Table 1 displays the general and sociodemographic characteristics of the participants in the study. The analysis ultimately encompassed 4,296 individuals with diabetes. Out of these, 2,821 (65.6%) were identified as Non-DKD, whereas 1,475 (34.4%) were diagnosed with DKD. Compared to the Non-DKD group, individuals with DKD showed significantly higher levels across multiple parameters, including age, male population, diabetes duration, SBP, HbA1c, total cholesterol (TC), triglycerides (TG), low-density lipoprotein (LDL), UACR, and ePWV. While high-density lipoprotein (HDL) and eGFR were notably lower in the DKD group compared to their Non-DKD counterparts.

Table 2 delineates the characteristics of participants, organized by quartiles of ePWV. Participants within the highest ePWV quartile, relative to those in the lowest quartile, were generally older, had a higher prevalence of being female and DKD, and showed increased levels of diabetes duration, SBP, DBP, HDL, and UACR. Moreover, in the highest ePWV quartile compared to the lowest, lower levels of HbA1c, TC, TG, LDL, and eGFR were observed.

The determinants of DKD in patients with T2DM are reported in Table 3. DKD correlated positively with age, the male gender, duration of diabetes, SBP, HbA1c, TC, TG, LDL, and ePWV, but negatively with HDL levels in univariate linear regression analysis. Further multivariate analysis showed that the increased prevalence of DKD was associated with being male, high duration of diabetes, high HbA1c, high TG, and high ePWV.

We devised two models to evaluate the independent impact of ePWV on the prevalence of DKD, UACR ≥ 30 mg/g, and eGFR < 60 mL/min per 1.73 m². As illustrated in Table 4, a higher ePWV was linked to a greater prevalence of DKD, UACR ≥ 30 mg/g, and eGFR < 60 mL/min per 1.73 m², even when adjusting for a range of confounding factors, including age, sex, duration of diabetes, HbA1c, TC, TG, HDL, and LDL. When comparing individuals in the highest ePWV quartile to those in the first quartile, the prevalence of DKD, and UACR ≥ 30 mg/g was significantly higher by 32%, and 39% respectively after adjusting for these confounding factors. Each unit increase in ePWV was associated with a 23%, 20% and 21% increase in the prevalence of DKD, UACR ≥ 30 mg/g, and eGFR < 60 mL/min per 1.73 m² in T2DM participants, respectively, after adjusting for the same confounding factors.

The distinct clinical phenotypes of DKD were then clearly defined and quantified, as shown in Table 5: patients with UACR ≥ 30 mg/g and eGFR ≥ 60 mL/min/1.73 m²; patients with UACR ≥ 30 mg/g and eGFR < 60 mL/min/1.73 m² (representing “classic” diabetic kidney disease); and patients with UACR < 30 mg/g and eGFR < 60 mL/min/1.73 m². The number and percentage of participants in these three phenotypes were 1072 (72.7%), 235 (15.9%), and 168 (11.4%), respectively. We found that ePWV was associated with DKD across all three phenotypes, and this association remained significant even after adjusting for confounding factors.

RCS curves were utilized to evaluate the dose-response relationship between ePWV and the prevalence of DKD, UACR ≥ 30 mg/g, and eGFR < 60 mL/min per 1.73 m². As illustrated in Fig. 1, a J-shaped relationship was observed between ePWV and the prevalence of DKD and eGFR < 60 mL/min per 1.73 m², even after adjusting for age, sex, duration of diabetes, HbA1c, TC, TG, HDL, and LDL. Additionally, the analysis revealed a linear association between ePWV and the prevalence of UACR ≥ 30 mg/g, after controlling for the same confounding factors.

RCS analysis of ePWV in relation to the prevalence of DKD, UACR ≥ 30 mg/g, and eGFR < 60 mL/min per 1.73 m². A: DKD B: UACR ≥ 30 mg/g C: eGFR < 60 mL/min per 1.73 m². Model 0 The model was not adjusted. Model 1 The model was adjusted for age, sex, duration of diabetes, HbA1c, TC, TG, HDL, and LDL

As illustrated in Fig. 2, we stratified the analyses by age (< 60 or ≥ 60 years), sex (male or female), duration of diabetes (< 10 or ≥ 10 years), and presence of hypertension (yes or no) to examine whether these potential confounders influenced the associations between ePWV and the prevalence of DKD, and to evaluate any interactions. The results of these stratified analyses demonstrated that the associations between ePWV and the prevalence of DKD were generally consistent across all sub-populations. However, significant interactions were observed between ePWV and age for DKD in participants and statistically significant associations were only observed in participants older than 60 years.

Subgroup analysis for the associations between ePWV and the prevalence of DKD