Abstract
Serum estradiol is a central biomarker of female reproductive biology and is also linked to bone, metabolic, and cardiovascular health. Population-level data on estradiol distribution among women in Kazakhstan remain limited, and it is unclear whether lower estradiol categories in national survey data reflect geographic clustering or individual age-metabolic composition. We performed a secondary analysis of a large national survey-based analytical sample from Kazakhstan using the WHO STEPS framework. The analytical sample included 3498 women aged 18–69 years with available serum estradiol measurements and core covariate data. The primary binary outcome was a lower-tail estradiol category defined as E2 < 200 pmol/L, corresponding approximately to the empirical lower quartile of the analytical female subsample. Descriptive statistics, age-specific quartile summaries, multivariable logistic regression, continuous-outcome regression for ln(E2), nonlinearity diagnostics, and regional heterogeneity analyses were used. Median serum estradiol was 313.00 pmol/L (interquartile range 207.25–457.00). In unweighted and unadjusted analytical-sample comparisons, there was little evidence of regional heterogeneity in estradiol distribution (ANOVA p = 0.583; Kruskal–Wallis p = 0.287; eta-squared = 0.0049; ICC for ln[E2] = 0.000861). In the multivariable model, compared with women aged 18–29 years, the odds of belonging to the lower-tail estradiol category increased from the 40–49-year group (aOR 1.82, 95% CI 1.31–2.53) to the 50–59-year group (aOR 3.72, 95% CI 2.60–5.33) and were highest in the 60–69-year group (aOR 7.48, 95% CI 5.23–10.70). Overweight and obesity were associated with higher odds of lower-tail estradiol, and low HDL cholesterol showed a modest independent association (aOR 1.26, 95% CI 1.04–1.53). In this large national survey-based analytical sample of women, lower-tail estradiol was more closely characterized by age, adiposity, and HDL cholesterol than by regional clustering. The lower estradiol category used in this study should not be interpreted as a clinical diagnosis of hypoestrogenism, particularly because reproductive-stage information was unavailable.