Background
Although assessing bone mass with dual-energy X-ray absorptiometry (DXA) is the gold standard for osteoporosis diagnosis, studies have shown that most fractures occur in individuals with a BMD T-score above the WHO operational threshold for osteoporosis[
1]. Recently, the use of clinical risk factors (CRFs) has been shown to enhance the performance of BMD in the prediction of hip and major osteoporotic fractures. In addition to a prior fragility fracture, CRFs include age, sex, body mass index (BMI), use of glucocorticoids, secondary osteoporosis, rheumatoid arthritis, parental history of hip fracture, current smoking, and alcohol intake of three or more units/day. The WHO fracture risk assessment tool (FRAX) allows for estimation of individual 10-year major osteoporotic and hip fracture probabilities[
2].
Vertebral fractures (VFs) are the most common type of osteoporotic fractures in older adults. It has been shown that VFs are usually asymptomatic (only one fourth to one third of these fractures come to medical attention)[
3] and that women with a VF are four to five times more likely to suffer another VF and are also at increased risk for hip fracture and other nonspine fractures compared with women without a VF;[
4,
5] thus their detection remains an important challenge for clinicians. Moreover, radiographically detected VFs are associated with reduced quality of life, increased morbidity and mortality[
6]. Consequently, the identification of asymptomatic VFs is of primordial importance especially in patients without densitometric osteoporosis, a common situation where all experts agree to recommend treatment[
7,
8]. Recently, Vertebral fracture assessment (VFA), which is a method for imaging the thoraco-lumbar spine using bone densitometers[
9] has been showed to have good accuracy and reliability. It can easily be performed at the time of bone mineral density (BMD) measurement, allowing integration of BMD and VF information in the clinical care of patients evaluated for osteoporosis[
1]. Advantages of VFA compared with spine radiographs include greater patient convenience (VFA can be done in association with BMD testing by DXA), smaller dose of ionizing radiation, and lower cost. Previous studies report that, using VFA, around 90–95% of vertebra are interpretable[
10‐
12]. The majority of uninterpretable vertebra occur above T7,[
13,
14] where the prevalence of fracture is low, preserving the negative predictive value of VFA[
15].
The performance characteristics of the FRAX tool have been validated in many independent cohorts[
16].However, most if not all of these cohorts concerned elderly women, usually over the age of 65 and have mainly focused on hip fractures[
17,
18]. There is some uncertainty as to whether this screening tool would have the same performances in younger postmenopausal women and in identifying asymptomatic VFs. Recently, using VFA, we found that 19.7% of a cohort of asymptomatic women show evidence of moderate/severe VFs[
19]. As a FRAX model was developed recently for Morocco based on a large epidemiological study of hip fractures,[
20] we aimed in the present study to evaluate the performance of FRAX scores in comparison with BMD measurement in identifying women with prevalent asymptomatic VFs.
Discussion
This study shows that FRAX risk with and without BMD can predict prevalent asymptomatic osteoporotic fractures in low risk postmenopausal women recruited from general population. In this population, ROC c-statistical analysis showed that the performance of FRAX risk without BMD was better than that of lumbar spine or femoral neck T-scores. Analysis of the BMD with the DXA technique for the axial skeleton has traditionally been considered as the best predictive test known to determine fragility VFs. Moreover, the strategy of intervention for VFs prevention in medical practice has been based on this test until the appearance of the importance of other risk factors for fracture.
The FRAX tool is not designed to examine the risk of asymptomatic VFs; however we performed this analysis because of the importance of this kind of fractures in the outcome of osteoporosis. It is now well established that identification of VFs change the patient’s diagnostic classification, estimation of fracture risk, and influence the decision for a pharmacological intervention as treatment of patients with prevalent VFs reduces the risk of future fractures even when the baseline T-score is above the osteoporosis diagnostic cutpoint of -2.5. Thus, regarding the important health consequences of osteoporotic VFs, together with the fact that most of them are undiagnosed, emphasizes the need for developing better methods to identify patients with asymptomatic VFs.
Few studies have focused on the predictive value of FRAX risk for identifying women with asymptomatic VFs. In Korea, So et al. in a cross-sectional study including 194 patients found that FRAX underestimated the risk of VFs compared to BMD[
24]. Only two studies did a longitudinal analysis. The first one was conducted in 3321 post-menopausal women with low bone mass (60% of them having a femoral neck T score ≤ -2.5) from the FIT (Fracture Intervention Trial) placebo group, of whom 30% had a radiographically detected vertebral fracture at baseline[
25]. The AUC was significantly greater for FRAX with femoral neck BMD (AUC =0.71) than FRAX without femoral neck BMD (AUC =0.68; p =0.002). The second study was conducted in a cohort of postmenopausal women (mean age 65.5 years), of whom 12.5% had a radiographically detected VF at baseline and a mean lumbar spine T-score of -0.95, FRAX risk with and without BMD discriminated patients with incident radiographic VFs (FRAX risk with and without BMD predicted VFs with an AUC of 0.66 and 0.62 respectively). This study showed that the strongest risk factor of future VFs was the combination of age, femoral neck BMD and the presence of a radiographic VF at baseline[
26].
FRAX has been included as a tool for identifying postmenopausal women in recently updated guidelines published by the NOF in the United States[
27] and by the National Osteoporosis Guideline Group [NOGG]), in the UK[
16,
28]. The NOF recommends using FRAX when the decision to treat or not to treat is uncertain. It is primarily intended for postmenopausal women and men 40 years of age and older who have T-scores between -1.0 and -2.5 and who are not on treatment, and who have not had spine or hip fractures. The recommended threshold for intervention is a 10-year hip fracture probability ≥3% or major fracture (humerus, forearm, hip or clinical vertebral fracture) probability ≥20%. These NOF guidelines are difficult to apply in all countries as they are based on cost-effectiveness that produce 35% prevention rate for 5 years according to frequency, mortality and morbidity in the USA. Recently, the Japanese committee recommended a cut-off value of 15% on FRAX as treatment threshold for major osteoporotic fractures in osteopenic patients as they noted that FRAX underestimated fractures in the Japanese population[
29]. The FRAX scores observed in our study were lower than these recommended thresholds: they were likely influenced by the relatively young age of the patients in this group (60.9 yr) and the high BMI (29.8). Therefore, guidelines should be adjusted according to the socioeconomic model of each country.
As VFs are asymptomatic in two thirds of the cases, our study shows that FRAX risk, even without BMD, can discriminate subjects in whom testing BMD and VFA at the same time would be worthy. Approximately 16% of these women (with osteopenia) and 8.5% of women with normal BMD who otherwise may not have been identified as being at greater fracture risk were found to have unappreciated evident VFs (grade 2 and 3). Thus, assessing BMD alone would have underestimate the number of women needing osteoporosis medications (presence of VFs without osteoporosis).
We found that parity, which is not included in the FRAX algorithm, was significantly associated to the presence of VFs, independent of BMD and of the other CRFs. This risk factor may be more specific of early post-menopausal women than of older women. Hence, we assessed whether adding parity to FRAX would improve the ability to identify young postmenopausal women at high risk of fracture. We found that the new score combining parity and FRAX did not significantly improve the sensitivity of FRAX, and did not have a better discriminant value than lumbar spine T-score alone.
The assessment of fracture was carefully conducted using standard procedures of acquisition, and standard reading of all VFAs. All the morphometric assessments were made by an experienced investigator after training sessions. Before diagnosis of fracture, a non-osteoporotic origin was considered for each deformity. The main limitation lies in the procedures used to select subjects, who were all volunteers and ambulatory, and presumably healthier than the general population which also probably explain the low FRAX scores observed in this study. The Rabat population may not be adequately representative of the whole population. However, since the population living in the area of Rabat is a balanced mixture of the various regions constitutive of the country, we believe the impact on prevalence estimate is limited.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
AEM conceived the study, and participated in its design and coordination, performed the statistical analysis and wrote the manuscript. SS and NJ participated in data collection and manuscript revision. AM participated in data collection, statistical analysis and manuscript review. IG participated in data collection, read the VFAs and participated in manuscript review. All authors read and approved the final manuscript.