3.1 | Comparisons of baseline demographics and clinical parameters
Of 449 participants recruited to this study, 104 (23.16%) developed NODAT during a median follow-up period of 28.03 (IQR 12.00–84.23) months. The mean age (±SD) of the participants in this study population was 38.90±11.94 years, and 316 (70.38%) were male. The demographics and clinical characteristics of the study participants are summarized in Table 1. Compared with the non-NODAT group, the NODAT group comprised significantly older patients who had higher BMI; higher preoperative TC, LDL-C, non-HDL-C, TC/HDL-C, non-HDL-C/HDL-C and FPG levels; and higher prevalence of polycystic kidney disease, CMV infection, family history of diabetes, and IL-2Ra usage (all p < 0.05). There was no statistically significant difference in the remaining variables among the two groups (p > 0.05). Moreover, based on the lipid index distribution of KTRs (Additional file 1), the preoperative TC level was higher than 5.18 mmol/L, the TG level was higher than 1.7 mmol/L, the HDL-C level was lower than 1.04 mmol/L, the LDL-C level was higher than 1.4 mmol/L, and the non-HDL-C level was higher than 2.2 mmol/L in 14.48%, 32.52%, 40.31%, 87.08%, and 75.50% of patients in the KTRs group, respectively.
3.2 | Univariate analysis
The results of the univariate analysis are shown in Table 2. These results revealed that age, BMI, family history of diabetes, CMV infection, polycystic kidney disease, duration of follow-up, and elevated preoperative FPG, TC, LDL-C, non-HDL-C, TG/HDL-C, TC/HDL-C and non-HDL-C/HDL-C levels were all associated with NODAT. Furthermore, we found that sex, preoperative use of lipid-lowering drugs, hypertension, preoperative TG and HDL-C levels, use of IL-2Ra, and maintenance therapy were unassociated with NODAT.
3.3 | Association between the preoperative lipid profiles and the risk of NODAT
The crude and multivariate-adjusted models are shown in Table 3. Overall, both continuous and categorical analyses showed that the preoperative TC, LDL-C, non-HDL-C, TC/HDL-C, and non-HDL-C/HDL-C levels were significantly associated positively with incident NODAT (all p<0.05). They were independent predictors of diabetes after adjusting for other covariates. In the crude model, the following variables were positively associated (HR, 95% CI) with NODAT: preoperative TC (1.21, 1.02–1.44, p=0.0262), TG (1.27, 1.04–1.54, p=0.0162), LDL-C (1.29, 1.05–1.59, p=0.0158), non-HDL-C (1.39, 1.15–1.69, p=0.0008), TG/HDL-C (1.23, 1.05–1.44, p=0.0088), TC/HDL-C (1.31, 1.14–1.51, p=0.0002), and non-HDL-C/HDL-C (1.30, 1.13–1.50, p=0.0002) levels. However, HDL-C levels were non-significantly negatively associated with NODAT (HR 0.67, 95% CI 0.41–1.10, p=0.1113). When TG and TG/HDL-C were assessed as tertiles, the p for trend through the tertiles were statistically insignificant. As shown in Table 3, further adjustments of Model I for sex did not substantially alter the results. In the multivariate analysis (Model II), the preoperative TC (HR 1.25, 95% CI 1.09–1.58, p=0.0495), LDL-C (HR 1.33, 95% CI 1.02–1.75, p=0.0352), non-HDL-C (HR 1.41, 95% CI 1.09–1.82, p=0.0084), TC/HDL-C (HR 1.28, 95% CI 1.06–1.54, p=0.0109), and non-HDL-C/HDL-C (HR 1.26, 95% CI 1.05–1.52, p=0.0138) were still significantly associated with NODAT, and further adjustments for additional covariates did not weaken the association. As shown in Table 3, when TC was assessed as tertiles, the adjusted HR of NODAT for participants in T3 (≥4.46 mmol/L) was 2.00 (95% CI 1.11–3.62) compared with those in T1 (≤3.60 mmol/L, p for trend=0.0163). Similarly, the risk of NODAT increased approximately twofold in participants in T3 (≥2.57 mmol/L) than in those in T1 (≤1.90 mmol/L) with LDL-C levels (HR 1.97, 95% CI 1.11–3.51, p for trend=0.0114). Subjects in T3 (≥3.32 mmol/L) had a higher risk of NODAT (HR 2.49, 95% CI 1.37–4.51) than those in T1 (≤2.46 mmol/L) with non-HDL-C levels (p for trend=0.0015) (Table 3). Similarly, a higher TC/HDL (≥4.01) was associated with a higher risk of NODAT (HR 2.17, 95% CI 1.21–3.90, p for trend=0.0090), and a similar association between a higher non-HDL-C/HDL-C (≥3.03) and NODAT risk was also found (HR 2.08, 95% CI 1.17–3.71, p for trend=0.0116). The results of these analyses were consistent with those for the continuous variables, confirming the positive associations. However, after further adjustment, the association between the risk of NODAT and preoperative TG (HR 1.08, 95% CI 0.87–1.35, p=0.4850) was not significant. The preoperative TG/HDL-C was positively associated with NODAT (HR 1.23, 95% CI 1.05–1.44, p=0.0112) when only adjusted for sex; however, after adjusting for other covariables, the association with NODAT was not significant (HR 1.34, 95% CI 0.97–1.85, p=0.0726). The association between preoperative HDL-C (HR 0.74, 95% CI 0.43–1.27, p=0.2744) and NODAT was negative although non-significant. This suggests that preoperative HDL-C is a protective factor of NODAT, but its effect is easily influenced by other factors.
Based on the results of the above analyses, the preoperative TC, LDL-C, non-HDL-C, TC/HDL-C, and non-HDL-C/HDL-C levels proved to be significant predictors of NODAT. Thus, we further analyzed the dose–response association between preoperative lipid profiles and the risk of NODAT. Consistently, a linear association was confirmed between the higher risk of NODAT development and preoperative TC, LDL-C, non-HDL-C, TC/HDL-C, and non-HDL-C/HDL-C, using spline smoothing fitting, while adjusting for age, BMI, sex, family history of diabetes, polycystic kidney disease, CMV infection, preoperative FPG and TG levels, use of IL-2Ra, and maintenance pharmacotherapy (Figure 2). These positive associations between preoperative lipid TC, LDL-C, non-HDL-C, TC/HDL-C, and non-HDL-C/HDL-C and the risk of NODAT were evident in all of the subgroups considered and persisted after careful adjustments (Additional file 2–6).