Peritonitis Rates of the Past Thirty Years: From Improvement to Stagnation

Ever since the introduction of peritoneal dialysis (PD) in the 1960s, peritonitis has been the most important complication of this life-saving treatment. Although fewer than 4% of peritonitis episodes now result in death, peritonitis continues to contribute, directly or indirectly, to the deaths of 16% of PD patients (1,2). In addition, peritonitis is the most common cause of technique failure leading to conversion to long-term hemodialysis (3,4).

Since the early 1980s, multiple advances in the technology of PD have been made, including the disconnect system, standardization of antibiotic therapy, improvements in exit-site care, and possibly new PD solutions. Do these advances translate into better clinical outcomes?

In this issue of Peritoneal Dialysis International, van Esch et al. (5) describe their experience with 731 consecutive new PD patients at a single center in Netherlands over a span of 30 years. Their message is simple: Peritonitis rates have declined significantly during that period. The incidence of antibiotic resistance has been increasing, but the overall rate of cure remains high and stable.

The results of their study are in line with previous findings from other parts of the world (6-9). In the United States, the overall incidence of peritonitis in continuous ambulatory PD patients during the 1980s and early 1990s averaged 1.1 - 1.3 episodes per patient-year of treatment (6). My group previously reported that, during 1994 - 2003, the incidence of peritonitis in Hong Kong declined to 0.46 episodes from 1.10 episodes per year (7). Since just after the year 2000, many centers have reported a peritonitis rate of 0.2 - 0.6 episodes per year of treatment, about 1 episode in 20 - 60 patient-months (8). Although the actual peritonitis rate reported by van Esch et al. (5) is somewhat higher than those in previous studies, the difference is probably a result of differences in patient selection and the prevalence of comorbidities; the overall observation is consistent with many (but not all) reports in which the peritonitis rate dropped gradually from the early 1980s to the 1990s, but remained static thereafter (6-8).

The data presented by van Esch et al. (5) clearly show that the improvement in the peritonitis rate coincides with the widespread use of disconnect systems. The contributions of mupirocin exit-site care and biocompatible dialysis solution are less obvious. It is important to note that, although antibiotic protocols against Staphylococcus aureus are effective in reducing the risk of S. aureus catheter infections, and a wealth of clinical studies support the use of topical antibiotics (either mupirocin or gentamicin) at the catheter exit site in all patients (10-12), the impact of those actions on the real-life peritonitis rate, as revealed by van Esch (5), is less pronounced than has been reported by clinical trials, possibly indicating that these simple measures for exit-site care are not often incorporated into routine clinical practice.

The role of disconnect systems in reducing the peritonitis rate is well reported and accords with findings in previous studies (7,10). It is generally believed that introduction of the Y-set and double-bag disconnect systems has reduced principally the incidence of peritonitis caused by gram-positive organisms; the incidence of gram-negative infections remains about the same as ever. That hypothesis is also supported by the van Esch study (5). Nonetheless, the results reported by those authors should be extrapolated to other parts of the world only with caution, because the distribution of causative organisms varies substantially from one country to another. Notably, in the study by van Esch and colleagues (5), S. aureus and coagulase-negative staphylococcal species together contributed to 45% of all peritonitis episodes, with polymicrobial episodes accounting for another 19%, and gram-negative organisms just 11%. That pattern is certainly different from those in many other parts of the world (6-8,13). Specifically, the incidence of polymicrobial peritonitis is much lower in many other centers, and gram-negative peritonitis episodes are much more common in tropical and subtropical areas (7). Moreover, Escherichia coli is the most common cause of gram-negative peritonitis episodes in the Western world (8,13), and Pseudomonas species are the foremost cause in some Asian countries (7); however, about two thirds of all gram-negative peritonitis episodes in the van Esch study were caused by neither organism—a unique observation that might be related to unidentified local patient or clinical factors.

Besides disconnect systems, has anything else potentially contributed to lowering peritonitis rates? Yes, an important confounding factor was not explored in detail in the van Esch study: the increase over time in the use of machine-assisted automated PD (APD). A number of previous studies suggested that the peritonitis incidence is lower with APD than with continuous ambulatory PD (14-16), although that finding is not confirmed by all studies (17,18). In the report from van Esch et al. (5), only 1 patient used APD during the period from 1979 to 1987 (and that patient was most likely treated with a traditional cycler machine using hanging bags); in contrast, nearly 70% of the patients after 2004 were on machine-assisted PD. It is almost certain that such a change in PD practice would have a profound effect on the incidence of peritonitis. From a global perspective, the use of APD varies from less than 10% in some Asian countries to more than 50% in North America (19). It remains uncertain how differences in the choice of PD modality contribute to observed differences in the peritonitis rate from country to country.

In addition to the spectrum of causative organisms and APD use, the low prevalence of patients with diabetes in the PD population studied by van Esch et al. might further limit the generalizability of the results to other countries. Specifically, about 22% of patients in the study had diabetic nephropathy, and that incidence remained remarkably stable throughout the study period. Although that incidence of diabetic nephropathy resembles the incidence in a recent report from the United Kingdom (9), it is substantially different from the incidences reported in many other parts of the world, where diabetic nephropathy often contributes 40% - 50% of new dialysis patients—an incidence that has continued to increase since the late 1990s. It seems unlikely that the Netherlands and the United Kingdom are the “odd countries out,” with distinctly low incidences of diabetic nephropathy. Instead, a different set of selection criteria must have been operating in the PD programs in the different countries.

In spite of the potential for bias discussed here, the results of the van Esch study are consistent with those of previous reports, and the overall message is clear. Taking a pessimistic viewpoint, the depressing fact revealed by all these studies is that, despite technologic advances, little improvement in the peritonitis rate has been achieved in the most recent decade. The availability of new biocompatible PD solutions has not appeared to further reduce the peritonitis rate—an observation that is also consistent with the recent meta-analysis reported by Srivastava et al. (20), which confirmed that patients randomized to biocompatible or conventional PD solutions show no difference in peritonitis rates or technique survival. The sobering truth is that calling for a large-scale randomized controlled trial of some novel technology is easier than performing the trial. Because the background peritonitis rate is now reasonably low in many countries, it might not be easy to devise effective measures to further lower the rate, and the required study would involve a huge sample size and incredible expense. Ironically, if any center were to have a peritonitis rate high enough to justify a randomized control trial, observers would be rather skeptical about the center’s standard of clinical care, and results would hardly be applicable to other countries.

Taking the optimistic viewpoint, a lesson to be learned from the van Esch study is that further research should be re-directed to the prevention and management of antibiotic resistance and the preservation of peritoneal function after peritonitis.

Disclosures

CCS receives a research grant from Baxter Healthcare.