ICAAC/IDSA/ASTMH 2003 Conference Coverage

The following summary of selected abstracts from 3 meetings will be published in multiple parts. The 43rd Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC) met in Chicago September 14-17, 2003. The Infectious Disease Society of America (IDSA) met in San Diego October 9-12, 2003. The American Society of Tropical Medicine and Hygiene met in Philadelphia December 3-7, 2003. — Stan Deresinski, MD, FACP

Gastrointestinal Infection

C difficile

Piperacillin-tazobactam exposure failed to induce C difficile toxin production in a human chemostat gut model. Increased use of piperacillin/tazobactam was reported to be associated with reduced rates of C difficile diarrhea. Separately, a 68% reduction in piperacillin-tazobactam use resulting from a nationwide shortage, with a concomitant increased use of other antibiotics, was associated with a 47% decrease in the incidence of C difficile colitis (ICAAC K-728, K-730, K-732).

In addressing the issue of how many tests for C diffi cile toxin are necessary to effectively rule out its presence, a retrospective review found that 97.4% of cases of C difficile disease were detected with the first 2 cytotoxicity assays performed (IDSA 336).

Given the apparently increased frequency of treatment failure and relapses of C difficile-associated disease, novel effective therapies are urgently required. Ramoplanin was active against strains of C difficile resistant to vancomycin or metronidazole. Five patients with protracted and/or recurrent diarrhea due to C difficile were given 1-6 doses of intravenous immunoglobulin. Three "had a good therapeutic response," and 1 had recurrence at 6 weeks. The patient who received 6 doses died of intractable C difficile-associated disease (ICAAC K-729, E-2188).

E coli 0157:H7

A retrospective review of 362 patients with E coli 0157:H7 enteritis found that 16% who received antibiotics developed hemolytic uremic syndrome, while 29% of those who did not developed this complication. This finding stands in contrast to previously published studies that have reported that antibiotic therapy of this infection is associated with an increased risk of hemolytic uremic syndrome (IDSA 864).


Multistate outbreaks of S newport infection were associated with fresh tomatoes from a single packinghouse and with imported honeydew melon. An outbreak of S javiana infections occurred in organ transplant recipients who had attended the US Transplant Games, a 4-day athletic competition among 1500 recipients held at a Florida theme park (IDSA 870, LB-9, 871).

Children with typhoid fever were treated with either oral azithromycin for 5 days (n = 32) or IV ceftriaxone for 5 days (n = 36). Blood cultures remained positive on day 3 in 37% of azithromycin recipients but in no ceftriaxione recipients. All day 8 blood cultures were sterile, and microbiological cure was achieved in 100% and 97%, respectively, and clinical cure in 94% and 97% (IDSA 869).

V cholerae

A single dose of ciprofloxacin was as effective as 12 doses of erythromycin in the treatment of childhood cholera in a randomized trial in South Africa (ICAAC G-1551).


Outbreaks of norovirus acute gastroenteritis occurred during several consecutive voyages of a cruise ship. Investigation suggested both foodborne and person-to-person transmission. Detection of identical strains of norovirus among case passengers before and after ship sanitization suggested environmental contamination (IDSA 880).

Despite early institution of control measures, an outbreak of norovirus on a ship continued throughout its cruise, affecting 399 of 1038 (38%) passengers. The strain of norovirus detected was genetically indistinguishable from that identified during an outbreak on the same ship that occurred 15 months earlier. Smaller outbreaks involving 3%, 13%, and 5% of passengers occurred on the next 3 voyages of the ship, but none has occurred since the ship was removed from service and aggressively cleaned. Epidemiologic investigation suggested person-to-person transmission (IDSA 879).

Cruise ships are not the only site of norovirus outbreaks. The overall mortality rate in a nursing home during a norovirus outbreak was 3 times higher than in the months preceding and following it (IDSA 524).

Traveler’s Diarrhea

Traveler’s diarrhea is not always a trivial self-limited event. Six of 62 students who developed diarrhea while in Mexico subsequently developed irritable bowel syndrome (IDSA 876).

Bloodstream Infection, Endocarditis

Patients with peripheral intravenous catheters were randomized on day 3 of catheterization to either have routine catheter replacement every 3 days or no replacement. Routine, peripheral IV-catheter change did not reduce the incidence of phlebitis or of a positive catheter culture (ICAAC K-2041).

Infants and children who had undergone cardiac surgery were randomized to have a transparent dressing alone or covering a chlorhexidine-impregnated dressing placed over their central venous catheter access sites. While the chlorhexidine dressing was associated with reduced site colonization, there was no difference in related bloodstream infections (4% in each group) (ICAAC K-580).

A meta-analysis of 40 studies concluded that the most accurate diagnostic method for intravascular device-related bloodstream infection was the use of paired quantitative blood cultures. The most accurate rapid test not requiring catheter removal was the acridine orange leukocyte cytospin. (For a recent prospective analysis of the latter technique, see JPEN J Parenter Enteral Nutr. 2003;27:146.) Separately, the use of quantitative blood cultures obtained simultaneously through a central venous catheter and from a peripheral vein had sensitivity, specificity, and negative predictive values of 94-98% when the CVC specimen had > 15 CFU (ICAAC K-2038, D-1698). It has also been recently reported that a differential time to blood culture positivity between central and peripheral samples ³ 120 minutes is highly accurate in the diagnosis of catheter-related bloodstream infection (Ann Intern Med. 2004;140:18).

The necessity for immediate IV catheter removal in patients with suspected infection of a vascular access device was addressed in a clinical study. Sixty-four patients with suspected catheter-related infection, but who were hemodynamically stable and did not have proven bacteremia, insertion-site infection, or an intravascular foreign body, were randomized to immediate catheter removal or "watchful waiting." There was no difference in outcome associated with the 2 approaches (ICAAC K-2039).

Valve surgery in patients with complicated left-sided native-valve endocarditis was independently associated with reduced mortality at 6 months, especially in patients with moderate-to-severe congestive heart failure (IDSA 492).


Deterioration or lack of improvement in hemodynamic status, fever, and leukocyte count during the first 48-72 hours of sepsis management is associated with a 62% likelihood that the empirical antibiotic therapy administered is inappropriate. Examination of 1342 patients entered in a multicenter, controlled trial of etanercept in the treatment of severe sepsis found that inappropriate initial antimicrobial therapy was independently associated with increased mortality (OR, 1.5; P = .02) (ICAAC L-118, L-112).

A meta-analysis of 64 randomized trials involving non-neutropenic patients with sepsis found that the combination therapy with an aminoglycoside plus a b-lactam was not associated with reduced mortality compared to monotherapy with a b-lactam (ICAAC L-621).


Maybe it’s OK to eat the salad after all! Twenty patients undergoing remission induction therapy for acute leukemia with selective decontamination of the digestive tract were randomized to receive either a low-bacterial or a standard hospital diet. There was no difference in fecal contamination with potential pathogens or in number of infections (ICAAC K-1372).

Ninety-two percent of 36 episodes of low-risk febrile neutropenia treated with orally administered gatifloxacin responded, with defervescence occurring at a median of 2 days. A meta-analysis of 15 trials comparing oral to intravenous antibiotic therapy for febrile neutropenia in the absence of hemodynamic instability, altered mental status, end-organ failure, inability to swallow, pregnancy, and lactation found that the 2 strategies yielded equivalent results (ICAAC L-115, IDSA 375).

Two randomized trials found that piperacillin/ tazobactam and cefepime yielded comparable results as initial empiric antibacterial therapy in febrile neutropenic patients with hematologic malignancies (ICAAC L-114, IDSA 373).

In a randomized trial, caspofungin and itraconazole, each administered intravenously as antifungal prophylaxis in patients with myelodysplastic syndrome undergoing induction chemotherapy, yielded similar results (ICAAC M-984).

Caspofungin was compared to liposomal amphotericin B (3 mg/kg/d) as empiric antifungal therapy in persistently febrile neutropenic patients in a large, randomized trial. Caspofungin proved comparable to the liposomal amphotericin B in overall success and was better tolerated (ICAAC M-1761).

Surgical Infections

The isolation of clindamycin-resistant B fragilis from the peritoneal cavity of patients with intra-abdominal infection was associated with worse outcome than was isolation of susceptible isolates in a case-control study. Cure rates at end of treatment in patients with intra-abdominal infections originating in the colon were 83% both in patients randomized to receive ertapenem 1 g q24h and in those randomized to receive piperacillin/tazobactam 3.375 g q6h (ICAAC K-563, IDSA 328).

A study of 30 cirrhotics with peritonitis and bacteremia due to E coli found that the expected increase in C-reactive protein, a protein synthesized by the liver, was intact but attenuated relative to patients without liver dysfunction (IDSA 733).

Only 2.45% of 937 patients undergoing bariatric surgery developed nosocomial infection, but 21.7% of the 23 infected patients died (IDSA 546).

Forty percent of surgical-site infections after mastectomy and breast reconstructive surgery were initially diagnosed > 30 days after surgery (ICAAC K-1302).

The use of nasal mupirocin was associated with a reduction in deep sternal wound infections due to S aureus in patients undergoing coronary artery bypass grafting, when compared to noncontemporaneous controls (IDSA 560).

A meta-analysis of 7 randomized trials found that vancomycin prophylaxis was not more effective than beta-lactam prophylaxis in patients undergoing cardiac surgery (ICAAC K-1295).

Orthopedic Infections

Sixteen of 509 (3%) patients undergoing revision arthroplasty were discovered to have previously unsuspected prosthetic joint infection. One-half were due to coagulase-negative staphylococci and one-quarter to Propionibacterium spp. Nine had antibiotic impregnated cement placed at revision, and 12 received IV antibiotics for a median of 28 days (range, 2-42). Nine received chronic oral suppression. One failure occurred 2.9 years after revision and consisted of infection with the "same" organism with a different antibiogram (IDSA 579).

The synovial fluid of patients with aseptic knee implant failure (n = 21) or infection (n = 21) who did not have underlying inflammatory joint disease was prospectively studied. Gross purulence was absent in the aseptic group but was present in two-thirds of the latter. The synovial fluid of the aseptic group had a lower mean WBC count (742/mm3 vs 56,422/mm3) and percentage of neutrophils (15% vs 94%). While there was overlap in the total WBC, there was none when the percentage of neutrophils was examined (ICAAC K-568).

Ninety-nine of 509 (19%) episodes of prosthetic joint infection were treated with debridement and retention of the prosthesis with a 2-year cumulative probability of success of 59%. IV antibiotics were administered for a median of 28 days, with suppressive oral antibiotics given for a median duration of 280 days to 89% of patients. Variables associated with an increased risk of treatment failure included S aureus infection, the presence of a sinus tract at presentation, and a duration of symptoms prior to debridement of more than 7 days (IDSA 493).

The median interval between resection and reimplantation arthroplasty in 232 patients with prosthetic joint infection treated with a 2-stage exchange was 67 days (range, 8-1900). Cultures were negative at the time of reimplantation in 91%. An antibiotic impregnated spacer was placed in 72%, and the median duration of IV antibiotic therapy was 42 days. The 2-year probability of success was 85% but was only 70% in 23 patients with acute inflammation present at the time of reimplantation (IDSA 283).

Twenty patients with staphylococcal or enterococcal infections involving orthopedic hardware were treated with linezolid, as well as debridement and, in 10 patients, device removal. Patients were followed for a mean of 9 months after completion of 5-422 days (median, 32 days) of therapy. Clinical cure or improvement occurred in 90%, but bacterial persistence was documented in 3 patients (15%). Eight (40%) developed reversible "myelosuppression," and 1 (5%) developed irreversible painful neuropathy. Separately, 53 patients with osteomyelitis were treated with linezolid for a mean duration of 54 days (range, 1-390 days). Clinical cure or improvement was achieved in 72% with follow-up of 6-24 months. Treatment was "well tolerated" (IDSA 318, 319).

A retrospective review of 639 patients with vertebral osteomyelitis found that the mean duration of IV antibiotic administration was 6.4 weeks, and total duration of antibiotic therapy (IV plus oral) was 13.4 weeks; 2 patients remained on suppressive therapy. Only 1 relapse of infection occurred—the affected patient had received IV antibiotic for 12 weeks (IDSA 322).

Stan Deresinski, MD, FACP Clinical Professor of Medicine, Stanford; Associate Chief of Infectious Diseases, Santa Clara Valley Medical Center, is Editor of Infectious Disease Alert.