Peritonitis management in the CAPD program at the University of Missouri -Columbia November 1983 - October 1984

Zbylut J. Twardowski, Karl D. Nolph, Ramesh Khanna, Leonor P. Ryan, Barbara F. Prowant

During the periodfrom November 1983 to October 1984 there were 52 episodes of clinical peritonitis in our 63 CAPD patients for a peritonitis rate of one episode every 14.5 patient months. Thirty-nine episodes (75%) were treated at home. All episodes were ultimately cured including 6 ( 12% ) after catheter removal for refractory peritonitis. Five patients ( 10% ) were permanently transferred to hemodialysis because of repeated bouts of peritonitis. There was no mortality associated with peritoneal infections. probably because of our policy of early catheter removal in refractory peritonitis.

Peritonitis continues to be a frequent complication of CAPD leading to morbidity, hospitalization, transfer to hemodialysis and sometimes death. The USA CAPD Registry reports that peritonitis was the reason for 27% of the transfers to other forms of therapy (I).

Our program started in January, 1977 (2), and has been steadily growing. More than half of our patients live further than 50 miles from Columbia; therefore, telephone contact and use of local hospital facilities constitute the basic method of peritonitis management in this segment of our patient population. Through the years our management of peritonitis has been evolving. Our peritonitis experiences from 1977-1983 have been published in several reports (3-6). This paper will present our current approach and report data covering the period from November 1, 1983 to October 31, 1984.

The decision-making trees used for peritonitis management are presented in Figures 1-9. Numbers in parentheses correspond to the number of patients in each category .The figures are mostly self-explanatory , however, some terms used in the figures are not generally accepted and require definitions.

The definitions are as follows:

Clinical peritonitis -abdominal symptoms and/or cloudy drainage with dialysate cell count equal to or over 100/mm3 and more than 50% neutrophils. We always assume that the peritonitis is infectious and start treatment with antibiotics.

Abating peritonitis -symptoms and signs begin to abate and dialysate is less cloudy within 24 hours of antibiotic treatment.

Unabating peritonitis -no improvement within 24 hours despite antibiotic treatment if symptoms do not qualify for refractory peritonitis. (See below.)

Delayed abating peritonitis -unabating peritonitis improving within 48 hours, usually afterpreliminary culture results and changing antibiotic.

Resistant peritonitis -peritonitis not improving within 48 hours, i.e. symptoms do not qualify for refractory peritonitis. (See below.)

Late abating peritonitis -resistant peritonitis improving after change of antibiotics according to sensitivity.

Refractory peritonitis -when any of the 4 following conditions are present refractory peritonitis is diagnosed:

Early successive episodes ofperitonitis -peritonitis episodes occurring within 3 months, but the organism may be different. If the same organism is identified, the occurrence time is more than 15 days.

During the period under consideration 78 patients were treated with CAPD at our institution for a total duration of 756 patient months (average patient population 63). Symptoms suggesting peritonitis occurred 55 times. The patients called a CAPD nurse on duty who made the preliminary assessment (Figure 1). At this initial contact no patient was considered to have severe enough symptoms to notify a physician immediately. Twenty-seven patients came to our out-patient clinic, 15 patients went to a local hospital emergency room and in 13 instances a family member took a cloudy dialysate bag to a local hospital with the patient remaining at home.

Figure 2 portrays the initial evaluation. Patients seen in the clinic or emergency room were clinically assessed by a nurse. Two in and out dialysis solution exchanges were performed while awaiting the results of dialysate cell count and Gram's stain. The patients visiting local hospitals performed exchanges after returning home. When dialysate cell count and Gram's stain results were available a physician was notified. Diagnosis of clinical peritonitis was made in 52 instances (Figure 2). In three instances dialysate was cloudy because of fibrin clots or red blood cells and abdominal symptoms were unrelated to peritonitis.

Gram's stain results dictated the choice of initial antibiotics (Figure 3). Forty-four instances with no organisms and 6 with a gram positive organism were treated with intraperitoneal Kefzol and intramuscular Tobramycin. Tobramycin alone was used for initial treatment in two instances when gramnegative organisms were seen. Maintenance intraperitoneal Kefzol was continued in 50 instan~es and Tobramycin in two. Eight patients were admItted to the hospital because they were unable to care for themselves and did not have an available helper. No patient met the criteria to qualify for the diagnosis of refractory peritonitis at the initial contact. All patients were continued on a regular CAPD e~change schedule with Kefzol and/or tobramycm (Figure 3). Within 24 hours symptoms improved ~n 37 episodes (abating peritonitis) and in 15 they dId not (unabating peritonitis).

Culture results for 37 episodes of abating peritonitis are shown in Figure 4. Patients with a gram positive organism or no growth were continued ~n maintenance Kefzol. Intraperitoneal Tobramycm was begun in two patients whose dialysate grew gram negative organisms on culture. Thirty-five patients continued to improve (abating peritonitis) and were treated with the same antibiotics for ten days. Two patients' dialysate gradually beca~e more cloudy within the next 24 hours and their peritonitis was classified as resistant.

Among the fifteen unabating peritonitis patients (Figure 5) , three had refractory peritonitis, eight improved (delayed abating peritonitis) and four developed resistant peritonitis.

Management of six patients with resistant peritonitis is shown in Figure 6. Two patients, who were on inappropriate antibiotic according to sensitivity testing, responded to treatment when proper antibiotic was given. Two patients who were on antibiotics to which organisms were sensitive did not improve even after the addition of a synergistic antibiotic. The two patients with culture negative peritonitis were treated with another antibiotic: one patient improved and one did not.

Eleven patients (Figure 4) had delayed or late abating peritonitis and were cured within ten days. All in all peritonitis was cured without catheter removal in 46 out of 52 instances.

Refractory peritonitis developed in 6 patients (Figure 7, Table I), with apparent exit/tunnel infection in five. The sixth patient did not have an apparent reason for the refractory peritonitis; cultures were negative, ultrasound and CT scan were negative for any intra abdominal pathology. In all these patients peritonitis resolved quickly when the catheters were removed and systemic antibiotics were continued. We assume that refractory peritonitis indicates tunnel infection, a ruptured viscus, or catheter seeding with microorganisms. We believe that in this situation early catheter removal decreases morbidity, reduces protein losses, and prevents malnutrition. Therefore, once the diagnosis of refractory peritonitis is made we always remove the catheter and the patient is transferred to hemodialysis. Two of our refractory peritonitis patients were transferred permanently to hemodialysis because of the previous multiple episodes of peritonitis.

Figure 8 presents our approach to a patient with a relapse or early successive episode of peritonitis. Relapses were not observed during the period, but early successive episodes occurred in 12 instances. In only two patients was the same organism identified in early successive episodes. After the second early episode the treatment was prolonged for 4 more days beyond the usual 10 days and the patient retrained.

Our approach in 4 instances of third early episodes of peritonitis is portrayed in Figure 9. Different organisms were grown in the dialysate during successive episodes of one patient and another patient had a culture negative peritonitis. This last patient was known to be non-compliant in the past and we had a justified suspicion that the sample was not delivered to the local hospital according to the protocol. This patient was permanently transferred to hemodialysis because of multiple previous episodes of peritonitis. The first patient was retrained, and the CAPD exchange system was altered. However, the patient developed a fourth early episode and peritonitis and was subsequently transferred permanently to hemodialysis (Figure 8) . The third patient (Figure 9) had the same grampositive organism with successive episodes which indicated the presence of an infectious nidus on the intraperitoneal part of the catheter or in the tunnel. Despite the fact that the peritonitis episode was cured the catheter in this patient was removed. The fourth patient with known extensive diverticulosis had gram-negative organisms in successive episodes. Because we had a strong suspicion of an internal source of infection this patient was transferred permanently to hemodialysis in spite of good response to treatment.

A summary of peritonitis episodes is given in Table I. Table II summarizes the hospital admissions related to peritonitis. Transfers to hemodialysis and catheter removals are summarized in Table III and IV respectively.

During the period under consideration our peritonitis rate was one episode every 14.5 patient months. All peritonitis episodes were ultimately cured, six after catheter removal. Five patients were permanently transferred to hemodialysis, one because of an internal source of infection and four because of poor and uncorrigible technique. There was no mortality associated with peitonitis which, we believe, may be related to our early catheter removal in refractory peritonitis.

The peritonitis management as presented above has served well for our patient population but is not considered as universal. We may safely predict that the decision making tree will be modified in the coming years.


From the Division of Nephrology. Department of Medicine. University of Missouri, Columbia, Missouri.