Portal de Periódicos da CAPES

Worldwide variation of dialysis-associated peritonitis in children

2007; Elsevier BV; Volume: 72; Issue: 11 Linguagem: Inglês

10.1038/sj.ki.5002523

1523-1755

Franz Schaefer, Reinhard Feneberg, Nejat Aksu, Osman Dönmez, Büşra Aksan Sadıkoğlu, Steven R. Alexander, S. Mir, Il Soo Ha, Michel Fischbach, Eva Šimková, Alan R. Watson, Kirsten Møller, Heike von Baum, Bradley A. Warady,

Health Systems, Economic Evaluations, Quality of Life

Peritonitis is the most common cause of dialysis failure in children on chronic peritoneal dialysis. We performed a prospective study of 501 peritonitis episodes in 44 pediatric dialysis centers located in 14 countries that examined peritonitis etiology, efficiency of opinion-based management guidelines, and final outcomes. Culture-negative incidence varied significantly from 11% in North America to 67% in Mexico. Argentina and North America had the highest rate of Gram-negative episodes. Pseudomonas-based peritonitis was eightfold more common in the United States than in Europe, and correlated with the frequency of exit site cleansing and topical mupirocin administration. Significant regional variation in antibiotic susceptibility was noted for the first generation cephalosporins and aminoglycosides. Initial response rates to standardized empiric antibiotic treatment did not differ between regions; however, final outcomes were significantly less favorable in Eastern Europe. The wide regional variation in culture-negative peritonitis, and the distribution and antibiotic susceptibilities of causative bacteria needs to be taken into consideration when the guidelines for empiric therapy of pediatric dialysis-associated peritonitis are revised. Peritonitis is the most common cause of dialysis failure in children on chronic peritoneal dialysis. We performed a prospective study of 501 peritonitis episodes in 44 pediatric dialysis centers located in 14 countries that examined peritonitis etiology, efficiency of opinion-based management guidelines, and final outcomes. Culture-negative incidence varied significantly from 11% in North America to 67% in Mexico. Argentina and North America had the highest rate of Gram-negative episodes. Pseudomonas-based peritonitis was eightfold more common in the United States than in Europe, and correlated with the frequency of exit site cleansing and topical mupirocin administration. Significant regional variation in antibiotic susceptibility was noted for the first generation cephalosporins and aminoglycosides. Initial response rates to standardized empiric antibiotic treatment did not differ between regions; however, final outcomes were significantly less favorable in Eastern Europe. The wide regional variation in culture-negative peritonitis, and the distribution and antibiotic susceptibilities of causative bacteria needs to be taken into consideration when the guidelines for empiric therapy of pediatric dialysis-associated peritonitis are revised. Pediatric end-stage renal disease is a rare disorder, with an annual incidence of 1–3 per million population in the Western world. Dialysis periods are usually shorter than in adults since renal transplantation is the preferred end-stage renal disease treatment modality in children. Preemptive living donorship programs and allograft allocation principles favoring pediatric recipients are being implemented in the United States and Europe, reducing the need for long-term maintenance dialysis and, as a consequence, the average size of pediatric dialysis populations. As a downside of this most welcome development, the individual expertise with complications of dialysis treatment is decreasing. On the other hand, new pediatric dialysis programs are rapidly growing in many parts of the world where renal replacement therapy is still in the process of being available to the general population. Both the vanishing pediatric dialysis expertise in some and the rapidly growing need for knowledge in other regions of the globe create a strong need for defining and implementing standards of pediatric dialysis care, and for the systematic, prospective collection of data regarding treatment complications and outcomes. Chronic peritoneal dialysis (CPD) is the dialysis modality of first choice in the majority of children and adolescents. Although the incidence of CPD-related infectious complications has decreased in children as well as in adults within the past two decades, bacterial peritonitis remains the most important cause of morbidity and irreversible technique failure in the pediatric population. In order to improve the knowledge regarding this complication, the International Pediatric Peritonitis Registry (IPPR) was formed by a global consortium of 44 pediatric dialysis centers comprised of 20 European, 9 Turkish, 2 Asian, and 13 American centers.1.Feneberg R. Warady B.A. Alexander S.R. et al.The international pediatric peritonitis registry: a global Internet-based initiative in pediatric dialysis.Perit Dial Int. 2005; 25: S130-S134PubMed Google Scholar The registry has collected detailed prospective information of CPD-related peritonitis in children in order to assess the validity of the International Society for Peritoneal Dialysis (ISPD) pediatric peritonitis treatment guidelines2.Warady B.A. Schaefer F. Holloway M. et al.Consensus guidelines for the treatment of peritonitis in pediatric patients receiving peritoneal dialysis.Perit Dial Int. 2000; 10: 610-624Google Scholar for the pediatric CPD population in different parts of the world. In the analysis presented here, we investigated the global variation of the clinical manifestations, the spectrum and antibiotic sensitivity patterns of causative organisms, and the outcomes of bacterial peritonitis in children undergoing CPD. Tenckhoff catheters with a straight ending were standard in Eastern Europe (80% of patients), Asia (100%), and Mexico (63%), whereas in the United States, Western Europe, Turkey, and Argentina, 81–96% of patients had coiled catheters (P<0.0001). In 85% of patients entered into the registry two-cuff catheters were used, with significantly lower usage prevalence of this variant in the United States (68%, P<0.0001). Swan-neck catheters were more common in Turkey (73%), Asia (92%), and Argentina (96%) than in Eastern and Western Europe (both 15%), and were not used at all in Mexico (0%) (P<0.0001). Daily exit site care was applied in 13 of 14 (93%) centers in America and Asia, in 64% (9/14) of centers in Eastern Europe and Turkey, but only in 1 of 13 (8%) Western European centers (P<0.001). Large regional differences existed also with regard to the choice of exit site cleansing agent (P<0.0001). Soap (6 of 11 sites) or sodium hypochlorite (3 of 11) was used in North America. Povidone iodine was used in 7 of 9 Turkish, 6 of 18 European, both Asian but none of the US centers. Seven of 18 European sites used β-octenidine. Topical antibiotic prophylaxis against Staphylococcus aureus was practiced by 34/41 centers (83%). Selective nasal decontamination was widely used in Europe, Turkey, and Asia (25/30 centers), but in only 3 of 11 US centers. Conversely, topical prophylaxis was administered to the exit site more commonly in the United States (63%) and Turkey (55%) than in the other regions (3/21=14%) (P<0.05). At the time of peritonitis, 43% of patients in the United States, 17% in Turkey, 14% in Western Europe, 6% in Eastern Europe, 4% in Argentina, and none in Mexico and Asia received S. aureus prophylaxis. With respect to microbiology practices, gram stains of dialysis effluent were performed routinely in patients with suspected peritonitis in 11/12 centers in the United States and Asia, but were not typically performed in one-third (10/27) of laboratories in Europe, Turkey, and Argentina. In 55% of centers, the effluent was cultured in bottles (BacT/Alert, BACTEC), in 30% on blood agar plates, and in 15% other media were used. Susceptibility testing was performed by disc diffusion in 64% of laboratories, by broth microdilution in 21% and by using other tests in 15%, without significant regional clustering. Between October 2001 and December 2004, a total of 548 peritonitis episodes were recorded in 392 pediatric patients, that is, 1.4±0.8 episodes per patient. An average of 12.7±11 episodes was documented per center (European 13.5±12, American 12±9.9, Asian 6.5±7.8; NS). Five hundred and one episodes (91.4%) met the diagnostic and therapeutic criteria defined by the guidelines and were included in the analyses. The 47 episodes (8.6%) in which treatment was not performed according to one of the regimens recommended in the guidelines were distributed randomly among the participating centers and regions. The distribution of patient age and clinical symptoms at peritonitis onset differed significantly between regions (Table 1). The fraction of patients who were younger than 2 years was significantly greater in the United States and Western Europe than in Turkey and Latin America. Reported disease severity at first presentation was highest in Argentina and mildest in Mexico. Multivariate analysis of variance demonstrated the regional differences of disease severity score at presentation to be independent of patient age, the spectrum of causative organisms, and regional differences in preferred PD modality, catheter, and tunnel configuration.Table 1Patient age and clinical symptoms at time of diagnosis in 491 episodes of non-fungal peritonitisAllUSAMexicoArgentinaWestern EuropeEastern EuropeTurkeyAsiaP-valueAge <2 years69 (13%)17 (19%)3 (11%)0 (0%)32 (18%)7 (14%)8 (5%)2 (15%)0.001Marked effluent cloudiness348 (70%)44 (60%)13 (48%)19 (76%)95 (62%)40 (85%)124 (77%)13 (100%) 38°C231 (47%)27 (40%)7 (26%)16 (64%)83 (55%)34 (72%)57 (35%)7 (54%)<0.0001Severe pain209 (42%)30 (42%)4 (15%)18 (72%)69 (45%)24 (51%)61 (38%)3 (23%)<0.0001Disease severity score2.02±1.151.98±1.121.41±0.93*P<0.05 vs Argentina, Western and Eastern Europe.2.72±1.02#P<0.05 vs USA, Turkey, Mexico, Asia.2.19±1.112.41±1.151.82±1.111.46±1.66*P<0.05 vs Argentina, Western and Eastern Europe.<0.0001P<0.05 in last column indicates significant heterogeneity between regions; individual regions differing significantly from average percentage are marked in bold.* P<0.05 vs Argentina, Western and Eastern Europe.# P<0.05 vs USA, Turkey, Mexico, Asia. Open table in a new tab P<0.05 in last column indicates significant heterogeneity between regions; individual regions differing significantly from average percentage are marked in bold. The distribution of causative organisms is given in Table 2 and Figure 1. Culture negative peritonitis was rarest in North America (11%) and Argentina (15%), but accounted for 42% of all episodes in Turkey and 67% in Mexico (overall rate 29%; heterogeneity among countries P<0.0001). North America (46%, P<0.05) and Argentina (70%, P<0.005) had higher rates of Gram-negative episodes than the other regions. Peritonitis episodes caused by Pseudomonas spp. were more likely to occur in the United States (OR 2.95, confidence interval 1.32–6.60, P<0.01) and episodes with other Gram-negative organisms were more frequent in Argentina (OR 2.90, confidence interval 1.24–6.79, P<0.005) than in the other regions. Conversely, Gram-positive infections were predominant in Europe. Patients in Eastern Europe were significantly more likely to develop peritonitis caused by coagulase-negative staphylococci ((OR 2.57, confidence interval 1.30–5.06, P<0.01). S aureus peritonitis was twice as common in Western Europe than in the United States (P<0.05). Enterococci as a cause of peritonitis tended to occur more frequently in Turkey than in other regions (P=0.06).Table 2Distribution of causative organisms by regionAllUSAMexicoArgentinaWestern EuropeEastern EuropeTurkeyAsiaP-valueN3787492312542969Fungal10 (3%)3 (4%)0 (0%)0 (0%)4 (3%)1 (2%)2 (2%)0 (0%)NSCoag-neg. staph.85 (22%)11 (15%)1 (11%)1 (4%)32 (26%)16 (38%)23 (24%)1 (11%)<0.01S. aureus78 (21%)10 (14%)4 (44%)5 (22%)34 (27%)6 (14%)19 (20%)0 (0%)<0.05Streptococci31 (8%)7 (9%)1 (11%)0 (0%)13 (10%)3 (7%)5 (5%)2 (22%)NSEnterococci20 (5%)1 (1%)0 (0%)0 (0%)8 (6%)2 (5%)9 (9%)0 (0%)0.05Other Gram-pos.21 (6%)8 (11%)0 (0%)1 (4%)4 (3%)2 (5%)5 (5%)1 (11%)NSPseudomonas28 (7%)12 (16%)0 (0%)3 (13%)2 (2%)1 (2%)7 (7%)3 (33%)<0.0005Other Gram-neg.105 (28%)22 (30%)3 (33%)13 (57%)28 (22%)11 (26%)26 (27%)2 (22%)NSP<0.05 in last column indicates significant heterogeneity between regions; individual regions differing significantly from average percentage are marked in bold. Open table in a new tab P<0.05 in last column indicates significant heterogeneity between regions; individual regions differing significantly from average percentage are marked in bold. Notably, peritonitis caused by Pseudomonas spp. was significantly more common in patients from centers where exit site care was performed more than twice per week (P<0.005), where mupirocin cream was administered to the exit site (P<0.005) and where non-sterile cleansing agents (saline, soap) were used (P<0.001). Assignment to high-risk (glycopeptide) vs low-risk (cefazolin) therapy varied significantly (P<0.0001) between regions, from 100% cefazolin (Asia) to 93% glycopeptide usage (Argentina). The rate of adherence to the empiric treatment schedule recommended by the guidelines ranged from 43% (Asia) to 89% (Argentina) (P<0.0001). Significant regional variability of Gram-positive sensitivity to the first generation cephalosporin cefazolin was observed, with sensitivity rates significantly below average in the United States and above average in Eastern Europe (Table 3). In contrast, glycopeptide sensitivity was 96–100% without regional differences. Approximately 80% of Gram-negative and 50% of Gram-positive organisms were sensitive to the third generation cephalosporin ceftazidime, without differences between the regions. Aminoglycoside sensitivity was 88% on average for Gram-negative and 77% for Gram-positive organisms. Aminoglycoside sensitivity was significantly below average in Eastern Europe (P<0.05). Eighteen percent of Gram-negative organisms in Eastern Europe were resistant to aminoglycosides as compared to only 8% being resistant to ceftazidime.Table 3Regional distribution of in vitro antibiotic sensitivitiesAll regionsUSAArgentinaWestern EuropeEastern EuropeTurkeyP-valueCefazolin (All)105 (55%)10 (30%)4 (36%)37 (60%)20 (80%)33 (55%)<0.005 (Gram-pos.)88 (70%)7 (50%)4 (100%)35 (70%)15 (94%)26 (63%)0.05 (Gram-neg.)16 (25%)3 (16%)0 (0%)2 (17%)4 (57%)7 (39%)0.06Ceftazidime (All)113 (69%)18 (82%)9 (75%)32 (67%)18 (72%)33 (65%)NS (Gram-pos.)31 (52%)2 (67%)1 (100%)12 (50%)6 (55%)10 (50%)NS (Gram-neg.)81 (80%)16 (84%)8 (73%)20 (83%)11 (92%)23 (77%)NSAminoglycosides (All)222 (81%)38 (79%)14 (88%)83 (89%)25 (69%)55 (82%)0.09 (Gram-pos.)116 (77%)12 (67%)2 (100%)56 (86%)15 (63%)29 (81%)0.09 (Gram-neg.)105 (88%)26 (87%)12 (86%)27 (96%)9 (82%)26 (87%)NSGlycopeptides (Gram-pos.)182 (97%)23 (100%)6 (100%)72 (96%)24 (96%)48 (98%)NSCiprofloxacin (All)169 (93%)19 (90%)15 (94%)80 (94%)18 (86%)29 (94%)NSMexico and Asia excluded due to insufficient numbers of available resistograms. Numbers indicate absolute numbers and percentage of organisms tested sensitive.P<0.05 in last column indicates significant heterogeneity between regions; individual regions differing significantly from the average percentage are marked in bold. Open table in a new tab Mexico and Asia excluded due to insufficient numbers of available resistograms. Numbers indicate absolute numbers and percentage of organisms tested sensitive. P<0.05 in last column indicates significant heterogeneity between regions; individual regions differing significantly from the average percentage are marked in bold. Ninety-four percent of patients showed clinical improvement, and 55% were completely asymptomatic with clear dialysis effluent to be within 3 days of empairic antibiotic therapy (Table 4). The rate of patients with a good clinical response appeared to be lower in patients from Argentina; however, this difference was largely accounted by the greater initial disease severity and the high incidence of Gram-negative infections in patients from this country. After 4 weeks, all patients from Argentina showed full functional recovery.Table 4Response to empiric therapy and final outcomes of peritonitis by regionsAllUSAArgentinaMexicoWestern EuropeEastern EuropeTurkeyAsiaP-valueN49172182614946159133-day disease severity score0.33±0.690.3±0.51.0±1.20.0±0.20.4±0.70.2±0.50.3±0.60.6±1.3<0.00013-day clinical response460 (94%)69 (96%)18 (72%)26 (96%)141 (95%)46 (100%)148 (93%)12 (92%)<0.00013-day complete symptom resolution271 (55%)34 (47%)8 (32%)25 (93%)86 (58%)26 (57%)87 (55%)5 (38%)<0.00014-wk full functional recovery429 (89%)63 (89%)25 (100%)23 (88%)133 (90%)34 (76%)139 (90%)12 (92%)<0.005Relapsing peritonitis52 (10%)6 (8%)2 (8%)2 (7%)13 (9%)8 (17%)21 (13%)0 (0%)NSTemporary PD discontinuation12 (2.5%)4 (6%)2 (8%)1 (4%)2 (1%)2 (4%)1 (1%)0 (0%)NSPermanent PD discontinuation (technique failure)39 (8%)7 (10%)0 (0%))2 (8%)8 (5%)9 (20%)12 (8%))1 (8%)0.05Fungal peritonitis was excluded.P<0.05 in last column indicate overall heterogeneity among regions; individual regions differing significantly from overall mean/percentage are marked in bold. Open table in a new tab Fungal peritonitis was excluded. P<0.05 in last column indicate overall heterogeneity among regions; individual regions differing significantly from overall mean/percentage are marked in bold. Conversely, full resolution of clinical symptoms was associated with Mexican origin and, independently, a low disease severity score and the absence of S. aureus, Pseudomonas spp. and other Gram-negative infections (all P<0.005). Significant sequelae affecting PD function were observed more commonly in Eastern Europe, where CPD was permanently discontinued in 20% of patients following the peritonitis episodes as compared with 8% in the total population (P<0.0001). Moreover, catheter exchange was required more frequently (12/46, i.e. 26% of all episodes) in patients from Eastern Europe than in the other regions (44/445, i.e. 10%, P=0.005). Poor outcome was independently associated with the Eastern European region and a straight tunnel configuration (both P<0.005), whereas the choice of empiric treatment, patient age, and the other risk factors assumed in the guidelines were not predictive. Among the PD practices preferentially applied in Eastern Europe, the use of Tenckhoff catheters with a straight ending was associated with an increased rate of post-peritonitis technique failure (13 vs 7%, P=0.05). In this paper, we present the results of a unique global collaboration of pediatric nephrologists dedicated to improve peritoneal dialysis care in children. The analysis of 501 pediatric peritonitis episodes treated in 44 pediatric dialysis units on three continents demonstrates significant regional variability of the clinical presentation, bacterial spectrum, and treatment outcomes, providing a strong rationale for allowing site-specific adaptations in general clinical practice guidelines. The most striking regional variation was observed with regards to the rate of culture-negative peritonitis episodes. In Mexico, Turkey, and the Asian centers, effluent cultures remained sterile in 36–67% of cases, in contrast to 11–23% in the other regions. A survey of the laboratory procedures applied in the individual centers did not reveal any systematic differences in culture techniques that would explain these marked differences. Hence, we assume that other issues such as incubation of insufficient effluent volumes, long transport times in rural areas, and extreme ambient temperatures may adversely affect culture results in these regions. Most previous peritonitis studies in adults and children, typically performed in experienced centers in well-developed countries, reported culture-negative rates of less than 20%,3.de Freitas D.G. Gokal R. Sterile peritonitis in the peritoneal dialysis patient.Perit Dial Int. 2006; 26: 284PubMed Google Scholar, 4.Troidle L. Finkelstein F. Treatment and outcome of CPD-associated peritonitis.Ann Clin Microbiol Antimicrob. 2006; 5: 6Crossref PubMed Scopus (68) Google Scholar, 5.Kent J.R. Almond M.K. A survey of CAPD peritonitis management and outcomes in North and South Thames NHS regions (UK): support for the ISPD guidelines. International Society for Peritoneal Dialysis.Perit Dial Int. 2000; 20: 301-305PubMed Google Scholar a number which has subsequently been used as an acceptable maximum in PD quality assurance programs. Our study demonstrates that on a global scale, many PD units are far from meeting these standards. International peritonitis treatment guidelines should give special consideration to the management of culture-negative peritonitis. Another important finding of this global survey is the large regional variation of the distribution of causative organisms. In comparison with a pediatric peritonitis survey performed in the 1990s in Mid Europe,6.Schaefer F. Klaus G. Müller-Wiefel D.E. et al.Intermittent versus continuous intraperitoneal glycopeptide/ceftazidime treatment in children with peritoneal dialysis-associated peritonitis.J Am Soc Nephrol. 1999; 10: 136-145PubMed Google Scholar we noted an overall shift from Gram-positive to Gram-negative causative organisms. In the view of the generally decreasing incidence of CPD-related peritonitis, we believe that the apparent increase in Gram-negative infections largely reflects a decreased absolute frequency of Gram-positive peritonitis. The latter may be explained by minimized touch contamination due to improved handling and connection technologies, and by the implementation of S. aureus prophylaxis programs.7.Tacconelli E. Carmeli Y. Aizer A. et al.Mupirocin prophylaxis to prevent Staphylococcus aureus infection in patients undergoing dialysis: a meta-analysis.Clin Infect Dis. 2003; 37: 1629-1638Crossref PubMed Scopus (141) Google Scholar However, the proportion of Gram-negative organisms varied widely between regions, accounting for 70% of culture-positive infections in Argentina and 46% in the United States, but only around 25% in the European countries. As Gram-negative peritonitis is typically associated with a severe clinical course, it is not surprising that patients in Argentina presented with the most severe clinical symptoms at onset of disease of all patients in the registry. The reasons for this marked geographical variation of the bacterial spectrum are most likely multifactorial. Environmental influences such as climate and humidity may play a role, as well as the age distribution of the patients. Gram-negative infections were more frequently observed in young infants (data not shown), the proportion of whom was highest in the United States. Moreover, we noted important regional differences in various aspects of PD practices, including catheter and tunnel configuration, exit site care, and the routine use of topical antibiotic prophylaxis. Some of these factors have been reported to affect the incidence of PD-related infections,8.Piraino B. Preventing peritoneal dialysis related infections.Minerva Urol Nefrol. 2006; 58: 161-169PubMed Google Scholar,9.Verrina E. Honda M. Warady B.A. et al.Prevention of peritonitis in children on peritoneal dialysis.Perit Dial Int. 2000; 20: 625-630PubMed Google Scholar and they may also impact on the distribution of peritonitis causative organisms observed in different parts of the world. For instance, the incidence of S. aureus peritonitis was low in the United States (14%), where most sites followed mupirocin prophylaxis protocols and 43% of the patients (as compared with 0–17% in other regions) received mupirocin at the time when the peritonitis episode occurred. Conversely, the incidence of pseudomonas peritonitis was eightfold higher in the United States than for instance in Western Europe. This mupirocin-resistant microorganism may preferentially colonize exit sites frequently humidified by daily washing (standard in the United States but not in Western Europe) and the administration of mupirocin cream (practiced in two-thirds of the United States but none of the Western European sites). The observed significant association of pseudomonas peritonitis frequency with frequent exit-site care, use of non-disinfectant cleansing agents and topical mupirocin administration is in keeping with this interpretation. A further challenge to future empiric antibiotic treatment guidelines is the local variation of antibiotic sensitivities. Whereas Gram-negative ceftazidime resistance averaged around 20% without regional differences, the sensitivity of Gram-positive organisms to first generation cephalosporins varied significantly from 50% in the United States to 94% in the Eastern European countries. Conversely, aminoglycoside resistance was slightly more common in Eastern Europe than in other regions. This variability most likely reflects antibiotic treatment preferences in different regions, which are in turn influenced by local experience and treatment guidelines, cost considerations and, at least for some drugs, availability issues and marketing activities. Future peritonitis treatment guidelines should allow for local resistance patterns. For instance, although the empiric use of an aminoglycoside instead of ceftazidime would have increased Gram-negative coverage by 13% in Argentina and Western Europe, the advantage was only 3% in the United States and in Eastern Europe, 10% more Gram-negative organisms were sensitive to ceftazidime than to aminoglycosides. Hence, regular assessments of local sensitivity patterns and 'informed' selection of empiric therapy should be a cornerstone of future peritonitis treatment algorithms. Treatment outcomes showed some geographical variation. Although the early course appeared favorable in Mexico and adverse in Argentina, these differences were largely explained by marked differences in the initial disease severity and the distribution of causative organisms. The late functional sequelae of peritonitis, assessed here for the first time in pediatric patients, were more evenly distributed, with significantly less favorable results in the Eastern European population where catheter exchange was required most frequently and PD ultimately failed after a single peritonitis episode in as many as 20% of patients. Our analysis suggests that the preferential use of straight rather than coiled catheters in Eastern Europe may play a role. In patients with post-peritonitic adhesions, such catheters may provide inferior drainage compared with pig-tail catheters. Notably, neither the choice of empiric antibiotic therapy nor the 'correct' assignment of antibiotic therapy according to the risk profile suggested in the pediatric peritonitis treatment guidelines contributed to the regional outcomes. Hence, certain PD practices may have a greater impact on PD outcomes than the particular choice of empiric antibiotic therapy. Our findings have important consequences for PD practices. Although the relevance of individual associations from observational registries should ideally be tested in prospective clinical trials, this is frequently not feasible in small populations and may not even be justified when apparently superior practices do not involve greater cost or potential risks to the patients. In turn and based on our findings, we propose to abandon the use of straight-end catheters, because they were associated with poor functional recovery. Local monitoring of resistance patterns should guide the choice of empiric antibiotic treatment of peritonitis, and should replace the treatment assignment by putative risk factors recommended in the previous pediatric guidelines. The observed increased risk of pseudomonas peritonitis associated with topical mupirocin also raises doubts about the current concept of topical prophylaxis. In addition, our study indicates the potential roles of the frequency of exit-site care and the choice of cleansing agents as peritonitis risk factors. Hence, in the field of infection prophylaxis, the results of the IPPR should stimulate comparative prospective trials to objectively define the risk-benefit ratios of individual strategies. The IPPR exclusively utilized online data entry via an Internet-based web platform (www.peritonitis.org).1.Feneberg R. Warady B.A. Alexander S.R. et al.The international pediatric peritonitis registry: a global Internet-based initiative in pediatric dialysis.Perit Dial Int. 2005; 25: S130-S134PubMed Google Scholar Data pertaining to basic patient and PD modality characteristics, clinical presentation with peritonitis, microbiological results, empiric treatment and its subsequent modifications, clinical treatment response, and final outcomes were submitted prospectively along the course of a peritonitis episode and automatically checked for accuracy and consistency. In addition, center-specific demographic data and PD practices were collected by means of an online questionnaire. Data protection was ensured as the data input was anonymized. The registry protocol was approved by the ethical committee/institutional review board at each participating center. Monitoring in the IPPR was restricted to those patients who developed peritonitis episodes. Patients remaining peritonitis-free were not registered, nor was the total duration of dialysis times in those who developed peritonitis. Hence, information pertaining to the number of peritonitis episodes per time at risk is not available. In accordance with the ISPD guidelines, the diagnosis of peritonitis required (i) cloudy effluent, (ii) an effluent cell count of at least 100 cells/μl, and (iii) at least 50% polymorphonuclear cells in the differential cell count. The catheter exit site appearance was characterized according to a standardized scoring system (ESS).3.de Freitas D.G. Gokal R. Sterile peritonitis in the peritoneal dialysis patient.Perit Dial Int. 2006; 26: 284PubMed Google Scholar Treatment of peritonitis was assumed to be in accordance with the ISPD guidelines if the diagnostic criteria for peritonitis were fulfilled and/or an organism was grown on culture and empiric antibiotic treatment was initiated according to the recommended risk stratification schemes.2.Warady B.A. Schaefer F. Holloway M. et al.Consensus guidelines for the treatment of peritonitis in pediatric patients receiving peritoneal dialysis.Perit Dial Int. 2000; 10: 610-624Google Scholar The guidelines suggested the use of ceftazidime combined with a first generation cephalosporin (cefazolin or cefalothin) in children with an assumed low risk, and combined with a glycopeptide (vancomycin or teicoplanin) in patients assumed to be at a high risk for a severe course of infection. An increased risk was assumed in children with severe clinical symptoms, younger than 2 years of age, recent or current exit-site infection and a history of an Methicillin-resistant S. aureus (MRSA) infection. The empiric use of one of these antibiotic combinations, but not treatment selection according to the suggested risk stratification, was a prerequisite for inclusion of patients in the registry. A standardized Disease Severity Score (range 0–5) was calculated as the sum of 3 score points for pain and 2 for fever.3.de Freitas D.G. Gokal R. Sterile peritonitis in the peritoneal dialysis patient.Perit Dial Int. 2006; 26: 284PubMed Google Scholar Empiric treatment response was considered satisfactory if the Disease Severity Score 72 h after the start of empiric antibiotic therapy was ≤2 and effluent cloudiness had visually improved. Final treatment outcome was judged according to the occurrence of a relapse, temporary or permanent PD discontinuation, and the attainment of full functional recovery. The latter was assumed if PD was continued without functional impairment (such as ultrafiltration loss or intraabdominal adhesions compromising fluid flow), irrespective of whether a relapse occurred or a catheter exchange was performed. Relapsing peritonitis was defined by recurrence of peritonitis with the same organism (defined by biochemical differentiation and resistogram, or the occurrence of two episodes remaining culture-negative) within 4 weeks after termination of antibiotic treatment. Antibiotic resistograms accompanied most, but not all positive cultures. Some resistograms included a priori assumptions (Gram-negative organisms regarded as resistant to glycopeptides; enterococci regarded as resistant to cephalosporins). For statistical analysis, the following regions were defined: Western Europe (nine sites in Germany, two in Italy, one each in France, Austria, Finland, Greece, UK), Eastern Europe (three sites in Poland, one in Czech Republic), Turkey (three sites in Ankara, two in Istanbul, two in Izmir, one each in Antalya, Bursa), North America (11 sites in the United States), Mexico (Mexico City) Argentina (Buenos Aires), and Asia (Seoul (Korea), Bangkok (Thailand)). Student's χ2 tests were used to assess differences between proportions, and analysis of variance followed by Newman–Keuls tests for multiple comparisons to assess differences between group means. The association of outcomes with potential effector variables was assessed using univariate and multivariate logistic regression models. The International Pediatric Peritonitis Registry has been supported by an unrestricted grant by Baxter Healthcare Ltd. The IPPR was kindly supported by Baxter Healthcare and the International Society for Peritoneal Dialysis. Participants of the IPPR (sites ordered by number of peritonitis episodes entered): N Aksu, Izmir; A Sirin, B Sadikoglu, Istanbul; M Zimmering, Berlin; F Schaefer, R Feneberg, Heidelberg; R Munoz, Mexico City; E Sojo, Buenos Aires; O Donmez, Bursa; KE Bonzel, Essen; N Majkowski, J Lane, Chicago; S Testa, G Ardissino, Milan; L Sever, Istanbul; D Drozdz, Krakov; SR Alexander, Stanford; S Mir, N Ozkayin, Izmir; DE Müller-Wiefel, Hamburg; I Mader, Wroclaw; I Salusky, Los Angeles; Il Soo Ha, Seoul; J Flynn, New York; K Rönnholm, Helsinki; A Zurowska, Gdansk; M Ekim, Ankara; E Verrina, Genoa; N Besbas, Ankara; G Offner, Hannover; BA Warady, Kansas City; K Arbeiter, Vienna; M Fischbach, Strasbourg; CK Blaszak, Little Rock; E Simkova, Prague; G Klaus, Marburg; P Waber, M Seikaly, Dallas; M Benfield; M Sharbono, Birmingham; AR Watson, Nottingham; A Gür Güven, Antalya; J Symon; S Seattle; S Bakkaloglu, Ankara; C Stefanidis, Athens; A Al-Uzri, Portland; P Kingwatanakul, Bangkok; U John, Jena; L Stapenhorst, Cologne; K Haluany, Leipzig; D Gipson, Chapel Hill.