International Journal of Gynaecology and Obstetrics · Nov 16, 2026
Investigating the association between blood lipid profiles and insulin resistance and β‐cell function indices in pregnant women
By Feng J, Xu X, Tian Z et al.
Objective: This study aimed to examine the associations between first‐trimester lipid markers and second‐trimester insulin resistance (IR) and pancreatic β‐cell function (BCF) indices in pregnant women. An exploratory aim was to explore the apparent discrimination of lipid markers for cohort‐defined low insulin sensitivity. Methods: This retrospective secondary analysis used prospectively collected cohort data from 1202 pregnant women in Beijing, China. Lipids were measured at 4–8 weeks of gestation, and a 75 g oral glucose tolerance test was performed at 24–28 weeks. The Homeostatic Model Assessment of Insulin Resistance (HOMA‐IR) and the Matsuda index were analyzed as continuous measures of IR and insulin sensitivity. BCF was assessed using surrogate indices: the Homeostatic Model Assessment of β‐cell function (HOMA‐β) and the 60‐min insulinogenic index (IGI60). Linear‐regression models adjusted for age and early‐pregnancy body mass index examined the associations between each lipid marker and continuous HOMA‐IR and the Matsuda index as the primary analyses and with HOMA‐β and IGI60 as the secondary analyses. Binary logistic‐regression and receiver operating characteristic analyses of cohort‐defined low insulin sensitivity were exploratory. Results: After adjustment for age and early‐pregnancy body mass index, all seven lipid markers were significantly associated with continuous HOMA‐IR and the Matsuda index (all nominal P < 0.05). No statistically significant associations were observed with the surrogate β‐cell‐function indices HOMA‐β or IGI60. In the exploratory analyses, the lipid markers were associated with the cohort‐defined low‐Matsuda category. Triglycerides/high‐density lipoprotein cholesterol had the highest observed area under the curve in the total population (0.723; 95% confidence interval 0.689–0.756). The exploratory cutoff selected by maximizing Youden's index was 0.65, with a sensitivity of 74.1% and a specificity of 62.5%. Conclusion: First‐trimester lipid markers were associated with second‐trimester IR and insulin‐sensitivity indices, whereas no statistically significant associations were observed with the surrogate BCF indices used in this study. Triglycerides/high‐density lipoprotein cholesterol showed moderate apparent discrimination for cohort‐defined low insulin sensitivity and should be considered a hypothesis‐generating candidate marker requiring external validation.