Patients with underlying liver disease and ascites are

Transcription

1 Peritoneal Dialysis International, Vol. 26, pp Printed in Canada. All rights reserved /06 $ Copyright 2006 International Society for Peritoneal Dialysis CONTINUOUS AMBULATORY PERITONEAL DIALYSIS IN PATIENTS WITH HEPATITIS B LIVER DISEASE Kai Ming Chow, Cheuk Chun Szeto, Alan Ka Lun Wu, Chi Bon Leung, Bonnie Ching Ha Kwan, and Philip Kam-Tao Li Department of Medicine & Therapeutics, The Chinese University of Hong Kong, Shatin, Hong Kong, SAR, China Objective: We hypothesized that patients with hepatitis B virus infection and cirrhosis are more susceptible to peritonitis as a complication of peritoneal dialysis (PD). Methods: A retrospective study was carried out to compare peritonitis rates between cirrhotic and non-cirrhotic patients with hepatitis B virus infection. Results: Between 1994 and 2004, 25 PD patients with hepatitis B cirrhosis and 36 patients with hepatitis B without cirrhosis were included for analysis. Mean follow-up duration was 52 months. Subjects with hepatitis B cirrhosis consisted of more males and had higher total body weight. No cirrhotic patients (20 of them being Child Pugh class A, 2 class B, and 3 class C) had undergone portosystemic shunting or liver transplantation. Cirrhotic patients had slightly higher bilirubin concentration than the non-cirrhotic group (22 ± 50 vs 9 ± 4 µmol/l, µµ p = 0.16). There was no difference in median peritonitis-free survival between cirrhotic and non-cirrhotic patients (40 vs 37 months, p = 0.64 by log-rank test). The average peritonitis rate was 1 episode every 19.2 patient-months in the cirrhotic group and 1 episode every 20.5 patient-months in the non-cirrhotic group. Time to first peritonitis did not differ between the two groups with respect to gram-negative organisms (p = 0.88) or gram-positive organisms (p = 0.52). Cirrhotic patients had more frequent Streptococcus species peritonitis, which accounted for 13% of all peritonitis episodes, as opposed to 2% among the non-cirrhotic patients (p = 0.01). Overall treatment response rate and outcome did not differ between patients with and patients without cirrhosis. Conclusions: Peritonitis-free survival of cirrhosis patients infected by hepatitis B virus compares favorably with that in patients without cirrhosis. The presence of liver cirrhosis does not appear to compromise PD outcome. Perit Dial Int 2006; 26: Correspondence to: C.C. Szeto, Department of Medicine & Therapeutics, The Chinese University of Hong Kong, Prince of Wales Hospital, Shatin, Hong Kong, China. ccszeto@cuhk.edu.hk Received 4 October 2004; accepted 22 July KEY WORDS: Hepatitis B; peritonitis; Streptococcus; cirrhosis. Patients with underlying liver disease and ascites are susceptible to spontaneous bacterial peritonitis, which is thought to involve translocation of bacteria from the intestinal lumen to the lymph nodes, followed by subsequent bacteremia and infection of the ascitic fluid (1). Since the conventional lactate-buffered peritoneal dialysis (PD) solution has been demonstrated to impair liver cell function in vitro (2), innate defense against the bacteria inside the abdominal fluid might be reduced in the PD population. To make matters worse, macrophage Fc gamma receptor function, an important host defense against bacterial infection, is impaired in both end-stage renal disease and cirrhosis patients (3,4). This retrospective study was undertaken to analyze the incidence and pattern of peritonitis complications in 25 consecutive continuous ambulatory peritoneal dialysis (CAPD) patients with hepatitis B cirrhosis, compared with 36 hepatitis B virus-infected dialysis patients without cirrhosis. PATIENTS AND METHODS Between 1994 and 2004, a total of 61 incident patients with hepatitis B virus infection underwent CAPD in our center. Based on clinical, biochemical, sonographic, and, in 5 cases, histological criteria, 41% of those patients were diagnosed with liver cirrhosis. Severity of liver cirrhosis was ascertained according to the Child Pugh classification: if the total Child Pugh score is 5 or 6, the cirrhosis is designated class A, class B if the score is 7 9, and class C if the score is 10 or higher (5). The first-line renal replacement therapy in all cases was CAPD by means of disconnect systems (Baxter Healthcare, Deerfield, Illinois, USA, and Fresenius Medical Care, Deutschland GmbH, Bad Homburg, 213

2 CHOW et al. MARCH 2006 VOL. 26, NO. 2 PDI Germany) with lactate-buffered glucose-containing dialysate solutions. The diagnosis of peritonitis complicating PD was based on at least two of the following criteria: abdominal pain or cloudy PD effluent, leukocytosis in peritoneal fluid effluent (white cell count 100/mm 3 ), or positive Gram stain or culture from effluent (6). Bacterial culture of PD effluent in our unit was performed using the BacTAlert bottles (Organon Teknika, Durkham, North Carolina, USA) following the manufacturer s instructions. Peritonitis episodes were treated with our center s standard antibiotic protocol. Initial antimicrobial therapy for peritonitis, in general, consisted of intraperitoneal administration of cefazolin plus netilmicin or ceftazidime; alternatively, a fourth-generation cephalosporin plus or minus vancomycin was used. The details of peritonitis management and definitions of response and relapse have been described in our previous reports (7,8). None of the patients were on prophylactic antimicrobial therapy. Statistical analysis was performed by SPSS version 11.5 (SPSS Inc., Chicago, Illinois, USA) for Windows operating system (Microsoft, Redmond, Washington, USA). All data are expressed as mean ± SD unless otherwise specified. Statistical comparisons were performed using Student s t-test; comparisons of percentages between groups were made with the chi-square test or Fisher s exact test, as appropriate. Actuarial survival curves were determined according to the Kaplan Meier life-table method. All probabilities were two-tailed and the level of significance was set at RESULTS A total of 25 patients (mean age 53 ± 12 years) with hepatitis B cirrhosis were identified over a study period of 10 years. They were compared with a control group of hepatitis B subjects without cirrhosis, comprised of 36 subjects whose mean age was 53 ± 14 years (Table 1). There were more male subjects in the cirrhotic group (76% vs 47%, p = 0.04), which had a significantly higher body weight than the group without cirrhosis. Although not statistically significant, patients with cirrhosis had higher bilirubin concentration than non-cirrhotic patients (22 ± 50 vs 9 ± 4 µmol/l, p = 0.16). As shown in Table 1, baseline characteristics and comorbidities of the 25 cirrhotic patients and 36 non-cirrhotic patients were otherwise similar. Both groups of dialysis patients had similar etiology of renal disease and dialysis adequacy. According to Child-Pugh scores, most patients had an early stage of liver cirrhosis: 20 patients were classified Child-Pugh class A, 2 patients belonged to Child-Pugh 214 class B, and 3 belonged to class C. No patients had undergone portosystemic shunting or liver transplantation. We observed no difference in the probability of remaining free of peritonitis between the cirrhotic and the non-cirrhotic group (Figure 1). There were 1617 patientmonths of PD treatment in the cirrhotic group and 1678 patient-months in the non-cirrhotic group. According to the Kaplan Meier analysis, patients with cirrhosis had a median peritonitis-free survival of 40 months, which was similar to the median peritonitis-free survival of 37 months in the non-cirrhotic group (p = 0.64). There was 1 episode every 19.2 patient-months and 1 episode every 20.5 patient-months in the cirrhotic and noncirrhotic group respectively. In other words, the peritonitis rate was 0.62 episodes of peritonitis per year in the cirrhotic group and 0.59 episodes per year in the noncirrhotic group. With respect to the individual organisms causing peritonitis on dialysis, the time to first peritonitis did not differ between the two groups in terms of gram-negative organisms (p = 0.88) or gram-positive organisms (p = 0.52). Possible bacterial causes of peritonitis were carefully investigated among the two groups. Gram-positive organisms accounted for 34% of episodes in the cirrhotic group, compared with 27% in the non-cirrhotic group (p = 0.31). As shown in Table 2, there was no difference in the causative organisms between the two groups except for a greater preponderance of Streptococcus species in the cirrhosis patients. Streptococcus species accounted for 13% of all peritonitis episodes in the cirrhotic patients, as opposed to 2% among the non-cirrhotic patients (p = 0.01). The two episodes of streptococcal peritonitis in the non-cirrhotic patients were secondary to Streptococcus anginosus and Streptococcus salivarius respectively, whereas all 10 cases among cirrhotic patients were due to α-hemolytic streptococci. Overall, the treatment response rate and outcome did not differ between patients with and without cirrhosis (Table 2). Among the cirrhotic patients, antibiotic treatment failure requiring Tenckhoff catheter removal occurred in 15% of the peritonitis episodes, versus 11% in the non-cirrhotic group (p = 0.43). Relapse of peritonitis occurred in 15% of all peritonitis episodes in the cirrhotic and 11% of the non-cirrhotic group (p = 0.75). No sclerosing encapsulating peritonitis ensued. Four patients in the cirrhotic group (2 patients with Child-Pugh class C, 1 with Child-Pugh class B, and 1 with class A) and 3 non-cirrhotic patients died of peritonitis complicating dialysis (p = 0.66). None were related to Streptococcus peritonitis. In the 3 non-cirrhotic patients with mortality attributable to peritonitis, culture from their PD effluent grew Mycobacterium tuberculosis, Pseudomo-

3 PDI MARCH 2006 VOL. 26, NO. 2 CAPD IN PATIENTS WITH HEPATITIS B LIVER DISEASE TABLE 1 Baseline Demographic Characteristics Cirrhotic group (n=25) Non-cirrhotic group (n=36) p Value Age (years) 53±12 53± Gender (M:F) 19:6 17: Body weight (kg) 63.0± ±8.0 <0.001 Body mass index (kg/m 2 ) 23.9± ± Duration of follow-up (months) a 47 (32 81) 43 (18 60) 0.14 Primary renal disease Diabetic nephropathy 5 (20%) 12 (33%) IgA nephropathy 5 (20%) 7 (19%) Mesangiocapillary glomerulonephritis 2 (8%) 1 (3%) 0.58 Other glomerular disease 9 (36%) 9 (25%) Diabetes mellitus 7 (28%) 13 (35%) 0.59 Cerebrovascular disease 2 (8%) 7 (19%) 0.29 Peripheral vascular disease 2 (8%) 5 (14%) 0.69 Coronary artery disease 1 (4%) 6 (17%) 0.22 Serum bilirubin (µmol/l) b 22±50 9± Serum albumin (mg/l) 30±8 31± Kt/V 1.9± ± a Median (interquartile range) days. b To convert values for bilirubin from micromoles per liter to milligrams per deciliter, divide by Values are mean±sd unless otherwise indicated. Figure 1 Probability of CAPD patients remaining free of peritonitis. nas aeruginosa, and Serratia species, respectively. In the 4 cirrhotic patients who died from peritonitis, the causative organisms included Escherichia coli (2 cases), Pseudomonas aeruginosa, and Enterococcus faecalis. DISCUSSION In the present study, we attempted to limit confounding comorbidity by restricting our analysis to a cohort of CAPD patients with positive hepatitis B surface antigen. Among them, we demonstrated neither significant increase in peritonitis infection complication nor excess peritonitis mortality in the cirrhotic subgroup. To date, all (9 12) but one of the PD patient series (13) with cirrhosis were limited either by the short-term follow-up duration or to anecdotal case reports. Our finding is therefore of clinical and therapeutic significance in planning renal replacement therapy for hepatitis B virus-infected patients with end-stage renal failure. Compared to our hepatitis B-negative counterparts, we observed a relatively high peritonitis rate in our cohort with positive hepatitis B surface antigen, irrespective of cirrhosis status. This could have been related to the lesser likelihood of receiving kidney transplant or transferal to hemodialysis among hepatitis B-positive patients, even in the context of recurrent peritonitis. Presumably, the presence of liver cirrhosis is important in the pathogenesis of peritonitis owing to the nutritious and protective environment of ascites. On the other hand, the untoward effect of ascites might have been canceled out in the context of continuous PD because all PD patients have ascites, irrespective of the presence or absence of liver cirrhosis. Another distinct feature of peritonitis complication among cirrhotic subjects on PD, in particular, is the much earlier diagnosis of infection episodes compared to liver patients with cirrhosis alone (thus requiring paracentesis for diagnostic purpose). 215

4 CHOW et al. MARCH 2006 VOL. 26, NO. 2 PDI TABLE 2 Distribution and Characteristics of Peritonitis and Causative Organisms Cirrhotic group Non-cirrhotic group Episodes (total peritonitis episodes=79) (total peritonitis episodes=82) p Value Escherichia coli peritonitis Klebsiella species peritonitis Serratia species peritonitis Pseudomonas peritonitis Polymicrobial peritonitis Streptococcus peritonitis Coagulase-negative staphylococcus peritonitis Staphylococcus aureus peritonitis Culture-negative peritonitis Fungal peritonitis Tuberculous peritonitis Peritonitis requiring Tenckhoff catheter removal Peritonitis relapse Peritonitis causing mortality That being said, it is also plausible that patients with a more advanced stage of liver cirrhosis than our cohort (with over 80% of cases being Child s class less than B) might entail a greater risk of bacterial peritonitis complication. One might argue that the distinction between the cirrhotic and non-cirrhotic groups is inconsequential in our series, since serum albumin did not differ significantly between the two groups. Furthermore, early death unrelated to peritonitis in the cirrhosis group could have concealed patients with susceptibility to dialysis-related peritonitis. There was no evidence that this survival bias occurred because only 1 patient in the cirrhotic group died of non-peritonitis causes within 1 year of PD. Our study further shows, for the first time, that Enterobacteriaceae peritonitis does not occur in excess among cirrhotic patients on PD, as might have been suggested by other small case series (12). Conversely, the interesting finding of increased incidence of α-hemolytic streptococcal peritonitis brings us to the putative link and mechanism of peritonitis in patients with cirrhosis. As opposed to the route of gut translocation, α-hemolytic streptococcus is not found in the gastrointestinal tract in most cases and probably originates from hematogenous spread. Since α-hemolytic streptococci are frequently present in human oral flora, damaged oral mucosa is a possible portal of entry for the bacteremia. Another attractive hypothesis is transient colonization of gastric flora by organisms such as Streptococcus pneumoniae among patients with low gastric acidity and oropharyngeal colonization. In particular, cirrhotic subjects are often subjected to endoscopic procedures, which might in turn favor transient abdominal colonization by streptococci (14). Last but not least, it is important to note the relatively small number of cases in this series; our findings by chance alone remains a possibility. Our study has several limitations. First, our results might not apply to patients on automated PD. Second, we cannot exclude misclassification bias because liver cirrhosis can be difficult to diagnose without histologic evidence. In conclusion, our data suggest that hepatitis B virusinfected patients with cirrhosis are not particularly susceptible to peritonitis complicating CAPD, compared to patients without cirrhosis. Despite a greater preponderance of α-hemolytic streptococcal peritonitis, it was not associated with increased mortality rate or treatment failures. REFERENCES 1. Such J, Runyon BA. Spontaneous bacterial peritonitis. Clin Infect Dis 1998; 27: Riegel W, Ulrich C, Friedrichsohn C, Passlick-Deetjen J, Kohler H. Liver cell reactive components in peritoneal dialysis fluids. Miner Electrolyte Metab 1999; 25: Ruiz P, Gomez F, Schreiber AD. Impaired function of macrophage Fcγ receptors in end-stage renal disease. N Engl J Med 1990; 322: Gomez F, Ruiz P, Schreiber AD. Impaired function of macrophage Fcγ receptors and bacterial infection in alcoholic cirrhosis. N Engl J Med 1994; 331: Pugh RN, Murray-Lyon IM, Dawson JL, Pietroni MC, Williams R. Transection of the oesophagus for bleeding 216

5 PDI MARCH 2006 VOL. 26, NO. 2 CAPD IN PATIENTS WITH HEPATITIS B LIVER DISEASE oesophageal varices. Br J Surg 1973; 60: Keane WF, Alexander SR, Bailie GR, Boeschoten E, Gokal R, Golper TA, et al. Peritoneal dialysis-related peritonitis treatment recommendations: 1996 update. Perit Dial Int 1996; 16: Szeto CC, Chow KM, Leung CB, Wong TY, Wu AK, Lui SF, et al. Clinical course of peritonitis due to Pseudomonas species complicating peritoneal dialysis: a review of 104 cases. Kidney Int 2001; 59: Szeto CC, Wong TY, Chow KM, Leung CB, Li PK. The clinical course of culture-negative peritonitis complicating peritoneal dialysis. Am J Kidney Dis 2003; 42: Marcus RG, Messana J, Swartz R. Peritoneal dialysis in endstage renal disease patients with preexisting chronic liver disease and ascites. Am J Med 1992; 93: Poulos AM, Howard L, Eisele G, Rodgers JB. Peritoneal dialysis therapy for patients with liver and renal failure with ascites. Am J Gastroenterol 1993; 88: Selgas R, Bajo A, Jimenez C, Sanchez C, Del Peso G, Cacho G, et al. Peritoneal dialysis in liver disorders. Perit Dial Int 1996; 16(Suppl 1):S Bajo MA, Selgas R, Jimenez C, Del Peso G, Fernandez-Reyes MJ, Dapena F, et al. CAPD for treatment of ESRD patients with ascites secondary to liver cirrhosis. Adv Perit Dial 1994; 10: De Vecchi AF, Colucci P, Salerno F, Scalamogna A, Ponticelli C. Outcome of peritoneal dialysis in cirrhotic patients with chronic renal failure. Am J Kidney Dis 2002; 40: Capdevila O, Pallares R, Grau I, Tubau F, Liñares J, Ariza J, et al. Pneumococcal peritonitis in adult patients: report of 64 cases with special reference to emergence of antibiotic resistance. Arch Intern Med 2001; 161: