Serum basic amino acids, prolactin, and insomnia severity in perimenopausal individuals: a cross-sectional sub-study

Abstract

Insomnia symptoms are common during the menopausal transition, but estrogen-centered explanations do not fully account for inter-individual variation. We examined whether serum prolactin was associated with insomnia severity and whether circulating basic amino acids modified this association in perimenopausal individuals. This cross-sectional sub-study used data from a parent project concerning polyunsaturated fatty acids, bone mass, and insomnia during perimenopause. Participants were recruited from the outpatient and inpatient departments of Urumqi Fourth People's Hospital between December 30, 2020 and December 30, 2021. The complete-case analytic sample included 192 biological female participants aged 40–55 years with uterus and ovaries present. Insomnia symptoms were assessed with the Insomnia Severity Index (ISI; past two weeks). Serum prolactin, ovarian hormones, and arginine, histidine, and lysine were measured from morning venous blood samples. Spearman analysis showed a weak positive association between ISI score and prolactin (rho = 0.145, p = 0.044). In adjusted moderation models, the log(prolactin) x lysine interaction was significant (standardized beta = 0.136, 95% CI 0.030 to 0.241; p = 0.012; FDR-adjusted p = 0.036; interaction Delta R2 = 0.031). The log(prolactin) x arginine interaction was nominally significant but did not remain significant after correction for three interaction tests (standardized beta = 0.157, 95% CI 0.012 to 0.303; p = 0.035; FDR-adjusted p = 0.052), and the histidine interaction was not significant ( p = 0.087). In descriptive ISI score-derived groups, 110 participants had ISI scores < 8 and 82 had ISI scores > = 8; arginine was lower in the ISI > = 8 group than in the ISI < 8 group (21.27 [13.64, 25.73] vs 22.83 [16.05, 30.21] umol/L; p = 0.037), whereas prolactin, histidine, and lysine did not differ significantly. Higher serum prolactin was associated with greater ISI score, and lysine showed the most consistent evidence of modifying this association. The arginine finding was exploratory and sensitive to multiple-testing and influential-observation analyses, whereas histidine was not supported as a moderator. These cross-sectional findings should be interpreted as associational and hypothesis-generating.