Faster Lumbar Spine Bone Loss in Midlife Predicts Subsequent Fracture Independent of Starting Bone Mineral Density

Abstract

Context: Bone mineral density (BMD) decreases rapidly during menopause transition (MT), and continues to decline in postmenopause.

Objective: This work aims to examine whether faster BMD loss during the combined MT and early postmenopause is associated with incident fracture, independent of starting BMD, before the MT.

Methods: The Study of Women's Health Across the Nation, a longitudinal cohort study, included 451 women, initially premenopausal or early perimenopausal, and those transitioned to postmenopause. Main outcome measures included time to first fracture after early postmenopause.

Results: In Cox proportional hazards regression, adjusted for age, body mass index, race/ethnicity, study site, use of vitamin D and calcium supplements, and use of bone-detrimental or -beneficial medications, each SD decrement in lumbar spine (LS) BMD before MT was associated with a 78% increment in fracture hazard (P = .007). Each 1% per year faster decline in LS BMD was related to a 56% greater fracture hazard (P = .04). Rate of LS BMD decline predicted future fracture, independent of starting BMD. Women with a starting LS BMD below the sample median, and an LS BMD decline rate faster than the sample median had a 2.7-fold greater fracture hazard (P = .03). At the femoral neck, neither starting BMD nor rate of BMD decline was associated with fracture.

Conclusion: At the LS, starting BMD before the MT and rate of decline during the combined MT and early postmenopause are independent risk factors for fracture. Women with a below-median starting LS BMD and a faster-than-median LS BMD decline have the greatest fracture risk.

Keywords: bone mineral density; fracture; general population studies; menopause.

Figure 1.

The menopause transition (MT) was defined as 1 year before to 2 years after the final menstrual period (FMP), and early postmenopause as the 3 years after that. The starting bone mineral density (BMD) measurement was obtained from the last study visit before T1 (T0). The exposure period over which bone loss was calculated was the combined MT and early postmenopause, spanning 1 year before (T1) to 5 years (T2) after the FMP. Specifically, rate of BMD decline during the combined MT and early postmenopause was computed as the percentage of BMD that was lost T0 to the first visit after T2 (T3), divided by the number of years from T1 to T3. BMD did not decrease significantly in the Study of Women’s Health Across the Nation before T1. The BMD measurement at T0 is thus a close approximation of peak areal BMD. In the first set of analyses, we used 3 steps to test whether rate of BMD decline during the combined MT and early postmenopause predicts incident fracture independent of starting BMD before the MT. Steps 1 and 2 separately examined starting BMD and rate of BMD decline, respectively, as predictors of fracture. Step 3 included both starting BMD and rate of BMD loss in the same model. In the second set of analyses, we examined whether the rate of BMD decline and starting BMD have a synergistic effect on fracture.

Figure 2.

Synergistic effects of starting bone mineral density (BMD) before the menopause transition (MT) and rate of BMD decline during combined the MT and early postmenopause on fracture. At both the lumbar spine and femoral neck, participants were categorized as having high vs low starting BMD (greater or less than site-specific sample median) and slow vs fast BMD decline (decline slower or faster than site-specific median rate). For the lumbar spine and femoral neck (each tested separately), we tested membership in 4 groups (high starting BMD and slow BMD decline [reference]; high starting BMD and fast BMD decline; low starting BMD and slow BMD decline; low starting BMD and fast BMD decline) as a predictor of time to fracture after early postmenopause using Cox proportional hazard regression. Covariates included age, body mass index, race/ethnicity, study site, and exposure to vitamin D, calcium, bone-detrimental medications, and bone-beneficial medications. *Hazard ratios compared to reference group (high starting BMD and slow BMD decline).