Summaries from the 36th Meeting of the Infectious Disease Society of America: Part I

Conference Coverage

Note: The following summaries represent a selection of papers from those presented at the 36th Meeting of the Infectious Disease Society of America (IDSA), held November 12-15, 1998, in Denver. It is important to recognize that many of these summaries are extracted only from the published abstract and it is possible that some of the material presented at the conference may have differed. The abstracts are available at http://www.marathonmultimedia.com/abstracts/idsa.— Stan Deresinski, MD, FACP

Infectious Disease Consultation/Antibiotic Optimization and Resistance

Two studies added to the growing body of data demonstrating the value of expert intervention in antibiotic use and the management of infectious diseases.

In a university hospital using both a restricted formulary and a computer-generated antibiotic order form as part of an antibiotic quality improvement program, 82% of antibiotic orders over a 16-day period were for empiric antimicrobial therapy. Monitoring demonstrated that the rationale for use was correct in 86% of cases in which an infectious disease consultation had been obtained and in only 64% in which it had not (P = 0.046). The antibiotic order was correct in 61% and 39% (P = 0.069) without ID consultation; dosing errors were present, respectively, in 0% and 32% (P = 0.0002). (Abstract 679.)

A retrospective analysis found that an antimicrobial management program at a large university hospital was associated with a total cost savings of approximately $1.8 million, of which $363,000 was directly due to reduced antibiotic costs. Furthermore, the failure rate, defined as the occurrence of clinical failure, microbiological failure, or superinfection, was 5% among patients subjected to the antibiotic management program and 19% among those subject to "usual practice" (P < 0.05). (Abstract 33.)

A retrospective study of 504 ICU patients found that only 6.9% of recipients of imipenem-based regimens had microbiologically inadequate therapy, as defined by the presence of a mismatch between the spectrum of activity of the antibiotic regimen used and the susceptibility of the organisms cultured. The comparable figures for those receiving piperacillin/tazobactam-based regimens was 8.4% while for ceftazidime-based regimens it was 14.3%. Both comparisons with ceftazidime were statistically significant. Although crude mortality was similar in those who received microbiologically adequate antibiotic therapy when compared to those who received inadequate therapy, the median ICU stay was three days shorter (P = 0.02) and the average ICU cost was approximately $15,000 greater (P = 0.003) for the former group. (Abstract 685.)

Cefepime use in ICU patients with nosocomial pneumonia was, when compared to ceftazidime use, associated with reduced administration of vancomycin and reduced costs. In addition, pathogen eradication was achieved in 77% of cefepime recipients and 55% of those given ceftazidime (P = 0.04). (Abstract 666.)

A multicenter prospective study of K. pneumoniae bacteremia found that isolates from 80 (20%) of 400 consecutive patients with this infection produced extended spectrum beta lactamase (ESBL). Among those infected with an ESBL-producer, mortality was significantly lower when a carbapenem was administered in the first five days when compared to use of a non-carbapenem (5% vs 43%; P = 0.01). However, only four patients were treated with cefepime to which 87% of isolates were susceptible; two of these died. The mortality rate among quinolone recipients was 21% when the isolate was susceptible to quinolones. There was no evidence that adding amikacin to an active beta-lactam or quinolone improved outcome. (Abstract 188.)

Seventeen (26%) of 65 patients developed severe arthralgias and myalgias during treatment of VRE infection with quinupristin/dalfopristin. (Abstract 608.) Rhone-Poulenc reported the occurrence of similar symptoms in 180 (15%) of 1179 recipients of the drug. Discontinuation of drug administration was rarely necessary as a consequence of this adverse event that had a peak onset after 5-6 days of therapy, with no new episodes occurring after 10 days. There was no chemical evidence of muscle damage and the symptoms were reversible in each case. (Abstract 237.)

VRE became resistant to quinupristin/dalfopristin during treatment of infection with this streptogramin in six (14.3%) of 42 patients. (Abstract 607.) On the other hand, a compilation presented by Rhone-Poulenc indicated that the during the worldwide emergency use program, emergence of resistance by any isolate was documented in only six (1.8%) of 338 cases and that treatment was, nonetheless, successful in two of the six cases with in vitro resistance. (Abstract 53.)

Microbiology Laboratory

The effect of inappropriately drawn blood cultures (IDBC) on vancomycin use was examined at the University of California at San Francisco (UCSF) and at Stanford. Stanford, but not UCSF, had in place a set of blood culture guidelines that required, whenever blood cultures were obtained, at least two blood cultures drawn within 24 hours of each other from at least two draws, one of which was from a venipuncture. Sixty-one percent of blood cultures drawn at UCSF and 41% of those at Stanford were IDBC (P = 0.067). Only 41% of blood cultures at UCSF were from venipuncture, compared to 61% of those at Stanford (P < 0.001). Most patients with IDBC were administered vancomycin, including 60% at Stanford and 74% at UCSF (P = 0.165). Thus, enforcement of guidelines for blood cultures may reduce vancomycin use, but additional measures dealing with overuse of this drug are necessary. (Abstract 686.)

Eighty-one isolates of Enterobacteriaceae from 41 hospitals identified locally by a variety of methods as being not susceptible (intermediate or resistant) to imipenem were tested at a central laboratory by broth microdilution where 91% proved to be, in fact, susceptible to this carbapenem. Similarly, 62 (27%) of 228 isolates of P. aeruginosa identified locally as not susceptible, were susceptible to imipenem by broth microdilution. These discrepancies were likely due to degradation of the drug due to inappropriate storage of imipenem-containing testing material. (Abstract 770.)

Nosocomial Infections

Increasing attention is being paid to infection control and transmission in outpatient settings. In a fascinating and instructive case, a receptionist in a dermatologist’s office developed toxic shock syndrome (TSS) with bacteremia due to Group A beta-hemolytic Streptococcus (GABHS). Three days later, and before the receptionist had returned to work, a young woman had a mole excised at the dermatologist’s office; 18 hours later, she developed erythema and pain at the biopsy site followed by evidence of TSS. Culture of the wound grew GABHS identical to that of the first patient by DNA analysis. Evaluation of office personnel and equipment failed to identify a source of infection. However, it was determined that the receptionist’s husband had been seen in the office because of a relapsing groin rash at the same time as the young woman. Culture of the rash, as well as the rectum of the receptionist’s husband, yielded GABHS identical to the other two isolates. (Abstract 631.)

Another outbreak of postpartum Group A streptococcal infection was traced to an obstetrician with rectal carriage of the organism. (Abstract 610.)

The results of one study in a surgical unit reflect poorly on either our methods of detection of nosocomial infection, the clinical acumen of surgical staff, or both. A prospective analysis of nosocomial infection rates based on the National Nosocomial Infections Surveillance (NNIS) definitions and interventions, such as the administration of antibiotic therapy, for infection was performed on a surgical service. The rates of infection determined by examination of interventions was approximately twice that of the NNIS rates. This difference was largely attributable to treatment of colonization and poor documentation. (Abstract 612.) This study opens an important area of investigation and intervention.

Handwashing continues to be often avoided by some healthcare workers. In some cases, this is due to hand irritation. The tolerability of standard soap and water handwashing was compared to the use of an alcoholic hand gel (Purellö) in a randomized crossover (2 weeks on each treatment) trial involving ward nurses. The use of the alcohol gel was associated with significantly less irritation and dryness. (Abstract 87.) The importance of handwashing even after the use of gloves was demonstrated in one study. Seventeen (39%) of 44 healthcare workers in contact with patients colonized by VRE acquired the patients’ VRE strain on their gloves and the same strain was also present on the hands of five of these workers after glove removal. Thus, acquisition of VRE is diminished, but not prevented, by gloving. (Abstract 599.)

Fungal pathogens may also be transmitted by the hands of healthcare workers. Forty-one (42.7%) of 96 of infants in a neonatal ICU examined over a period of three months were colonized at multiple skin sites with Malassezia pachydermatis; three-fourths of the infants weighed less than 1300 g at birth. While colonization of the hands of healthcare workers was absent at the time they entered the nursery, 11 (17%) of 64 brief encounters with colonized infants resulted in hand contamination. The near environment of colonized infants was also contaminated. (Abstract 587.)

It has been suggested that a benefit of contact isolation is improved handwashing, while critics suggest that it interferes with patient care. It turns out that both sides of the argument may be correct. Patient encounters in an ICU were observed to determine the effect of a policy of contact isolation for patients colonized or infected with resistant organisms. When compared to patients not on contact isolation, those on contact isolation precautions were almost two times less likely to have contact with a healthcare worker, although these contacts were slightly longer and were more likely to be accompanied by handwashing by the worker. The authors appropriately conclude that, "The overall impact of contact isolation on patient care appears complex and warrants further study to determine whether its net effect is positive or negative." (Abstract 680.)

Devices used in patient care may serve as a source of nosocomial pathogens. An outbreak of severe infections due to Baciuus cereus occurred in a neonatal ICU with the isolates appearing to represent a single strain as determined by DNA analysis. A case-control study identified mechanical ventilation as a major risk factor for infection. The outbreak strain was recovered in cultures from the hands of healthcare workers and in balloons used for manual ventilation and the outbreak ceased after the balloons were sterilized. (Abstract 627.)

The use of topical petrolatum ointment was found, in a case-control study, to be associated with the occurrence of systemic candidiasis in low birth weight infants. (Abstract 42.)

Semiquantitative culture of removed intravascular catheter tips may be of value when concomitant blood cultures are obtained in a patient suspected of having catheter-related bacteremia. In a review of 215 catheter tip cultures obtained in 85 patients, one or both of these criteria were absent 40% of the time. Seventeen percent of cultures were performed at the time of discontinuation of intravenous therapy in patients in whom there was no suspicion of catheter-related bacteremia. Blood cultures were obtained concomitantly with catheter tip cultures only 59% of the time. Antibiotic use was never influenced by negative catheter cultures. Positive catheter cultures resulted in no change in antibiotic therapy 81% of the time because existing therapy was adequate (49%) or because the results were regarded as insignificant (32%). In 17% of instances, antibiotics were started because of positive catheter cultures despite negative or no blood cultures. (Abstract 621.) This study documents a problem that negatively affects patient care and costs.

Increasing evidence supports the potential for human to human transmission of Pneumocystis carinii infection. P. carinii var. hominis DNA was detected by PCR in nasopharyngeal secretions of an 8-year-old boy with PCP as well as in individuals with more than five minutes of contact with him but not in any of 30 controls. (Abstract 85.) No adverse effects of acquisition of infection are seen (usually reinfection) with this organism in individuals who are immunologically intact. However, these observations provide robustness to the recommendation to prevent exposure of immunocompromised patients not receiving PCP prophylaxis to individuals with PCP.

A retrospective cohort study of 3753 inpatients found that patients who received ampicillin/sulbactam were at greater risk of colonization or infection with nosocomial pathogens, particularly Gram negative rods, than were patients who had received ceftriaxone. (Abstract 640.) This observation may be consistent with the concept of the role of gut anaerobes in the maintenance of colonization resistance.

Diarrhea is a common occurrence in hospitalized patients. That due to Clostridium difficile toxin production is frequently encountered and usually readily diagnosed. Evidence is accumulating that Bacteroides fragilis may produce an enterotoxin that also causes diarrhea. B. fragilis enterotoxin gene sequences were detected in seven (6.7%) stools from 104 patients with hospital-acquired diarrhea and in two (2.8%; P = 0.27) in an outpatient control group. Only one patient was simultaneously positive for this gene and for toxigenic C. difficile. (Abstract 625.)

A study involving five hospitals in Mexico found that 7.3% (range, 0.6-15.7%) of intravenous infusates in use were contaminated with bacteria. (Abstract 637.)

Gastrointestinal Tract Infections

A 27-year-old woman taking high dosages of ibuprofen developed epigastric pain and fever after exposure to an individual with streptococcal pharyngitis. Laparotomy revealed "diffuse inflammation and thickening of the stomach wall and biopsy revealed transmural suppuration with mucosal inflammation and necrosis and submucosal microabscesses and necrosis," as well as gram positive cocci in chains. ASO titer increased from 384 to 3320. The presumed diagnosis was acute streptococcal gastritis. (Abstract 67.) This was a new one to me—but a number of previous reports of this entity are in the literature.

The means of transmission of Helicobacter pylori are a matter of great interest. H. pylori was cultured from saliva of three of 12 infected volunteers; after induced emesis, it was recovered from five of the 12 with two individuals being positive both before and after emesis. H. pylori was obtained in culture from all 65 samples of vomitus obtained from these subjects. Air sampled at a distance of one foot from subjects had H. pylori detectable by molecular technique. (Abstract 16.) In addition, H. pylori infection of children in Peru is associated with an increased risk of diarrhea, which could represent an additional source of environmental contamination with this organism. (Abstract 204.)

Bacterial diarrhea continues to represent the most frequent illness acquired by travelers to developing countries. Enterotoxigenic E. coli accounted for 78.9% of 658 stool pathogens isolated from U.S. students with diarrhea while visiting Guadalajara during the summer months between 1989 and 1998. This was followed in frequency by Shigella (12.8%), Salmonella (5.0%), and Campylobacter (0.9%). The remaining 2.4% were comprised of eight cases of Pleisomonas, four of Aeromonas, and four of Edwardsiella. There has been a progressive increase in antimicrobial resistance over the 10 years culminating in 1998 with 100% of strains being resistant to doxycycline, furazolidone, and ampicillin. In addition, 84.7% were resistant to trimethoprim-sulfamethoxazole and 27.8% to fluoroquinolones. The highest previous frequency of resistance to fluoroquinolones in any one year previously studied was only 4.2%. (Abstract 204.)

Salmonella outbreaks were associated with homemade Mexican-style cheese (S. typhimurium), sprouts (S. anatum), chicken (S. enteritidis), eggs (S. heidelberg), toasted oats cereal (S. agona), and reptiles. (Abstract 18; Abstracts 538-542.)

Several outbreaks of E. coli 0157:H7 infection were described, with sources including contaminated cole slaw and a clover-alfalfa sprout mix, as well as visiting a water park. (Abstracts 535-537.)

Fecal contamination, as determined by the presence of E. coli, was detected in 47 (66.2%) of 71 samples of hot sauce, including green, guacamole, "pico de gallo," and red sauces, served in restaurants in Guadalajara, Mexico, in the summer of 1998. The mean level of contamination was 5866 colonies per gram of hot sauce. In samples obtained from Mexican restaurants in Houston, 10 (40%) of 25 were contaminated but with a mean of only 3.6 colonies per gram. (Abstract 532.)

Hepatitis

Vaccination with the MSD hepatitis A vaccine beginning at 2 months of age was safe and was immunogenic in seronegative infants, but the presence of maternal antibodies impaired the response to the vaccine, an effect that was not overcome by increasing the dose.

A study of female prostitutes without a history of injection of illicit drugs found, by multivariate analysis, that independent risk factors for HCV seropositivity were sexual practices involving trauma (such as bleeding, biting, or piercing), lack of condom use in the previous five years, and increasing lifetime number of paying male sex partners. (Abstract 24.)

Ear Infections

Children (129) with acute otitis media were randomized to receive either a single IM dose of ceftriaxone (50 mg/kg) or 40 mg/kg/d amoxicillin given orally in three divided doses for 10 days. The failure rate after ceftriaxone therapy was 13.6% and, after amoxicillin therapy, was 10.9% (P = NS). (Abstract 94.)

Mycoses

Of 149 infants ages 5 days to 1 year who succumbed to sudden infant death syndrome in Chile, 51 (34%) had small clusters of P. carinii in their lungs vs. 10 (3%) of 356 age-matched controls (P = 0.0000001). While the association is strong, the small numbers of organisms detected were insufficient to be the direct cause of death. (Abstract 747.)

Fluconazole prophylaxis of vaginal candidiasis in HIV-infected patients was associated with an increased frequency of recovery of C. glabrata. (Abstract 322.) In a (U.S.) nationwide study of candidemia, the attributable mortality was 10%. There was a temporal trend toward an increasing frequency of recovery of non-albicans species. (Abstract 320.) Investigators were able to correlate in vitro susceptibility tests by both the NCCLS method and the E-test with elimination of candidemia in humans. The E-test performed somewhat better than did the NCCLS method. (Abstract 98.)

After a loading dose of 400 mg, fluconazole was administered orally in a dose of 50 mg daily to patients with reduced renal function with a mean creatinine clearance of approximately 20.4 mL/min (range: 9.9-27 mL/min). Peak plasma concentrations of the drug ranged from 12.6-33.7 mcg/mL while the mean urine concentration over 24 hours was 30.9 ± 6.5 mcg/mL (range: 21.7-43.4 mcg/mL), indicating this to be a reasonable dosage regimen in patients with severely reduced renal function. (Abstract 356.)

Itraconazole was compared to fluconazole in the treatment of 191 patients with coccidioidomycosis, including 70 with pulmonary, 71 with soft tissue, and 50 with bone/joint infection. The response (50% improvement by 8 months) rates in the total group were 50% in the fluconazole and 63% in those assigned itraconazole (P = 0.07). There was also a trend toward statistical significance in those with bone/joint infection, with response rates of 26% in the fluconazole and 52% in the itraconazole group (P = 0.06). (Abstract 100.)

Spirochetal Infections

LYMErix, a vaccine based on recombinant outer surface protein A (ospA) of Borrelia burgdorferi, was safe and immunogenic in 250 children ages 5-15 years. (Abstract 75.) Examination of data from efficacy studies in adults, which have been previously reported (N Engl J Med 1998;339:209-215), found that high-level (approximately 95%) protection was associated with serum antibody response of at least 1200 ELISA u/mL, a level achieved in more than 90% of vaccines after three doses of vaccine given on a 0-, 1-, and 12-month schedule. (Abstract 704.) This vaccination schedule is, however, less than optimal, requiring a full year for efficacy. A study of 400 adults randomized to this schedule or to one in which the vaccine was given monthly for three months (0, 1, and 2 months). The latter accelerated schedule was as well tolerated and as immunogenic as the former, with 93% achieving titers of more than 1200 u/mL. (Abstract 705.)