Demographic Data and Clinical Characteristics
We initially enrolled 155 patients, including 69 HD patients and 86 PD patients. All of them were performed CT test at the enrollment. After 2 years, there were 120 patients (57 in HD, 63 in PD) finished the follow-up CT test. Reasons of elimination including kidney transplantation, transferation, death, motion artefacts or stents of CT scans, and bypass surgery. In HD group, primary causes of ESRD were predominantly chronic glomerulonephritis (n=38, 55.1%), followed by diabetic nephropathy (n=14, 20.2%), chronic tubulointerstitial nephropathy (n=6, 8.7%), hypertensive nephropathy (n=4, 5.8%), and others (n=7, 10.1%); in PD group, there were chronic glomerulonephritis (n=40, 46.5%), diabetic nephropathy (n=22, 25.6%), hypertensive nephropathy (n=12, 14.0%), chronic tubulointerstitial nephropathy (n=8, 9.3%), and others (n=4, 4.7%) (Table 1).
In baseline, the mean age of HD group was 52.1±13.3 years, 47 (68.1%) were male, median dialysis vintage was 38 (12, 75) months and 17 (24.6%) had diabetes mellitus (DM); and in PD group, the mean age was 54.2±11.7 years, 39 (45.4%) were male, median dialysis vintage was 26 (12.8, 58.0) months and 36 (41.9%) had DM. Compared with HD patients, patients in PD group tend to have higher proportions of female and DM, higher levels of serum cCa, ALP, LDL-C and T-Chol, and lower levels of serum Alb, Scr and UA (Table 1).
Time–averaged clinical and biochemical data were shown in table 2. Differences between two groups were similar to baseline data. Compared to HD patients, PD patients had a higher proportion of female, higher levels of time-averaged ALP, LDL-C, HDL-C and T-Cho, and lower levels of time-averaged serum Alb and UA.
Coronary artery calcification
The median of baseline CAC score in HD group was 97 (1,744), and 95 (0,324) in PD group (Table 1). There was no significant difference (P= 0.361) in the baseline CAC score between 2 groups. Compared with baseline, CAC score of each group showed significant progress after 2-year follow-up (Table 3). But between the 2 groups, there was no significant difference in ∆CAC score: the median ∆CAC score in HD group was 119 (0, 389), and 136 (1, 377) in PD group (P=0.766) (Table 3). In figure 1, we stratified patients by dialysis modality, and depicted the baseline CAC score and the progression trend of each patient as individual trajectories of CAC scores. And figure 2 was the comparison of ∆CAC scores in HD and PD patients.
In Tobit regression, CAC score progressed with 92.17 per year in HD patients (95% CI –16.01 to 200.37) and with 126.80 per year in PD patients (95% CI 28.54 to 225.07). In unadjusted model of Tobit regression, HD was not significantly associated with higher CAC progression comparing with PD (unadjusted difference -32.73 per year; 95% CI -174.86 to 109.41; P=0.649). We performed 3 adjusted models in this part. When fully adjusted for age, gender, dialysis vintage, diabetes, albumin and total cholesterol, HD was also not significantly associated with faster progression of CAC than PD (adjusted difference 70.96 per year; 95% CI –82.30 to 224.23; P=0.361) (Table 4).
Subgroup analysis
Supplemental table 1 summarized results of subgroup analyses for different conditions of CAC progression in HD and PD patients. CAC progressed significantly faster in patients with DM than in patients without DM, which can be seen in both HD and PD groups (in HD group, ∆CAC scores of patients with DM and without DM were 415 (198, 931) and 24 (0, 292) respectively, P=0.004; in PD group, these were 280 (118, 734) and 25 (0, 278), respectively, P=0.006;). But no significant difference of CAC progression between HD and PD groups (P >0.05). For patients with dialysis vintage≤60 or >60 months, there weren’t significant differences of CAC progression in both HD and PD patients (Supplement table 1, P >0.05). For HD group, older patients (age >55) tend to has faster progression of CAC than younger patients (∆CAC scores in older patients were 172 (0, 474) and 51 (0, 347) in younger patients, P=0.040). However, the different speeds of calcification progression in different age groups weren’t seen in PD patients, and there also no significant different between HD and PD groups (P>0.05).
Influencing factors of CAC progression
To explore the factors that influence the progression of CAC, we analyzed the variables in ∆CAC score ≤100 and >100 groups. ∆CAC score >100 were considered to be a fast progression of CAC, while ∆CAC score ≤100, slow progression. Compared with the slow progression group, patients with fast CAC progression exhibited older age, higher proportion of DM and use of calcium-based phosphate binder, higher BMI, time-averaged ALP and CRP (P < 0.05; Supplemental table 2). In Logistic regression, after adjusted for these confounders, the result showed that older, DM and higher time-averaged serum P were independent risk factors of fast progression of CAC (P< 0.05; Supplemental table 3), but dialysis modality wasn’t associated with faster progression of CAC (OR=0.231, 95%CI: 0.042-1.261, P=0.091, Supplemental table 3).