We analyzed in this study the associations between serum 25(OH)D levels, BMD, and prevalence of VFs in a population of asymptomatic postmenopausal women. Our study demonstrated that increasing age, osteoporosis and hypovitaminosis D were the most important independent factors associated to the presence of prevalent VFs.
In agreement with our results, Gerdhem et al. evaluated 986 women and determined that patients with vitamin D levels below 20 ng/ml (50 nmol/L) had a twofold increased risk of presenting any type of osteoporotic fractures compared to women with levels above this value . However, only symptomatic osteoporotic fractures were assessed in this study and it is well known that the majority of VFs is actually asymptomatic and mostly undiagnosed . More recently and as shown in our study, Lopes et al. . found similar results: vitamin D insufficiency, age, and femoral neck BMD were the most important factors for prevalent moderate/severe vertebral fractures in a series of 415 eldery Brazilian women assessed with VFA.
Evaluation of the relationship between 25(OH)D and hip fractures by several investigators has shown that low serum levels of serum 25(OH)D is a relevant risk factor for hip fractures [26, 27]. In fact, 50% of women presenting hip osteoporotic fractures in the United States had serum vitamin D values lower than 12 ng/ml (30 nmol/L) . Similar results were found among different populations which suggested that vitamin D serum concentrations could constitute a useful predictor of hip fracture risk in the elderly [29–31].
In contrast to our study, other authors evaluating the incidence of VFs observed no significant association with vitamin D insufficiency . However, the authors failed to exclude women who had a history of VFs from the control group, which detracts from a definitive conclusion. A population-based cohort study by Garnero et al. . showed that serum 25(OH)D was not associated with incident fracture in postmenopausal French women (mean age 62.2 years). Roddam et al. ., using a nested case–control design, showed no such association in a British study population (mean age 52.6 years). However, the mean age of the participants in these studies was low, and for younger patients vitamin D status may play a less important role in fracture risk. A study targeting postmenopausal Japanese-American women living in Hawaii who had high 25(OH)D concentrations (mean value, 32 ng/ml) showed no association between vitamin D status and fracture risk . However, this study evaluated all types of fracture and relied upon X-ray films of asymptomatic patients to diagnose vertebral fracture; thus, it cannot be directly compared to our study.
Age and low BMD are independent risk factors for prevalent VFs. This evidence lends credence to the new recommendation by the ISCD that VFA should be utilized in all women with low BMD. We determined the best model of logistic-regression was with hip BMD and not with lumbar spine BMD, which may be due to established effects of aging on this site. It is well known that degenerative alterations that accompany ageing, such as osteophytes, and facet joint osteoarthritis and/or extraskeletal calcification around the lumbar spine may falsely increase the measurement of BMD at this site.
It is now well demonstrated that prevalent VFs are associated with an increase in morbidity and mortality. Moreover, post-menopausal women with severe VFs are at the highest risk of subsequent vertebral and nonvertebral fracture . In fact, the severity of a previous fracture was found to be a better predictor of future nonvertebral fracture risk than BMD. In this context, the finding that low (<10 ng/mL) 25(OH)D concentrations are associated with a high prevalence of moderate/severe VFs (30% in our study) supports the inclusion of this modifiable risk factor as part of the screening, and the identification of postmenopausal women with a high risk of VFs.
The interpretation of a serum level of 25(OH)D in the “insufficient” range is challenging for several reasons. Although there are several ways to measure 25(OH)D (radioimmunoassays, enzyme-linked assays, and liquid chromatography with mass spectrometry), the precision and accuracy of the assays, especially in nonreference laboratories, remain problematic. In one study, identical serum samples were provided to 6 different laboratories, and the chemiluminescent assay tended to return higher values for 25(OH)D . The 25OHD assay used in our study was the Roche electrochemiluminescence assay that has now been withdrawn from the market, in part because it underestimated 25(OH)D levels, especially at low levels and it only measures 25(OH)D3. Thus, one can speculate that the 25(OH)D reported levels are lower than the true levels, which may partly explain the very high proportion of women with 25(OH)D < 10 ng/ml. However, none of our patients was under D2 therapy and a cross-calibration data with another assay (Liaison, Diasorin) in 30 patients showed high correlation (r = 0.92; p < 0.0001) (data not shown). Moreover, the same observation of high prevalence of vitamin D deficiency had been reported in a previous study where the prevalence of values of 25(OH)D lower than 15 ng/mL were found in 43% of a series of young healthy Moroccan women during the summer season and using the technique of chemiluminescence (Liaison, Diasorin) . In our study, the prevalence of extremely low vitamin D status (52%) is indeed quite striking. Symptomatic vitamin D deficiency is still a common problem in Morocco and cases of rickets and osteomalacia are still observed. However, the serum concentrations of calcium and phosphorus in our series were within the reference range and none of our patients had clinical evidence of osteomalacia.
Finally, 25(OH)D levels change with the seasons, exposure to sunlight, and dietary intake. For example, in northern latitudes, serum levels of 25(OH)D decline by 20% from late summer to midwinter, whereas 30 minutes of full-body exposure to the sun during the summer rapidly generates vitamin D. Rabat which is located at the 34th latitude is almost sunny through all the summer season. Cultural habits may be the explanation that limits sun exposure in many Moroccan women as well as low vitamin D intake. In our country, food is not supplemented with vitamin D. Hence, cutaneous synthesis would be the major source of the vitamin in ambulatory women of this age group. The influence of clothing style on 25(OH)D level, has been showed in a previous study where in the summer season and after a multilinear regression analysis, wearing a veil was found to be an independent predictor of hypovitaminosis D  and similar conclusions has been reported in Lebanon , Saudi Arabia  and Turkey .
Our study had strengths and limitations. 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 two experienced investigators after training sessions and after a previous global visualization. Before diagnosis of fracture, a non-osteoporotic origin was considered for each deformity as vertebral deformities can be due to developmental abnormalities, Scheuermann’s disease sequelae and degenerative changes. Attention has been paid to osteoporotic depressions of the central end plates of the vertebrae, as osteoarthritic changes occur only on the anterior part of the vertebrae. However, even though history of trauma was inquired, we cannot exclude that some subjects did not report remote traumas. The main limitations lie in the cross-sectional nature of the study and in the procedures used to select subjects, who were all volunteers and ambulatory. 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. Another limitation is the dosage of PTH which could not be performed in this study. Indeed, in a population like this, with a so high prevalence of low levels of vitamin D, we would expect to have a secondary hyperparathyroidism, which, in turn, contributes to bone loss.