To identify avoidable risk factors for central venous catheter (CVC) infections in patients undergoing short-term catheterization.

Prospective multicenter cohort study.

Two university teaching hospitals and five large nonteaching hospitals.

Patients admitted to intensive care units or surgical units and exposed to short-term CVCs.

Of 623 catheterization episodes, 9.3% were associated with catheter-related infections (CRI). The skin at the catheter site was frequently colonized (16.2%) and was the potential source of infection in 56.1% of the cases, mostly local infections. The hub was colonized less frequently (3.5%) but was responsible for systemic infections more frequently

The following variables were independently associated with CRI: duration of catheterization (for 7 to 14 days, odds ratio [OR], 3.9; 95% confidence interval [CI]95, 1.4 to 10.7; and for >14 days, OR, 5.1; CI95, 1.7 to 15.4), coronary care unit service (OR, 6.7; CI95, 1.1 to 42.9) or surgery service (OR, 4.4; CI95, 1.03 to 18.5), second episode of catheterization (OR, 7.6; CI95, 1.8 to 32.3), skin colonization at the insertion site (OR, 56.5; CI95, 10.8 to 296), and hub colonization (OR, 17.9; CI95, 2.4 to 132).

The risk associated with skin colonization varied with use of jugular access or simultaneous colonization of the hub. When only symptomatic CRI was considered, the risk associated with hub colonization was consistently higher (OR, 36.6; 7 to 190) than that associated with skin colonization (OR, 3.2; CI95, 0.7 to 14).

Age, transparent dressing, jugular insertion, male gender, duration of catheterization, and hub colonization were independent risk factors for skin colonization. The effect of age varied by type of dressing and vice versa; the effect of jugular access varied by sex; and the effect of transparent dressing varied by length of catheter-ization and vice versa.

Total parenteral nutrition and skin colonization were independently associated with an increased risk of hub colonization.

Skin and hub colonization are the two major determinants for endemic CRIs; colonization of the hub, however, is more frequently associated with more severe infections. In order to reduce CRIs, more efforts should be focused on understanding which factors increase the risk of colonization both of the skin and of the hub.

To study the epidemiological, microbiological, and clinical features of infections due to Acinetobacter baumannii in a complex endemic situation over an 18-month period and to determine the clinical usefulness of plasmid DNA analysis of A baumannii in epidemiological investigations.

Review of medical and laboratory records. Antibiotic resistance patterns, biotyping, and plasmid profile analysis were used to characterize clinical and environmental isolates. Pulsed-field gel electrophoresis (PFGE) of chromosomal DNA was performed to verify results obtained with the other typing methods.

Four different intensive care units of an 800-bed tertiary care center in Cologne, Germany.

240 patients were colonized or infected with A baumannii during the study period. No seasonal variations were observed. The majority of isolates (53%) were recovered from the respiratory tract. Major infections occurred in 61 patients; these included 48 bacteremias and eight pulmonary infections. Five different epidemic strains were identified: one each was A baumannii biotype 2 and 6, and three were biotype 9. A baumannii biotype 9 accounted for the vast majority of isolates (88%), which were clustered into three epidemic strains demonstrating distinct plasmid profiles. Two of these were considered genetically related as shown by PFGE. Epidemic strains were multidrug resistant, being uniformly susceptible to imipenem only. An epidemiological investigation failed to identify any point source of infection. Barrier precautions and improved handwashing were instituted in three of the four units and significantly reduced the incidence of colonization and infection in these units. Attack rates remained unchanged, however, in the burn unit where control measures were not implemented.

Acinetobacter strains representing multiple biotypes and plasmid types were present in this endemic setting. Multidrug resistance in A baumannii is an important concern. Plasmid DNA analysis proved to be useful in epidemiological typing of A baumannii strains and may serve as a complementary typing system to traditional epidemiological methods.

To better define the role of multiple risk factors for cytotoxic Clostridium difficile-associated diarrhea.

Case-control study.

A Veterans Affairs Medical Center.

Thirty-three case patients with C difficile -associated diarrhea. Two control groups were used: one group consisted of 32 patients from the same ward as the case patients, and one group consisted of 34 patients with nosocomial diarrhea and negative C difficile toxin assays.

None.

Multivariate analyses revealed that exposure to second- or third-generation cepha-losporins was the most important independent risk factor, even after controlling for other antimicrobial use (odds ratio [OR]=8.3, 95% confidence interval [CI95] = 1.4 to 48.9 compared to ward controls; OR= 9.6, CI95 = 2.1 to 44.1 compared with diarrhea controls). Persons exposed to two or more antimicrobials simultaneously were at substantially elevated risk (OR= 18.7, CI95 = 4.1 to 85.8 compared with ward controls; OR=21.5, CI95=3.2 to 141.9 compared with diarrhea controls).

Physicians should consider carefully the appropriateness of second- and third-generation cephalosporin use and combination antimicrobial therapy, especially during nosocomial C difficile -associated diarrhea outbreaks.

The objectives of this quality standard are 1) to provide an implementation mechanism that will facilitate the reliable administration of prophylactic antimicrobial agents to patients undergoing operative procedures in which such a practice is judged to be beneficial and 2) to provide a guideline that will help local hospital committees formulate policies and set up mechanisms for their implementation. Although standards in the medical literature spell out recommendations for specific procedures, agents, schedules, and doses, other reports document that these standards frequently are not followed in practice.

We have specified the procedures in which the administration of prophylactic antimicrobial agents has been shown to be beneficial, those in which this practice is widely thought to be beneficial but in which compelling evidence is lacking, and those in which this practice is controversial. We have examined the evidence regarding the optimal timing of drug administration, the optimal dose, and the optimal duration of prophylaxis.

The intended outcome is more uniform and reliable administration of prophylactic antibiotics in those circumstances where their value has been demonstrated or their use has been judged by the local practicing medical community to be desirable. The result should be a reduction in rates of postoperative wound infection with a limitation on the quantities of antimicrobial agents used in circumstances where they are not likely to help.

Many prospective, randomized, controlled trials comparing placebo with antibiotic and comparing one antibiotic with another have been conducted. In addition, some trials have compared the efficacy of different doses or methods of administration. Other papers have reported on the apparent efficacy of administration at different times and on actual practice in specific communities. Only a small group of relevant articles found through 1993 are cited herein. When authoritative reviews are available, these-rather than an exhaustive list of original references-are cited.

We assumed that reducing rates of postoperative infection was valuable but that reducing the total amount of antimicrobial agents employed was also worthwhile. The cost of and morbidity attributable to postoperative wound infections should be weighed against the cost and potential morbidity associated with excessive use of antimicrobial agents.

More reliable administration of antimicrobial agents according to recognized guidelines should prevent some postoperative wound infections while lowering the total quantity of these drugs used. No harms are anticipated. The costs involved are those of the efforts needed on a local basis to design and implement the mechanism that supports uniform and reliable administration of prophylactic antibiotics.

All patients for whom prophylactic antimicrobial agents are recommended should receive them. The agents given should be appropriate in light of published guidelines. A short duration of prophylaxis (usually < 24 hours) is recommended.

More than 50 experts in infectious disease and 10 experts in surgical infectious disease and surgical subspecialties reviewed the standard. In addition, the methods for its implementation were reviewed by the American Society of Hospital Pharmacists.

The Quality Standards Subcommittee of the Clinical Affairs Committee of the Infectious Disease Society of America (IDSA) developed the standard. The subcommittee was composed of representatives of the IDSA (Drs. Gross and McGowan), the Society for Hospital Epidemiology of America (Dr. Wenzel), the Surgical Infection Society (Dr. Dellinger), the Pediatric Infectious Disease Society (Dr. Krause), the Centers for Disease Control and Prevention (Dr. Martone), the Obstetrics and Gynecology Infectious Diseases Society (Dr. Sweet), and the Association of Practitioners of Infection Control (Ms. Barrett). Funding was provided by the IDSA and the other cooperating organizations. The standard is endorsed by the IDSA.

The objective of this quality standard is to optimize the treatment of bacteremia in hospitalized patients by ensuring that the antibiotic given is appropriate in terms of the blood culture susceptibility of the pathogen. Although this standard may appear to be minimal in scope, it is needed because appropriate antimicrobial treatment is not given in 5% to 17% of cases. To implement the standard, physicians, pharmacists, and microbiologists will need to devise a coordinated strategy.

We considered criteria for appropriate dosing, most cost-effective selection, proper antibiotic levels in serum, least toxicity, narrowest spectrum, specific clinical indications, and optimal duration of treatment. All these criteria were rejected as the basis for the standard because they were too controversial and too difficult to be applied by a nonphysician chart reviewer. In contrast, the selection of an antibiotic to which the pathogen is sensitive is a non-controversial criterion and easy for a chart reviewer to apply.

The standard is designed to reduce the incidence of adverse outcomes of septicemia such as renal failure, prolonged hos-pitalization, and death.

Several well-designed clinical trials without randomization as well as case-controlled studies have confirmed the benefit of using an antibiotic that is appropriate in light of the susceptibility of the isolate in blood culture. Prospective, randomized, placebo-controlled trials are not available.

Our premise is that the presence of bacteremia is a risk factor for serious adverse outcomes. We also believe that the administration of antibiotics must always be guided by the susceptibility report for the pathogen(s) obtained from blood cultures. This concern is more critical for pathogens from the blood than for those from most other body sites. We had evidence that susceptibility reports for pathogens from positive blood cultures were not always used properly. We used group discussion to reach a consensus among the members of the Quality Standards Subcommittee.

Through the implementation of this standard, at least 5% of bacteremias could be treated more appropriately. An unknown number of deaths would likely be prevented, and mortality from bacteremia treated inappropriately would probably be reduced. The primary undesirable feature of the standard is an increased workload of pharmacists and microbiologists.

Treatment of bacteremia with an antibiotic that is appropriate in terms of the pathogen's blood-culture susceptibility is a minimal standard of care for all patients.

We consulted more than 50 experts in infectious diseases from the fields of medicine, surgery, pediatrics, obstetrics and gynecology, nursing, epidemiology, pharmacology, and government. In addition, the methods for its implementation were reviewed by the American Society of Hospital Pharmacists and were tested by one of the members of the Quality Standards Subcommittee.

The Quality Standards Subcommittee of the Clinical Affairs Committee of the Infectious Diseases Society of America (IDSA) developed the standard. The subcommittee was composed of representatives of the IDSA (Drs. Gross and McGowan), the Society for Hospital Epidemiology of America (Dr. Wenzel), the Surgical Infection Society (Dr. Dellinger), the Pediatric Infectious Diseases Society (Dr. Krause), the Centers for Disease Control and Prevention (Dr. Martone), the Obstetrics and Gynecology Infectious Diseases Society (Dr. Sweet), and the Association of Practitioners of Infection Control (Ms. Barrett). Funding was provided by the IDSA and the other cooperating organizations. This standard is endorsed by the IDSA.

To determine the prevalence of tuberculous infection among a sample of physicians at Barnes Hospital and to determine the frequency of tuberculin skin testing and the adequacy of follow-up for physicians with positive tuberculin skin tests.

Convenience sample.

1,000-bed, university-affiliated tertiary care hospital.

Physicians attending departmental conferences were screened for tuberculosis. Prior history of tuberculosis, antituberculous therapy, BCG vaccination, and previous tuberculin skin test results were obtained with a standardized questionnaire. Tuberculin skin tests were performed on those who were previously skintest negative.

Tuberculosis infection, prophylactic therapy.

Eighty-six (24.5%) of 351 physicians in the study were skin test positive by history or currently performed skin test. Of 61 who reported a previously reactive skin test, 40 (66%) had been eligible for isoniazid prophylaxis, but only 15 (37.5%) of 40 had completed at least six months of therapy. Of 290 physicians reporting a previously negative skin test, 25 conversions (8.6%) were identified. Previously undiagnosed, asymptomatic pulmonary tuberculosis was identified in one physician.

Infection with Mycobacterium tuberculosis is common among physicians. Physicians were screened irregularly for tuberculosis, and the use of prophylactic therapy was inconsistent. Aggressive tuberculosis screening programs for healthcare workers should be instituted.