Treatment of Peritonitis and Other Clinical Complications of Peritoneal Dialysis in the Critically Ill Patient

Pathophysiology and Microbiology of Peritonitis

Microorganisms can access the peritoneal cavity through several routes. Contamination of PD fluid and the luminal tract is usually the result of touch contamination and breach of sterile technique during inflow. Migration along the catheter surface often is seen in conjunction with exit site and tunnel infections. Translocation through contaminated viscera such as bowel and bladder occurs during inflammation or infection of these organs, and this route may be germane in the critically ill patient. Hematogenous spread from distant sites of infection rarely occurs. Interestingly, secondary bacteremia from PD peritonitis appears to be essentially unheard of in the absence of underlying abdominal pathology.

Most peritonitis episodes are caused by bacterial pathogens. Gram-positive organisms, particularly Staphylococcus aureus (SA) and coagulase-negative staphylococci (CNS), are seen most commonly and are usually the result of skin flora contamination during PD fluid instillation or catheter infection. Gram negatives are less frequently cultured, although the relative rate of infections from organisms such as Pseudomonas may be increasing. Polymicrobial or anaerobic peritonitis often is due to intraabdominal or gut pathology and should prompt an evaluation for underlying surgical pathology. Fungal peritonitis, most commonly from Candida species, is a serious complication, often preceded by antibiotic use, especially for a previous bacterial peritonitis. Culture-negative peritonitis is not uncommon and was seen in 20% of cases in one observational study.

Clinical Features, Evaluation of Suspected Peritonitis, and Diagnosis

The cardinal symptoms of PD peritonitis are abdominal pain and change in effluent color from clear to cloudy. Another important clue is a reduction in ultrafiltration consequent to peritoneal inflammation. Evolving signs of sepsis such as fever and hemodynamic instability may be seen. Exit site infection or tunnel abscess, with purulent discharge from the exit site or overlying cellulitis, also may be present and should be sought actively during any evaluation for PD peritonitis.

Patients with suspected peritoneal infection should undergo prompt drainage of all peritoneal fluid. A sample of the effluent should be sent for cell count and differential, gram stain, and culture. Samples ideally should have dwelled for 2 hours in the peritoneal cavity. Patients who are empty of PD fluid should have 1 L of PD fluid instilled for 2 hours, after which it should be drained and analyzed as outlined above. For optimal pathogen detection, the International Society for Peritoneal Dialysis (ISPD) recommends 5 to 10 mL of PD fluid be sent in aerobic and anaerobic blood culture bottles. Any catheter exit-site purulence also should be cultured. As mentioned, blood cultures are infrequently positive in PD-associated peritonitis but may more likely isolate pathogens when surgical pathology is present and can be sent in the proper clinical context.

The diagnosis, as established by the ISPD guidelines, relies on the presence of any two of the following features: (1) abdominal pain or cloudy effluent, (2) PD cell count of >100 cells/µL or PMNs >50% after a 2-hour dwell, and (3) positive dialysate culture. Patients on automated peritoneal dialysis (APD) with rapid cycles, such as acutely ill patients being managed for volume overload, may not have sufficient dwell times to reach an absolute PD fluid leukocyte count of more than 100 cells/µL, in which case a differential of more than 50% consisting of neutrophils is indicative of infection. A predominance of eosinophils usually indicates noninfectious inflammation such as allergic reactions to icodextrin-based PD fluids or after catheter insertion, although infection must be ruled out. Patients occasionally may have nondiagnostic cell counts or negative cultures, and repeat testing after 12 to 24 hours may be indicated if clinical suspicion remains high. If signs of peritonitis persist despite negative cultures after 3 to 5 days, cultures in special media for mycobacteria, fungi, and fastidious organisms can be considered. Last, because critically ill patients are at high risk of developing peritonitis and may not be able to complain of abdominal symptoms, careful monitoring of effluent is critical. In such cases, daily PD cell counts for screening may be useful for the rapid detection of peritonitis. This is particularly important in acute PD and may be accomplished in resource-poor settings by daily effluent testing for leukocytes with urinalysis Chemstrips.

Certain findings may raise suspicion for underlying abdominal or surgical pathology. These include localized abdominal pain or tenderness, the isolation of multiple enteric organisms, greater severity of presentation, and persistent signs of infection despite appropriate initial treatment. Peritoneal fluid analysis may aid in the diagnosis of secondary peritonitis, with amylase levels above 50 IU/L or lipase levels greater than 15 being suggestive. Peritoneal free air often is seen in PD patients and is not necessarily indicative of bowel perforation. Abdominal computed tomography is rarely useful and does not definitely eliminate an underlying process. Surgical evaluation should not be delayed when secondary peritonitis is suspected, irrespective of imaging findings.

Antibiotic Therapy

The cornerstone of effective PD peritonitis management is the prompt administration of empiric antibiotic therapy once a presumptive diagnosis has been made. Given the frequency of gram-positive skin flora and gram-negative enteric organisms, empiric therapy should be simultaneously directed toward both. For gram-positive coverage, recommended agents include first-generation cephalosporins or vancomycin. The latter is preferred in centers where methicillin resistance is common. Gram-negative coverage usually is obtained with third- or fourth-generation cephalosporins or aminoglycosides. A typical empiric regimen may consist of cefazolin and ceftazidime administered intraperitoneally (IP).

Concern over accelerated loss residual renal function (RRF) may lead to avoidance of aminoglycosides, although there is some evidence that short-term aminoglycoside therapy does not significantly affect RRF. It is nonetheless our practice to avoid aminoglycosides when patients produce more than 100 mL of urine daily, given the association of RRF with long-term outcomes.

Aminoglycosides are associated further with significant ototoxicity, and if selected as initial therapy, switching antibiotics as soon as susceptibility results are available may be ideal. Other appropriate agents for gram-negative coverage include aztreonam in penicillin-allergic patients and, local resistance patterns permitting, fluoroquinolones. Therapy eventually should be narrowed based on results of cultures and sensitivities. Re-culture and serial PD cell counts can be monitored to follow the response to therapy. Duration of antibiotic therapy depends on clinical context and the organism isolated. The reader is referred to the 2016 ISPD guidelines on management of PD peritonitis for further details ( www.ispd.org ).

In atypical cases, initial antibiotic choice may be individualized. A recent Cochrane review was unable to demonstrate any consistent benefit of one antibiotic regimen over another because of the general poor quality and heterogeneity of available studies. In cases of suspected surgical peritonitis, the ISPD recommends initial therapy with metronidazole, IP vancomycin, and gram-negative coverage with either ceftazidime or an aminoglycoside. Suspected fungal peritonitis should be treated with azole or echinocandin therapy. Most antibiotics appear to be compatible with coadministration in a single dwell, with the possible exception of penicillins and aminoglycosides. Vancomycin and ceftazidime should be mixed in a solution with a volume of at least 1L. Icodextrin has been shown to be compatible with vancomycin, aminoglycosides, cefazolin, and ceftazidime. In the case of multi-drug–resistant organisms, consultation with an infectious diseases specialist is advised.

The ideal route for antibiotic administration appears to be IP, because inflammation is likely limited to few cell layers beyond the peritoneum and bacteremia is rare. IP administration results in high levels at the site of infection, and enough antibiotic is thought to be absorbed systemically during peritoneal inflammation to provide continuous peritoneal drug delivery. Fluoroquinolones, if deemed appropriate, are administered orally.

Antibiotics may be given in one prolonged dwell daily (intermittent administration) or in each exchange (continuous administration). Vancomycin in particular is best dosed intermittently. One recommended regimen would be to monitor vancomycin blood levels daily and redose when levels fall below 15 µg/mL, or alternatively readminister doses every 4 to 5 days. The higher the RRF, the more frequently the drug will have to be given. Aminoglycosides appear to be equally effective with both methods of dosing, but intermittent administration may be preferable to avoid toxicity. For all other antibiotics, we prefer intermittent over continuous dosing because of ease of administration and efficacy. Antibiotics should be allowed to dwell for at least 6 hours. As mentioned earlier, continuous peritoneal antibiotic delivery with intermittent dosing is obtained through diffusion of the antibiotic back into the cavity from the blood. In patients on automated PD, enhanced antibiotic clearance may be seen, which can be circumvented either through continuous administration where possible or switching to manual exchanges for the duration of antibiotic therapy.