deformity with subsequent fracture was equivalent between men and women.
The effect of vertebral deformity on incidence of symptomatic vertebral fracture observed in this study was high in both sexes (HR, 6.5; 95% CI, 2.7–16), but was well within the range of relative risk previously reported (between 4.7- and 12.6-fold).(6, 8, 9) The mechanism of the relationship between vertebral deformity and subsequent low-trauma fracture risk is unknown, but it is perhaps equally not surprising, because individuals with vertebral deformity have reduced BMD as observed in this study. However, this study further showed that the effect of vertebral deformity on subsequent fracture risk was independent of BMD and age. Thus, it is possible that vertebral deformity is a marker of poor bone health such as height loss and increased bone loss. Furthermore, the fact that the presence of a vertebral deformity increased the risk of subsequent symptomatic vertebral fracture implies the roles of reduced mobility and mechanical factors in the relationship.
In contrast to previous studies,(8–13) in this study, there was no significant association between vertebral deformity and hip fracture risk in either men or women. However, the absolute incidence of hip fracture in our study was small (two cases in men and eight cases in women). Although the relative hazard of hip fracture associated with a vertebral deformity was not statistically significant, the probability that an HR is >1 and 2 was estimated to be 0.77 and 0.41, respectively.
Past studies have taken into account the effect of BMD on fracture risk and vertebral deformity. In this study, we found that the effect of vertebral deformity on fracture risk was independent of age and BMD. The relative hazard of fracture associated with a vertebral deformity (around 2) is equivalent to the relative hazard associated with 1 SD lower femoral neck BMD. Interestingly, the association between vertebral deformity and fracture risk was dependent on BMD levels. For example, the relative hazard of fracture associated with a vertebral deformity was highest in those with osteoporotic BMD levels (T score ≤ −2.5), although in those with osteopenia, the risk of fracture was also increased with the presence of a vertebral deformity. This suggests that a combination of vertebral deformity and low BMD predicts subsequent fracture more accurately than either alone.
It is known that vertebral deformity and fracture are associated with increased mortality risk.(8, 10, 14, 15) However, it is not clear whether the vertebral deformity-mortality association is mediated through the vertebral deformity-fracture association. In this study, the risk of death was found to be highest in men and women who had a baseline vertebral deformity and then sustained a subsequent symptomatic fracture. It is important to note that this association was independent of BMD and age. It is not possible in this study to pinpoint the mechanism of the excess of mortality associated with vertebral deformity; however, the effects of lifestyle factors and other concomitant diseases can not be ruled out.
Because of the relatively small number of deaths and fractures, no subgroup analysis was done. Nevertheless, among the 17 deaths after fracture, 7 had symptomatic vertebral fracture, 2 had hip fractures, 5 had major fractures, and 3 had Colles' fracture. These fractures have been previously reported to be associated with increased mortality.(16, 17, 28) It could be argued that the vertebral deformity-mortality relationship was mainly attributable to the fracture associated with a vertebral deformity.
The cause of death in this study was not ascertained. However, previous studies have suggested that low BMD and co-morbidity/ disability was each associated with increased risk of death.(29–34) In this study, the effect of difference in mortality rate between vertebral and nonvertebral deformity groups was only evident after 6 years. This suggests that short time studies are less likely to observe the effect of vertebral deformity on mortality, and the median of time of fracture to the time of death was 1.9 years. It also suggests that the time after fracture presents a window of opportunity for intervention to save lives.
The prevalence of vertebral deformity in this population was between 17% and 30%, which is comparable with the prevalence of low BMD (10% in men and 25% in women). The prevalence of vertebral deformity or low BMD was 38%, and given their independent effects on fracture risk, this high prevalence suggests that up to 40% of elderly were at high but unrecognized risk of fracture, because only about one-third of vertebral deformities have been reported to come to clinical attention.(4, 5) The data presented here have important clinical implications, in that individuals who had vertebral deformity combined with low BMD (two risk factors) had the highest risk of fracture (83%) and death (70%) compared with individuals with normal BMD and without prior vertebral deformity (28% for fracture and 23% for death). On the other hand, for individuals with either a baseline vertebral deformity or low BMD, the risk of fracture and death was intermediate and only modestly, 1.6- and 2-fold, above those without either risk factor in both men and women. This finding suggests that early identification and targeting of interventions in elderly men and women with a vertebral deformity with or without low BMD may reduce subsequent fractures as well as the burden of morbidity and mortality.
Despite a prospective and long duration of follow-up, the study has some potential limitations. Vertebral deformity was defined using a morphometric approach that does not completely exclude deformities because of congenital changes and Scheuermann's disease; therefore, some misclassification could have occurred. However, the effect of this type of misclassification on the result would decrease the association observed. For the mortality data, we did not collect the cause of death; therefore, it was difficult to make inference regarding the link between fracture and death. However, this study used a fracture ascertainment system that completely documented all fractures and objectively verified the fracture type.
In conclusion, a vertebral deformity predicts subsequent fracture risk particularly symptomatic vertebral fracture, and is associated with increased fracture-associated mortality in elderly men and women. Vertebral deformity should be considered a primary risk factor for fracture and mortality, particularly in combination with low BMD. Intervention in this subgroup may be cost effective and beneficial to the community.
The authors thank the expert assistance of Janet Watters and Donna Reeves in the interview, data collection, and measurement bone densitometry and the invaluable help of the staff of Dubbo Base Hospital. This work was supported by the National Health and Medical Research Council of Australia.
- 11993 Prevalence and incidence of vertebral deformities. Osteoporos Int 3: 113–119., , , , ,
- 21996 The prevalence of vertebral deformity in European men and women: The European Vertebral Osteoporosis Study. J Bone Miner Res 11: 1010–1018., , , , ,
- 32001 Vertebral deformation in urban Swedish men and women: Prevalence based on 797 subjects. Acta Orthop Scand 72: 273–278., , , , ,
- 41992 Incidence of clinically diagnosed vertebral fractures: A population-based study in Rochester, Minnesota, 1985–1989. J Bone Miner Res 7: 221–227., , ,
- 51997 Clinical consequences of vertebral fractures. Am J Med 103: 30S–43S.
- 61999 Prevalent vertebral deformities predict hip fractures and new vertebral deformities but not wrist fractures. Study of Osteoporotic Fractures Research Group. J Bone Miner Res 14: 821–828., , , ,
- 72001 Risk of new vertebral fracture in the year following a fracture. JAMA 285: 320–323., , , , , , , , , , ,
- 82003 The effect of vertebral fracture as a risk factor for osteoporotic fracture and mortality in a Spanish population. Osteoporos Int 14: 520–524., , , , ,
- 91999 Vertebral fractures predict subsequent fractures. Osteoporos Int 10: 214–221., , , ,
- 102003 Prevalent vertebral deformities predict increased mortality and increased fracture rate in both men and women: A 10-year population-based study of 598 individuals from the Swedish cohort in the European Vertebral Osteoporosis Study. Osteoporos Int 14: 61–68., , , ,
- 112003 Hip fracture patients have more vertebral deformities than subjects in population-based studies. Bone 32: 180–184., , , , , , ,
- 121997 Risk factors for hip fracture in a Japanese cohort. J Bone Miner Res 12: 998–1004., , ,
- 132001 Prevalent vertebral deformity predicts incident hip though not distal forearm fracture: Results from the European Prospective Osteoporosis Study. Osteoporos Int 12: 85–90., , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , ,
- 141998 Mortality associated with vertebral deformity in men and women: Results from the European Prospective Osteoporosis Study (EPOS). Osteoporos Int 8: 291–297., , , , , , , , , , , , , , , , , , , ,
- 152003 Incident vertebral fractures and mortality in older women: A prospective study. Osteoporos Int 14: 589–594., , , , , ,
- 161999 Mortality after all major types of osteoporotic fracture in men and women: An observational study. Lancet 353: 878–882., , , ,
- 172003 Adverse outcomes of osteoporotic fractures in the general population. J Bone Miner Res 18: 1139–1141.
- 181993 Prediction of osteoporotic fractures by postural instability and bone density. BMJ 307: 1111–1115., , , , , ,
- 191993 Long-term fracture prediction by bone mineral assessed at different skeletal sites. J Bone Miner Res 8: 1227–1233., , , ,
- 202002 Epidemiology and outcomes of osteoporotic fractures. Lancet 359: 1761–1767.,
- 211996 Prevalent vertebral deformities: Relationship to bone mineral density and spinal osteophytosis in elderly men and women. Osteoporos Int 6: 233–239., , , , ,
- 221991 Classification of vertebral fractures. J Bone Miner Res 6: 207–215., , , ,
- 231997 Source of variability in bone density: Implication for study design and analysis. J Bone Miner Res 12: 124–135., ,
- 241990 SAS/STAT: User's Guide, Release 6.03 ed. SAS Institute, Cary, NC, USA, pp. 549–640.
- 252001 Effect of vertebral fracture on health related quality of life in a Spanish population older than 54 years. Med Clin (Barc) 116: 533–535., , , , ,
- 261998 The association of radiographically detected vertebral fractures with back pain and function: A prospective study. Ann Intern Med 128: 793–800., , , , , , , ,
- 271996 Vertebral fracture and other predictors of physical impairment and health care utilization. Arch Intern Med 156: 2469–2475., ,
- 281993 Population-based study of survival after osteoporotic fractures. Am J Epidemiol 137: 1001–1005., , , ,
- 292000 Prevalent vertebral deformities predict mortality and hospitalization in older women with low bone mass. Fracture Intervention Trial Research Group. J Am Geriatr Soc 48: 241–249., , , , , , ,
- 301998 Bone mineral density is a predictor of survival. Calcif Tissue Int 63: 190–196., , , ,
- 312001 Bone mineral density at the hip predicts mortality in elderly men. Osteoporos Int 12: 259–265.,
- 322002 Bone mineral density and mortality in elderly men and women: The Rotterdam Study. Bone 30: 643–648., , , , ,
- 332000 Rate of bone loss is associated with mortality in older women: A prospective study. J Bone Miner Res 15: 1974–1980., , , ,
- 341996 Mortality following fractures in older women. The study of osteoporotic fractures. Arch Intern Med 156: 1521–1525., , ,