Bone health in women with premature ovarian insufficiency/early menopause: a 23-year longitudinal analysis

Abstract

STUDY QUESTION What is the frequency of, and predictors for, osteoporosis, fractures, and osteoporosis management (investigation, treatment) in women with premature ovarian insufficiency (POI; menopause <40 years) and early menopause (EM; menopause 40–44years)? SUMMARY ANSWER Over the 23-year follow-up duration, at a mean age of 68 years, women with POI/EM had higher osteoporosis/fracture risk and prevalence, higher osteoporosis screening and anti-osteoporosis medication use compared to women with usual age menopause; increasing age was predictive of increased risk of osteoporosis/fracture and menopause hormone therapy (MHT) prior to or at study entry (aged 45–50 years) was protective. WHAT IS KNOWN ALREADY Women with POI/EM have increased risk of osteoporosis and fractures with limited data regarding risk factors for reduced bone density and fractures. Clinical guidelines recommend screening with dual X-ray absorptiometry (DXA) and treatment with MHT for most women with POI/EM to reduce osteoporosis and fracture risk; however, studies indicate gaps in osteoporosis knowledge, guideline uptake, and management adherence by clinicians and women. STUDY DESIGN, SIZE, DURATION The Australian Longitudinal Study on Women’s Health is a prospective longitudinal study of Australian women. This study uses the cohort of women born between 1946 and 1951, surveyed nine times between 1996 and 2019. Data from the Australian administrative health records, including hospital admissions data (fractures, osteoporosis), Medicare Benefits Schedule (DXA), and the Pharmaceutical Benefits Scheme (PBS; MHT, anti-osteoporosis medication, available only from 2002) were linked to survey data. PARTICIPANTS/MATERIALS, SETTING, METHODS Survey respondents with self-reported age of menopause were included. POI/EM was defined as menopause <45 years. T-test or chi-square were used for comparisons at baseline (P < 0.05 indicates significance). Generalized estimating equations for panel data explored predictors for the longitudinal outcomes of osteoporosis, fractures, DXA rates, MHT use, and anti-osteoporosis medication (in women with osteoporosis/fracture, from Survey 4 onwards only). Univariable regression was performed, and variables retained where P < 0.2, to form the multivariable model, and bootstrapping with 100 repetitions at 95% sampling of the original dataset to ensure robustness of results. MAIN RESULTS AND THE ROLE OF CHANCE Eight thousand six hundred and three women were included: 610 (7.1%) with POI/EM. Mean (SD) baseline age was 47.6 (1.45) years in the entire cohort and mean (SD) age of menopause was 38.2 (7.95) and 51.3 (3.04) years in women with POI/EM and usual age menopause, respectively (P < 0.001). Over the 23 years, of women with POI/EM, 303 (49.7%) had osteoporosis/fractures, 421 (69.0%) had DXA screening, 474 ever used MHT (77.7%), and 116 (39.1%) of those with osteoporosis/fractures used anti-osteoporosis medication. Of women with usual age menopause, 2929 (36.6%) had osteoporosis/fractures, 4920 (61.6%) had DXA screening, 4014 (50.2%) used MHT, and 964 (33.0%) of those with osteoporosis/fractures used anti-osteoporosis medication. Compared to women with menopause at age ≥45 years and after adjusting for other risk factors, women with POI/EM had increased risk of osteoporosis (odds ratio [OR] 1.37; 95% CI 1.07–1.77), fractures (OR 1.45; 1.15–1.81), DXA testing (OR 1.64; 1.42–1.90), MHT use (OR 6.87; 5.68–8.30), and anti-osteoporosis medication use (OR 1.50; 1.14–1.98). In women with POI/EM women, increasing age was associated with greater risk of osteoporosis/fracture (OR 1.09; 1.08–1.11), and MHT prior to or at study entry (aged 45–50 years), was protective (OR 0.65, 0.45–0.96). In women with POI/EM, age (OR 1.11; 1.10–1.12), fractures (OR 1.80, 1.38–2.34), current smoking (OR 0.60; 0.43–0.86), and inner (OR 0.68; 0.53–0.88) or outer regional (OR 0.63; 0.46–0.87) residential location were associated with DXA screening. In women with POI/EM, increasing age (OR 1.02; 1.01–1.02), and currently consuming alcohol (OR 1.17; 1.06–1.28), was associated with having ever used MHT. In the 299 women with POI/EM and osteoporosis/fractures, only 39.1% ever received treatment with an anti-osteoporosis medication. Increasing age (OR 1.07; 1.04–1.09) and lower BMI (OR 0.95; 0.92–0.98) were associated with greater likelihood of treatment with anti-osteoporosis medication. LIMITATIONS, REASONS FOR CAUTION Survey data including age of menopause were self-reported by participants; fracture questions were not included in the 2001 survey, and location or level of trauma of self-reported fractures was not asked. Additional risk/protective factors such as vitamin D status, calcium intake, and exercise were not able to be included. Due to sample size, POI and EM were combined for all analyses, and we were unable to differentiate between causes of POI/EM. PBS data were only available from 2004, and hospital admissions data were state-based, with all of Australia were only available from 2007. WIDER IMPLICATIONS OF THE FINDINGS This study supports previous literature indicating increased risk of osteoporosis and fractures in women with POI, and adds evidence for women with POI/EM, where there was a relative paucity of data. This is the first study to analyse a variety of clinical and demographic risk factors for osteoporosis and fractures in women with POI/EM, as well as analysing investigation and treatment rates. In these women, using MHT prior to or at study entry, aged 45–50 years, was protective for osteoporosis/fractures; however, having ever used MHT was not, highlighting the importance of early treatment with MHT in these women to preserve bone strength. Although women with POI/EM and osteoporosis or fractures were more likely to use anti-osteoporosis medications than those with usual age menopause, overall treatment rates are low at <40%, demonstrating a significant treatment gap that should be addressed to reduce future fracture risk. STUDY FUNDING/COMPETING INTEREST(S) This study was funded by The Australian NHMRC Centre of Research Excellence Women’s Health in Reproductive Life (CRE-WHIRL, project number APP1171592). A.R.J. is the recipient of a National Health and Medical Research Council post-graduate research scholarship (grant number 1169192). P.R.E. is supported by a National Health and Medical Research Council grant 1197958. P.R.E. reports grants paid to their institution from Amgen, Sanofi, and Alexion, honoraria from Amgen paid to their institution, and honoraria from Alexion and Kyowa-Kirin. TRIAL REGISTRATION NUMBER N/A.