Although there is wide recognition that type 2 diabetes mellitus (DM2) is a risk factor for fractures (1), the effects of insulin resistance (a key pathophysiologic mechanism in DM2) on bone remain uncertain. In vitro, insulin signaling promotes osteoblast differentiation, proliferation, and function (2-4). However, in the in vivo models of insulin resistance, insulin signaling leads to expansion of bone marrow adipose tissue, decreased trabecular bone mineral density (BMD), and decreased cortical thickness (5). In states of insulin resistance, osteoblasts may also be resistant to insulin signaling (6, 7). Results from human studies of the relation between insulin resistance and BMD are similarly inconclusive, with studies reporting positive (8-11), negative (12-16), or no association (17)(18)(19)(20). Notably, to our knowledge, all published human investigations on this topic are cross-sectional (8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20).The objective of this study was, therefore, to examine the longitudinal associations of insulin resistance with BMD in midlife women before, during, and after the menopause transition (MT). BMD decreases rapidly in a 3-year window spanning 1 year before to 2 years after the final menstrual period (FMP). We define this period as the MT; premenopause (more than 1 year prior to the FMP) precedes the MT, and postmenopause BACKGROUND. The effects of insulin resistance on bone mineral density (BMD) are unclear.
METHODS.In Study of Women's Health Across the Nation (SWAN) participants, we used multivariable regression to test average insulin resistance (homeostatic model assessment of insulin resistance, HOMA-IR) and rate of change in insulin resistance as predictors of rate of change in lumbar spine (LS) and femoral neck (FN) BMD in 3 stages: premenopause (n = 861), menopause transition (MT) (n = 571), and postmenopause (n = 693). Models controlled for age, average BW, change in BW, cigarette use, race and ethnicity, and study site.
RESULTS.The relation between HOMA-IR and BMD decline was biphasic. When average log 2 HOMA-IR was less than 1.5, greater HOMA-IR was associated with slower BMD decline; i.e., each doubling of average HOMA-IR in premenopause was associated with a 0.0032 (P = 0.01, LS) and 0.0041 (P = 0.004, FN) g/cm 2 per year slower BMD loss. When greater than or equal to 1.5, average log 2 HOMA-IR was not associated with BMD change. In women in whom HOMA-IR decreased in premenopause, the association between the HOMA-IR change rate and BMD change rate was positive; i.e, slower HOMA-IR decline was associated with slower BMD loss. In women in whom insulin resistance increased in premenopause, the association was negative; i.e, faster HOMA-IR rise was associated with faster BMD decline. Associations of average HOMA-IR and HOMA-IR change rate with BMD change rate were similar in postmenopause, but weaker during the MT.
CONCLUSION.When it decreases, insulin resistance is associated with BMD preservation; when it increases, insulin resistance is associated with BMD loss.