In a population-based cohort of approximately 5000 women (mean age 68 years), Micha€ elsson et al. [1] recently showed a marked seasonal difference in the association between serum 25-hydroxyvitamin D (S-25OHD) levels and areal bone mineral density (BMD) at the total hip, femoral neck and lumbar spine as well as of the total body. There was positive association between S-25OHD concentrations up to 40 nmol L À1 and areal BMD during summer. By sharp contrast, during winter, areal BMD was not associated with S-25OHD concentrations and similar between even below 20 nmol L À1 and over 80 nmol L
À1. I agree with one reasonable explanation of the seasonal difference that a season-related dynamic process influenced the association between S-25OHD levels and areal BMD [1] and would also like to discuss this important evidence from another point of view.Of note, S-25OHD levels were predominantly linked to season-related ultraviolet type B (UV-B) radiation doses, and the difference during summer in areal BMD was relatively greater at the total hip and femoral neck compared with the lumbar spine [1]. The degree of mechanical loading is higher at the former sites than at the latter site, and skeletal loading from activity outdoors can be a major determinant of areal BMD. Consequently, the obvious seasonal discrepancy [1] seems to suggest that the association during summer could result from physical activity, whilst S-25OHD levels might not directly affect areal BMD.In consistent with the above logic, recent evidence is unlikely to support a cause and effect relationships between vitamin D status and musculoskeletal as well as nonmusculoskeletal [2] health in observational studies. For example, a Mendelian randomization study suggested that the association between genetically low vitamin D status and areal BMD was not causal [3], which agrees with the effect of vitamin D supplementation on areal BMD [4] and would be compatible with compensatory adaptation of bone to mechanical strain that is a prominent mechanism in the skeleton [5][6][7]. Furthermore, in the randomized, double-blind, placebo-controlled ViDA trial, monthly high-dose vitamin D supplementation of 100 000 IU colecalciferol for 3-4 years did not change the risk of fractures, and the result was independent of baseline S-25OHD levels including even less than 25 nmol L À1 and over 75 nmol L À1 [8]. The same interpretation can be applied to observational studies showing the association between use of proton pump inhibitors and an increased risk of hip fracture [9,10].As discussed by the authors [1], the effect of vitamin D supplementation on areal BMD appears to be indirect. On the basis of suggestion by the Scientific Advisory Committee on Nutrition (SACN) in adults aged 50 years or older that vitamin D supplementation improves muscle strength/function and increases areal BMD, especially at the femoral neck rather than total hip and notably not at the lumbar spine [11], we suggested that an increase in areal BMD by vitamin D supplementation can result indirectly fro...