Peritonitis is a serious complication of peritoneal dialysis

Peritoneal Dialysis International, Vol. 30, pp. 311 319 doi: 10.3747/pdi.2008.00258 0896-8608/10 $3.00 +.00 Copyright 2010 International Society for Peritoneal Dialysis STAPHYLOCOCCUS AUREUS PERITONITIS IN AUSTRALIAN PERITONEAL DIALYSIS PATIENTS: PREDICTORS, TREATMENT, AND OUTCOMES IN 503 CASES Sridevi Govindarajulu, 1,2 Carmel Mary Hawley, 1,2 Stephen P. McDonald, 1,3 Fiona G. Brown, 1,4 Johan B. Rosman, 1,5 Kathryn J. Wiggins, 1,6 Kym M. Bannister, 1,7 and David W. Johnson 1,2 Australia and New Zealand Dialysis and Transplant Registry, 1 Adelaide; Department of Renal Medicine, 2 University of Queensland at Princess Alexandra Hospital, Brisbane; Department of Nephrology & Transplantation Services, 3 University of Adelaide at the Queen Elizabeth Hospital, Adelaide; Department of Nephrology, 4 Monash Medical Centre, Clayton, Victoria, Australia; Renal Department, 5 Middlemore Hospital, Otahuhu, Auckland, New Zealand; Department of Medicine, 6 University of Melbourne, St Vincent s Hospital, Fitzroy, Victoria; Department of Nephrology, 7 Royal Adelaide Hospital, Adelaide, Australia Correspondence to: D.W. Johnson, Department of Nephrology, Level 2, ARTS Building, Princess Alexandra Hospital, Ipswich Road, Woolloongabba, Brisbane Qld 4102, Australia. david_johnson@health.qld.gov.au Received 12 November 2008; accepted 1 July 2009. Staphylococcus aureus peritonitis is a serious complication of peritoneal dialysis (PD). Since reports of the course and treatment of S. aureus peritonitis have generally been limited to small, single-center studies, the aim of the current investigation was to examine the frequency, predictors, treatment, and clinical outcomes of this condition in all 4675 patients receiving PD in Australia between 1 October 2003 and 31 December 2006. 3594 episodes of peritonitis occurred in 1984 patients and 503 (14%) episodes of S. aureus peritonitis occurred in 355 (8%) individuals. 273 (77%) patients experienced 1 episode of S. aureus peritonitis, 52 (15%) experienced 2 episodes, 19 (5%) experienced 3 episodes, and 11 (3%) experienced 4 or more episodes. The predominant antibiotics used as initial empiric therapy were vancomycin (61%) and cephazolin (31%). Once S. aureus was isolated and identified, the prescription of vancomycin did not appreciably change for methicillin-sensitive S. aureus (MSSA) peritonitis (59%) and increased for methicillin-resistant S. aureus (MRSA) peritonitis (84%). S. aureus peritonitis was associated with a higher rate of relapse than non-s. aureus peritonitis (20% vs 13%, p < 0.001) but comparable rates of hospitalization (67% vs 70%, p = 0.2), catheter removal (23% vs 21%, p = 0.4), hemodialysis transfer (18% vs 18%, p = 0.6), and death (2.2% vs 2.3%, p = 0.9). MRSA peritonitis was independently predictive of an increased risk of permanent hemodialysis transfer [odds ratio (OR) 2.11, 95% confidence interval (CI) 1.17 3.82] and tended to be associated with an increased risk of hospitalization (OR 2.00, 95% CI 0.96 4.19). The initial empiric antibiotic choice between vancomycin and cephazolin was not significantly associated with clinical outcomes, but serious adverse outcomes were more likely if vancomycin was not used for subsequent treatment of MRSA peritonitis. In conclusion, S. aureus peritonitis is a serious complication of PD, involves a small proportion of patients, and is associated with a high rate of relapse and repeat episodes. Other adverse clinical outcomes are similar to those for peritonitis overall but are significantly worse for MRSA peritonitis. Empiric initial therapy with either vancomycin or cephazolin results in comparable outcomes, provided vancomycin is prescribed when MRSA is isolated and identified. Perit Dial Int 2010; 30:311 319 epub ahead of print: 26 Feb 2010 www.pdiconnect.com doi: 10.3747/pdi.2008.00258 KEY WORDS: Antibiotics; bacteria; fungi; microbiology; peritonitis; outcomes. Peritonitis is a serious complication of peritoneal dialysis (PD), accounting for 30% of technique failures and 21% of infectious deaths in Australian and New Zealand PD patients (1). A recent study comparing infectious mortality between hemodialysis (HD) and PD reported that PD was associated with a 66% increased risk of death from infection and that this excess risk of death was primarily accounted for by bacterial and fungal peritonitis (2). The most common microbial causes of PD-associated peritonitis are gram-positive organisms, particularly coagulase-negative staphylococci (3,4). Staphylococcus aureus accounts for a smaller proportion of peritonitis episodes (12% 20%) but is overrepresented in the more severe forms of peritonitis leading to hospitalization and 311

GOVINDARAJULU et al. MAY 2010 VOL. 30, NO. 3 PDI catheter removal (5 8). The International Society for Peritoneal Dialysis (ISPD) peritonitis guidelines recommend that PD catheter-related peritonitis due to S. aureus is unlikely to respond to antibiotic therapy without catheter removal (9). There is also a high risk of repeat peritonitis during the first 6 months after treatment of S. aureus peritonitis (10,11). However, the clinical course and treatment of S. aureus peritonitis and the outcomes for methicillin-resistant versus methicillinsensitive organisms remain unclear because most available studies are based primarily on older small clinical studies involving single centers (often where PD expertise is concentrated) (5 8,10,12). The aim of the current study was to examine the frequency, predictors, treatment, and clinical outcomes of S. aureus peritonitis in all Australian PD patients. Subgroup comparisons were also made between peritonitis episodes due to methicillin-sensitive S. aureus (MSSA) and methicillin-resistant S. aureus (MRSA). PATIENTS AND METHODS STUDY POPULATION The study included all Australian adult patients from the ANZDATA Registry who were receiving PD between 1 October 2003 (when detailed peritonitis data started to be collected) and 31 December 2006. The data collected included demographic data, cause of primary renal disease, comorbidities at the start of dialysis (coronary artery disease, peripheral vascular disease, cerebrovascular disease, chronic lung disease, diabetes, hypertension, and smoking status), body mass index, late referral (defined as commencement of dialysis within 3 months of referral to a nephrologist), microbiology of peritonitis episodes (up to 3 organisms for polymicrobial episodes), and initial and subsequent antibiotic treatment regimens. In cases of polymicrobial peritonitis, S. aureus peritonitis was recorded if S. aureus was at least one of the isolated organisms. Center size was categorized according to quartiles of the numbers of patients cared for by individual units over the duration of the study: small (<11 patients), small medium (11 38 patients), medium large (39 98 patients), and large (>99 patients). The outcomes examined were peritonitis relapse, repeat peritonitis, peritonitis-associated hospitalization, catheter removal, temporary or permanent transfer to HD, and patient death. Peritonitis relapse was defined as an episode of peritonitis occurring within 4 weeks of the last antibiotic dose (or within 5 weeks if intermittent vancomycin was used) for peritonitis due to the 312 same organism. Relapsed peritonitis was not recorded as a separate peritonitis episode on ANZDATA (i.e., it was counted as a continuation of the preceding episode). Repeat peritonitis was defined as an episode of peritonitis occurring more than 4 weeks after the last antibiotic dose (or more than 5 weeks if intermittent vancomycin was used) for peritonitis due to the same organism. Repeat peritonitis was recorded as a separate peritonitis episode on the ANZDATA Registry. Peritonitis-related death was recorded if the patient s death was directly attributable to peritonitis in the clinical opinion of the treating nephrologist. STATISTICAL ANALYSIS Results are expressed as frequencies and percentages for categorical variables, mean ± standard deviation for continuous variables, and median and interquartile range (IQR) for nonparametric data. Differences between two groups of patients were analyzed by chi-square test for categorical data, unpaired t-test for continuous parametric data, and Mann Whitney test for continuous nonparametric data. The independent predictors of S. aureus peritonitis were determined by multivariate Poisson regression using backward stepwise elimination (13). To account for the structure of the data (specifically for the fact that outcomes of multiple peritonitis episodes within the same patient are correlated), a multilevel hierarchical model was created with a random effect for state of residence, treating unit, and individual patient. Predictors of outcomes of peritonitis (including hospitalization, relapse, and permanent HD transfer) were determined by multivariate binary logistic regression, utilizing the hierarchical model incorporating random effects as for the Poisson models. First-order interaction terms between the significant covariates were examined for all analyses. Data were analyzed using the software packages SPSS release 12.0 (SPSS Inc., North Sydney, Australia) and Stata/SE 10.0 (College Station, TX, USA). A p value less than 0.05 was considered statistically significant. RESULTS POPULATION CHARACTERISTICS A total of 4675 patients received PD in Australia during the study period (1 October 2003 to 31 December 2006). In this group, 3594 episodes of peritonitis occurred in 1984 (42%) patients; 503 episodes of S. aureus peritonitis occurred in 355 individuals, accounting for 14% of all peritonitis episodes in 8% of all patients. The

PDI MAY 2010 VOL. 30, NO. 3 STAPHYLOCOCCAL PERITONITIS rates of all peritonitis and S. aureus peritonitis were 0.60 and 0.08 episodes per patient-year of treatment. Additional organisms were isolated in 39 (8%) episodes of S. aureus peritonitis, including coagulase-negative staphylococci (n = 7), streptococci (n = 6), enterococci (n = 3), other gram-positive organisms (n = 2), pseudomonas (n = 2), Acinetobacter (n = 2), E. coli (n = 2), Klebsiella (n = 2), Proteus (n = 2), Enterobacter (n = 5), other gram-negative organisms (n = 5), and fungi (n = 5). PREDICTORS OF S. AUREUS PERITONITIS The characteristics of patients that did and did not experience S. aureus peritonitis are shown in Table 1. On univariate analysis, patients that experienced S. aureus peritonitis during the study period were more likely to be referred late (within 3 months of dialysis commencement), be current smokers, and have diabetic nephropathy, peripheral vascular disease, and baseline PET results available than those individuals that did not experience S. aureus peritonitis. On multivariate analysis of all peritonitis episodes using a multilevel hierarchical Poisson regression model, S. aureus peritonitis was significantly and independently predicted by younger age [first quartile: reference; second quartile: adjusted odds ratio (OR) 0.69, 95% confidence interval (CI) 0.27 1.73; third quartile: OR 0.27, 95% CI 0.09 0.86; fourth quartile: OR 0.13, 95% CI 0.03 0.54). The development of S. aureus peritonitis was not associated with gender, racial origin, body mass index, kidney function at dialysis commencement, late referral, smoking status, peritoneal transport status, chronic lung disease, coronary artery disease, peripheral vascular disease, cerebrovascular disease, diabetes mellitus, cause of end-stage renal failure, center size, or state where treated. TREATMENT OF S. AUREUS PERITONITIS TABLE 1 Characteristics of All Australian Peritoneal Dialysis (PD) Patients That Did and Did Not Experience Staphylococcus aureus Peritonitis (SAP) at Any Stage During the Period 2003 2006 SAP Non-SAP Characteristic (n=355) (n=4320) p Value Age (years) 60.9±16.0 61.6±16.8 0.5 Women 160 (45%) 1966 (46%) 0.9 White 263 (74%) 3318 (77%) 0.3 BMI (kg/m 2 ) 26.0±5.0 25.9±6.5 0.7 egfr (ml/min/1.73 m 2 ) 6.8±3.8 7.1±4.5 0.5 Late referral 101 (28%) 1010 (23%) 0.03 ESRF cause 0.03 CGN 86 (24%) 1273 (29%) Diabetic nephropathy 128 (36%) 1191 (28%) Renovascular disease 46 (13%) 593 (14%) Polycystic kidneys 13 (4%) 243 (6%) Reflux nephropathy 16 (5%) 181 (4%) Other 43 (12%) 602 (14%) Unknown 23 (6%) 272 (6%) Current smoker 55 (15%) 502 (12%) 0.03 Chronic lung disease 52 (15%) 547 (13%) 0.3 Coronary artery disease 134 (38%) 1533 (35%) 0.4 Peripheral vascular disease 95 (27%) 951 (22%) 0.04 Cerebrovascular disease 53 (15%) 555 (13%) 0.3 Diabetes mellitus 158 (45%) 1580 (37%) 0.003 Peritoneal transport status 0.01 High 45 (13%) 426 (10%) High average 128 (36%) 1583 (37%) Low average 98 (28%) 980 (23%) Low 15 (4%) 183 (4%) Unknown/not specified 69 (19%) 1148 (27%) Center size (no. PD patients) 0.4 Small ( 10) 5 (1%) 50 (1%) Small medium (11 38) 20 (6%) 301 (7%) Medium large (39 98) 69 (19%) 959 (22%) Large ( 99) 261 (74%) 3010 (70%) State 0.07 New South Wales 158 (45%) 1674 (39%) Northern Territory 8 (2%) 77 (2%) Queensland 60 (17%) 895 (21%) South Australia 15 (4%) 280 (6%) Tasmania 3 (1%) 74 (2%) Victoria 68 (19%) 892 (21%) Western Australia 43 (12%) 428 (10%) BMI = body mass index; egfr = effective glomerular filtration rate at dialysis start; ESRF = end-stage renal failure; CGN = chronic glomerulonephritis. The vast majority of patients with PD-associated peritonitis were treated with either intraperitoneal vancomycin or cephazolin in combination with gentamicin (Table 2). Episodes of S. aureus peritonitis were more likely than other forms of peritonitis to be treated initially with vancomycin rather than cephazolin prior to knowledge of the dialysate effluent culture results. In 19 (4%) S. aureus peritonitis cases, vancomycin and cephazolin were used in combination from the outset. The initial antibiotic regimen was changed in 256 (51%) S. aureus peritonitis episodes after a median period of 3 days, with the most common changes being substitution of vancomycin for cephazolin and cessation of gentamicin. Oral antibiotics were used in a small minority of patients, with the most common choice being cephalexin followed by isoxazolyl penicillins (flucloxacillin, diclox- 313

GOVINDARAJULU et al. MAY 2010 VOL. 30, NO. 3 PDI TABLE 2 Antimicrobial Agents Prescribed in Initial, Second, and Third Antibiotic Regimens for Staphylococcus aureus Peritonitis Episodes (n=503) in Australian Peritoneal Dialysis Patients 2003 2006 1st 2nd 3rd regimen regimen regimen Antibiotic (n=503) (n=256) (n=68) Cephazolin 158 (31%) 25 (10%) 5 (7%) Vancomycin 309 (61%) 144 (56%) 26 (38%) Gentamicin 335 (67%) 21 (8%) 2 (3%) Ciprofloxacin 18 (4%) 7 (3%) 2 (3%) Cefoxitin 9 (2%) 0 (0%) 0 (0%) Ceftazidime 25 (5%) 2 (0.8%) 0 (0%) Ceftriaxone 17 (3%) 6 (2%) 0 (0%) Cephalothin 35 (7%) 10 (4%) 2 (3%) Cephalexin 14 (3%) 22 (9%) 11 (16%) Other cephalosporin 1 (0.2%) 1 (0.4%) 0 (0%) Flucloxacillin/dicloxacillin/ cloxacillin 13 (3%) 4 (2%) 15 (22%) Amoxicillin±clavulanate 10 (2%) 1 (0.4%) 11 (16%) Rifampicin 2 (0.4%) 4 (2%) 3 (4%) Teicoplanin 1 (0.1%) 0 (0%) 0 (0%) Amphotericin/fluconazole 4 (0.8%) 4 (2%) 3 (4%) Other a 13 (3%) 36 (14%) 3 (4%) a Other includes amikacin, ampicillin, azlocillin, aztreonam, clindamycin, erythromycin, fleroxacin, imipenem, isoniazid, ketoconazole, metronidazole, miconazole, moxalactam, netilmicin, ofloxacin, piperacillin, pyrazinamide, sulfamethoxazole/trimethoprim, ticarcillin, tobramycin, and unknown. Results represent number of episodes treated with antibiotic (% of total treated with 1st-, 2nd-, or 3rd-line regimen). Note that values within each column add to more than 100% because of the use of combination antimicrobial regimens. acillin, or cloxacillin), and amoxicillin with or without clavulanic acid. These oral antibiotics were mostly introduced as a third antimicrobial regimen after a median treatment period of 7 days (Table 3). Overall, the median total antibiotic course duration for S. aureus peritonitis was 14 days. In terms of adjunctive therapy, rifampicin was added to the treatment regimen in 9 (1.8%) cases of S. aureus peritonitis and 8 (0.3%) cases of non-s. aureus peritonitis (p < 0.001). Tenckhoff catheter instillation with either urokinase or streptokinase was performed in 3 (0.6%) cases of S. aureus peritonitis and 27 (0.9%) cases of non-s. aureus peritonitis (p = 0.5). Heparin was administered to dialysate in 75 (15%) episodes of S. aureus peritonitis and 665 (22%) episodes of non- S. aureus peritonitis (p = 0.001). Antifungal chemoprophylaxis (principally oral nystatin) was co-prescribed 314 in 39 (8%) cases of S. aureus peritonitis and 221 (7%) cases of non-s. aureus peritonitis (p = 0.6). OUTCOMES OF S. AUREUS PERITONITIS Of the 355 individuals that experienced S. aureus peritonitis during the study period, 82 (23%) experienced repeat peritonitis, 112 (32%) underwent catheter removal (4 patients had catheter removals on 2 separate occasions), and 11 (2.2% ) patients died. With respect to repeat peritonitis, 273 (77%) patients experienced 1 episode, 52 (15%) experienced 2 episodes, 19 (5%) experienced 3 episodes, and 11 (3%) experienced 4 or more episodes. The median (IQR) time period between repeat S. aureus peritonitis episodes [62 (38 124) days] tended to be shorter than between episodes of non-s. aureus peritonitis [77 (34 182) days], although this difference was not statistically significant (p = 0.10). Of the 26 patients that had their catheters removed for S. aureus peritonitis and subsequently returned to PD with new catheters, 4 (15%) experienced repeat episodes of peritonitis after a median (IQR) period of 127 (10 700) days (p = 0.9 vs patients that did not have their catheters removed). Relapse occurred in 100 (20%) episodes in 62 (17%) patients and was significantly more frequent than relapse following non-s. aureus peritonitis episodes (402 or 13%, p < 0.001; Table 3) in 198 (12%) patients (p = 0.009). Patients that experienced a relapse of S. aureus peritonitis were significantly more likely to have their catheters removed than patients that did not experience a relapse (33% vs 21%, p = 0.008; OR 1.90, 95% CI 1.17 3.07). Using a multilevel hierarchical logistic regression model, an increased risk of S. aureus peritonitis relapse was significantly associated with the presence of peripheral vascular disease (adjusted OR 2.75, 95% CI 1.43 5.29) and the use of vancomycin compared with cephalosporins (OR 2.62, 95% CI 1.32 5.21). A lower risk of relapse was independently predicted by female gender (OR 0.43, 95% CI 0.24 0.79) and the middle tertile of age (OR 0.49, 95% CI 0.25 0.95). Hospitalization occurred in similar proportions of patients with S. aureus and non-s. aureus peritonitis, although the hospital stay for S. aureus peritonitis was significantly longer by a median period of 1 day (Table 3). The significant independent predictors of hospitalization were female gender (OR 2.97, 95% CI 1.58 5.56) and Aboriginal and Torres Strait Islanders (OR 0.34, 95% CI 0.13 0.91). Isolation of MRSA also tended to be associated with an increased risk of hospitalization (OR 2.00, 95% CI 0.96 4.19).

PDI MAY 2010 VOL. 30, NO. 3 STAPHYLOCOCCAL PERITONITIS TABLE 3 Treatment Characteristics and Clinical Outcomes of Peritoneal Dialysis-Associated Peritonitis Due to Staphylococcus aureus or Other Organisms in Australia 2003 2006 S. aureus peritonitis Non-S. aureus peritonitis Outcome (n=503 episodes) (n=3091 episodes) p Value Treatment Change to 2nd antibiotic regimen 256 (51%) 1754 (57%) 0.002 Time to 2nd antibiotic regimen 3 (2 7) 3 (2 5) <0.001 Change to 3rd antibiotic regimen 68 (13%) 429 (14%) 0.8 Time to 3rd antibiotic regimen 7 (5 11) 6 (4 10) <0.05 Total antibiotic treatment duration 14 (9 21) 14 (8 19) 0.06 Peritonitis relapse 100 (20%) 402 (13%) <0.001 Hospitalization [n (%)] 338 (67%) 2166 (70%) 0.2 Duration 7 (4 14) 6 (3 11) <0.001 Catheter removal [n (%)] 116 (23%) 659 (21%) 0.4 Time to occurrence 7.5 (3 15) 6 (3 12) 0.2 Temporary hemodialysis [n (%)] 20 (4%) 132 (4%) 0.8 Time to occurrence 9 (4 30) 6 (3 11) 0.08 Duration 49 (12 112) 68 (29 103) 0.4 Permanent hemodialysis [n (%)] 93 (18%) 542 (18%) 0.6 Time to occurrence 8 (4 13) 7 (4 12) 0.4 Death [n (%)] 11 (2.2%) 71 (2.3%) 0.9 Time to death 16 (2.5 48.5) 11 (4.25 21.75) 0.5 Results are expressed as number (%) or median (interquartile range) days. A p value less than 0.05 was considered statistically significant. There were no significant differences between S. aureus and other forms of peritonitis with respect to catheter removal (frequency or timing), HD transfer (frequency, timing, or duration), or death (frequency or timing) (Table 3). Permanent HD transfer was independently predicted by isolation of MRSA (OR 2.11, 95% CI 1.17 3.82). The low number of deaths precluded meaningful analysis by multivariate regression. Overall, 314 (62%) episodes of S. aureus peritonitis in 211 (59%) patients were successfully treated with antibiotics without experiencing relapse, catheter removal, or death. Of these 211 patients, 69 (33%) experienced repeat episodes of S. aureus peritonitis with the same catheter. EFFECT OF TIMING OF CATHETER REMOVAL IN S. AUREUS PERITONITIS Tenckhoff catheter removal occurred in 116 of the 503 episodes of S. aureus peritonitis after a median period of 7.5 days. When outcomes were compared between early (within 5 days) and late (after 5 days) catheter removal, no significant differences were observed with respect to permanent HD transfer (79% vs 78%, p = 0.9) or death (4% vs 3%, p = 0.7). MSSA AND MRSA PERITONITIS Methicillin-sensitive S. aureus (MSSA) caused 394 (78%) peritonitis episodes, while 109 (22%) were due to MRSA. Treatment characteristics and outcomes of MRSA and MSSA peritonitis are shown in Table 4. The initial empiric treatment of MSSA and MRSA peritonitis was similar, with approximately two thirds of patients receiving vancomycin (Table 4). Following the availability of culture results, the administration of vancomycin did not change appreciably for MSSA peritonitis episodes, but increased significantly to 84% for MRSA peritonitis. One patient with MRSA peritonitis received teicoplanin, leaving 15% of MRSA peritonitis episodes treated with antibiotics other than glycopeptides. The most common combination of these other antibiotics was cephazolin and gentamicin (46%), followed by ceftriaxone (15%), cephazolin alone (8%), cephazolin and cephalexin (8%), amoxicillin and clavulanate (8%), cephadrine and piperacillin (8%), ceftazidime and cephazolin, and an unknown antibiotic (8%). Relapse rates for MRSA and MSSA peritonitis were comparable. However, MRSA peritonitis was generally more severe than MSSA peritonitis, as evidenced by a significantly greater frequency and duration of hospitalization 315

GOVINDARAJULU et al. MAY 2010 VOL. 30, NO. 3 PDI TABLE 4 Treatment Characteristics and Clinical Outcomes of Peritoneal Dialysis-Associated Peritonitis Due to Methicillin-Sensitive Staphylococcus aureus (MSSA) or Methicillin-Resistant Staphylococcus aureus (MRSA) in Australia 2003 2006 Outcome MSSA (n=394 episodes) MRSA (n=109 episodes) p Value Treatment Empiric vancomycin 240 (61%) 71 (65%) 0.5 Vancomycin prescribed after culture results 233 (59%) 92 (84%) <0.001 Total antibiotic treatment duration 14 (8 21) 14 (10 23) 0.3 Peritonitis relapse 79 (20%) 21 (19%) 0.9 Peritonitis repeat 122 (31%) 26 (24%) 0.2 Hospitalization [n (%)] 256 (65%) 82 (75%) <0.05 Duration 7 (4 13) 9 (5 17) <0.05 Catheter removal [n (%)] 82 (21%) 34 (31%) <0.05 Time to occurrence 7 (3 15) 10 (3 17) 0.6 Temporary hemodialysis [n (%)] 14 (4%) 6 (6%) 0.4 Time to occurrence 9 (4 36) 8 (4 24) 0.9 Duration 34 (10 98) 83 (21 122) 0.3 Permanent hemodialysis [n (%)] 66 (17%) 27 (25%) 0.06 Time to occurrence 7 (3 12) 11 (5 20) <0.05 Death [n (%)] 6 (2%) 5 (5%) 0.05 Time to death 9 (1 55) 22 (9 41) 0.5 Results are expressed as number (%) or median (interquartile range) days. and by increased frequencies of catheter removal, permanent HD transfer, and death (Table 4). Treatment of MRSA peritonitis episodes with empiric initial regimens that did not include vancomycin was not associated with significant increases in the rates of relapse (17% vs 23%), hospitalization (77% vs 72%), catheter removal (30% vs 32%), permanent HD transfer (13% vs 30%), or death (7% vs 4%). Failure to include vancomycin at any stage in the treatment of MRSA peritonitis was associated with generally higher rates of relapse (23% vs 20%), hospitalization (89% vs 72%), catheter removal (50% vs 28%), permanent HD transfer (36% vs 22%), and death (7% vs 4%), although none of these differences reached statistical significance. DISCUSSION The present study represents the largest examination to date of the frequency, predictors, treatment, and clinical outcomes of PD-associated S. aureus peritonitis. Staphylococcus aureus was found to account for 14% of all PD-related peritonitis episodes and was associated with higher rates of relapse (20%) and repeat episodes (29%) than cases of peritonitis due to other organisms. However, S. aureus and non-s. aureus peritonitis resulted in comparable rates and timing of hospitalization, catheter removal, HD transfer, and death, although the duration of hospitalization was slightly longer for S. aureus 316 peritonitis than for non-s. aureus peritonitis. Only 8% of patients experienced S. aureus peritonitis during the study period. In S. aureus peritonitis episodes where the catheter was removed, the rates of permanent HD transfer and death were comparable whether catheters were removed early (within the first 5 days) or late. MRSA peritonitis was generally more severe than MSSA peritonitis, as evidenced by a significantly greater frequency and duration of hospitalization and by increased frequencies of catheter removal, HD transfer, and death. The initial empiric antibiotic choice between vancomycin and cephazolin did not generally influence the clinical outcomes of S. aureus peritonitis, but serious adverse outcomes were more likely if vancomycin was not included in the antibiotic regimen at any stage for the treatment of MRSA peritonitis. The results of our investigation are in keeping with previous findings by Szeto and colleagues (10), who reported 245 episodes of S. aureus peritonitis in 152 patients at a single center in Hong Kong between 1994 and 2005. Similar to our study, Szeto et al. observed that S. aureus peritonitis accounted for 11.9% of all peritonitis episodes, occurred at a rate of 0.072 episodes per year, and was caused by MRSA on 18.4% of occasions. The investigators also observed that the initial antibiotic regimen (cephazolin vs vancomycin) had no significant effect on the risk for relapse or repeat peritonitis and that MRSA peritonitis was generally more severe and re-

PDI MAY 2010 VOL. 30, NO. 3 STAPHYLOCOCCAL PERITONITIS sulted in more frequent need for hospitalization than MSSA peritonitis. However, in contrast to our findings, Szeto et al. observed that S. aureus peritonitis was associated with lower rates of relapsed (8.6%) and repeat (24%) peritonitis, lower rates of hospitalization overall (16% vs 67%), lower rates of catheter removal (5.7%), lower rates of HD transfer, and a slightly higher rate of death (4.9%). It is possible that the lower rates of S. aureus relapse in the Hong Kong Prince of Wales Hospital PD patient population compared with Australian PD patients reflect the longer antibiotic durations employed (21 vs 14 days) and the more frequent prescription of adjuvant rifampicin therapy. Szeto et al. reported that rifampicin therapy was associated with a 49.9% relative risk reduction in relapse or repeat S. aureus peritonitis, whereas, in the present study, rifampicin was associated with a 6.5-fold increased relative risk of peritonitis relapse. Since prescription of rifampicin was nonrandom in both studies, the apparent disparity in results most likely reflects differences in patient selection. Similarly, the lower rates of catheter removal and HD transfer in the Hong Kong population probably reflect the predominant (>80%) utilization of PD as a renal replacement therapy in that country and the fact that HD transfer occurred as a last resort, such that patients switched to HD only when they had ultrafiltration failure or peritoneal sclerosis. The clinical outcomes of S. aureus peritonitis in Australia were also generally comparable with those of the Network 9 (Indiana, Ohio, Kentucky, ESRD Tri-State Renal Network) study of 149 episodes of S. aureus PD peritonitis at 68 centers during 1991 (8). Hospitalization, catheter removal, HD transfer, and death occurred in 30%, 18%, 9%, and 4% of cases, respectively. These peritonitis outcomes were significantly worse than those of coagulase-negative staphylococci but better than those of non-pseudomonal gram-negative peritonitis. The predominant antibiotics used as initial empiric therapy for covering gram-positive organisms in PDassociated peritonitis were vancomycin (61%) and cephazolin (31%). Interestingly, once S. aureus was isolated and identified as either MSSA or MRSA, the prescription of vancomycin did not appreciably change in the MSSA group (59%) and increased only to 84% in the MRSA group. Although it is desirable to minimize the use of vancomycin to curtail the possible development of vancomycin-resistant enterococci (14), the continued prescription of vancomycin following identification of MSSA may have been due to the greater convenience of intermittent dosing with this agent as opposed to more frequent dosing with cephalosporins. Continued vancomycin prescription for treatment of S. aureus (predominantly MSSA) peritonitis was associated with a higher rate of relapse compared with cephalosporins, although this may have reflected the relatively higher rate of vancomycin treatment of S. aureus peritonitis in the ambulatory (non-hospital) setting compared with cephalosporin therapy. The ISPD guidelines on peritonitis treatment also recommend that MRSA peritonitis in PD patients must be treated with vancomycin (9). In the present study, treatment of MRSA peritonitis episodes with empiric regimens that did not include vancomycin were not associated with significant increases in adverse outcomes, provided vancomycin was added after MRSA culture results became known. However, failure to include vancomycin at any stage in the treatment of MRSA peritonitis was associated with increased rates of relapse, hospitalization, catheter removal, permanent HD transfer, and death. Our results therefore support the ISPD recommendations. It is also noteworthy that the median antibiotic treatment course for S. aureus peritonitis in Australia was only 14 days, despite ISPD recommendations that episodes should be treated with vancomycin or cephazolin, as appropriate, for 3 weeks (9). Since treatment duration was defined in the ANZDATA Registry as the time between the first and the last antibiotic dose received, it is likely that the effective treatment duration was longer than 14 days (especially for drugs such as vancomycin). On multivariate Poisson regression analysis, treatment course duration was not significantly related to any of the clinical outcome measures examined. Nevertheless, it is conceivable that the generally higher rates of outcomes in our study compared with those of Szeto and associates (10) may have reflected the shorter course of treatment in Australia. Another novel finding of the present investigation was that no clear evidence could be found for a center effect and specifically that center size (as a surrogate marker of PD expertise) was not predictive of clinical outcomes of S. aureus peritonitis. Center effects have been described for HD (15) and renal transplantation (16) but have received scant attention in PD. Further research in this field would seem warranted. Interestingly, younger age was the only independent predictor of S. aureus peritonitis in the present study. This finding is in keeping with community studies demonstrating that younger age is a strong independent predictor of throat and nasal carriage of S. aureus (17). The strengths of this study include its very large sample size and inclusiveness. We included all patients receiving PD in Australia during the study period, such that a variety of centers were included with varying approaches to the treatment of peritonitis. This greatly enhanced the 317

GOVINDARAJULU et al. MAY 2010 VOL. 30, NO. 3 PDI external validity of our findings. These strengths should be balanced against the study s limitations, which include limited depth of data collection. ANZDATA does not collect important information such as the presence of concomitant exit-site and tunnel infections, patient compliance, individual unit management protocols (including the use of prophylactic mupirocin), laboratory values (such as C-reactive protein and dialysate white cell counts), severity of comorbidities, antibiotic dosages, or routes of antibiotic administration. Although we adjusted for a large number of patient characteristics, the possibility of residual confounding could not be excluded. In common with other registries, ANZDATA is a voluntary registry and there is no external audit of data accuracy, including the diagnosis of peritonitis and the attribution of death to peritonitis. Consequently, the possibility of coding/classification bias cannot be excluded. Furthermore, the reasons underpinning antibiotic choices, changes, and treatment durations were not captured by ANZDATA, making it difficult to draw firm conclusions about the relationship between prescribed antibiotics and outcomes. For example, in the absence of information about laboratory sensitivities or early response to treatment, it is difficult to determine whether antibiotic changes were appropriate or otherwise, and whether they represented a step down or step up due to satisfactory or unsatisfactory progress respectively. Similarly, the surprising association of rifampicin therapy with peritonitis relapse may reflect predominant use of this agent in high risk situations, since it was primarily prescribed as a second- or third-line agent. In conclusion, S. aureus peritonitis is a common and serious complication of PD and is associated with a high rate of relapse (20%) and repeat episodes (29%). Adverse clinical outcomes, such as hospitalization, catheter removal, hemodialysis transfer, and death, are no more frequent with S. aureus peritonitis than with non- S. aureus peritonitis but are significantly worse with MRSA than with MSSA peritonitis. Empiric initial therapy with either vancomycin or cephazolin results in comparable outcomes, provided vancomycin is prescribed when MRSA is isolated and identified. Center size was not predictive of clinical results. DISCLOSURES David Johnson is a consultant for Baxter Healthcare Pty Ltd and has previously received research funds from this company. He has also received speakers honoraria and research grants from Fresenius Medical Care. Kym Bannister is a consultant for Baxter Healthcare Pty Ltd. Fiona Brown is a consultant for Baxter and Fresenius and 318 has received travel grants from Amgen and Roche. Stephen McDonald has received speaking honoraria from Amgen Australia, Fresenius Australia, and Solvay Pharmaceuticals and travel grants from Amgen Australia, Genzyme Australia, and Jansen-Cilag. The remaining authors have no competing financial interests to declare. ACKNOWLEDGMENTS The authors gratefully acknowledge the substantial contributions of the entire Australia and New Zealand nephrology community (physicians, surgeons, database managers, nurses, renal operators, and patients) in providing information for and maintaining the ANZDATA Registry database. REFERENCES 1. Johnson DW, Chang S, Excell L, Livingston B, Bannister K, McDonald SP. Peritoneal dialysis. In: McDonald SP, Excell L, eds. ANZDATA Registry Report 2006. Adelaide, South Australia: Australian and New Zealand Dialysis and Transplant Registry; 2007: 87 103. 2. 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; 53:290 7. 3. Szeto CC, Leung CB, Chow KM, Kwan BC, Law MC, Wang AY, et al. Change in bacterial aetiology of peritoneal dialysisrelated peritonitis over 10 years: experience from a centre in South-East Asia. Clin Microbiol Infect 2005; 11: 837 9. 4. Mujais S. Microbiology and outcomes of peritonitis in North America. Kidney Int Suppl 2006; 103:S55 62. 5. de Lourdes Ribeiro de Souza da Cunha M, Montelli AC, Fioravante AM, Neves Batalha JE, Teixeira Caramori JC, Barretti P. Predictive factors of outcome following staphylococcal peritonitis in continuous ambulatory peritoneal dialysis. Clin Nephrol 2005; 64:378 82. 6. Peacock SJ, Howe PA, Day NP, Crook DW, Winearls CG, Berendt AR. Outcome following staphylococcal peritonitis. Perit Dial Int 2000; 20:215 19. 7. Kim D, Tapson J, Wu G, Khanna R, Vas SI, Oreopoulos DG. Staph aureus peritonitis in patients on continuous ambulatory peritoneal dialysis. Trans Am Soc Artif Intern Organs 1984; 30:494 7. 8. Bunke CM, Brier ME, 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. 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. Szeto CC, Chow KM, Kwan BC, Law MC, Chung KY, Yu S, et al. Staphylococcus aureus peritonitis complicates peritoneal

PDI MAY 2010 VOL. 30, NO. 3 STAPHYLOCOCCAL PERITONITIS dialysis: review of 245 consecutive cases. Clin J Am Soc Nephrol 2007; 2:245 51. 11. Troidle L, Gorban-Brennan N, Kliger A, Finkelstein F. Differing outcomes of gram-positive and gram-negative peritonitis. Am J Kidney Dis 1998; 32:623 8. 12. Piraino B, Perlmutter JA, Holley JL, Bernardini J. Staphylococcus aureus peritonitis is associated with Staphylococcus aureus nasal carriage in peritoneal dialysis patients. Perit Dial Int 1993; 13(Suppl 2):S332 4. 13. McCullagh P, Nelder JA. Generalised Linear Models. London: Chapman and Hall; 1989. 14. Bartley PB, Schooneveldt JM, Looke DF, Morton A, Johnson DW, Nimmo GR. The relationship of a clonal outbreak of Enterococcus faecium vana to methicillin-resistant Staphylococcus aureus incidence in an Australian hospital. J Hosp Infect 2001; 48:43 54. 15. Pisoni RL, Greenwood RN. Selected lessons learned from the Dialysis Outcomes and Practice Patterns Study (DOPPS). Contrib Nephrol 2005; 149:58 68. 16. Briganti EM, Wolfe R, Russ GR, Eris JM, Walker RG, McNeil JJ. Graft loss following renal transplantation in Australia: is there a centre effect? Nephrol Dial Transplant 2002; 17:1099 104. 17. Mertz D, Frei R, Periat N, Zimmerli M, Battegay M, Fluckiger U, et al. Exclusive Staphylococcus aureus throat carriage: at-risk populations. Arch Intern Med 2009; 169:172 8. 319