Original Articles. Peritoneal dialysis outcomes after temporary haemodialysis transfer for peritonitis

Transcription

1 Nephrol Dial Transplant (2014) 29: doi: /ndt/gfu050 Advance Access publication 3 March 2014 Original Articles Peritoneal dialysis outcomes after temporary haemodialysis transfer for peritonitis Yeoungjee Cho 1,2, Sunil V. Badve 1,2, Carmel M. Hawley 1,2, Stephen P. McDonald 1,3, Fiona G. Brown 1,4, Neil Boudville 1,5, Philip Clayton 1,6 and David W. Johnson 1,2 1 Australia and New Zealand Dialysis and Transplant Registry, Adelaide, Australia, 2 Department of Renal Medicine, University of Queensland at Princess Alexandra Hospital, Brisbane, Australia, 3 Department of Nephrology and Transplantation Services, University of Adelaide at the Queen Elizabeth Hospital, Adelaide, Australia, 4 Department of Nephrology, Monash Medical Center, Clayton, Victoria, Australia, 5 School of Medicine and Pharmacology, University of Western Australia, Perth, Australia and 6 Department of Renal Medicine, Royal Prince Alfred Hospital, Sydney, Australia Correspondence and offprint requests to: David W. Johnson; david.johnson2@health.qld.gov.au ABSTRACT Background. There has not been a comprehensive examination to date of peritoneal dialysis (PD) outcomes after temporary haemodialysis (HD) transfer for peritonitis. Methods. The study included all incident Australian patients who experienced peritonitis between 1 October 2003, and 31 December 2011, using Australia and New Zealand Dialysis and Transplant Registry data. Patients were grouped into three categories: Interim HD, Permanent HD and Never HD based on HD transfer status after the first peritonitis. The independent predictors of HD transfer and subsequent return to PD were determined by multivariable, multilevel mixed-effects logistic regression analysis. Matched case control analyses were performed to compare clinical outcomes (e.g. patient survival) between groups. Results. Of the 3305 patients who experienced peritonitis during the study period, 553 episodes (16.7%) resulted in transfer to HD and 101 patients subsequently returned to PD. HD transfer was significantly and independently predicted by inpatient treatment of peritonitis [odds ratio (OR) 11.45, 95% confidence interval (CI) ] and the recovered microbiologic profile of organisms recognized to be associated with moderate (20 40%) to high (>40%) rates of catheter removal (moderate: OR 2.45, 95% CI ; high: OR 8.63, 95% CI ). Matched case control analyses yielded comparable results among Interim, Permanent and Never HD groups in terms of patient survival (P = 0.28), death-censored technique survival [hazard ratio (HR) 0.87, 95% CI ; P = 0.48] and peritonitis-free survival (HR 0.84, 95% CI , P = 0.49). Conclusions. In an observational registry study of first peritonitis episodes, temporary HD transfer was not associated with inferior patient-level clinical outcomes when compared with others who either never required HD transfer or remained on HD permanently if all patient-level and peritonitisrelated factors were considered equal. Therefore, return to PD after a temporary HD due to peritonitis should not be discouraged in appropriate PD patients. Keywords: haemodialysis, patient survival, peritoneal dialysis, peritonitis, technique survival INTRODUCTION Peritonitis is a serious complication of peritoneal dialysis (PD) responsible for significant morbidity, and accounts for 22% of technique failures and 14% of infectious deaths in Australian and New Zealand PD patients [1]. Approximately 20% of peritonitis episodes require cessation of PD and transfer to haemodialysis [2], although this figure varies depending on the causative organism [3 13]. While return to PD is a viable option for many patients transferring to HD for severe peritonitis, only a small minority (<20%) actually do so [2]. Studies of PD outcomes after temporary haemodialysis (HD) transfer for peritonitis have been limited. Troidle et al. The Author Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. 1940

2 [14] reported poor technique survival in those who return to PD after an episode of peritonitis, which resulted in catheter removal (n = 189; 20% 12-month technique survival). However, this study was limited by its single-centre, retrospective study design, lack of adjustment for potential confounding covariates [e.g. differences in dialysis vintage (study duration extended from 1990 to 2002), comorbid conditions] or comparison against matched controls. Moreover, due to restriction by their study design, the authors were unable to explicitly describe whether there were any patient-level or peritonitis-related factors that could assist clinicians to identify patients who should preferentially remain on HD or to return to PD after an episode of peritonitis, which resulted in a transfer to HD. In particular, there has not been a comprehensive examination to date of PD outcomes after temporary haemodialysis transfer for peritonitis across multiple centres. The aim of the current study was to examine the predictors and outcomes of return to PD following temporary HD transfer for first peritonitis in all incident Australian adult PD patients, as recorded in the ANZDATA registry. The second aim was to compare these outcomes in patients who permanently or temporarily transferred to haemodialysis following the first peritonitis and those who did not transfer to haemodialysis following first peritonitis using a propensity score matched case control analysis. SUBJECTS AND METHODS Study population The study included all Australian adult patients from the ANZDATA registry who experienced PD-related peritonitis between 1 October 2003 (when detailed peritonitis data started to be collected) and 31 December Collection and analysis of ANZDATA registry data were approved by the Princess Alexandra Hospital Human Research Ethic Committee. Permission to analyse the data was also granted by an ANZDATA executive. The analyses were performed on deidentified data extracted from ANZDATA. Only incident patients who commenced PD from 1 October 2003 were included to capture the first peritonitis episodes. Patients were grouped into three categories: Interim HD (PD patients who returned to PD after a time on HD postperitonitis), Permanent HD (PD patients who were recorded as remaining permanently on HD post-peritonitis), and Never HD (PD patients who experienced peritonitis but did not require a transfer to HD). Return to PD was defined as any episode of PD following HD, as reported by the individual treating unit on the ANZDATA peritonitis registry form. The data collected included demographic data, cause of primary renal disease, comorbidities at the start of dialysis, smoking status, body mass index (BMI), late referral (defined as commencement of dialysis within 3 months of referral to a nephrologist), microbiology of peritonitis episode, initial and subsequent antibiotic treatment regimens, and subsequent outcomes including the occurrence and timing of peritonitisassociated relapse, recurrence, hospitalization, catheter removal, temporary or permanent haemodialysis transfer and death. These data were collected by medical and nursing staff in each renal unit at the time of peritonitis and submitted to the ANZDATA registry. Peritonitis rates were calculated according to the standardized recommendations made by the International Society for PD [15]. The outcomes examined were technique survival, death-censored technique survival, peritonitis-free survival and patient survival. Statistical analysis Results were expressed as frequencies and percentages, mean ± standard deviation or median (25th 75th percentile), as appropriate. Differences between groups were analysed by χ 2 test for categorical data, unpaired t-test for continuous normally distributed data and Kruskal Wallis test for continuous non-normally distributed data. The independent predictors of HD transfer (temporary or permanent) and subsequent return to PD (versus permanent HD transfer) following an individual s first episode of peritonitis were determined by multivariable, multilevel mixed-effects logistic regression analyses. In order to account for the structure of the data, a multilevel hierarchical model was created with a random effect for treating unit and the peritonitis vintage according to the year of peritonitis occurrence (Category 1: ; Category 2: ; Category 3: ). For the purpose of statistical analysis, the microbiologic causes of peritonitis were categorized into three groups according to their averaged risk of catheter removal based on previous publications (catheter removal rate: low: <20%; medium: 20 40%; high: >40% or greater; Table 1)[3 13]. Overall survival curves (i.e. patient survival, technique survival and peritonitis-free survival) after return to PD in the Interim HD patients were generated according to the Kaplan Meier method. The independent predictors of technique survival (including death-censored), overall patient survival and peritonitis-free survival were determined by multivariate Cox proportional hazards model analysis using backward stepwise elimination to determine the most parsimonious model. Firstorder interaction terms between the significant covariates were examined, where appropriate. Adjusted survival curves were estimated using the Cox average covariate method, which calculates predicted survival probabilities at the mean levels of Table 1. Categories of peritonitis microbiology according to severity defined by catheter removal rate Severity index Catheter removal rate Microbiology Low <20% Coagulase-negative Staphylococci [9] Culture-negative [10] Streptococci [13] Moderate 20 40% Corynebacteria [7] Enterococci [8] Staphylococcus aureus (all) [11] Non-pseudomonas Gram-negative organisms [4] High >40% Polymicrobial [6] Fungi [5] Pseudomonas [3] PD after peritonitis 1941

3 the covariates. Technique survival was censored for renal transplantation, recovery of dialysis-independent renal function and end of study (31 December 2011). Peritonitis-free survival was censored for technique failure and end of study (31 December 2011). Patient death was regarded as a technique failure event, except where specified as death-censored technique failure. Overall patient survival was censored for renal transplantation, recovery of dialysis-independent renal function, permanent HD transfer and end of study (31 December 2011). In order to account for the risk of immortal time bias [16] in the Interim HD patients (i.e. Interim HD patients had to have survived long enough to return to PD compared with no HD or permanent HD patients), a matched case control analysis was performed whereby one Interim HD patient was matched with two Permanent HD patients or never HD patients. The pre-specified matching criteria were age (±5 years) and diabetic status. However, using these criteria, adequate matching was not feasible. Therefore, a propensity score (PS) was derived where PS was calculated using a multivariate ordinal logistic regression incorporating all available patientlevel covariates. Then, patients were stratified by hospitalization status (i.e. to indicate severity of peritonitis episode) to undergo 1:1 nearest neighbour match using the derived PS score. In order to mitigate against immortal time bias, only controls with PD duration greater than the time spent by matched cases to return to PD (i.e. time from peritonitis plus time on temporary HD) were selected for the analysis. Time-to-event analyses were performed where time zero was defined as the time of return to PD after the first peritonitis for each subject analysed (duration defined by the matched case subject). Patient survival was analysed in all matched patient groups, whereas only Interim HD and Never HD patients were included for technique survival (including deathcensored technique survival) and peritonitis-free survival. Data were analysed using the software package Stata/SE12.0 (College Station, TX). P-values <0.05 were considered statistically significant. RESULTS Of the incident PD patients who commenced PD from 1 October 2003, there were 3305 patients who experienced peritonitis during the study period (1 October December 2011). They were followed for 8023 patient-years (mean follow-up 2.43 years per patient), of which 6898 patient-years were spent on PD (mean PD duration 2.09 years per patient). Overall peritonitis rate in all Australian PD patients during the study period was 0.57 episodes per patient-year (95% CI ). The microbiologic causes of peritonitis were categorized into three groups according to severity defined catheter removal rate (Table 1). Temporary versus permanent HD transfer for severe peritonitis Of the 3305 episodes of peritonitis that occurred during the study period, 553 (16.7%) resulted in transfer to HD, either on a temporary (n = 101 or 3%) or permanent basis (n = 452 or 13.7%). Temporary HD transfer occurred after a median [interquartile range] period of 7 [3.5 10] days following peritonitis onset and was required for a median period of 69.5 [35 108] days. Predictors of return to PD following transfer to HD for severe peritonitis The characteristics of patients according to the requirement of HD transfer after the first episode of peritonitis are summarized in Table 2. Patients in the Interim HD group were significantly less likely to be females, whereas Permanent HD and Interim HD patients were significantly more likely to experience peritonitis caused by microbes known to lead to a higher risk of catheter removal (e.g. pseudomonas species, fungi) compared with Never HD patients. Furthermore, patients who ever required HD transfer were significantly more likely to undergo hospitalization as part of their peritonitis treatments. No differences were observed among the three groups with respect to age, racial origin, BMI, cause of primary renal disease, co-morbidities (chronic lung disease, coronary artery disease, peripheral vascular disease, cerebrovascular disease and diabetes mellitus), late referral to a renal unit within 3 months of dialysis commencement, and current smoking status. Using multilevel logistic regression analyses, the independent predictors of HD transfer (temporary or permanent) for peritonitis were inpatient treatment of peritonitis [odds ratio (OR) 11.45, 95% confidence interval (CI) ] and microbiology profile known to lead to moderate to high rates of catheter removal (compared with Low group as reference: moderate: OR 2.45, 95% CI ; high: OR 8.63, 95% CI ). There were no identified significant predictors of return to PD after temporary HD. PD outcomes following return to PD Of the 101 patients in the Interim HD group, the median time from PD restart to subsequent technique failure was 1.09 years (95% CI years) and these patients were on PD for a median period of 0.80 years prior to their temporary HD transfer. Technique survival rates after return to PD were 55 and 38%, at 1 and 2 years, respectively. Using multivariate Cox regression, the only significant independent predictor of technique failure following temporary HD transfer for peritonitis and subsequent return to PD was peripheral vascular disease [adjusted hazard ratio (HR) 2.03, 95% CI ]. The median time from PD restart to death-censored technique failure was 2.01 years (95% CI years), and it was also associated with the presence of peripheral vascular disease (HR 3.03, 95% CI ). Death-censored technique survival rates after return to PD were 74 and 60%, at 1 and 2 years, respectively. The median time from PD restart to patient death was 3.09 years [95% CI 1.96 years-(not estimable)], and coronary artery disease was predictive of a shorter time to death (HR 3.99, 95% CI ). Forty patients experienced another episode of peritonitis after a return to PD, and the median time from PD restart to the next peritonitis event was 1.35 years (95% CI years). There was no independent predictor of subsequent peritonitis episode Y. Cho et al.

4 Table 2. Characteristics of all Australian PD patients who experienced peritonitis for the first time according to the haemodialysis transfer status during the period Characteristics All patients (n = 3305) Never HD (n = 2752) Interim HD (n = 101) Permanent HD (n = 452) P-values Age (year) 63 (51 73) 62 (51 73) 65 (53 74) 63.5 (54 73) 0.22 Women 1393 (42.2) 1139 (41.4) 37 (36.6) 217 (48.0) 0.02 Racial origin White 2423 (73.3) 2013 (73.2) 76 (75.3) 334 (73.9) 0.33 ATSI 370 (11.2) 307 (11.2) 11 (10.9) 52 (11.5) MPI 106 (3.2) 95 (3.5) 5 (4.9) 6 (1.3) Asian 213 (6.4) 181 (6.6) 4 (4.0) 28 (6.2) Other 193 (5.8) 156 (5.7) 5 (4.9) 32 (7.1) BMI <18.5 kg/m (4.8) 128 (4.7) 5 (5.0) 26 (5.8) to <25 kg/m (41.2) 1144 (41.9) 40 (40.0) 170 (37.8) kg/m (32.0) 871 (31.9) 39 (39.0) 142 (31.6) >30 kg/m (21.9) 590 (21.6) 16 (16.0) 112 (24.9) ESRD Chronic GN 833 (25.2) 685 (24.9) 24 (23.8) 124 (27.4) 0.56 Diabetic nephropathy 1146 (34.7) 963 (34.9) 40 (39.6) 143 (31.6) Renovascular disease/hypertensive 483 (14.6) 401 (14.6) 16 (15.8) 66 (14.6) nephrosclerosis Polycystic kidneys 192 (5.8) 155 (5.6) 3 (2.9) 34 (7.5) Reflux nephropathy 109 (3.3) 94 (3.4) 4 (3.9) 11 (2.4) Other/unknown 542 (16.4) 454 (16.5) 14 (13.9) 74 (16.4) Comorbid conditions Chronic lung disease 389 (11.8) 320 (11.6) 11 (10.9) 58 (12.8) 0.73 Coronary artery disease 1171 (35.4) 967 (35.1) 39 (38.6) 165 (36.5) 0.68 Peripheral vascular disease 582 (17.6) 492 (17.9) 19 (18.8) 71 (15.7) 0.51 Cerebrovascular disease 426 (12.9) 361 (13.1) 13 (12.9) 52 (11.5) 0.64 Diabetes mellitus 1486 (44.9) 1241 (45.1) 49 (48.5) 196 (43.4) 0.61 Late referral 705 (21.4) 590 (21.5) 22 (21.8) 93 (20.6) 0.91 Current smoking status 452 (13.7) 381 (13.8) 19 (18.8) 52 (11.5) 0.13 Cather removal rate based on microbiology Low (<20%) 1677 (52.8) 1537 (58.2) 27 (28.4) 113 (25.4) <0.001 Medium (20 40%) 1081 (34.0) 877 (33.2) 37 (39.0) 167 (37.5) High (>40%) 421 (13.2) 225 (8.5) 31 (32.6) 165 (37.1) Hospitalization for treatment of peritonitis 2062 (69.7) 1592 (64.6) 80 (91.9) 390 (96.5) <0.001 ATSI, Aboriginal and Torres Strait Islander; BMI, body mass index; ESRD, end-stage renal disease; GN, glomerulonephritis; MPI, Maori and Pacific Islander. Matched case control analysis In an attempt to minimize the impact of selection bias and immortal time bias on clinical outcomes (e.g. technique survival, patient survival and peritonitis-free survival) comparisons between those who did and did not undergo HD transfer for peritonitis, matched case control analyses were performed. This was important because Interim HD patients had to survive a median period of 0.8 years on PD prior to temporary HD transfer and a median period of 0.19 years on temporary HD prior to return to PD). The baseline characteristics of the temporary HD transfer cases and controls were well matched (Table 3). Patient survival was not significantly different across the three patient groups (P = 0.28; Figure 1). The median time to death was shorter in the Interim HD patients (3.45 years) compared with Never HD patients (5.88 years), although the difference did not reach statistical significance (HR 1.25, 95% CI , P = 0.28). Similarly, technique survival (HR 0.96, 95% CI ; P = 0.81; Figure 2) and death-censored technique survival (HR 0.87, 95% CI ; P = 0.48; Figure 3) were comparable between matched Interim HD and Never HD patients. Consistent results were also observed for the peritonitis-free survival analysis (HR 0.84, 95% CI , P = 0.49; Figure 4). These results were similarly replicated when analyses were repeated in subgroups according to the categories of catheter removal based on the microbiologic profiles of the peritonitis episodes (i.e. low, medium and high). DISCUSSION The present study represents the largest observational study to date examining the effect of temporary HD transfer for peritonitis on a number of patient-level clinical outcomes. A modest proportion of patients (n = 553; 16.7% of all peritonitis) required transfer to HD following peritonitis, and an even smaller number of patients (n = 101) eventually returned to PD after a period of temporary HD. Transfer to HD was predicted by the characteristics of peritonitis itself (i.e. microbiologic profile and inpatient treatment), while no predictors of return to PD were identified. Matched case control analyses yielded comparable patient-level clinical outcomes (i.e. patient survival, technique survival and peritonitis-free survival) among patient groups defined by their HD transfer status post-peritonitis. PD after peritonitis 1943

5 Table 3. Characteristics of cases (PD patients who restarted PD following temporary HD transfer for peritonitis) compared with matched controls Variable All Patients (n = 455) Never HD (n = 182) Interim HD (n = 91) Permanent HD (n = 182) Age (year) 64 (53 72) 62.5 (52 72) 66 (53 74) 63.5 (53 72) 0.55 Women 209 (45.9) 87 (47.8) 34 (37.4) 88 (48.4) 0.19 Racial origin Non-indigenous 403 (88.6) 162 (89.0) 82 (90.1) 159 (87.4) 0.78 Indigenous 52 (11.4) 20 (11.0) 9 (9.9) 23 (12.6) BMI <18.5 kg/m 2 31 (6.8) 12 (6.6) 5 (5.5) 14 (7.7) to <25 kg/m (40.4) 85 (46.7) 35 (38.5) 64 (35.2) kg/m (30.8) 45 (24.7) 36 (39.6) 59 (32.4) >30 kg/m (21.9) 40 (22.0) 15 (16.5) 45 (24.7) ESRD Chronic GN 123 (27.0) 53 (29.1) 22 (24.2) 48 (26.4) 0.24 Diabetic nephropathy 149 (32.8) 45 (24.7) 37 (40.7) 67 (36.8) Renovascular disease/hypertensive nephrosclerosis 71 (15.6) 34 (18.7) 15 (16.5) 22 (12.1) Polycystic kidneys 31 (6.8) 15 (8.2) 3 (3.3) 13 (7.1) Reflux nephropathy 17 (3.7) 6 (3.3) 3 (3.3) 8 (4.4) Other/unknown 64 (14.1) 29 (15.9) 11 (12.1) 24 (13.2) Comorbid conditions Chronic lung disease 53 (11.7) 18 (9.9) 9 (9.9) 26 (14.3) 0.36 Coronary artery disease 160 (35.2) 65 (35.7) 37 (40.7) 58 (31.9) 0.35 Peripheral vascular disease 70 (15.4) 25 (13.7) 17 (18.7) 28 (15.4) 0.57 Cerebrovascular disease 38 (8.4) 10 (5.5) 12 (13.2) 16 (8.8) 0.09 Diabetes mellitus 205 (45.1) 72 (39.6) 46 (50.6) 87 (47.8) 0.14 Late referral 106 (23.3) 49 (26.9) 20 (21.9) 37 (20.3) 0.31 Current smoking status 65 (14.3) 24 (13.2) 18 (19.8) 23 (12.6) 0.24 Cather removal rate based on microbiology Low (<20%) 120 (26.4) 48 (26.4) 25 (27.5) 47 (25.8) 0.96 Medium (20 40%) 188 (41.3) 78 (42.9) 35 (38.5) 75 (41.2) High (>40%) 147 (32.3) 56 (30.8) 31 (34.1) 60 (33.0) Hospitalization for treatment of peritonitis 420 (92.3) 168 (92.3) 84 (92.3) 168 (92.3) 1.00 Two controls were matched with each case for hospitalization status of first peritonitis and propensity scores derived using logistic regression based on all recorded patient-level covariates. P-values FIGURE 1: Kaplan Meier survival curves for patients returning to PD following temporary HD transfer for peritonitis (cases, n = 87) compared with matched controls who either never required HD (n = 174) or remained on HD permanently (n = 174). Two controls were matched with each case for hospitalization status for treatment of peritonitis episode and propensity scores derived using all available patient-level covariates. Survival was timed from return to PD (duration defined by the matched case subject; P = 0.28). In the present study, the proportion of patients who required transfer to HD due to peritonitis (16.7%) was somewhat lower compared with the previous ANZDATA publication (20%; FIGURE 2: Kaplan Meier technique survival curves for patients returning to PD following temporary HD transfer for peritonitis (cases, n = 89) compared with matched controls who never required HD (n = 178). Two controls were matched with each case for hospitalization status for treatment of peritonitis episode and propensity scores derived using all available patient-level covariates. Technique survival was timed from return to PD (duration defined by the matched case subject; P = 0.81). study duration ), and was comparable to the results reported by others (16%) [14]. In contrast to these relatively consistent results, the probability of returning to PD after a 1944 Y. Cho et al.

6 FIGURE 3: Kaplan Meier death-censored technique survival curves for patients returning to PD following temporary HD transfer for peritonitis (cases, n = 89) compared with matched controls who never required HD (n = 178). Two controls were matched with each case for hospitalization status for treatment of peritonitis episode and propensity scores derived using all available patient-level covariates. Technique survival was timed from return to PD (duration defined by the matched case subject; P = 0.48). FIGURE 4: Kaplan Meier peritonitis-free survival curves for patients returning to PD following temporary HD transfer for peritonitis (cases, n = 60) compared with matched controls who never required HD (n = 120). Two controls were matched with each case for hospitalization status for treatment of peritonitis episode and propensity scores derived using all available patient-level covariates. Peritonitis-free survival was timed from return to PD (duration defined by the matched case subject; P = 0.49). HD transfer was markedly decreased in the current study (18.3%), compared with the previously published results. For instance, Troidle et al. [14] in their single-centre, retrospective observational study, reported a catheter reinsertion rate of 47%, whereas 51% was reported by another study from Hong Kong (n =100)[17]. Such a large discrepancy in the data might have stemmed from differences in practices unique to the treating centre or to the country. This is particularly relevant for Hong Kong where the PD first policy is adopted and unless there are medical contraindications present for PD, patients are only reimbursed financially for PD [18]. In contrast to these studies, the present investigation was a multi-centre observational study based on the data obtained from a national registry, which mitigates the risk of centre-related bias. Although there are no direct financial incentives in place to promote PD as the preferred dialysis modality in Australia (e.g. compared with home HD), there has been an increase in promotion of home dialysis uptake in the past few years which may result in an increase in the proportion of patients returning to PD in the future. There are a number of clinical indications for catheter removal due to PD-related peritonitis endorsed by the International Society of Peritoneal Dialysis (ISPD). These largely relate to lack of treatment response and the microbiologic profile of the peritonitis (e.g. fungi) [15]. In keeping with these recommendations, the transfer to HD post-peritonitis was associated with peritonitis microbiology in the present study. In addition, hospitalization for treatment of peritonitis, which might reflect infection severity, was another significant and independent predictor of transfer to HD. To date, it remains uncertain what impact a transfer to HD post-peritonitis has on clinical outcomes, and whether a return to PD should be encouraged. Szeto et al. [17] aimed to examine the feasibility of resuming PD in patients who required HD post-peritonitis in their observational study involving 100 prevalent PD patients who underwent attempted reinsertion of a PD catheter, 4 weeks after the initial removal. The mean follow-up duration was 18.5 months, and catheter reinsertion and resumption of PD was successful in 51 of the 100 patients ( success group). A history of severe peritonitis was the only significant predictor of failure of successful catheter reinsertion. More importantly, they reported superior patient survival in the success group compared with the fail group at 24 months (80.3 and 56.2%, respectively; P = 0.01). Furthermore, technique survival in the success group was 56.3% at 24 months, which is similar to the 24-month death-censored technique survival rate of 60% in the present investigation. Nevertheless, the conclusions that can be drawn from the study are limited by its small sample size (n = 100), singlecentre and retrospective study design and a lack of adjustment for potential confounding covariates (e.g. comorbidities) in the analysis. In addition, these patients were prevalent PD patients (at risk of Neyman bias), and although the differences did not reach a statistical significance, the mean duration of PD was shorter in the success group compared with the fail group (29.7 months versus 41 months), which could have contributed to the observed survival advantage ( patient and technique survival). In contrast to these encouraging results, Troidle et al. [14] reported a mean technique survival of only 15.4 months after catheter reinsertion in their patients. This is shorter than the median time-to-death-censored technique failure from time of PD restart in the current study (2.01 years). However, since the study by Troidle et al. [14] was limited by the absence of a comparison cohort in the context of a number of limitations in the study design (e.g. singlecentre, lack of adjustment for confounding covariates), the impact of return to PD on technique survival was uncertain. In the present study, comparable patient-level outcomes (i. e. patient survival, technique survival, peritonitis-free survival) were observed following a return to PD after temporary HD PD after peritonitis 1945

7 treatment when compared with patients who never transferred to HD or remained permanently on HD after the first peritonitis episode. These findings are strengthened by well-matched cases and controls (for demographic, comorbid conditions as well as peritonitis-related characteristics), as well as accounting for the risk of immortal time bias by matching to only the controls with PD duration greater than the time spent by matched cases on PD and HD prior to return to PD post-peritonitis. Further subgroup analyses according to microbiologic profile (i.e. catheter removal rate: low, medium and high) observed consistently comparable outcomes across the groups. The strengths of this study include its very large sample size and inclusiveness. We included all patients who experienced their first peritonitis in Australia during the study period, such that a variety of centres were included with varying approaches to the microbiological diagnosis and treatment of peritonitis. This greatly enhanced the 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 information such as the presence of concomitant exit-site and tunnel infections, antimicrobial susceptibilities of isolated micro-organisms, patient compliance, individual unit management protocols, laboratory values (such as dialysate white cell counts), severity of comorbidities, disconnect systems used, prescribed PD dialysate (especially icodextrin), proximity of patient residence to dialysis unit, antibiotic dosages or routes of antibiotic administration, peritoneal dialysate culture methodology or previous antibiotic exposure for any indication. Even though we adjusted for a large number of patient characteristics, the possibility of residual confounding could not be excluded. Given the relatively low proportion of patients who ever returned to PD after initial HD transfer, the possibility of selection bias and confounding by indication cannot be excluded. The ANZDATA registry does not collect information on reasons for non-return to PD. In common with other registries, ANZDATA is a voluntary registry and there is no external audit of data accuracy, including the diagnosis of peritonitis. Consequently, the possibility of coding/classification bias cannot be excluded. In conclusion, a modest proportion of patients required transfer to HD following an episode of peritonitis, and an even smaller number eventually returned to PD after a temporary HD. Peritonitis microbiologic profile and hospitalization were associated with transfer to HD. Nevertheless, return to PD after a period of time spent on HD should not be discouraged as it was associated with non-inferior patient-level clinical outcomes compared with other patient groups who either never required HD or stayed permanently on HD post-peritonitis following adjustment for patient- and peritonitis-related covariates. ACKNOWLEDGEMENTS 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. CONFLICT OF INTEREST STATEMENT D.J. 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. He is a current recipient of a Queensland Government Health Research Fellowship. C.M.H. is a consultant for Fresenius Medical Care and received a travel grant from Fresenius, and has received research grant from Baxter Healthcare, and research grants and travel grants from AMGEN Australia. F.B. is a consultant for Baxter and Fresenius and has received travel grants from Amgen and Roche. S.M. has received speaking honoraria from AMGEN Australia, Fresenius Australia and Solvay Pharmaceuticals and travel grants from AMGEN Australia, Genzyme Australia and Jansen-Cilag. Y.C. is a current recipient of the Australian Postgraduate Award. The remaining authors have no competing financial interests to declare. The results presented in this paper have not been published previously in whole or part, except in abstract format. (See related article by Donovan and Carrington. Peritoneal dialysis outcomes after temporary haemodialysis for peritonitis influence on current practice. Nephrol Dial Transplant 2014; 29: ) REFERENCES 1. Australia & New Zealand Dialysis & Transplantation (ANZDATA) Registry, Chapter 6: Peritoneal dialysis, anzdata/anzdatareport/34threport/2011c06_peritoneal_v1.8.pdf (22 June 2012, date last accessed). 2. Ghali JR, Bannister KM, Brown FG et al. Microbiology and outcomes of peritonitis in Australian peritoneal dialysis patients. Perit Dial Int 2011; 31: Siva B, Hawley CM, McDonald SP et al. Pseudomonas peritonitis in Australia: predictors, treatment, and outcomes in 191 cases. Clin J Am Soc Nephrol 2009; 4: Jarvis EM, Hawley CM, McDonald SP et al. Predictors, treatment, and outcomes of non-pseudomonas Gram-negative peritonitis. Kidney Int 2010; 78: Miles R, Hawley CM, McDonald SP et al. Predictors and outcomes of fungal peritonitis in peritoneal dialysis patients. Kidney Int 2009; 76: Barraclough K, Hawley CM, McDonald SP et al. Polymicrobial peritonitis in peritoneal dialysis patients in Australia: predictors, treatment, and outcomes. Am J Kidney Dis 2010; 55: Barraclough K, Hawley CM, McDonald SP et al. Corynebacterium peritonitis in Australian peritoneal dialysis patients: predictors, treatment and outcomes in 82 cases. Nephrol Dial Transplant 2009; 24: Edey M, Hawley CM, McDonald SP et al. Enterococcal peritonitis in Australian peritoneal dialysis patients: predictors, treatment and outcomes in 116 cases. Nephrol Dial Transplant 2010; 25: Fahim M, Hawley CM, McDonald SP et al. Coagulase-negative staphylococcal peritonitis in Australian peritoneal dialysis patients: predictors, treatment and outcomes in 936 cases. Nephrol Dial Transplant 2010; 25: Fahim M, Hawley CM, McDonald SP et al. Culture-negative peritonitis in peritoneal dialysis patients in Australia: predictors, treatment, and outcomes in 435 cases. Am J Kidney Dis 2010; 55: Govindarajulu S, Hawley CM, McDonald SP et al. Staphylococcus aureus peritonitis in Australian peritoneal dialysis patients: predictors, treatment, and outcomes in 503 cases. Perit Dial Int 2010; 30: Y. Cho et al.

8 12. Johnson RJ, Ramsey PG, Gallagher N et al. Fungal peritonitis in patients on peritoneal dialysis: incidence, clinical features and prognosis. Am J Nephrol 1985; 5: O Shea S, Hawley CM, McDonald SP et al. Streptococcal peritonitis in Australian peritoneal dialysis patients: predictors, treatment and outcomes in 287 cases. BMC Nephrol 2009; 10: Troidle L, Gorban-Brennan N, Finkelstein FO. Outcome of patients on chronic peritoneal dialysis undergoing peritoneal catheter removal because of peritonitis. Adv Perit Dial 2005; 21: Li PK, Szeto CC, Piraino B et al. Peritoneal dialysis-related infections recommendations: 2010 update. Perit Dial Int 2010; 30: Shariff SZ, Cuerden MS, Jain AK et al. The secret of immortal time bias in epidemiologic studies. J Am Soc Nephrol 2008; 19: Szeto CC, Chow KM, Wong TY et al. Feasibility of resuming peritoneal dialysis after severe peritonitis and Tenckhoff catheter removal. J Am Soc Nephrol 2002; 13: Li PK, Szeto CC. Success of the peritoneal dialysis programme in Hong Kong. Nephrol Dial Transplant 2008; 23: Received for publication: ; Accepted in revised form: Nephrol Dial Transplant (2014) 29: doi: /ndt/gfu248 Advance Access publication 24 July 2014 Effects of intradialytic cycling compared with pedometry on physical function in chronic outpatient hemodialysis: a prospective randomized trial Clara Bohm 1, Krista Stewart 2, Jennifer Onyskie-Marcus 2, Dale Esliger 3, Dean Kriellaars 4 and Claudio Rigatto 1 1 Section of Nephrology, University of Manitoba, Winnipeg, MB, Canada, 2 Winnipeg Regional Health Authority, Winnipeg, MB, Canada, 3 Loughborough University, Leicestershire, UK and 4 School of Medical Rehabilitation, University of Manitoba, Winnipeg, MB, Canada Correspondence and offprint requests to: Clara Bohm; cbohm@hsc.mb.ca ABSTRACT Background. Individuals on hemodialysis have low physical function and activity levels. Clinical trials have shown improvements in these parameters with exercise programming. Pedometers have not been extensively evaluated in individuals on hemodialysis. This randomized clinical trial compared the effects of intradialytic cycling versus a pedometer program on physical function, physical activity and quality of life. Methods. Sixty patients were randomly assigned to two study groups. The ergometer group cycled during each hemodialysis session for 24 weeks. Pedometer participants followed a homebased walking program for 24 weeks. The primary outcome was aerobic capacity [VO 2peak and 6-minute walk (6MW) test]. Secondary outcomes included lower extremity strength [sit-to-stand (SS) test], flexibility [sit-and-reach (SR) test], physical activity (accelerometer) and health-related quality of life. Measurements were collected at baseline and at 12 and 24 weeks. Results. At 12 and 24 weeks, there was no significant change in the VO 2peak or 6MW test between or within study groups. SS testing in the ergometer group improved from 10.2 (SD 3.4) to 11.4 (SD 2.5) cycles from baseline to 24 weeks (P < 0.005). Similarly, in the pedometer group, SS cycles improved from 10.1 (SD 3.3) to 12.2 (SD 3.5) (P < 0.005). The SR test also significantly improved over time in both the study groups. No significant changes were noted for other secondary outcomes. Conclusions. Both intradialytic cycling and pedometer programming improved aspects of physical function. Neither intervention had a significant effect on aerobic capacity. No significant differences in any outcomes were identified between interventions groups. Keywords: exercise, hemodialysis, pedometers, physical activity, physical function INTRODUCTION Individuals with end-stage renal disease (ESRD) have low levels of physical function and activity [1 3]. Both physical activity and function decline over time on dialysis [4]. Self- The Author Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. 1947