Gonadal and thyroid hormone profiles in women with hormone-sensitive bipolar disorder: an exploratory cross-sectional study

To compare gonadal and thyroid hormone profiles between women with bipolar disorder with hormone-sensitive (HS) and non-hormone-sensitive (NHS) patterns of reproductive-related mood worsening, and to examine differences in affective symptom burden and severe suicidal ideation. In this exploratory cross-sectional study, 75 women aged 18–45 years with bipolar disorder who were not experiencing a current manic or severe depressive episode were classified as HS ( n = 30) or NHS ( n = 45) using a structured retrospective assessment of reproductive-related mood worsening. Between-group endocrine differences were examined using unadjusted comparisons and multivariable linear regression adjusting for age, BMI, menstrual phase, bipolar subtype, and mood stabilizer use. Multiple comparisons were addressed using the Benjamini-Hochberg false discovery rate (FDR) procedure. Sensitivity analyses examined phenotype heterogeneity and menstrual-phase composition. LH was higher in the HS group than in the NHS group in the unadjusted analysis ( P = 0.006; FDR-adjusted q = 0.030). After covariate adjustment, HS status was nominally associated with higher lnLH (B = 0.571, standardized β = 0.270, 95% CI 0.012–1.131, P = 0.045), but this association did not survive FDR correction across the five adjusted gonadal hormone models (q = 0.225). Effect estimates were generally similar across sensitivity analyses, although precision decreased with additional restrictions. No clear between-group differences were observed for other gonadal or thyroid measures. Depressive and anxiety symptom scores were higher in the HS group, and a higher proportion of HS participants had severe suicidal ideation at assessment (7/30 vs. 1/45, P = 0.006); however, lnLH was not significantly correlated with symptom severity. The HS phenotype was characterized by an exploratory LH signal and greater depressive and anxiety symptom burden. The adjusted LH association did not survive FDR correction and therefore requires prospective replication. Severe suicidal ideation was more frequent in the HS group, but this cross-sectional finding does not establish independent suicide-related risk.