In peritoneal dialysis (PD) patients, peritonitis is a serious

Proceedings of the ISPD 2006 The 11th Congress of the ISPD 0896-8608/07 $3.00 +.00 August 25 29, 2006, Hong Kong Copyright 2007 International Society for Peritoneal Dialysis Peritoneal Dialysis International, Vol. 27 (2007), Supplement 2 Printed in Canada. All rights reserved. THE ACHILLES HEEL OF PERITONEAL DIALYSIS? Cheuk-Chun Szeto and Kai-Ming Chow Department of Medicine and Therapeutics, Prince of Wales Hospital, The Chinese University of Hong Kong, Shatin, Hong Kong SAR, PR China Peritonitis caused by gram-negative bacteria is a serious complication of peritoneal dialysis. Antibiotic resistance is common, and response to medical treatment is often poor. In the present article, we review recent advances in the understanding of the pathogenesis and treatment of this serious condition. Perit Dial Int 2007; 27(S2):S267 S271 www.pdiconnect.com KEY WORDS: Renal failure; antibiotic resistance; extended-spectrum β-lactamase. In peritoneal dialysis (PD) patients, peritonitis is a serious complication (1 3) and probably the most important cause of technique failure (2 5). Although gram-positive organisms are the most common bacteriologic cause of PD-related peritonitis (1,6), the incidence of gram-positive infections is falling because of the advances in PD connectology (5,7,8). Gram-negative peritonitis may result from touch contamination, exitsite infection, or possibly a bowel source such as constipation, colitis, or transmural migration, but the cause is often unclear (9 12). Gram-negative organisms now account for 20% 30% of all PD-related peritonitis (5,13). CLINICAL OUTCOMES The clinical outcomes of gram-positive and gramnegative peritonitis are markedly different. In fact, when peritonitis rates are used to predict outcome, the suggestion has been made that the two groups of organisms should be examined separately (12). For example, in a retrospective study, Troidle et al. (11) reviewed 415 episodes of peritonitis that occurred in 375 PD patients between 1993 and 1995. As compared with Correspondence to: C.C. Szeto, Department of Medicine & Therapeutics, Prince of Wales Hospital, The Chinese University of Hong Kong, Shatin, Hong Kong SAR, PR China. ccszeto@cuhk.edu.hk episodes of gram-positive peritonitis, episodes caused by gram-negative organisms were more likely to require hospital admission (74% versus 24%, p < 0.001) or catheter removal (18% versus 4%, p < 0.001) and to lead to relapse (32% versus 9%, p < 0.05). In the Network 9 Peritonitis Study (10), the risks of catheter loss, hospitalization, and technique failure in 136 episodes of non-pseudomonas gram-negative peritonitis were significantly worse than in 530 peritonitis episodes caused by gram-positive organisms. More recently, Choi et al. (14) reviewed 490 episodes of peritonitis that occurred between 1995 and 2002. In that series, 32% of gram-negative and 10% of gram-positive peritonitis episodes required catheter removal (p < 0.0001). In a previous series by our group, gram-negative peritonitis accounted for 43% of the episodes that required catheter removal (4). Although PD could be resumed after catheter re-insertion in about half of the cases, loss of solute clearance or ultrafiltration was common following an episode of severe peritonitis (4). SPECIFIC ORGANISMS P. aeruginosa, Escherichia coli, and Klebsiella species are the bacteria that most commonly cause gram-negative peritonitis, identified in 7.1%, 6.8%, and 5.2% of culture-positive cases respectively (15). Amongst all gram-negative organisms, Pseudomonas species are probably the most important cause of serious peritonitis in PD patients. We previously reported on 104 episodes of Pseudomonas peritonitis (16), which accounted for 13% of all peritonitis in our dialysis unit from 1995 to 1999. The major risk factors were a history of antibiotic therapy within 30 days before onset (66% of cases) and concomitant exit-site infection (45% cases). The overall primary response rate was 61%; the rate of complete cure was 22%. Presence of exit-site infection was associated with a lower primary response rate (47% vs 72%, p < 0.01) and S267

PROCEEDINGS OF THE 11TH CONGRESS OF THE ISPD JUNE 2007 VOL. 27, SUPPL 2 PDI a lower rate of complete cure (11% vs 32%, p < 0.02). The episodes that followed recent antibiotic therapy had a significantly lower rate of complete cure than did the de novo cases (12% vs 43%, p < 0.001). More importantly, among the 24 cases that required catheter removal, the chance of returning to PD was higher when the catheter was removed before day 10 than when it was removed on day 15 (64% vs 50%), although that result was not statistically significant. These findings suggest that, when response to antibiotic therapy is suboptimal, early removal of the catheter is important. Among all gram-negative organisms, those from the family Enterobacteriaceae, a large heterogeneous group of gram-negative bacteria, are most commonly involved in PD-related peritonitis. We recently reviewed 210 episodes of Enterobacteriaceae peritonitis, which constituted 12% of all peritonitis in our dialysis unit from 1995 to 2004 (17). The most common species was E. coli, which accounted for 111 episodes (53%). The overall primary response rate was 85%; the complete cure rate was 58%. Presence of exit-site infection was associated with a lower rate of complete cure (43% vs 61%, p = 0.034), but not primary response (89% vs 84%, p = 0.3). The episodes that followed recent antibiotic therapy had a marginally lower rate of complete cure than the others did (49% vs 63%, p = 0.06), but the response rate for primary infections was similar. A total of 82 episodes (39%) did not respond to treatment with a single antibiotic despite sensitivity in vitro, and a second antibiotic had to be added. Interestingly, patients treated with two antibiotics had a marginally lower risk of relapse and recurrence than did those treated with one antibiotic (21% vs 36%, p = 0.051). EVOLUTION IN EPIDEMIOLOGY Since the late 1980s, a substantial change has occurred in the pattern of causative organisms of PDrelated peritonitis. Zelenitsky et al. (15) reviewed 546 episodes of peritonitis that occurred in 374 patients during 1991 1998. The rate of gram-positive peritonitis declined significantly to 0.28 episodes from 0.75 episodes per patient year during that period (p = 0.02), but the occurrence of gram-negative peritonitis remained constant at approximately 0.16 episodes per patient year. We recently reviewed 1787 episodes of PD-related peritonitis in 544 patients treated at our center during 1994 2003 (5). From 1994 to 1998, the incidence of peritonitis caused by coagulase-negative staphylococci species declined to 0.06 episodes from 0.21 episodes per year on PD, paralleling a reduction in the use of the spike S268 set for PD during that period. In contrast, the incidence of E. coli and Klebsiella peritonitis remained static during the same period (Figure 1). Some evidence suggests that the severity of gramnegative peritonitis is worsening. For example, Valdes Sotomayor et al. (12) studied 49 episodes of E. coli peritonitis that occurred in two university hospitals during 1980 2001. The duration of peritoneal inflammation was 3.1 ± 1.7 days for the episodes occurring before 1996, but 5.2 ± 2.7 days for those occurring after 1996 (p = 0.002). Catheter removal was needed in 19% of episodes before 1996 and in 39% episodes from 1996 onward (p = 0.012). It is distinctly possible that the increasing severity of peritonitis is at least partly related to increasing resistance to commonly used antibiotics. ANTIBIOTIC RESISTANCE Gram-negative bacteria have adapted to many antibiotics, especially the β-lactams, by modifying the substrate spectrum of common plasmid-mediated and chromosomal β-lactamases. With the liberal use of antibiotics in many countries, the incidence of resistance to common antibiotics is increasing. At our center, the response rate of gram-negative peritonitis to antibiotic therapy remained static for 10 years (5), but the percentage of cases that needed an alteration of the antibiotic regimen rose from 13.6% in 1994 to 58.7% in 2003 (Figure 2), indicating increasing prevalence of clinical antibiotic resistance. In the series by Zelenitsky et al. (15), the incidence of resistance to ciprofloxacin increased to 47.8% in 1998 from 5.4% in 1992 (p = 0.003). Our recent data on Enterobacteriaceae peritonitis also show that a gradual increase occurred in the prevalence of resistance to several commonly used antibiotics (17). Figure 1 Change in the incidence of gram-negative infections over 10 years.

PDI JUNE 2007 VOL. 27, SUPPL 2 PROCEEDINGS OF THE 11TH CONGRESS OF THE ISPD Figure 2 Change in the response rate to antibiotic treatment and percentage of cases that required an alteration of antibiotic regimen over 10 years. From 2001 to 2004, we observed a gradual increase in the prevalence of resistance to cephalothin and cefuroxime, but not to ceftazidime. A slight increase also occurred in prevalence of resistance to netilmicin or gentamicin, but the prevalence of resistance to ciprofloxacin remained constant. Recent antibiotic therapy was associated with resistance to cefotaxime, ceftazidime, cefoperazone/sulbactam, and piperacillin/tazobactam, accounting for 57%, 75%, 67%, and 80% of the resistant isolates respectively. It should be noted that an in vitro sensitivity test is often not a reliable guide for clinical practice, especially in the presence of bacterial biofilm adherent to a foreign body. Sepandj et al. (18) compared the minimum inhibitory concentration (MIC) and minimum biofilmeliminating concentration (MBEC) of 8 isolates of E. coli and another 8 isolates of Pseudomonas species. Antibiotic sensitivities were found to be significantly higher for the planktonic organisms as tested by MIC assay than for the same organisms in their biofilm state as tested by MBEC assay. For example, none of the isolates was resistant to ciprofloxacin by MIC assay, but 10 were resistant by MBEC assay. This observation indicates that clinical judgment concerning therapeutic response is important, and the presence of exit-site infection, which is a surrogate marker of biofilm formation, is a strong indicator of a need for more aggressive therapy. Extended-spectrum β-lactamase (ESBL) producing bacteria have been attracting special concern in recent years (19). In general, ESBL confers resistance to ceftazidime, cefotaxime, ceftriaxone, aztreonam, and other oxyimino-β-lactams and is found most often in Klebsiella species and E. coli, although it has also been detected in many other gram-negative pathogens. The prevalence of ESBL is probably underestimated because detection in clinical laboratories is imperfect (19). Traditionally, carbapenems are the ideal agents for therapy (19). Yip et al. (20) identified 11 episodes of ESBL-producing E. coli peritonitis over a period of 10 years. In their series, use of cephalosporins and gastric acid inhibitors were risk factors for ESBL-producing E. coli peritonitis. Compared with patients having non-esbl-producing E. coli peritonitis, patients with the ESBL-producing variant more often developed treatment failure (46% vs 13%, p = 0.02) and more often died of sepsis (27% vs 4%, p = 0.02). The peritoneal failure rate was also higher in the group infected with ESBL-producing organisms, although the difference was not statistically significant (18% vs 4%, p = 0.12). In our center, 7 of 103 isolates of E. coli and Klebsiella species from 1999 to 2003 were positive for ESBL (17). Of those 7 cases, 4 had recently received antibiotic therapy; 3 cases attained complete cure. Although carbapenems are the surest agents for therapy, the variety of β-lactamases that confer resistance to carbapenems is increasing, and overuse of any single class of antibiotic is likely to be followed by selection of pathogens resistant to that agent (19). STRATEGIES FOR PREVENTION As discussed earlier, the presence of exit-site infection, which is a surrogate marker of biofilm formation, is a strong indicator of need for more aggressive therapy. Exit-site mupirocin is effective in preventing Staphylococcus aureus exit-site infection. However, it does not reduce gram-negative infections. Bernardini et al. (21) recently reported a multicenter double-blind randomized controlled trial that compared daily gentamicin or mupirocin cream to the catheter exit site. Gentamicin cream applied daily to the catheter exit site reduced P. aeruginosa and other gram-negative catheter infections by more than 50%, and reduced peritonitis by 35%, particularly gram-negative peritonitis. In addition, gentamicin cream was as effective as mupirocin in preventing S. aureus infections. Moreover, gentamicin cream is notably less costly than is mupirocin, to which the incidence of resistance is rising. This study is particularly relevant to countries where gram-negative peritonitis is common. Once gram-negative peritonitis develops, suboptimal response to antibiotic therapy should lead to consideration of early catheter removal. Although the contemporary guideline recommends that catheter removal be considered by day 4, our recent data suggest that a dialysate cell count exceeding 1000/mm 3 on day 3 is a strong predictor of treatment failure and that early catheter removal should then be planned (22). S269

PROCEEDINGS OF THE 11TH CONGRESS OF THE ISPD JUNE 2007 VOL. 27, SUPPL 2 PDI Relapsed peritonitis is often caused by the persistence of bacterial biofilm on the catheter. Catheter exchange for patients with exit-site infection concurrent with peritonitis is a logical and effective measure to prevent relapse. Lui et al. (23) reviewed 37 patients with refractory Pseudomonas exit-site infection treated with catheter exchange in 1994 and 2003. In all patients, effective antibiotics were continued for 7 days, and PD was resumed 2 weeks after the surgery. None of the patients had recurrent exit-site infection at 4 weeks, and no case of Pseudomonas peritonitis occurred within 1 year. Similarly, in 14 episodes of Pseudomonas peritonitis with concomitant exit-site infection, we performed a catheter exchange procedure after the dialysis effluent cleared (16). In no case did relapse peritonitis occur within 1 year after the catheter exchange. These studies make a strong case for early elective catheter exchange in patients with persistent gram-negative exit-site infection. CONCLUSIONS Gram-negative peritonitis is a common cause of morbidity and treatment failure in PD patients. The incidence of peritonitis caused by gram-negative organisms has not decreased despite advances in connectology. In fact, evidence suggests that cases are becoming more severe, which is likely a result of an increasing prevalence of antibiotic resistance. Biofilm formation around the dialysis catheter is probably the major cause of relapse. In selected cases, early catheter removal and elective catheter exchange are effective strategies for preserving the peritoneum and preventing relapse. Further studies should focus on exit-site care, optimal antibiotic regimens, and measures to prevent or eradiate biofilm formation. ACKNOWLEDGMENTS This study was supported in part by the CUHK research accounts 6901031 and 7101215. REFERENCES S270 1. Piraino B. Peritonitis as a complication of peritoneal dialysis. J Am Soc Nephrol 1998; 9:1956 64. 2. Oreopoulos DG, Tzamaloukas AH. Peritoneal dialysis in the next millennium. Adv Ren Replace Ther 2000; 7:338 46. 3. Szeto CC, Wong TY, Leung CB, Wang AY, Law MC, Lui SF, et al. Importance of dialysis adequacy in mortality and morbidity of Chinese CAPD patients. Kidney Int 2000; 58: 400 7. 4. Szeto CC, Chow KM, Wong TY, Leung CB, Wang AY, Lui SF, et al. Feasibility of resuming peritoneal dialysis after severe peritonitis and Tenckhoff catheter removal. J Am Soc Nephrol 2002; 13:1040 5. 5. Szeto CC, Leung CB, Chow KM, Kwan BC, Law MC, Wang AY, et al. Change in bacterial aetiology of peritoneal-dialysis-related peritonitis over ten years: experience from a center in South-East Asia. Clin Microbiol Infect 2005; 10:837 9. 6. Prowant B, Nolph KD, Ryan L, Twardowski Z, Khanna R. Peritonitis in continuous ambulatory peritoneal dialysis: analysis of an 8-year experience. Nephron 1986; 43:105 9. 7. Burkart JM, Hylander B, Durnell Figel T, Roberts D. Comparison of peritonitis rates during long-term use of standard spike versus Ultraset in continuous ambulatory peritoneal dialysis. Perit Dial Int 1990; 10:41 3. 8. Li PK, Szeto CC, Law MC, Chau KF, Fung KS, Leung CB, et al. Comparison of double-bag and Y-set disconnect systems in continuous ambulatory peritoneal dialysis a randomized prospective multi-center study. Am J Kidney Dis 1999; 33:535 40. 9. Piraino B, Bailie GR, Bernardini J, Boeschoten E, Gupta A, Holmes C, et al. Peritoneal dialysis related infections recommendations: 2005 update. Perit Dial Int 2005; 25: 107 31. 10. Bunke CM, Brier M, Golper TA. Outcomes of single organism peritonitis in peritoneal dialysis: gram negatives versus gram positives in the Network 9 Peritonitis Study. Kidney Int 1997; 52:524 9. 11. Troidle L, Gorban Brennan N, Liger A, Finkelstein F. Differing outcomes of gram-positive and gram-negative peritonitis. Am J Kidney Dis 1998; 32:623 8. 12. Valdes Sotomayor J, Cirugeda A, Bajo MA, del Peso G, Escudero E, Sanchez Tomero JA, et al. Increased severity of Escherichia coli peritonitis in peritoneal dialysis patients independent of changes in in vitro antimicrobial susceptibility testing. Perit Dial Int 2003; 23:450 5. 13. Szeto CC, Li PK. Peritoneal dialysis-related infections. In: Pereira BJ, Sayegh MH, Blake P, eds. Chronic kidney disease, dialysis, and transplantation a companion to Brenner & Rector s the kidney. 2nd ed. Philadelphia: Saunders; 2005. 14. Choi P, Nemati E, Banerjee A, Preston E, Levy J, Brown E. Peritoneal dialysis catheter removal for acute peritonitis: a retrospective analysis of factors associated with catheter removal and prolonged postoperative hospitalization. Am J Kidney Dis 2004; 43:103 11. 15. Zelenitsky S, Barns L, Findlay I, Alfa M, Ariano R, Fine A, et al. Analysis of microbiological trends in peritoneal dialysis-related peritonitis from 1991 to 1998. Am J Kidney Dis 2000; 36:1009 13. 16. Szeto CC, Chow KM, Leung CB, Wong TY, Wu AK, Wang AY, et al. Clinical course of peritonitis due to Pseudomonas species complicating peritoneal dialysis: a review of 104 cases. Kidney Int 2001; 59:2309 15. 17. Szeto CC, Chow VC, Chow KM, Lai RW, Chung KY, Leung CB, et al. Enterobacteriaceae peritonitis complicating peritoneal dialysis: a review of 210 consecutive cases. Kidney Int 2006; 69:1245 52.

PDI JUNE 2007 VOL. 27, SUPPL 2 PROCEEDINGS OF THE 11TH CONGRESS OF THE ISPD 18. Sepandj F, Ceri H, Gibb A, Read R, Olson M. Minimum inhibitory concentration (MIC) versus minimum biofilm eliminating concentration (MBEC) in evaluation of antibiotic sensitivity of gram-negative bacilli causing peritonitis. Perit Dial Int 2004; 24:65 7. 19. Jacoby GA, Munoz Price LS. The new β-lactamases. N Engl J Med 2005; 352:380 91. 20. Yip T, Tse KC, Lam MF, Tang S, Li FK, Choy BY, et al. Risk factors and outcomes of extended-spectrum β-lactamase producing E. coli peritonitis in CAPD patients. Perit Dial Int 2006; 26:191 7. 21. Bernardini J, Bender F, Florio T, Sloand J, PalmMontalbano L, Fried L, et al. Randomized, double-blind trial of antibiotic exit site cream for prevention of exit site infection in peritoneal dialysis patients. J Am Soc Nephrol 2005; 16: 539 45. 22. Chow KM, Szeto CC, Cheung KK, Leung CB, Wong SS, Law MC, et al. Predictive value of dialysate cell counts in peritonitis complicating peritoneal dialysis. Clin J Am Soc Nephrol 2006; 1:768 73. 23. Lui SL, Yip T, Tse KC, Lam MF, Lai KN, Lo WK. Treatment of refractory Pseudomonas aeruginosa exit-site infection by simultaneous removal and reinsertion of peritoneal dialysis catheter. Perit Dial Int 2005; 25:560 3. S271