Preventing Peritoneal Dialysis Infections

Many patients with end-stage kidney disease prefer peritoneal dialysis (PD) to hemodialysis to replace kidney function¹ but avoid PD due to fears of serious complications, such as peritonitis or catheter exit site and tunnel infections, which often have higher incidences than those recommended by the International Society for Peritoneal Dialysis (ISPD).² Factors that increase the risk of PD-related peritonitis include patient nasal Staphylococcus aureus carriage, diabetes, advanced age, obesity, depression, cardiovascular disease, and catheter connection methods or exit site infection (ESI).³ A 2004 Cochrane systematic review reported 2 different significant (P < .05) effects of antimicrobial agents on patients receiving PD.⁴ One randomized clinical trial³ (RCT) on 2716 patients reported that, compared with placebo, intranasal mupirocin, bactericidal to S aureus, reduced PD catheter ESI and tunnel infections but not peritonitis rates. A meta-analysis of 4 RCTs⁵ on 335 patients revealed preoperative intravenous antibiotics reduced the incidence of early peritonitis, but it did not affect catheter ESI or tunnel infections. These results suggest peritonitis and ESI or tunnel infections are independent outcomes potentially responsive to different interventions. Herein, this Evidence Corner examines what can be learned about these 2 different PD-related infectious outcomes from an update³ to the earlier Cochrane review and a study of topical antimicrobial agents.⁵ 

RELATED CONTENT

Validation of a Clinically Relevant Diagnostic Marker for Early Diagnosis of Infection in Chronic Wounds

Off-label Usage of Absorbable Beads Containing Antibiotics for Prevention of Surgical Site Infections

Improvements in Calcific Uremic Arteriolopathy Wound Healing During SNF472 Treatment: Results of a Phase 2 Trial

Reference: Campbell D, Mudge DW, Craig JC, Johnson DW, Tong A, Strippoli GF. Antimicrobial agents for preventing peritonitis in peritoneal dialysis patients. Cochrane Database Syst Rev. 2017;4:CD004679. doi: 10.1002/14651858.CD004679.pub3.

Rationale: Preventing PD-related peritonitis and catheter exit site or tunnel infections is a high priority, yet practice patterns to do so vary widely. These include various oral, intravenous, intranasal, or catheter exit site antimicrobial interventions. 

Objective: A systematic review updated evidence supporting benefits and harms of antimicrobial strategies used to prevent peritonitis in patients with PD.

Methods: The authors searched the Cochrane Kidney and Transplant Specialised Register up to October 4, 2016, for RCTs and quasi-RCTs comparing the effects on patients with PD catheter placement of (1) oral or topical antibiotics versus placebo or no treatment or other antibiotics; (2) nasal antibiotics versus placebo or no treatment; (3) preoperative or perioperative systemic antibiotic prophylaxis versus placebo, no treatment, or other antibiotic; or (4) intravenous or intraperitoneal or nasal or topical antibiotics or antimicrobial agents or topical antiseptic interventions on peritonitis or catheter exit site or tunnel infections or catheter removal or replacement (primary outcomes). Secondary outcomes explored were time to first peritonitis, incidence of peritonitis relapse, death of any cause, technique failure, or antimicrobial agent toxicity. Studies were assessed for quality of evidence using GRADE ratings, risk of bias, and homogeneity of data. For outcomes reported as event numbers, relative risk (RR) of the outcome was calculated as the ratio of the number of patients in the experimental treatment group experiencing 1 or more episodes of the outcome divided by the corresponding number of patients in the control group. If the 95% confidence interval for the calculated RR excluded 1.0, the 2 groups compared were considered to differ in RR. 

Results: Thirty-nine parallel RCTs on 4435 participants with chronic PD were included in the review. The only significant meta-analysis results reported were that preoperative and perioperative intravenous vancomycin reduced the risk of peritonitis compared with placebo or cefazolin but had no effect on PD exit site or tunnel infections within 14 days after PD insertion (1 RCT, 177 patients). Most episodes of fungal peritonitis were caused Candida sp after a course of intravenous or oral antibiotics and were effectively treated using oral nystatin or fluconazole (2 RCTs, 817 patients). Most studies enrolled fewer than 200 patients and were of low quality with high risk of bias because blinding of interventions was not feasible.

Authors’ Conclusions: Preoperative and perioperative intravenous vancomycin reduced peritonitis risk compared with placebo or cephazolin, but it did not affect early incidence of PD exit site or tunnel infections. There were uncertain effects on peritonitis risk of intranasal antibiotic used to eradicate patient nares S aureus burdens. More adequately powered, high-quality RCTs are needed to inform peritonitis prevention practice decisions.

Reference: Zhang L, Badve SV, Pascoe EM, Beller E, Cass A, Clark C, de Zoysa J, Isbel NM, McTaggart S, Morrish AT, Playford EG, Scaria A, Snelling P, Vergara LA, Hawley CM, Johnson DW; HONEYPOT Study Collaborative Group. The effect of exit-site antibacterial honey versus nasal mupirocin prophylaxis on the microbiology and outcomes of peritoneal dialysis-associated peritonitis and exit-site infections: a sub-study of the HONEYPOT Trial [published online July 29, 2015]. Perit Dial Int. 2015;35(7):712–721.

Rationale: Prior meta-analyses have found topical mupirocin prevented S aureus-related peritonitis or exit-site and tunnel infections but can lead to development of resistant organisms. The HONEYPOT study reported that antibacterial honey applied daily to the PD exit site had comparable effects with those of intranasal mupirocin on PD-related infections.

Objective: The authors conducted a planned secondary analysis of the HONEYPOT study, comparing effects of intranasal mupirocin with topically applied honey to PD exit sites on ESI, PD-infection-related hospitalizations, microbiology, or technique failure.

Methods: With informed consent, adults or children with end-stage kidney disease undergoing PD at 1 of 10 Australian nephrology centers, with no ESI tunnel infection or peritonitis during the previous month, were stratified by study site, PD incidence status, or nasal carriage then assigned 1:1 to receive usual care plus patient-applied honey wound gel to the PD exit-site daily (N = 186). If patients harbored intranasal S aureus, bilateral intranasal mupirocin was administered twice daily for 5 consecutive days each month (N = 185) for up to 24 months. Though the trial was open label, microbial swab and effluent samples of patients with infection signs were evaluated blindly to treatment. The primary outcome was time to any PD-related ESI or tunnel or peritonitis infection, displayed as Kaplan-Meier survival curves for each treatment group. This planned substudy analysis compared organism-specific ESI or peritonitis cases between the 2 groups. Secondary outcomes were time to first peritonitis-related hospitalization or time to PD infection-related withdrawal or conversion from PD to hemodialysis. Intent-to-treat analyses with likelihood of a type 1 error < .05 for statistical significance compared organism-specific infection rate ratios for the 2 groups, expressed as episodes per patient-year at risk. 

Results: The 2 groups were comparable at baseline on all characteristics, with both groups having 22% of patients as nasal carriers of S aureus. The only 2 patients who developed mupirocin-resistant strains of S aureus were in the nasal mupirocin group. Each of the groups developed 0.41 episodes of PD-related peritonitis per patient year (total 82 episodes) with no significant differences in observed gram-positive, gram-negative, culture-negative, or polymicrobial peritonitis or ESI rates (total 73 episodes). There were no significant differences between groups in first peritonitis-related hospitalization (25%–26%) or first infection-related hospitalization (34%–38%) or causes of PD withdrawal, respectively: 18% for exit-site honey; 29% for the intranasal mupirocin group (P = .11), which also experienced 1 death. 

Authors’ Conclusions: There were no significant differences between groups receiving topical exit-site honey or intranasal mupirocin in the effects on PD catheter-related infections, hospitalizations, or technique failure or on microbial profiles of the infections observed.

Though few significant effects were reported, there is much to learn from these works and their derivative references.^3,5 A meticulous Australian RCT⁶ of 101 patients undergoing PD placement reported 0% ESI, despite 12%–18% nasal staphylococcal carriage and 6%–10% PD-related peritonitis before PD catheter insertion. Every patient received the same standard care, including preoperative systemic cephazolin antibiotic prophylaxis, with 10% povidone iodine applied to the PD exit site at catheter insertion and during thrice weekly exit site dressing changes. Catheter-related bacteremia incidence was 12% for both groups with catheter exit sites topically treated daily using either 10% mupirocin (n = 50) or medical-grade honey (n = 51). This was less than half of the 25% PD-related infection incidence reported by Zhang et al,⁵ whose patients at the 10 participating clinical centers received less rigorously controlled usual care. What differences in practice between the usual care of the HONEYPOT centers⁵ and the standardized care in the Australian RCT⁶ may underlie the latter RCT’s lower incidence of PD-related bacteremia? Similar variability in standards of care across studies or centers in the Cochrane meta-analysis³ also may have obscured significant effects of agents studied. Clinical practice and studies of patients undergoing PD require rigorously controlled systemic and local infection prevention procedures.^2,7

Additional lines of research that merit further RCT study emerged from closer examination⁸ of significant (P < .05) results in 1 or more RCTs cited in the 2017 Cochrane review.³ Promising topical PD catheter exit site interventions that reduced ESI incidence included 2% mupirocin cream or ointment^9-11 or honey.⁶ Cleansing the PD catheter exit site with sodium hypochlorite 10% aqueous solution reduced ESI incidence more than cleansing it with pH-neutral soap.¹² Incidence of PD-related peritonitis decreased in patients randomly assigned to receive systemic oral rifampin 300 mg twice daily for 5 days every 3 months¹³ after PD catheter insertion as compared with no oral antibiotic. 

Another lesson from this literature is that clearer answers are derived from studies that test each intervention in the context of rigorously controlled standardized care varying only 1 intervention or 1 route of administration per study. Comparing intranasal mupirocin to honey applied to the PD catheter exit site⁵ confounded route of administration with compound studied. Clearer answers will come from carefully designed studies exploring dosing parameters for each route of administration with valid, homogeneous outcomes analyses appropriately adjusted for appropriate risk factors. Slowly, but surely, science can clarify how to optimize each intervention in the context of a well-controlled ISPD² standard of care protocol for PD placement and use. Each successive finding will light the path to better outcomes for patients with end-stage renal disease.