Article ID: CJ-25-0382
Background: Although hemodialysis (HD) is a significant risk factor for infective endocarditis (IE), risks associated with peritoneal dialysis (PD) and specific types of HD access remain underexplored.
Methods and Results: This retrospective study analyzed 215,965 patients with end-stage renal disease (ESRD) and 1,076,532 age- and sex-matched individuals without ESRD from the general population (control group). ESRD patients were grouped into those undergoing HD or PD, with the HD group further subdivided by access type: arteriovenous fistula (AVF), catheter, or graft. The incidence of IE was lower in the PD than HD group; among patients on HD, the incidence of IE was lowest for those with access via an AVF.
Conclusions: PD and AVF access lower the risk of IE among patients with ESRD. Thus, the clinical use of both should be prioritized.
Infective endocarditis (IE) is a severe inflammatory disease of the endocardium. Patients with end-stage renal disease (ESRD) undergoing chronic dialysis are at increased risk of IE.1 These patients require renal replacement therapy, including hemodialysis (HD),2 peritoneal dialysis (PD), or transplantation. Although chronic HD is a known risk factor for IE, the association between PD and IE remains less explored. Moreover, few studies have investigated the impact of HD access type on IE risk.
This retrospective cohort study used Taiwan’s National Health Insurance Research Database (2001–2013).3 The study was performed in accordance with the Declaration of Helsinki, and was approval from the Chang Gung Memorial Hospital Institutional Review Board.
After excluding patients aged <18 years and with prior IE, we identified 214,643 patients with ESRD and 1,076,532 age- and sex-matched controls without ESRD. Patients with ESRD were classified by initial dialysis into HD (n=190,159) and PD (n=24,484) groups, with HD patients further divided into groups according to access type: arteriovenous fistula (AVF; n=81,785), catheter (n=90,640), or graft (n=17,734). The primary outcome was hospitalization for incident IE. Follow-up continued until IE, death, or study end (including modality/access switch for dialysis comparisons). To minimize confounding, inverse probability of treatment weighting using propensity scores was applied, and IE risk was assessed using the Fine and Gray competing risk model.
Over a mean (±SD) follow-up of 8.3±5.8 years, the incidence rate of IE was 2.63 per 1,000 person-years (95% confidence interval [CI] 2.54–2.73) in the ESRD group and 0.12 (95% CI 0.12–0.13) for the control group, with a significant hazard ratio (HR) for ESRD of 20.21 (95% CI 18.86–21.65). Among patients with ERSD, the incidence of IE was lower in the PD than HD group (1.2/1,000 person-years [95% CI 1.1–1.3] vs. 2.7/1,000 person-years [95% CI 2.6–2.8], respectively), with a subdistribution HR for PD of 0.36 (95% CI 0.33–0.39; Figure A). The incidence of IE among patients on HD according to access type was 3.5, 3.1, and 1.8 per 1,000 person-years for those with access via a tunneled catheter, graft, and AVF, respectively. The risk was lowest for those with access via an AVF (Figure B).
Cumulative event rates of incident infective endocarditis (IE) in patients according to (A) dialysis modality and (B) type of hemodialysis (HD) access in the inverse probability of treatment weighting-adjusted cohort. CI, confidence interval; PD, peritoneal dialysis; SHR, subdistribution hazard ratio.
This population-based study found that patients with ESRD on chronic dialysis have a higher risk of IE than the general population, with patients on HD, especially those with access via tunneled catheters or grafts, facing greater risk than those with access via an AVF. Although prior studies reported no significant increase in IE risk with PD,4 our analysis showed PD patients have a significantly elevated risk, although peritonitis, not bacteremia, is the more common PD-related infection.5 These findings emphasize the need to prioritize PD as the initial modality and highlight the need for ongoing research and tailored clinical strategies for this vulnerable population.
This work was supported by grants from the Chang Gung Memorial Hospital (CORPG3P0531, CMRPG3P0801, CORPG3P0511, CORPG3P0541, CORPG3N0281, CORPG3M0371, BMRPD95 to S.-W.C.) and the National Science and Technology Council (NSTC-112-2314-B-182A-107 and NSTC-113-2314-B-182A-087).
The authors have no relationships relevant to the contents of this paper to disclose.
This study was approved by the Institutional Review Board of Chang Gung Memorial Hospital (202400614B0).
This study uses data from the National Health Insurance Research Database, which is subject to restrictions prohibiting data sharing.
