7863 Polycystic Ovary Syndrome in Relation to Clinical, Biochemical, and Functional Musculoskeletal Health Outcomes During Midlife: A Longitudinal Cohort Study

Abstract Disclosure: M. Kazemi: None. S. Rifas-Shiman: None. E. Yu: None. S. Wang: None. M. Hivert: None. W. Perng: None. V. Fitz: None. J. Shifren: None. E. Oken: None. J. Chavarro: None. Background: Research about musculoskeletal composition and function in women with polycystic ovary syndrome (PCOS) is conflicting and primarily derived from cross-sectional analyses in clinical populations, raising concerns about the generalizability of findings and the temporal relation of observed associations. We tested the hypothesis that PCOS status is associated with adverse musculoskeletal outcomes during the menopausal transition. Methods: We enrolled 405 women in early pregnancy in 1999-2002 and followed them for ∼17y. Women with a PCOS clinical diagnosis were classified as “diagnosed PCOS” (N=26) 3y post-enrollment. In women without a PCOS diagnosis, we defined “probable PCOS” (N=40) as the presence of ≥2 PCOS traits, including ovulatory dysfunction (cycle length<21, ≥35d, or too irregular to determine length), elevated testosterone (>75th%) or elevated anti-Müllerian hormone (>75th%). We classified the remaining women as “no PCOS" (N=339). Musculoskeletal health outcomes (lean body mass [LBM], bone mineral density [BMD], and fat mass by dual-energy X-ray absorptiometry; plasma procollagen type I N-propeptide [PINP] and C-terminal telopeptide of type I collagen [CTX-I]; handgrip strength; and fractures) were assessed at 17y visit. We evaluated associations using linear regression models with log-transformed outcomes adjusted for age (mean [SD] 37 [5]y) and BMI (25.7 [5.4] kg/m2) at 3y visit. Further, we stratified analyses by menopausal status at 17y visit (self-reported amenorrhea for 12mos) to account for known effects of menopause on musculoskeletal health. Results: At the 17y follow-up, women were 51 (5)y; BMI, 27.9 (6.3) kg/m2. Diagnosed and probable PCOS were unrelated to spine (log β [95%CI]: 0.03 [-0.01, 0.07]) and hip (0.02 [-0.01, 0.06]) BMD (g/cm2) vs. no PCOS. However, diagnosed and probable PCOS were associated with increases in lower limbs (0.04 [0.01, 0.06]) and total (0.03 [0.004, 0.05]) BMD (g/cm2), without differences between PCOS groups. Compared to no PCOS, probable (-0.26 [-0.41, -0.11]) but not diagnosed PCOS (0.00 [-0.18, 0.18]) was inversely related to plasma CTX-I (ng/dL); a similar pattern was observed for PINP. In analyses stratified by menopausal status, PCOS was unrelated to BMD in postmenopausal women but remained positively associated with lower limb (0.04 [0.01, 0.06]) BMD (g/cm2) in premenopausal women. Conversely, associations between probable PCOS and plasma CTX-I were stronger in postmenopausal (-0.41 [-0.71, -0.11]) than in premenopausal women (-0.16 [-0.33, 0.01] ng/dL). Similarly, probable PCOS was inversely associated with PINP (-0.40 [-0.78, -0.01] mg/L) in postmenopausal, unlike premenopausal women. Diagnosed and probable PCOS were unrelated to LBM, fat mass, handgrip strength, or fracture risk. Conclusions: Any BMI-independent protective effects of PCOS on BMD may be offset during menopausal transition. Presentation: 6/2/2024