Findings in the case-control study
A final total of 61 GDM patients and 122 paired controls were enrolled into the study. Demographics characteristics of the study subjects are shown in Table 1. Before PSM, maternal age, weight, BMI, FBG, OGTT 1h, OGTT 2h and blood pressure was significantly higher in the GDM group than that in the control group. With the use of PSM, there were no significant differences in maternal age, weight, BMI status and blood pressure status between two groups. Therefore, the differences in demographics characteristics between two groups were eliminated following PSM. Table 2 shows the features of the clinical cardiometabolic profiles for both two categories of participants and no significant differences were detected between the two groups in the circulating levels of the clinical cardiometabolic parameters between the two groups.
As shown in Figure 1, regardless before and after PSM, the circulating levels of CCD80 in the GDM subjects was over 19% lower than that in the control group (0.31 ± 0.17 vs. 0.25 ± 0.10, P = 0.009; 0.31 ± 0.12 vs. 0.25 ± 0.10, P = 0.003, respectively).
Conditional multi‑logistic regression analysis to identify the relationship between the concentration of CCDC80 and GDM
As shown in Table 3, the conditional multi-logistic regression analysis revealed that patients with elevated plasma levels of CCDC80 had a significantly reduced risk of GDM, the unadjusted OR was 0.60 (95%CI: 0.42-0.85, P = 0.005). We categorized CCDC80 levels into quartile and found that study participants with high CCDC80 (Q4: > 0.361 vs. Q1: ≤ 0.207) had significantly decreased risk of GDM (OR = 0.36, 95%CI: 0.14-0.92, P = 0.033),and that there was a significant linear trend across categories (trend test, P = 0.013). After adjustment for ALT, AST and creatinine (model 2), the adjusted OR was 0.60 (95%CI: 0.41–0.89, P = 0.011), the linear trend across categories was also significant (trend test, P = 0.023). After adjustment for IL-6 and CRP (model 3), the adjusted OR was 0.61 (95%CI: 0.42–0.87, P = 0.006). Compared with the lowest level group, the OR of highest level CCDC80 was 0.36 (95%CI: 0.14–0.94, P = 0.036; trend test, P = 0.015). After adjustment for triglyceride, cholesterol, HDL-C, LDL-C, apoA1 and apoB (model 4), the adjusted OR for categorized CCDC80 levels was 0.52 (95%CI: 0.34–0.78, P = 0.002; trend test, P = 0.013) (highest vs. lowest quartile). When all participants (before PSM) were enrolled into this analysis, the results were not substantially altered (Additional file 1: Tables S1). Using the conditional multi-logistic regression analysis, we disclosed that CCDC80 was a strong independent predictor of GDM.
ROC curves analysis for the relationship between circulating CCDC80 level and GDM
To esteem the predictive value of CCDC80 for GDM, we undertaken ROC curve analysis. The area under the ROC curve was 0.612 (95%CI: 0.539-0.685) in unmatched full samples (Figure 2A) and 0.633 (95%CI: 0.550-0.716) in PSM samples (Figure 2B). We also determined the predictive ability of simple measures routinely available at booking visit (maternal age, gestational age, BMI, SBP and DBP) in the unmatched whole samples. This demonstrated an AUC of 0.724, which increased significantly to 0.744 and 0.735 with the addition of CCDC80 and HDL-C, respectively (Additional file 2). Thus, CCDC80 exhibited acceptable capacity to distinguish the GDM patients from general population.
The univariate and multivariate associations between CCDC80 and cardiometabolic parameters
As presented in Table 4, among all demographics and clinical cardiometabolic determinants, bivariate correlation analyses showed that CCDC80 levels were positively associated with AST (r = 0.281, P < 0.001), MAO (r = 0.252, P = 0.001), complement C1q (r = 0.170, P = 0.022), LDL-C (r = 0.169, P = 0.022), ApoA1 (r = 0.219, P = 0.003), ApoB (r = 0.207, P = 0.005), and negatively associated with OGTT at 1 h blood glucose (r = -0.150, P = 0.042).
To ensure whether plasma levels of CCDC80 were independently correlated with these markers, the multivariate linear regression analysis with adjustment for maternal age and gestational age were performed. The analyses demonstrated that the CCDC80 levels could independently predicted the values of OGTT at 1 h blood glucose (β = -2.161, 95%CI: -4.269- -0.052), AST (β = 13.529, 95%CI: 6.886-20.171), MAO (β = 5.312, 95%CI: 2.291-8.333), complement C1q (β = 52.258, 95%CI: 6.490-98.025), LDL-C (β = 1.062, 95%CI: 0.103-2.021), ApoA1 (β = 0.784, 95%CI: 0.265-1.303) and ApoB (β = 0.432, 95%CI: 0.121-0.743) (Table 4). Additionally, unmatched whole samples were enrolled into this analysis, the results are shown in Additional file 1: Table S2.