Baseline Characteristics and Comorbidities
Figure 1 showed the number of participants enrolled and those excluded. The number of potential participants was 3073, with 42 excluded due to age less than 18 years, and 92 excluded because of less than 3-month dialysis duration. Ultimately, the remaining 2939 patients were eligible for the present analysis.
All variables with less than 5% missing data were imputed before the data analysis, and there was no missing data for outcomes. Of 2939 with a median age of 50.0 (IQR 39.0-61.0), 1697 (57.7%) were men, 549 (18.7%) had DM, 410 (14.0%) had pre-existing CVD, and 1915 (65.2%) had hypertension. All patients were divided into four group: DM plus CVD group (n=177, 6.0%), DM group (n=372, 12.7%), CVD group (n=233, 7.9%), and the control group (n=2157, 73.4%, Table 1). Compared to the control group, DM plus CVD group tended to be elderly, with higher body mass index, systolic BP, hemoglobin, cholesterol, and triglyceride, but lower diastolic BP, serum uric acid, as well as more likely to be hypertension, hyperlipidemia, taking calcium channel blockers, beta blockers, diuretics, ACEI/ARBs, aspirin, and statins. Additionally, DM patients were likely to be higher percentiles of men and hypertension, higher levels of body mass index and eGFR, and lower levels of diastolic BP and low density lipoprotein, compared with those with pre-existing CVD.
Variables at Baseline and Co-existence of DM and Pre-existing CVD, DM, and Pre-existing CVD.
We analyzed the factors associated with co-existence of DM and pre-existing CVD, DM, and pre-existing CVD versus the control group at baseline using the multinomial logistic regression (Table 2). When adjusting for confounding factors, we found that elderly age, men, hypertension, hyperlipidemia, and lower levels of serum uric acid were associated with high risk of co-existence of DM and pre-existing CVD, DM, and pre-existing CVD versus the control group. Of note, hypertension was powerfully associated with 13.72 (6.14 to 30.63)-time risk of co-existence of DM and pre-existing CVD versus the control group. In addition, higher levels of body mass index, systolic BP, hemoglobin, and cholesterol, but lower levels of diastolic BP were associated with higher risk of co-existence of DM and pre-existing CVD.
Observational Period and Mortality
The median observational period was 35.1 (IQR 17.9-61.7) months. During this period, 519 (17.7%, 95% CI 16.4% to 19.1%) of 2939 patients died, including 258 (43.4%, 95% CI 39.0% to 47.6%) CVD deaths, 54 (10.4%, 95%CI 6.1% to 15.3%) infection deaths, 9 (1.7%, 95%CI 0.7% to 2.9%) gastrointestinal bleeding deaths, 19 (3.1%, 95%CI 1.6% to 4.8%) tumor deaths, 93 (17.9%, 95%CI 14.6% to 21.3%) other death causes, and 82 (15.8%, 95%CI 12.6% to 18.9%) unknown death causes. In addition, 353 (12.0%, 95%CI 10.8% to 13.2%) transferring to hemodialysis, 153 (5.2%, 95% CI 4.4% to 6.1%) receiving renal transplants, 26 (0.9%, 95% CI 0.5% to 1.2%) transferring to other dialysis centers, and 100 (3.4%, 95% CI 2.7% to 4.1%) loss of follow-up. The number of all-cause mortality was 89 (50.3%, 95% CI 42.6% to 57.4% ), 117 (31.5%, 95% CI 27.2% to 36.4%), 57 (24.5%, 95% CI 19.3% to 29.8%), and 256 (11.9%, 95% CI 10.6% to 13.2%) in the DM plus CVD, DM, CVD, and control groups, respectively. The number of CVD mortality was 42 (23.7%, 95% CI 17.5% to 30.8% ), 57 (15.3%, 95% CI 12.1% to 19.1%), 34 (14.6%, 95% CI 10.1% to 18.9%), and 125 (5.8%, 95% CI 4.9% to 6.8%) in the DM plus CVD, DM, CVD, and control groups, respectively.
The incidence of all-cause mortality was 51.3/1000 patient-years in the study population, with 25.5/1000 patient-years of CVD mortality incidence (Table 3). The incidence of all-cause mortality was 179.4, 102.4, 72.6, and 33.3/1000 patient-years, and CVD mortality incidence was 84.6, 49.4, 43.3, and 16.2/1000 patient-years among the DM plus CVD, DM, CVD, and control groups, respectively.
Co-existence of DM and Pre-existing CVD, DM, and Pre-existing CVD and Mortality
Survival analysis found that the DM plus CVD group had poorer cumulative survival (P<0.001) and CVD mortality-free survival (P<0.001) compared to the control group (Figure 2). The association between comorbidities and mortality was evaluated by the different Cox proportional hazards regression models (Table 4). When comparing to the control group, the DM plus CVD, DM, and CVD groups had 2.85 (95% CI 2.18 to 3.72), 1.89 (95%CI 1.50 to 2.38), and 1.43 (95%CI 1.07 to 1.92)-time risk of all-cause morality, and 2.79 (95%CI 1.91 to 4.08), 1.88(95%CI 1.35 to 2.61), and 1.82 (95%CI 1.23 to 2.68)-time risk for CVD mortality compared to the control group in the model 4, respectively.
Similar trends of the association between comorbidities and mortality were observed among subgroups of men, women, hypertension, non-hypertension, hyperlipidemia, and non-hyperlipidemia (Figure 3). Of note, when comparing to pre-existing CVD patients, DM patients had 1.44 (95%CI 1.04 to 1.98)-time all-cause mortality and 1.11 (95%CI 0.72 to 1.71) CVD mortality in the Cox model 4. In the study population, there was no interaction between DM and pre-existing CVD on all-cause and CVD mortality (β=0.203, P=0.292; β=0.281, P=0.123).
Sensitivity Analysis
When conducting competing risk analyses with hemodialysis or renal transplants as the competing risk factors, the DM plus CVD, DM, and CVD groups had 2.18 (95% CI 1.67 to 2.86), 1.81 (95%CI 1.05 to 3.10), and 1.49 (95%CI 1.19 to 1.86)-time risk of all-cause morality compared to the control group, respectively, in the Fine and Gray model 4. Similarly, when using non-CVD mortality, hemodialysis or renal transplants as the competing risk factors, the DM plus CVD, DM, and CVD groups had 2.56 (95% CI 1.76 to 3.72), 1.81 (95%CI 0.82 to 3.99), and 1.54 (95%CI 1.12 to 2.10)-time risk of all-cause morality, respectively, compared to the control group in the Fine and Gray model 4.
By the end of study, 1022 (34.7%) adult patients were follow up less than 24 months, with 238 (23.3%, 95%CI 20.8% to 26.2%) deaths. The remaining 1917 (65.2%) adult patients were follow up for at least 24 months, with 281 (14.7%, 95% CI 13.0% to 16.1%) of all-cause mortality and 135 (7.0%, 95% CI 5.9% to 8.2%) of CVD mortality. When comparing to the control group, the DM plus CVD, DM, and CVD groups had 2.17 (95% CI 1.46 to 3.23), 1.69 (95% CI 1.23 to 2.31), and 1.44 (95% CI 1.08 to 1.92)-time risk of all-cause mortality, and 2.97 (95% CI 1.70 to 5.21), 2.15 (95% CI 1.37 to 3.37), and 1.46 (95% CI 1.04 to 1.98)-time risk of CVD mortality in the Cox regression model 4, respectively, among adult those with at least 24-month follow-up period.