Remnant cholesterol in early pregnancy as a predictor of gestational diabetes mellitus beyond triglyceride levels: a retrospective cohort study

Abstract

Background: Recent findings suggest remnant cholesterol (RC) may contribute to gestational diabetes mellitus (GDM). This study aimed to explore the link between RC and GDM in early pregnancy to assess RC's risk independently of triglycerides (TG).

Methods: We performed a retrospective cohort study involving pregnant women admitted to the Suzhou Ninth Hospital affiliated to Soochow University. Data were obtained from the electronic medical record system. RC (mg/dL) was determined by subtracting high-density lipoprotein cholesterol and low-density lipoprotein cholesterol from total cholesterol. GDM was diagnosed using a 75-g oral glucose tolerance test diagnosed at 24-28 weeks of pregnancy. Various analyses, including logistic regressions, restricted cubic spline (RCS), subgroup, sensitivity, and receiver operating characteristic (ROC) analyses, were performed to evaluate the independent association between RC levels and GDM. We conducted discordance analyses to assess GDM risk in RC versus TG discordant and concordant groups using various cut points. The link between RC and different GDM subtypes was also further examined.

Results: Of the 1,361 women studied, 353 (25.9%) developed GDM. After adjusting for multiple variables, RC was linked to a higher risk of GDM, with an odds ratio (OR) of 1.05 and a 95% confidence interval (CI) of 1.02-1.08. Individuals in the highest RC quartile were more likely to develop GDM (OR: 2.27, 95% CI: 1.37-3.74) than those in the lower quartile. Additionally, the adjusted RCS analysis revealed a significant linear-dose-response link between RC and GDM risk (P for all < 0.001; P for nonlinear = 0.357). Various sensitivity and subgroup analyses confirmed the robustness of our results. The discordant low TG and high RC group revealed an association with GDM, whereas the high TG and low RC group did not when compared with the low TG and low RC group. Even after adjusting for variables, the association remained significant. Similar results were observed using varying clinical cut points. In early pregnancy, RC emerged as the strongest diagnostic marker for GDM, with an ROC curve area of 0.687, surpassing other traditional lipid markers. Further analysis revealed a strong connection between early pregnancy RC levels and GDM subtypes marked solely by high fasting glucose.

Conclusions: Higher RC levels were linked to GDM independently of traditional risk factors, especially TG levels, and may be more useful in diagnosing specific GDM subtypes.

Keywords: Gestational diabetes mellitus; Remnant cholesterol; Triglycerides.

Fig. 1

Flowchart of patient selection. OGTT, oral glucose tolerance test

Fig. 2

Association between RC and GDM risk using multivariable-adjusted RCS, with four knots at the 5th, 35th, 65th, and 95th percentiles. Adjusted for age, BMI, family history of diabetes and hypertension, ethnicity, reproductive technology, multiparity‌, hemoglobin, FPG, TSH, ALT, and TC. The reference point (green dashed line) was set at the median RC level of 16.62 mg/dL for odds ratio interpretation. The linear trend was significant (P < 0.001), with no evidence of nonlinearity (P = 0.357). RC, remnant cholesterol

Fig. 3

Subgroup analyses of the association between RC and GDM. The model was adjusted for age, early-pregnancy BMI, family history of diabetes, assisted reproductive technology, multiparity, hemoglobin, FPG, TSH, ALT, TG, TC, family history of hypertension, and Han Chinese. RC, remnant cholesterol; OR, odds ratio; CI, confidence interval; BMI, body mass index; ALT, alanine aminotransferase

Fig. 4

RC was assessed for its predictive capacity of GDM across all participants using ROC analysis. The analysis revealed that the AUC for the RC ratio was 0.687. This measure outperformed traditional lipid markers such as TG, HDL-C, LDL-C, and TC in predicting GDM. RC, remnant cholesterol; HDL-C, high-density lipoprotein cholesterol; TC, total cholesterol; TG, triglycerides; LDL-C, low-density lipid cholesterol

Fig. 5

Comprehensive analysis of the associations between GDM, TG, and RC levels using upset plots. The four plots below analyze the overlap between TG levels, RC levels, and the occurrence of GDM. Each plot uses different cut-off values for TG and RC to explore their individual and combined effect on GDM incidence. These plots demonstrate the intersection of high and low TG and RC categories and their corresponding relationship with GDM occurrence. A TG and RC are dichotomized by their median values. B TG and RC are dichotomized using cut-off thresholds of 175 and 25 mg/dL, respectively. C TG and RC are dichotomized using cut-off thresholds of 150 and 23 mg/dL, respectively. D TG and RC are dichotomized using cut-off thresholds of 129 and 16 mg/dL, respectively. This study concentrated on clinically significant TG cut points of 175, 150, and 129 mg/dL. These cut points were derived from global guideline recommendations, the upper limit of the regional reference range, and the optimal cut-off value identified through the ROC curve analysis (for both TG and RC), respectively. Corresponding risk RC cut points were established by determining equivalent population percentiles from the pooled cohort that matched these TG values. RC, remnant cholesterol; TG, triglycerides