Acinetobacter Peritoneal Dialysis Peritonitis: Description and Relation to the spice Family of Organisms

Materials and Methods

Two search strategies were employed in order to ensure the most relevant publications on the topic were captured. First, a PubMed search was conducted using the key words “Acinetobacter,” “peritonitis,” and “dialysis.” This search was then limited to the past 25 years, encompassing the period January 1982 to May 2007. Second, the references cited in all articles identified as pertinent following the first search were reviewed.

Results and Discussion

The PubMed search identified 35 publications, of which 7 principally addressed the subject of Acinetobacter PD peritonitis. Of these 7, 2 were excluded as they predated the inclusion period of study: 1 described 2 patients with Acinetobacter PD peritonitis treated with peritoneal antibiotic lavage (1), the other described an Acinetobacter PD peritonitis outbreak due to water bath contamination of the dialysate (2). One additional paper was identified following the review of cited references. The format of all 6 documents was either as case report or as case series.

Lye et al., from Singapore, reviewed 13 episodes occurring in 11 patients on CAPD over an 18-month period from 1989 to 1990 (3). This accounted for 14.3% of the total number of episodes of PD peritonitis and was the most common cause of gram-negative bacillary PD peritonitis. Antibiotic treatment of 2 weeks’ duration was successful in 12 of the episodes. Specifically, IP gentamicin alone was used in 9 (including 1 repeat episode), oral pefloxacin alone in 2, and combination IP ceftazidime and oral pefloxacin in 1. In the 1 unsuccessfully treated episode, the peritonitis was catheter related, with a concomitant exit-site infection and infected pleural effusion; consequently, a repeat episode transpired and an eventual conversion to hemodialysis ensued (3). Three years later, the same group from Singapore described 32 episodes of PD-related Acinetobacter infections, of which 29 were PD peritonitis, occurring in 28 patients on CAPD from 1989 to 1992 (4). Nearly 80% were community acquired. Once again, this accounted for 10.7% of the total number of episodes of PD peritonitis and was the most common cause of gram-negative bacillary PD peritonitis. The majority of peritonitis episodes, specifically 21, were treated with IP gentamicin alone, which was successful in 72.2% of cases. The remaining 8 peritonitis episodes were treated with oral pefloxacin alone, IP ceftazidime alone, IP ampicillin-sulbactam alone, or a combination of IP gentamicin, oral pefloxacin, or IP ceftazidime with variable success (4). It is unclear what role the development of new resistance played in the beta-lactam treatment failures.

Ruiz et al., from Spain, presented their experience following 127 CAPD patients over 8 years, in which only 1.6% of the 377 peritonitis episodes were due to Acinetobacter species (5). All 6 of these episodes represented recurrences occurring an average of 10 days following peritonitis episodes due to other bacteria. Treatment with a 10-day course of IP tobramycin with or without cephalothin was successful in only 1 patient; catheter removal was employed in the remaining 5 patients because of relapsing or refractory peritonitis (5). Galvao et al., from Michigan, described 23 patients with Acinetobacter PD peritonitis during the period 1979 – 1987 (6). This accounted for the second most common cause of gram-negative bacillary PD peritonitis. Treatment with IP antibiotics was successful in 21 patients: IP amino-glycoside therapy alone was used in 16 episodes (including 1 relapse) and combination IP aminoglycoside and beta-lactam in the remaining 5 episodes. However, 2 episodes required catheter removal as 1 was catheter related with concomitant tunnel infection and pericatheter leak and the other was associated with Acinetobacter bacteremia (6). Valdez et al., from New York, described 19 episodes of Acinetobacter PD peritonitis between 1982 and 1989 (7). All but 1, in which the infection was refractory despite a 12-day course of IP gentamicin and required catheter removal for cure, were successfully treated with either IP aminoglycoside therapy alone for a mean of 10.7 days in 14 cases (including 1 relapse following a 10-day course of oral ciprofloxacin), IP gentamicin for 4 days stepped down to oral ciprofloxacin for 8 days in 2 cases, or intravenous ceftriaxone for 10 days in the remaining 1 case (7). More recently, Borras et al., from Spain, presented a case of Acinetobacter junii PD peritonitis in a patient on automated PD that was refractory to IP tobramycin and oral ciprofloxacin and then IP cefotaxime and oral levofloxacin, necessitating catheter removal (8).

Beta-lactamase production by gram-negative bacteria is a well-established mechanism of beta-lactam antibiotic resistance resulting in treatment failure or recurrence. The emergence of two specific types of beta-lactamase has received growing attention, namely, plasmid-mediated extended-spectrum beta-lactamases (ESBLs) and inducible chromosomally mediated beta-lactamases, which vary in antibiogram profiles. ESBLs are produced by various Enterobacteriaceae (e.g., Escherichia coli, Klebsiella species); typically, they demonstrate in vitro resistance to oxyimino beta-lactams (e.g., cefotaxime, ceftazidime, ceftriaxone, cefepime) and susceptibility to beta-lactam/beta-lactamase inhibitor combinations, cephamycins (e.g., cefoxitin, cefotetan, cefmetazole), and carbapenems. However, in vivo, the bacteria may be resistant to such antibiotics because of either an inoculum effect (ESBL hyperproduction) or other resistance mutations (e.g., porin protein). On the other hand, the acronym SPICE denotes a family of several unique gram-negative bacteria with inducible chromosomally mediated beta-lactamases (Serratia, Pseudomonas/Providencia, indole-positive Proteus/Acinetobacter/Morganella, Citrobacter, Enterobacter or Hafnia). As evident, Acinetobacter is one of these SPICE organisms. Initial in vitro susceptibility may be followed by the selection of beta-lactamase-producing mutants due to induction beta-lactam therapy. At this point, the bacteria demonstrate in vitro resistance to oxyimino beta-lactams, beta-lactam/beta-lactamase inhibitor combinations, and cephamycins, and susceptibility to carbapenems. This may very well serve as an explanation for the index patient's rebound peritoneal leukocytosis, despite the negative repeat culture, that occurred following an initial period of therapeutic response. While identification of Acinetobacter with an antibiogram now demonstrating beta-lactam antibiotic resistance would have been more definitive, one may surmise the coexistence of two subpopulations: the original beta-lactam-sensitive subpopulation appropriately suppressed on ceftazidime and an evolving beta-lactamase-producing subpopulation below the laboratory's sensitivity for identification of positive culture. This may be substantiated by the fact that such mutant organisms may not be easily detected at the standard inocula used in susceptibility testing.

Testing for the presence or absence of such beta-lactamases is not as straightforward, however, because of enzyme heterogeneity, variable substrate activity, coexistence of other beta-lactamases, and changing criteria for detection over time. The Clinical and Laboratory Standards Institute (CLSI) makes specific recommendations regarding resistance testing for ESBL-producing E. coli and Klebsiella pneumoniae/oxytoca only. Laboratories are encouraged to test with larger than normal inocula, use lower breakpoints for determination of resistance, and test for synergy between beta-lactams and beta-lactamase inhibitors. Further, a high index of clinical suspicion with frequent susceptibility testing of repeat isolates is warranted (9).

The current body of data addressing the topic of antimicrobial therapy for both ESBL- and inducible beta-lactamase-producing organisms, Acinetobacter notwithstanding, is sparse. Specifically, there are no clinical trials but rather case reports and limited retrospective epidemiologic studies. The surest therapy for such organisms remains a carbapenem; other therapies may be tried and, as demonstrated in the specific case series above of Acinetobacter PD peritonitis, many patients respond to IP aminoglycoside and/or oral fluoroquinolone. However, many of these bacteria are multidrug resistant (especially to aminoglycosides and fluoroquinolones) through other concomitant antibiotic resistance mechanisms; moreover, even in one report describing outcomes among patients with ciprofloxacin-sensitive ESBL-positive Klebsiella pneumoniae bacteremia, imipenem performed better than ciprofloxacin (10).

While carbapenems are the antimicrobials of choice, therapy with an IP aminoglycoside, oral fluoroquinolone, or, perhaps even better, double coverage that includes a non-beta-lactam antibiotic may be tried and is more feasible among PD patients. Although the 2005 International Society for Peritoneal Dialysis (ISPD) guidelines emphasize the need to monitor for resistance among various gram-negative bacteria, the use of double antibiotic therapy is recommended only for Pseudomonas and Stenotrophomonas (11). However, perhaps double coverage may also be more efficacious for such potentially multidrug-resistant organisms in which Acinetobacter is included. Clearly though, the body of data to defend this empiric suggestion is evolving but is as yet insufficient.