We assessed the efficacy of perioperative antibiotic prophylaxis for surgery in a randomized, double-blind trial of 1218 patients undergoing herniorrhaphy or surgery involving the breast, including excision of a breast mass, mastectomy, reduction mammoplasty, and axillary-node dissection. The prophylactic regimen was a single dose of cefonicid (1 g intravenously) administered approximately half an hour before surgery. The patients were followed up for four to six weeks after surgery. Blinding was maintained until the last patient completed the follow-up and all diagnoses of infection had been made. The patients who received prophylaxis had 48 percent fewer probable or definite infections than those who did not (Mantel-Haenszel risk ratio, 0.52; 95 percent confidence interval, 0.32 to 0.84; P = 0.01). For patients undergoing a procedure involving the breast, infection occurred in 6.6 percent of the cefonicid recipients (20 of 303) and 12.2 percent of the placebo recipients (37 of 303); for those undergoing herniorrhaphy, infection occurred in 2.3 percent of the cefonicid recipients (7 of 301) and 4.2 percent of the placebo recipients (13 of 311). There were comparable reductions in the numbers of definite wound infections (Mantel-Haenszel risk ratio, 0.49), wounds that drained pus (risk ratio, 0.43), Staphylococcus aureus wound isolates (risk ratio, 0.49), and urinary tract infections (risk ratio, 0.40). There were also comparable reductions in the need for postoperative antibiotic therapy, non-routine visits to a physician for problems involving wound healing, incision and drainage procedures, and readmission because of problems with wound healing. We conclude that perioperative antibiotic prophylaxis with cefonicid is useful for herniorrhaphy and certain types of breast surgery.

The intestinal protozoan cryptosporidium is known to cause diarrhea in immunocompromised patients, but few cases have been reported in detail in immunocompetent persons. During a 12-month period, we identified cryptosporidium in the stools of 43 immunocompetent patients. The numbers of cases were increased in those under 4 years old and in those from 30 to 39 years old. Of 30 index cases, 23 (77 per cent) were diagnosed in the late summer or the fall. Fifteen of the 43 patients (35 per cent) had other gastrointestinal pathogens, of which only Giardia lamblia was statistically associated with cryptosporidium. In the 28 patients in whom other gastrointestinal pathogens were not identified, the clinical manifestations were predominantly watery, nonbloody diarrhea and, less commonly, abdominal discomfort, anorexia, fever, nausea, and weight loss. The infection was self-limited in all 43 patients. Clustering of cases occurred in a day-care center and in two families. These clinical observations confirm worldwide findings and suggest that cryptosporidium is a relatively common nonviral cause of self-limited diarrhea in immunocompetent persons in the northeastern United States.

catheters that are sources of sepsis and those that are not; however, this method requires removal of the catheter to establish the diagnosis. The

We conducted a study to determine the usefulness of the Gram stain in the detection of intravascular catheter-associated infection. A total of 330 intravascular catheters were prospectively collected from adults and children suspected of having such an infection. Semiquantitative solid-agar cultures of the distal catheter tip were correlated with blood cultures. Catheter-associated bacteremia occurred in 34 per cent of cases in which catheter tips were colonized (greater than or equal to 15 colonies per agar plate). There were no cases of catheter-associated bacteremia in patients with uncolonized catheters. Immediately after culture, whole catheter segments were stained by the Gram technique. Gram-negative and gram-positive bacteria and yeast were easily identifiable under oil immersion (X 1000), located predominantly on external catheter surfaces. Any catheter with at least one organism per 20 oil-immersion fields was designated as positive by Gram stain, but the majority of the 41 positive catheters had much larger numbers of organisms. The Gram stain of the catheter tip was 100 per cent sensitive and 96.9 per cent specific for the detection of catheter-tip colonization, with positive and negative predictive values of 83.9 and 100 per cent, respectively. We conclude that a Gram stain of the distal catheter tip is a simple, inexpensive, and accurate test for the rapid diagnosis of intravascular catheter-associated infection.