This retrospective cohort study used the TriNetX electronic health record network to identify adults with liver cirrhosis who initiated chronic dialysis for end-stage kidney disease (ESKD). Patients who started peritoneal dialysis (PD) were compared with those starting hemodialysis (HD). The analysis reported results for matched pairs and used a 30-day landmark to begin follow-up after dialysis initiation.
Investigators conducted a 1:1 propensity matching on 43 covariates to balance measured baseline characteristics between PD and HD groups. The matched analytic sample comprised 1,534 pairs of patients. The study design was retrospective and observational; follow-up commenced at the 30-day landmark to reduce early-treatment selection bias.
The primary endpoint was all-cause mortality. Secondary endpoints included peritonitis, sepsis, gastrointestinal bleeding that required endoscopic hemostasis, and hospitalization frequency. Exploratory outcomes were cardiovascular and included a composite of major adverse cardiovascular events (MACE), non-fatal myocardial infarction, and non-fatal stroke. The report presents outcomes through a 3-year horizon for survival comparisons; additional follow-up duration for other endpoints is reported in the abstract where applicable.
Among the 1,534 matched pairs, 3-year survival was lower for patients treated with peritoneal dialysis compared with hemodialysis (39.5% vs 46.6%). PD was associated with higher mortality: hazard ratio (HR) 1.23 (95% confidence interval [CI], 1.10 to 1.38; P < 0.001). This indicates that, in this propensity-matched cohort, PD patients experienced a significantly increased risk of death over the observation window.
Infectious events were more frequent among PD patients. The risk of peritonitis was markedly higher with PD (HR 3.85; 95% CI, 2.86 to 5.19; P < 0.001). PD was also associated with an increased risk of sepsis (HR 1.51; 95% CI, 1.23 to 1.84; P < 0.001). These results quantify a substantially greater infectious burden for PD in patients with cirrhosis and ESKD in this dataset.
Gastrointestinal bleeding requiring endoscopic hemostasis did not differ significantly between PD and HD groups (HR 1.20; P = 0.38). The abstract reports no statistically significant reduction in severe GI bleeding with PD relative to HD.
Patients receiving PD had more hospital admissions on average than those on HD. Mean hospital admissions were 4.94 for PD versus 4.32 for HD (P = 0.02), indicating a higher healthcare utilization burden among PD patients in this cohort.
In exploratory analyses, PD was associated with a higher incidence of the composite major adverse cardiovascular events (MACE) (HR 1.36; 95% CI, 1.08 to 1.72; P = 0.01). There was a borderline higher risk of non-fatal myocardial infarction with PD (HR 1.30; 95% CI, 1.00 to 1.69; P = 0.05). Non-fatal stroke did not differ significantly between PD and HD (HR 1.32; P = 0.10). These cardiovascular findings were presented as exploratory.
The authors conclude that among patients with liver cirrhosis and ESKD, initiation of peritoneal dialysis was associated with higher mortality, a greater infectious complication burden (notably peritonitis and sepsis), and more hospitalizations compared with hemodialysis. PD did not reduce the incidence of severe gastrointestinal bleeding requiring endoscopic hemostasis. In exploratory analyses, PD was also associated with higher composite MACE.
Clinicians treating patients with cirrhosis and ESKD should consider the differential risks identified in this propensity-matched EHR cohort when discussing modality selection; the results suggest that PD carries important infection- and survival-related disadvantages in this specific population.
This summary is based on the study abstract. The abstract does not report several details that are important for full interpretation, including granular baseline characteristic tables, causes of death, center-level practice patterns, duration of follow-up for nonmortality endpoints beyond the 3-year survival window, sensitivity analyses, or potential residual confounding. Those details were not reported in the source abstract and would need to be reviewed in the full article for comprehensive appraisal.
Overall, the study provides propensity-matched comparative outcome data from a large EHR network indicating higher mortality and infectious complications with PD versus HD in patients who have both cirrhosis and ESKD.