Although long-term outcomes of hemodialysis and

Transcription

1 Peritoneal Dialysis International, Vol. 31, pp doi: /pdi /11 $ Copyright 2011 International Society for Peritoneal Dialysis SIMILAR PERITONITIS OUTCOME IN CAPD AND APD PATIENTS WITH DIALYSIS MODALITY CONTINUATION DURING PERITONITIS Wim Rüger, 1 Frans J. van Ittersum, 1 Luiz F. Comazzetto, 2 Sanne E. Hoeks, 3 and Pieter M. ter Wee 1 Department of Nephrology, 1 VU University Medical Center; VU University, 2 Amsterdam; Department of Clinical Epidemiology, 3 Erasmus Medical Center, Rotterdam, The Netherlands Background: As few data exist on treatment of peritonitis in patients on automated peritoneal dialysis (APD), and as pharmacokinetics of several antibiotics are reported to be unfavorable in APD, some favor switching to continuous ambulant PD (CAPD) while treating APD-related peritonitis. We explored whether treating peritonitis with patients continuing their usual PD modality had an effect on outcome. Methods: We performed a retrospective analysis of the 508 episodes of PD-associated peritonitis seen in 205 patients in our center from January 1993 to January During this period, the standard initial therapy for PD-related peritonitis was a combination of intraperitoneal gentamicin and rifampicin. Results: There was no difference in cure rate between CAPD and APD groups. Likewise, initial and maximal leukocyte counts in the PD fluid (PDF), relapse rates, catheter removal rates, and death during treatment of peritonitis were similar in the CAPD and APD groups. Median (interquartile range) duration of elevated leukocyte count in PDF was longer in APD: 5.0 ( ) days versus 4.0 ( ) days in CAPD (p <0.001). APD patients were treated with antibiotics longer than CAPD patients: 16.0 ( ) versus 15.0 ( ) days (p = 0.036). Also, after correction for possible confounders, odds ratios for death and for the combined end point death or catheter removal showed no difference when patients treated for peritonitis stayed on their own modality. Conclusion: Regarding rate of relapse, mortality, or the combined end point mortality plus catheter removal, we found no difference between CAPD and APD patients continuing their own PD modality during treatment of PDrelated peritonitis. Intermediate end points such as duration of elevated PDF leukocyte count and duration of antibiotic treatment were longer in APD patients. Perit Dial Int 2011; 31:39-47 epub ahead of print: 17 Jun doi: /pdi Correspondence to: W. Rüger, Department of Nephrology, Room OD53, VU University Medical Center, PO Box 7057, 1007 MB Amsterdam, The Netherlands. j.ruger@vumc.nl Received 23 November 2009; accepted 16 April KEY WORDS: APD; CAPD; outcome; peritonitis. Although long-term outcomes of hemodialysis and peritoneal dialysis (PD) are considered similar, at least during the first years (1,2), infection-related mortality in PD patients exceeds its equivalent in hemodialysis patients. This difference is attributable mainly to mortality from PD-related peritonitis (3), which is possibly a cause of the present unpopularity of PD (4). Continuing efforts are being made to reduce the occurrence of peritonitis to a minimum. In past decades, the introduction especially of the Y-set and twin bag have reduced the incidence of PD-related peritonitis (5). Several reports [summarized by Diaz-Buxo (6)] have shown lower peritonitis incidence in patients performing automated PD (APD) compared to patients on continuous ambulant PD (CAPD). Because of other advantages offered by APD, such as freedom from performing exchanges during daytime hours, decreased mechanical complications (hernias, back pain) (7), and body figure issues, it is nowadays the preferred PD modality for many patients and is gaining territory on CAPD in many parts of the world (8,9). The combination of lower peritonitis incidence in APD patients and the shift toward APD may contribute to the declining overall incidence of PD-related peritonitis. Clinical data on treatment of APD-related peritonitis are scarce. Pharmacokinetic studies of intraperitoneally administered antibiotics showing more clearance of antibiotics in patients on APD have led to the recommendation to switch APD patients temporarily to CAPD during PD-related peritonitis (6,10,11). Whether this strategy improves clinical outcome has not been established. For the past 15 years, APD and CAPD patients in our department have been treated for PD-related peritonitis with similar antibiotic treatment protocols while continuing their PD modality. The goal of the present study is to compare peritonitis outcomes of patients on APD and CAPD continuing their dialysis modality during peritonitis episodes. 39

2 RÜGER et al. JANUARY 2011 VOL. 31, NO. 1 PDI METHODS PATIENTS This retrospective single-center study was conducted at the Department of Nephrology of the VU University Medical Center (Amsterdam, The Netherlands), a mixed secondary and tertiary referral center. We acquired all results of abdominal fluid cultures performed in our hospital in adult patients between January 1993 and January 2007 by an automated search of the electronic database of the Department of Medical Microbiology. Before further data collection, we assessed whether these patients were indeed suffering from a PDrelated peritonitis at the time of the culture. The diagnosis of PD-related peritonitis was based on the standard criteria of presence of two or three of the following: signs of peritoneal inflammation, positive Gram stain and/or culture of PDF, and leukocyte count in PDF more than /L (12). Patient records were reviewed to obtain patient characteristics (e.g., age, gender, PD modality, start date of PD), data on the peritonitis episode (start date of the episode, start and end date of antibiotic treatment, organisms cultured, initial and maximal leukocyte counts in PDF, number of days with PDF leukocyte count above /L), and outcome data (recovery, catheter replacement, catheter removal, death). In addition, infections were classified as relapse, repeat, or recurrent peritonitis. By definition, a relapse is an infection with either the same organism or a negative culture occurring within 4 weeks of completion of therapy of a prior (culture positive) episode. Peritonitis caused by the same organism but later than 4 weeks was classified as repeat. If the infection occurred within 4 weeks of completion of therapy of a prior episode but with a different organism it was classified as recurrent (13). PERITONITIS TREATMENT PROTOCOL As soon as PD-related peritonitis was suspected on clinical grounds, patients were started on empirical broad-spectrum antibiotic treatment. The standard initial therapy for PD-related peritonitis in our center is a combination of intraperitoneal gentamicin (20 mg/l dialysis fluid in the first exchange and thereafter once daily in the long dwell) and rifampicin (50 mg/l in all exchanges). In our hands, this regimen has proven to be efficacious (14). Empirical aminoglycoside use for PD peritonitis has been shown to be safe with respect to both residual renal function (15) and ototoxicity (16). Some- times a different treatment was chosen; for example, in case of a known allergy or when culture results of a recent peritonitis episode or exit-site infection required different antibiotics. Also, patients could have been treated with intravenous antibiotics; for example, if systemic infection was suspected. From October 1996 to October 1999, patients could also have been treated according to the protocol of the CIPPER trial (ciprofloxacin 50 mg/l and rifampicin 50 mg/l vs cephradine 250 mg/l) (17). Subsequent antibiotic treatment was adapted as seemed fit by the clinician, based on culture results and clinical response. During treatment for peritonitis, patients stayed on their own PD modality. Rarely, APD patients were switched to CAPD; for example, in case of outflow problems due to fibrin clots. If the clinical situation allowed, patients were preferably treated on an outpatient basis. Follow-up of leukocyte counts in PDF was performed with a 2- to 3-day interval, or more often if necessary. In CAPD patients, leukocyte counts were performed on the effluents of the overnight dwells. APD patients collected a sample from the first effluents from the long dwell immediately before starting APD. If patients were treated at home, these samples were kept refrigerated overnight and delivered to the hospital the next morning. Antibiotic treatment was continued until 1 week after the leukocyte count in the effluent was lower than /L, with a minimum of 2 weeks (3 weeks if Staphylococcus aureus was the causative organism). Reasons for removal of the PD catheter were refractory peritonitis, fungal peritonitis, or suspicion of bowel perforation or other abdominal crisis. Concomitant exitsite infection with certain micro-organisms (e.g., Pseudomonas) could also result in catheter removal. OUTCOME Outcomes of peritonitis episodes were categorized as initial cure (indicating resolution of the infection with antibiotic treatment alone), catheter replacement (resolution with antibiotics plus simultaneous catheter removal and replacement without the need to temporarily transfer the patient to hemodialysis), catheter removal, or death. Catheter removal and death were combined as unfavorable outcome. The occurrences of relapse, repeat, and recurrent peritonitis in CAPD- and APD-related peritonitis were compared. In addition, we assessed peritonitis duration by reviewing duration of elevated leukocyte count in PDF as well as duration of antibiotic treatment. Because follow-up of leukocyte count was not feasible after cath- 40

3 PDI JANUARY 2011 VOL. 31, NO. 1 CONTINUATION OF PD MODALITY DURING PERITONITIS TREATMENT eter removal, episodes resulting in catheter removal were excluded from this analysis. STATISTICAL ANALYSIS Continuous data are expressed as mean ± standard deviation (SD) or median with interquartile range (IQR). Comparisons were performed using Student s t-tests or Mann Whitney U tests, as appropriate. Dichotomous data were analyzed with chi-square tests. A p value less than 0.05 was considered statistically significant. The database has a hierarchical structure with multiple peritonitis episodes occurring in patients. To account for the different sources of variation (episode level and patient level) in observed outcome, a random effects model for logistic regression analysis was constructed. We used a hierarchical model with patient fit as a random effect, thereby correcting for the multiple episodes per patient. The CAPD group was used as reference. In the first step of the multilevel model we included age and gender. Time on PD and initial leukocyte count in PDF were added second. We tested for interaction between the variables in the multivariable model. None of these interaction terms were statistically significant and were therefore not included in the final model. Analysis was performed using SPSS 15.0 (SPSS Inc., Chicago, IL, USA) and R statistical software (R Foundation for Statistical Computing, Vienna, Austria). Hierarchical regression models were constructed with the Laplace method, using the LME4 package of R. RESULTS PATIENT AND PERITONITIS EPISODE CHARACTERISTICS Characteristics of APD and CAPD patients and peritonitis episodes are presented in Table 1. From January 1993 to January 2007, we registered 508 episodes of PDrelated peritonitis in 205 different adult patients. In 65% of the APD-related episodes, the patient was male, whereas 44% were males (p < 0.001) in CAPD-related peritonitis. Patients on CAPD were older (57.5 ± 16.6 vs 52.2 ± 16.0 years for APD, p < 0.001) and had been treated with PD for a shorter period of time (30.1 ± 30.0 vs 37.7 ± 33.6 months for APD, p = 0.005) when peritonitis was diagnosed. In patients first peritonitis episodes, the assigned causes of renal failure in the CAPD and APD groups were similar. Looking at all peritonitis episodes, glomerulonephritis was the cause of renal failure in 27% of APD-related episodes versus 13% of all episodes in CAPD patients (p < 0.001). This difference was attributable to a few individual patients on APD who had recurrent episodes of peritonitis. The median (IQR) leukocyte count in PDF at presentation with peritonitis [1.0 ( ) 10 9 in CAPD vs 1.4 ( ) 10 9 /L in APD, p = 0.103] and the median maximal (IQR) leukocyte count in PDF during treatment [1.7 ( ) 10 9 in CAPD vs 1.9 ( ) 10 9 /L in APD, p = 0.162] were similar in the two PD modalities. Also, peritonitis frequency and time from start of PD to the patient s first peritonitis episode were similar for CAPD- and APD-related peritonitis. Causative organisms cultured from PDF are shown in Table 2. In 64.4% of peritonitis episodes, one or more gram-positive micro-organisms, primarily coagulasenegative staphylococci (CoNS), were cultured from the PDF. Gram-negative micro-organisms, primarily Enterobacteriaceae, were cultured in 33.9% of cases. Polymicrobial peritonitis occurred in 18.5% and cultures remained negative in 9.4% of episodes. More cases with CoNS were seen in CAPD patients (32.3% vs 24.3% in APD, p = 0.044). There was a trend toward more yeast or fungal infections in the APD group (8.4% vs 4.8% in CAPD, p = 0.107). PERITONITIS OUTCOME We registered an overall initial cure rate of 76.0% and another 3.5% of patients recovered with antibiotics plus simultaneous catheter removal and replacement with continuation of PD, as described by Posthuma et al. (18). Catheter removal without simultaneous replacement was performed in 13.8%. In 6.7% of episodes, the patient died during the course of treatment of PD-related peritonitis. There was no difference in crude outcome between CAPD and APD groups (Table 3). Odds ratios for mortality and for unfavorable outcome are shown in Table 4. We could not find any difference in mortality or unfavorable outcome of PD-related peritonitis between APD and CAPD patients. Also, after correction for possible confounding factors and for differences between CAPD and APD groups, outcome was similar. In 7.9% of episodes, the peritonitis was classified as relapse; 14.4% were repeat and 7.1% recurrent peritonitis. Here also no statistically significant difference between CAPD and APD groups was found (Table 5). Analysis of the duration of peritonitis episodes showed a longer duration in APD patients. Median (IQR) duration of elevated leukocytes in PDF was 4.0 ( ) days in CAPD- and 5.0 ( ) days in APD-related peritonitis (p < 0.001). Median (IQR) duration of antibiotic administration was 15.0 ( ) days in CAPD patients and 16.0 ( ) days in APD patients (p = 41

4 RÜGER et al. JANUARY 2011 VOL. 31, NO. 1 PDI TABLE 1 Patient and Peritonitis Episode Characteristics Characteristic All episodes CAPD APD p Value Episodes (n) Male (%) 54% 44% 65% <0.001 Mean (range) age (years) 55.0 (21 90) 57.5 (21 90) 52.2 (21 79) <0.001 Mean (range) time on dialysis (months) 33.6 (0 200) 30.1 (0 144) 37.7 (0 200) Initial leukocyte count in PDF ( 10 9 /L) a 1.1 ( ) 1.0 ( ) 1.4 ( ) Maximal leukocyte count in PDF ( 10 9 /L) a 1.8 ( ) 1.7 ( ) 1.9 ( ) Patients (n) Male (% at 1st peritonitis) 51.2% 44.6% 59.1% Assigned cause of ESRF at 1st peritonitis b Glomerulonephritis 30 (15%) 13 (12%) 17 (18%) 0.63 Nephrosclerosis 39 (19%) 24 (21%) 15 (16%) Diabetes mellitus 25 (12%) 15 (13%) 10 (11%) Other 85 (42%) 46 (41%) 39 (42%) Unknown 26 (13%) 14 (13%) 12 (13%) Mean (range) time to 1st peritonitis (months) 21.6 (0 142) 22.9 (0 142) 20.3 (0 106) 0.42 Peritonitis frequency (1 episode/n months) PDF = peritoneal dialysis fluid; ESRF = end-stage renal failure; CAPD = continuous ambulatory peritoneal dialysis; APD = automated peritoneal dialysis. a Median (interquartile range). b Number (%). TABLE 2 Causative Micro-Organisms All episodes CAPD APD p Value Episodes (n) Gram positive CoNS 145 (28.5%) 87 (32.3%) 58 (24.3%) Staphylococcus aureus 73 (14.4%) 33 (12.3%) 40 (16.7%) Streptococci 51 (10.0%) 29 (10.8%) 22 (9.2%) 0.56 Other gram positive 92 (18.1%) 45 (16.7%) 47 (19.7%) 0.39 Any gram positive 327 (64.4%) 177 (65.8%) 150 (62.8%) 0.48 Gram negative Enterobacteriaceae a 121 (23.8%) 63 (23.4%) 58 (24.3%) 0.82 Nonfermenters b 36 (7.1%) 18 (6.7%) 18 (7.5%) 0.71 Other gram negative 36 (7.1%) 18 (6.7%) 18 (7.5%) 0.71 Any gram negative 172 (33.9%) 91 (33.8%) 81 (33.9%) 0.99 Yeast or fungus 33 (6.5%) 13 (4.8%) 20 (8.4%) Culture negative 48 (9.4%) 22 (8.2%) 26 (10.9%) 0.30 Polymicrobial 94 (18.5%) 49 (18.2%) 45 (18.8%) 0.86 CoNS = coagulase-negative staphylococci; CAPD = continuous ambulatory peritoneal dialysis; APD = automated peritoneal dialysis. a Citrobacter, E. coli, Enterobacter, Klebsiella, Morganella, Proteus, or Serratia species. b Pseudomonas, Stenotrophomonas, or Acinetobacter species. Values expressed as number (%) ). Subgroup analysis of the peritonitis episodes with a single causative micro-organism cultured is shown in Table 6. For gram-positive monoculture and culturenegative peritonitis we found a longer median duration of elevated leukocyte count in PDF of APD patients (p = and respectively). In case of monoculture 42

5 PDI JANUARY 2011 VOL. 31, NO. 1 CONTINUATION OF PD MODALITY DURING PERITONITIS TREATMENT TABLE 3 Crude Peritonitis Outcome All episodes CAPD APD p Value Initial cure 386 (76.0%) 203 (75.5%) 183 (76.7%) 0.98 Catheter replacement 18 (3.5%) 10 (3.7%) 8 (3.3%) Catheter removal 70 (13.8%) 37 (13.8%) 33 (13.8%) Death 34 (6.7%) 19 (7.1%) 15 (6.3%) Total CAPD = continuous ambulatory peritoneal dialysis; APD = automated peritoneal dialysis. Values expressed as number (%). TABLE 4 Crude and Adjusted Odds Ratios (OR) for Mortality and Unfavorable Outcome by Peritoneal Dialysis (PD) Modality Parameter All episodes CAPD APD Deaths (n) Crude OR for death (95% CI) Reference 0.91 ( ) Age- and gender-adjusted OR (95% CI) Reference 1.28 ( ) OR adjusted for previous plus time on PD and initial leukocytes (95% CI) Reference 2.09 ( ) Unfavorable outcome a (n) Crude OR for unfavorable outcome (95% CI) Reference 0.95 ( ) Age- and gender-adjusted OR (95% CI) Reference 1.08 ( ) OR adjusted for previous plus time on PD and initial leukocytes (95% CI) Reference 1.02 ( ) CI = confidence interval; CAPD = continuous ambulatory PD; APD = automated PD. a Defined as death or catheter removal. TABLE 5 Peritonitis Episode Classification All episodes CAPD APD p Value Primo infection 359 (70.7%) 186 (69.1%) 173 (72.4%) 0.86 Relapse 40 (7.9%) 22 (8.2%) 18 (7.5%) Recurrence 36 (7.1%) 21 (7.8%) 15 (6.3%) Repeat 73 (14.4%) 40 (14.9%) 33 (13.8%) Total CAPD = continuous ambulatory peritoneal dialysis; APD = automated peritoneal dialysis. Values expressed as number (%). of a nonfermenting organism, a trend toward longer duration of APD-related peritonitis was seen (p = 0.059). DISCUSSION The main finding of the present analysis is that continuing PD patients on their own treatment modality during treatment for peritonitis results in similar clinical outcomes in CAPD and APD. APD treatment was associated with longer duration of elevated leukocyte count in the PDF but, in our study, no causality can be established. Although our analysis was performed retrospectively, its findings are important since there have been only a few studies addressing the issue of outcome of APD- and CAPD-related peritonitis on the original PD modality. The available literature on this subject is flawed in various ways. Some reports describe only peritonitis treated on an outpatient basis. Consequently, their results are biased by the exclusion of the serious, that is, clinically treated, episodes (19,20). In other studies, although not always stated explicitly, most APD patients 43

6 RÜGER et al. JANUARY 2011 VOL. 31, NO. 1 PDI TABLE 6 Duration (in Days) of Elevated Leukocytes in Peritoneal Dialysis (PD) Fluid by Causative Micro-Organism and PD Modality a All episodes CAPD APD CAPD:APD (n:n) p Value All episodes b 5.0 ( ) 4.0 ( ) 5.0 ( ) 213:186 <0.001 Gram positive CoNS 3.0 ( ) 3.0 ( ) 4.0 ( ) 70: Staphylococcus aureus 6.0 ( ) 6.0 ( ) 6.0 ( ) 22: Streptococci 6.0 ( ) 4.0 ( ) 7.5 ( ) 15: Other gram positive 5.0 ( ) 5.0 ( ) 5.0 ( ) 14: All gram positive 4.0 ( ) 4.0 ( ) 5.0 ( ) 121: Gram negative Enterobacteriaceae c 6.0 ( ) 5.0 ( ) 6.0 ( ) 36: Nonfermenters d 6.5 ( ) 5.0 ( ) 10.0 ( ) 5: Other gram negative 4.0 ( ) 4.5 ( ) 3.5 ( ) 6: All gram negative 5.0 ( ) 5.0 ( ) 6.0 ( ) 47: Yeast or fungus No data 0:0 Culture negative 5.0 ( ) 3.0 ( ) 7.0 ( ) 17: Polymicrobial 5.0 ( ) 4.5 ( ) 5.0 ( ) 28: CoNS = coagulase-negative staphylococci; CAPD = continuous ambulatory PD; APD = automated PD. a Only episodes without catheter removal and with a single micro-organism cultured. b Includes polymicrobial and culture-negative episodes. c Citrobacter, E. coli, Enterobacter, Klebsiella, Morganella, Proteus, or Serratia species. d Pseudomonas, Stenotrophomonas, or Acinetobacter species. Values expressed as median (interquartile range) duration in days. were switched to CAPD during peritonitis (21,22). Furthermore, adjustments for patient characteristics, such as age and gender, had not been made (23,24) or results were compared with historical data (20,25). Our retrospective study showed several differences in characteristics of our CAPD and APD populations. In APD-related peritonitis, patients were predominantly male and, on average, younger, and they had been on PD for a longer time than their CAPD counterparts. This might be explained by the fact that younger patients are more attracted to the freedom from daytime exchanges in APD, whereas older patients might be afraid of the technical aspects of APD and therefore prefer CAPD. The difference in time on PD might reflect the fact that, in case of high peritoneal transport, APD is used as an alternative to CAPD. The initial and maximal leukocyte counts in PDF were similar in CAPD- and APD-related peritonitis. The duration of elevated leukocyte counts in PDF and the duration of antibiotic treatment in patients in whom the PD catheter remained in situ were different. The latter is probably related to the first as, according to the recommendations of the ISPD on PD-related peritonitis, antibiotic treatment was continued until at least 1 week after clearing of the PDF (13). With incorporation of interaction terms for age, gender, and dialysis vintage in the logistic regression model, we found no statistically significant effect of these parameters on leukocyte count or its evolution. There can be several explanations for the longer duration of elevated leukocyte counts in the effluent of APD patients. First, differences in pharmacokinetics have been shown for several antibiotics in APD and CAPD treatment (10). In spite of these differences, Blunden et al. found no correlation of serum concentrations of gentamicin and vancomycin with peritonitis outcome (23). Whether differences in antibiotic levels result in differences in the duration of elevation of leukocytes in PDF is not known. Another possible explanation is that, although the initial leukocyte count in the effluent in APD patients is not statistically different from that in CAPD patients, peritonitis tends to be discovered in a later stadium in APD patients. Due to the frequent nighttime exchanges and effluent being drained directly into the sewage by many APD patients, symptoms of peritoneal inflammation may be mitigated and cloudiness of effluent can be missed (26). One can hypothesize that this sets back APD patients at the start of peritonitis treatment already and that, even with equally potential treatment, APD patients have longer lasting peritonitis. It would be interesting to see whether APD patients that do switch to CAPD dur- 44

7 PDI JANUARY 2011 VOL. 31, NO. 1 CONTINUATION OF PD MODALITY DURING PERITONITIS TREATMENT ing peritonitis treatment also have longer lasting leukocyte elevation in their PDF. Finally, it can be speculated that differences in leukocyte counts in PDF in CAPD and APD patients do not so much reflect the severity of the peritonitis but are more an indication of differences in immune system responsiveness. Peritoneal macrophage function is reported to be affected by dwell time. The long day dwell in APD is thought to be beneficial for host defense and may contribute to reported lower peritonitis incidence. As macrophages are capable of recruiting neutrophils, it is not definitely known whether PDF leukocyte counts can be used as a comparator of peritoneal inflammation (27,28). In our study, CAPD patients had more infections in which CoNS was the causative organism. Peritonitis with CoNS is, in general, clinically milder (13), has good initial cure rates (21,23,29,30), and is reported to have lower initial and cumulative dialysate leukocyte counts (31). However, in a subgroup analysis (Table 6) of the peritonitis episodes with CoNS monoculture, we did not find a difference in duration of elevated PDF leukocytes between the two PD-modalities. Statistical significance may have been lost due to the relatively small subgroups. Although analysis of larger subgroups (all gram-positive or all gram-negative episodes showed longer duration of peritonitis episodes in APD patients when a gram-positive monoculture was found, we did not find a difference in occurrence of gram-positive peritonitis, in monoculture or in polymicrobial infections with gram-positive micro-organisms cultured. For culture-negative peritonitis episodes, there was a statistically significant difference in duration of leukocyte count elevation in PDF. For the small group of nonfermenting organisms, a trend toward longer duration in APD peritonitis was seen. In most episodes with cultures revealing yeasts, the catheter was removed. As episodes resulting in catheter removal were excluded from analysis of peritonitis duration, yeast peritonitis cannot explain the difference in duration of elevated leukocyte counts in PDF. Whether the longer duration of elevated leukocytes is disadvantageous for APD patients is uncertain. Selgas et al. showed in CAPD patients that the total number of days with PDF leukocyte count >100/mL was related to changes in ultrafiltration capacity (32). There are no such studies in APD patients. Nevertheless, our crude peritonitis outcome data showed no difference in diverse outcome categories (Table 3). An initial cure rate of 76.0% is similar to those reported in the literature (21,23 25). Higher initial cure rates are reported in patients treated on an outpatient basis (19,20). Our mortality rate of 6.7% is similar to mortality rates in other reports (21,22,24,33,34). Statistical correction for possible confounding factors and observed differences in characteristics of the CAPD and APD groups, such as age, gender, and dialysis vintage, did not reveal any difference in outcome between these groups. Theoretically, a better peritonitis outcome for APD patients that do switch to CAPD during peritonitis episodes could have been nullified by their staying on APD. Due to the observational and retrospective nature of our study, this possibility cannot be completely excluded but we think it is unlikely as better peritonitis outcome for APD patients compared to CAPD patients has never been described in other studies. In conclusion, when treating APD and CAPD patients for PD-related peritonitis with rifampicin and gentamicin as empirical antibiotics while patients continued their own PD modality, we found no effect on rate of relapse, mortality, or the combined end point mortality plus catheter removal. We found a difference in intermediate end points such as the duration of elevated PDF leukocyte count or, and probably related to this, a longer course of antibiotic treatment. Causality could not be established in this study. Whether the longer peritonitis duration can have consequences for the individual APD patient in the long term is uncertain. Nevertheless, in our view, it is not necessary to switch APD patients to CAPD while treating PD peritonitis. There are several study limitations worth mentioning. First, there is the possibility of limited generalizability, as our empirical antibiotic regimen is not according to current ISPD recommendations and as causative organisms and sensitivities may differ from other centers. Furthermore, this was a nonrandomized retrospective study with bias lurking. For example, patient characteristics making CAPD or APD the more appropriate treatment for a certain patient could have induced selection bias. A comparison between APD patients switching or not switching to CAPD during peritonitis treatment would be a better study design but, because of our treatment protocol and the retrospective nature of our study, these data were not available. Also, vintage bias due to the long study period cannot be excluded. Despite the long study period, there were relatively few events. This makes it hard to confidently exclude differences in outcomes and to correct for all possible confounders as it limits the number of covariates incorporated in the multivariate regression model. However, this study is one of the largest studies comparing outcome in APD- and CAPD-related peritonitis and is, we think, as we did correct for demographic differences, the best available evidence on the subject. Ideally, a properly designed, well-powered randomized 45

8 RÜGER et al. JANUARY 2011 VOL. 31, NO. 1 PDI study comparing patients with APD-related peritonitis that do switch to CAPD with patients that continue APD during peritonitis treatment should be done. DISCLOSURES The authors do not have any conflicts of interest to declare. REFERENCES 1. Vonesh EF, Snyder JJ, Foley RN, Collins AJ. Mortality studies comparing peritoneal dialysis and hemodialysis: what do they tell us? Kidney Int Suppl 2006; (103):S Sanabria M, Munoz J, Trillos C, Hernandez G, Latorre C, Diaz CS, et al. Dialysis Outcomes in Colombia (DOC) study: a comparison of patient survival on peritoneal dialysis vs hemodialysis in Colombia. Kidney Int Suppl 2008; (108): S Johnson DW, Dent H, Hawley CM, McDonald SP, Rosman JB, Brown FG, et al. Associations of dialysis modality and infectious mortality in incident dialysis patients in Australia and New Zealand. Am J Kidney Dis 2009; 53: van Biesen W, Veys N, Vanholder R, Lameire N. Peritonealdialysis-related peritonitis: the art of rope-dancing. Nephrol Dial Transplant 2002; 17: Strippoli GF, Tong A, Johnson D, Schena FP, Craig JC. Catheter-related interventions to prevent peritonitis in peritoneal dialysis: a systematic review of randomized, controlled trials. J Am Soc Nephrol 2004; 15: Diaz-Buxo JA, Crawford TL, Bailie GR. Peritonitis in automated peritoneal dialysis: antibiotic therapy and pharmacokinetics. Perit Dial Int 2001; 21(Suppl 3):S Mahale AS, Katyal A, Khanna R. Complications of peritoneal dialysis related to increased intra-abdominal pressure. Adv Perit Dial 2003; 19: U.S. Renal Data System. USRDS 2008 Annual Data Report: Atlas of Chronic Kidney Disease and End-Stage Renal Disease in the United States. Bethesda, MD: National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases; Johnson D, Chang SH, Excell L, Livingston B, Bannister K, McDonald SP. Peritoneal Dialysis: ANZDATA Registry Report Adelaide, South Australia: Australia and New Zealand Dialysis and Transplant Registry; 2008: Manley HJ, Bailie GR. Treatment of peritonitis in APD: pharmacokinetic principles. Semin Dial 2002; 15: Manley HJ, Bailie GR, Frye RF, McGoldrick MD. Intravenous vancomycin pharmacokinetics in automated peritoneal dialysis patients. Perit Dial Int 2001; 21: Fried L, Piraino B. Peritonitis. In: Gokal R, Khanna R, Krediet R, Nolph KD, eds. Textbook of Peritoneal Dialysis. 2nd ed. London: Kluwer Academic Publishers; 2000: 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: Brulez HF, Moncasi EP, Posthuma N, Choy K, ter Wee PM. The efficacy of intraperitoneally administered gentamicin and rifampin as initial treatment of peritoneal dialysis-related peritonitis. Adv Perit Dial 1995; 11: Baker RJ, Senior H, Clemenger M, Brown EA. Empirical aminoglycosides for peritonitis do not affect residual renal function. Am J Kidney Dis 2003; 41: Mars RL, Moles K, Pope K, Hargrove P. Use of bolus intraperitoneal aminoglycosides for treating peritonitis in endstage renal disease patients receiving continuous ambulatory peritoneal dialysis and continuous cycling peritoneal dialysis. Adv Perit Dial 2000; 16: de Fijter CW, ter Wee PM, Oe LP, Verbrugh HA. Intraperitoneal ciprofloxacin and rifampicin versus cephradine as initial treatment of (C)APD-related peritonitis: a prospective randomized multicenter comparison (CIPPER trial). Perit Dial Int 2001; 21: Posthuma N, Borgstein PJ, Eijsbouts Q, ter Wee PM. Simultaneous peritoneal dialysis catheter insertion and removal in catheter-related infections without interruption of peritoneal dialysis. Nephrol Dial Transplant 1998; 13: Yishak A, Bernardini J, Fried L, Piraino B. The outcome of peritonitis in patients on automated peritoneal dialysis. Adv Perit Dial 2001; 17: Troidle L, Gorban-Brennan N, Kliger A, Finkelstein F. Oncedaily intraperitoneal cefazolin and oral ciprofloxacin as empiric therapy for the treatment of peritonitis. Adv Perit Dial 1999; 15: Kavanagh D, Prescott GJ, Mactier RA. Peritoneal dialysisassociated peritonitis in Scotland ( ). Nephrol Dial Transplant 2004; 19: Rodriguez-Carmona A, Perez FM, Garcia FT, Fernandez RC, Valdes F. A comparative analysis on the incidence of peritonitis and exit-site infection in CAPD and automated peritoneal dialysis. Perit Dial Int 1999; 19: Blunden M, Zeitlin D, Ashman N, Fan SL. Single UK centre experience on the treatment of PD peritonitis antibiotic levels and outcomes. Nephrol Dial Transplant 2007; 22: Goldberg L, Clemenger M, Azadian B, Brown EA. Initial treatment of peritoneal dialysis peritonitis without vancomycin with a once-daily cefazolin-based regimen. Am J Kidney Dis 2001; 37: Fielding RE, Clemenger M, Goldberg L, Brown EA. Treatment and outcome of peritonitis in automated peritoneal dialysis, using a once-daily cefazolin-based regimen. Perit Dial Int 2002; 22: Steele M, Kwan JT. Potential problem: delayed detection of peritonitis by patients receiving home automated peritoneal dialysis (APD). Perit Dial Int 1997; 17: Brunkhorst RR. Host defenses in APD. Semin Dial 2002; 15:

9 PDI JANUARY 2011 VOL. 31, NO. 1 CONTINUATION OF PD MODALITY DURING PERITONITIS TREATMENT 28. de Fijter CW, Verbrugh HA, Oe LP, Peters ED, van der Meulen J, Donker AJ, et al. Peritoneal defense in continuous ambulatory versus continuous cyclic peritoneal dialysis. Kidney Int 1992; 42: Szeto CC, Kwan BC, Chow KM, Lau MF, Law MC, Chung KY, et al. Coagulase negative staphylococcal peritonitis in peritoneal dialysis patients: review of 232 consecutive cases. Clin J Am Soc Nephrol 2008; 3: Mujais S. Microbiology and outcomes of peritonitis in North America. Kidney Int Suppl 2006; (103):S Davies SJ, Bryan J, Phillips L, Russell GI. Longitudinal changes in peritoneal kinetics: the effects of peritoneal dialysis and peritonitis. Nephrol Dial Transplant 1996; 11: Selgas R, Fernandez-Reyes MJ, Bosque E, Bajo MA, Borrego F, Jimenez C, et al. Functional longevity of the human peritoneum: how long is continuous peritoneal dialysis possible? Results of a prospective medium longterm study. Am J Kidney Dis 1994; 23: Perez FM, Rodriguez-Carmona A, Garcia-Naveiro R, Rosales M, Villaverde P, Valdes F. Peritonitis-related mortality in patients undergoing chronic peritoneal dialysis. Perit Dial Int 2005; 25: Fried LF, Bernardini J, Johnston JR, Piraino B. Peritonitis influences mortality in peritoneal dialysis patients. J Am Soc Nephrol 1996; 7: