Peritonitis Caused by Haemophilus parainfluenzae in a Patient Undergoing Continuous Ambulatory Peritoneal Dialysis

Abstract

We report a case of peritonitis in a patient undergoing continuous ambulatory peritoneal dialysis. Haemophilus parainfluenzae biotype III was recovered in pure culture from dialysis fluid.

CASE REPORT

Peritonitis is a common complication of continuous ambulatory peritoneal dialysis (CAPD), with two-thirds of patients developing peritonitis during the first year of dialysis (8). Three cases of CAPD peritonitis due to Haemophilus parainfluenzae have been reported (1, 3, 10), and none of these mentions bacteremia.

A 44-year-old patient with chronic renal failure, probably due to nephrosclerosis, started on CAPD in April 1997. He had a history of coronary disease treated by angioplasty and stent implantation, peripheral arteriopathy with left femoral stenosis, and long-term arterial hypertension. The patient was in good health from the beginning of CAPD treatment and had no episodes of peritonitis. He was admitted to our Peritoneal Dialysis Unit in November 1998 with abdominal pain and cloudy peritoneal fluid. The day before admission he developed diarrhea. On admission he was afebrile, and the peritoneal fluid was cloudy and contained 1,860 leukocytes/mm³ with 80% polymorphonuclear neutrophils (PMN). Physical examination revealed marked abdominal distension, increased bowel sounds, and abdominal tenderness. The peritoneal catheter exit site was of normal appearance. A blood test revealed 10,710 leukocytes/mm³ with 66% PMN, a hematocrit value of 42%, and a hemoglobin concentration of 13.1 g. As the patient was afebrile, blood cultures were not collected. After taking samples of dialysis fluid for Gram staining and culture, teicoplanin and cefotaxime were administered intraperitoneally.

H. parainfluenzae biotype III was recovered in pure culture from dialysis fluid, although no organisms were detected on a Gram stain of the centrifuged sample. Five-milliliter samples of fluid were inoculated in each of two blood culture bottles of the Bactec system NR 860 (Johnston Laboratories, Inc., Towson, Md.) Subcultures were performed on blood agar and chocolate agar, which were incubated in a CO₂-enriched atmosphere at 37°C, and on Brucella agar incubated in an anaerobic atmosphere. Gram-negative coccobacilli grew after 24 h on chocolate agar. They were identified as H. parainfluenzae on the basis of V-factor requirement and absence of hemolysis. Identification of the isolate to species level was confirmed and biotyping was performed by using growth-independent rapid biochemical tests of the API NH system (bioMérieux, Marcy l’Etoile, France). Susceptibility to antimicrobial agents was determined by a broth microdilution procedure (Sensititre; Radiometer, Copenhagen, Denmark) performed according to the guidelines of the National Committee for Clinical Laboratory Standards (6). The isolate was β-lactamase positive and susceptible to amoxicillin-clavulanate, cefotaxime, and chloramphenicol at the following MICs: 2/1, ≤0.06 μg/ml, and ≤2 μg/ml, respectively. With the culture identification of H. parainfluenzae, cefotaxime was continued for a total of 10 days, but teicoplanin was stopped. The patient responded well to this treatment, and symptoms disappeared 2 days after therapy was started. We investigated the presence of H. parainfluenzae in the patient’s feces and throat swab. This organism was not isolated.

Unlike surgical peritonitis, CAPD peritonitis is usually caused by a single organism, and bacteremia is extremely rare (8, 9, 11). Microorganisms causing CAPD peritonitis usually originate in the skin but could also originate in either the upper respiratory tract or the bowel. Coagulase-negative staphylococci are the most commonly identified agents, accounting for 40 to 60% of all positive cultures, followed by S. aureus and streptococci (10 to 20% each). Members of the family Enterobacteriaceae accounted for 5 to 20% of all positive cultures, nonfermentative gram-negative rods accounted for 3 to 15%, and gram-positive rods accounted for 2 to 4%. Values for mixed bacteria, fungi, mycobacteria, and anaerobes are generally <5% (11).

H. parainfluenzae is a commensal organism of the upper respiratory tract and has also been isolated from the urogenital tracts of children and adults. H. parainfluenzae can colonize the gastrointestinal tract; this fact is supported by some observations (2, 4, 5). The characterization of the iron-repressible outer-membrane proteins of H. parainfluenzae showed that they are closely related to those of enteric bacteria (5); these proteins may serve as adhesins, allowing colonization of the intestinal tract mucosa (2). Furthermore, H. parainfluenzae is dependent for growth upon V factor, which is supplied by intestinal flora. In a study of Haemophilus species in the human gastrointestinal tract, Megraud et al. (4) found that most of the strains isolated from the appendices were H. influenzae (77%) whereas 60% of the strains from the stools were H. parainfluenzae. This microorganism is an uncommon agent of human infections and has occasionally been associated with upper respiratory tract infections, endocarditis, osteomyelitis, psoas abscess, retroperitoneal abscess, peritonitis, urinary infections, liver abscess, and biliary tract infections. It is a very rare cause of CAPD peritonitis (1, 3, 10).

The three sites most frequently associated with CAPD infections are the catheter exit site, the subcutaneous peritoneal catheter tunnel in the abdominal wall, and the peritoneum itself (11). This episode is an unusual case of CAPD peritonitis. The source of H. parainfluenzae was not identified. Considering that this organism is a normal inhabitant of the human respiratory tract and may also colonize the gastrointestinal tract, we suggest that, in our case, the entry of H. parainfluenzae into the peritoneal cavity could have been by two routes: (i) through touch contamination during dialysate exchanges, or (ii) by transmural migration of H. parainfluenzae from the gastrointestinal tract, as some studies have indicated for Escherichia coli (7).

We believe that Haemophilus species should be considered possible etiologic agents of CAPD peritonitis. These organisms have fastidious nutritional requirements. H. parainfluenzae requires V factor (NAD) and growth in the laboratory only when it is provided with medium, such as chocolate agar. The low incidence of Haemophilus species in CAPD peritonitis could be explained, in part, by the fact that the majority of laboratories do not include plating the peritoneal fluid to a medium to ensure the growth of these microorganisms.