Vertebral fracture status and the World Health Organization risk factors for predicting osteoporotic fracture risk in Japan

Abstract

Introduction

Vertebral fractures are the most common osteoporotic fracture and the prevalence of vertebral fracture is commonly assessed in clinical practice in Japan. The objective of this study was to evaluate potential risk factors for osteoporotic fractures, including morphometric spine fracture status and the WHO risk factors for predicting 4-year fracture risk.

Methods

A population-based community cohort, the Adult Health Study, consisting of 2613 men and women with mean age of 65 enrolled in Hiroshima was followed prospectively for 4 years. The prevalence and incidence of spine fractures were identified from lateral and posterior–anterior spine radiographs using a semiquantitative method. Information on incident nonvertebral fragility fractures (hip, proximal humeral, and forearm) was collected at interviews by trained nurses and physicians during biennial health examinations.

Results

A model, including spine fracture status in addition to the WHO risk factors, appeared to provide greater prognostic information regarding future fracture risk (gradient of risk/standard deviation: GR/SD = 2.73) than a model with the WHO risk factors alone (GR/SD = 2.54). In univariate analyses, age, bone mineral density (BMD), prior clinical fracture, and spine fracture status had the highest gradient of risk. The presence of multiple prevalent spine or non-spine fractures significantly increased fracture risk, but, their contributions to the gradient of risk were similar to those when fracture status was categorized as a binary variable. A model considering those four risk factors yielded GR/SD = 2.67, indicating that it could capture most of the predictive information provided by the model with spine fracture status plus the WHO risk factors.

Conclusion

The use of age, BMD, prior clinical fracture and spine fracture predicted future fracture risk with greater simplicity and higher prognostic accuracy than consideration of the risk factors included in the WHO tool.

Introduction

Prediction of future fracture risk can provide clinicians and patients with important information for their decisions on life style and treatments. Recently, a fracture risk assessment tool (FRAX) was developed by the World Health Organization (WHO) [1]. The WHO fracture risk assessment tool considers clinical risk factors for future fracture, including age, prior clinical fracture, current smoking, alcohol use, parental history of hip fracture, glucocorticoid use, rheumatoid arthritis, and bone mineral density (BMD) in order to assign a 10-year absolute fracture risk [2].

Vertebral fractures are the most common fragility fracture in postmenopausal women with osteoporosis [3], [4], [5]. Many studies have demonstrated that prevalent vertebral fractures increased the risk of new vertebral and nonvertebral fractures in postmenopausal women [6], [7], [8], [9], [10], [11]. Cauley et al. [12] found that women with a prevalent vertebral fracture at baseline were greater than 4 times more likely to experience an incident vertebral fracture over 15 years of follow-up compared with women without a prevalent vertebral fracture. Furthermore, Siris et al. [13] demonstrated that, at any particular value for BMD, spine fracture status increased future vertebral or nonvertebral fragility fracture risk by up to 7-fold.

Vertebral fracture prevalence is higher among Japanese women than Caucasian women [14]. Fujiwara et al. reported that the risk of subsequent vertebral fracture increased 3 times for women with a prevalent vertebral fracture, which is similar to other findings [15]. Since X-ray is recommended to diagnose osteoporosis in Japanese guidelines, prevalent vertebral fracture status is commonly assessed in clinical practice in Japan. However, prevalent vertebral fractures are not identified as a distinct risk factor in the FRAX tool in Japan [16]. Recently, Chen et al. demonstrated the importance of prevalent vertebral fractures for predicting the future fracture risk in the Canadian Multicentre Osteoporosis Study (CaMos) which was one of nine cohorts used for the development and validation of the FRAX tool [17]. The results from CaMos used prevalent vertebral fracture status along with age and BMD to better predict future fracture risk than the WHO risk factors, with greater simplicity for Caucasians in the CaMos adult cohort [17]. Donaldson et al. reported that a combination of radiographic vertebral fracture, femoral neck BMD, and age could predict future vertebral fracture risk as well as the WHO risk factors for Caucasians in the Fracture Intervention Trial (FIT) [18]. Ensrud et al. also reported that simple models based on age and BMD alone or age and fracture history alone predicted 10-year risk of fracture as well as more complex FRAX models in the Study of Osteoporosis Fracture (SOF) [19].

It was acknowledged by the authors of the WHO tool [20] that a prior clinical vertebral fracture was an especially strong risk factor. It was also acknowledged that a fracture detected as a radiographic observation alone (a morphometric vertebral fracture) should be counted as a previous fracture [20]. However, most of the epidemiology studies from which this tool was developed did not include spine imaging, and so spine fracture status information was not available for study or for inclusion in the tool.

The objective of this analysis was to evaluate and compare potential risk factors, including morphometric vertebral fracture status and the WHO fracture risk factors for predicting 4-year fracture risk in a Japanese population-based cohort which was also used for the development and validation of the FRAX tool. Furthermore, because spine fracture status is an important determinant of future fracture risk, we hypothesized that consideration of morphometric vertebral fracture status would lead to a simple risk prediction tool.