To define the extent of nosocomial transmission of methicillin-resistant Staphylococcus aureus (MRSA) in patients admitted to a tertiary-care hospital.

A blinded, prospective surveillance culture study of patients admitted to the hospital to determine the transmission (acquisition) rate of MRSA Risk factors associated with the likelihood of MRSA colonization on admission were investigated.

Tertiary-care military medical facility.

All patients admitted to the medicine, surgery, and pediatric wards, and to the medical, surgical, and pediatric intensive care units were eligible for inclusion.

Five hundred thirty-five admission and 374 discharge samples were collected during the study period. One hundred forty-one patients were colonized with methicillin-susceptible S. aureus (MSSA) and 20 patients (3.7%) were colonized with MRSA on admission. Of the 354 susceptible patients, 6 acquired MRSA during the study for a transmission rate of 1.7%. Patients colonized with MRSA on admission were more likely to be older than non-colonized or MSSA-colonized patients, to have received antibiotics within the past year, to have been hospitalized within the prior 3 years, or to have a known history of MRSA. Patients acquiring MRSA had an average hospital stay of 17.7 days compared with 5.3 days for those who did not acquire MRSA. Pulsed-field gel electrophoresis of the 6 MRSA isolates from patients who acquired MRSA revealed 4 distinct band patterns.

Most patients colonized with MRSA were identified on admission samples. Surveillance cultures of patients admitted may help to prevent MRSA transmission and infection.

To evaluate the usefulness of a new alcohol sprayer by comparing it with an individual bottle of alcohol. Patterns of use and perceptions among healthcare personnel were compared for the two products.

Observational study recording the volume of alcohol used and the compliance rate (frequency of hand antisepsis per number of opportunities), and a survey of healthcare workers' perceptions of the different hand hygiene strategies.

A 20-bed medical unit in a public hospital in Marseille, France.

Healthcare workers of an infectious disease unit.

Hand hygiene alcohol systems (sprayer vs individual bottle, 70% ethyl alcohol).

The sprayer was used more frequently than the individual bottle (12.6 vs 9.7 hand washes per day). With the sprayer system, compliance was 91% for physicians, 28% for nurses, and 8% for housekeeping personnel. Alcohol hand antisepsis was preferred to washing hands with soap and water in low-risk situations such as simple entrance into a room (91% vs 36%; P < 10-6) or simple contact with a patient (69% vs 40%; P < .005). The sprayer system was considered easier to use (95%), more hygienic (90%), and faster (92%), with a better tolerance than the individual bottle.

The new alcohol sprayer should improve rapid hand antisepsis.

To assess the effect of a standard decontamination and sterilization protocol employed during reuse of cardiac electrophysiology (EP) catheters on human immunodeficiency virus (HIV).

Public health viral research laboratory.

Studies were performed on distal, electrode-containing segments of 40 EP catheters previously used in up to 10 clinical EP procedures. EP catheter segments were immersed for 1 hour in blood contaminated with a high titer of HIV. After air drying for 2 hours, subgroups of 8 EP catheters were subjected to either (1) no treatment, (2) washing in general purpose detergent, (3) washing in enzyme cleaner, (4) sterilization in ethylene oxide, or (5) the full protocol of sequential detergent-enzyme cleaner-ethylene oxide exposure. HIV infectivity after treatment was determined by measuring HIV RNA and, in cell culture studies, assessing HIV-induced cytopathic effects (CPEs) and supernatant HIV-specific p24 antigen content.

With no treatment, all catheters had high HIV RNA levels associated with CPEs and high p24 antigen levels. After washing in detergent, 5 of 8 catheters had HIV RNA detected, but without CPEs or p24 antigen. HIV RNA was detected in all catheters after washing in enzyme cleaner, with CPEs and a high p24 antigen level in 1 of 8 catheters. HIV RNA, CPEs, and p24 antigen were absent after ethylene oxide. After the full protocol, HIV RNA levels were undetectable (n = 7) or low (n = 1), without evidence of CPEs or p24 antigen.

Appropriate decontamination and sterilization of EP catheters during reuse is highly effective in inactivating HIV.

In 1998, the California Department of Health Services invited all healthcare facilities in California (n = 2,532) to participate in a statewide, voluntary sharps injury surveillance project The objectives were to determine whether a low-cost sharps registry could be established and maintained, and to evaluate the circumstances surrounding sharps injuries in California.

Approximately 450 facilities responded and reported a total of 1,940 sharps-related injuries from January 1998 through January 2000. Injuries occurred in a variety of healthcare workers (80 different job titles). Nurses sustained the highest number of injuries (n = 658). In hospital settings (n = 1,780), approximately 20% of the injuries were associated with drawing venous blood, injections, or assisting with a procedure such as suturing. As expected, injuries were caused by tasks conventionally related to specific job classifications. The overall results approximate those reported by the Centers for Disease Control and Prevention's National Surveillance System for Health Care Workers and the University of Virginia's Exposure Prevention Information Network.

These data further support findings from previous studies documenting the complex and persistent nature of sharps-related injuries in healthcare workers. In the future, mandated reporting using standardized forms and consistent application of decision rules would facilitate a more thorough analysis of injury events.

To evaluate excess mortality in critically ill patients with Escherichia coli bacteremia after adjustment for severity of illness.

Retrospective (1992-2000), pairwise-matched (1:2), risk-adjusted cohort study.

Fifty-four-bed ICU in a university hospital including a medical and surgical ICU, a unit for care after cardiac surgery, and a burns unit.

ICU patients with nosocomial E. coli bacteremia (defined as cases; n = 64) and control-patients without nosocomial bloodstream infection (n = 128).

Case-patients were matched with control-patients on the basis of the Acute Physiology and Chronic Health Evaluation (APACHE) II system: an equal APACHE II score (± 2 points) and diagnostic category. In addition, control-patients were required to have an ICU stay at least as long as that of the respective case-patients prior to onset of the bacteremia.

The overall rate of appropriate antibiotic therapy in patients with E. coli bacteremia was high (93%) and such therapy was initiated soon after onset of the bacteremia (0.6 ± 1.0 day). ICU patients with E. coli bacteremia had more acute renal failure. No differences were noted between case-patients and control-patients in incidence of acute respiratory failure, hemodynamic instability, or age. No differences were observed in length of mechanical ventilation or length of ICU stay. In-hospital mortality rates for cases and controls were not different (43.8% and 45.3%, respectively; P = .959).

After adjustment for disease severity and acute illness and in the presence of adequate antibiotic therapy, no excess mortality was found in ICU patients with E. coli bacteremia.

To determine the prevalence of and risk factors for tuberculin skin test positivity and conversion among New York City Department of Health and Mental Hygiene employees.

Point-prevalence survey and prospective cohort analysis. Sentinel surveillance was conducted from March 1,1994, to December 31, 2001.

HCWs in high-risk and low-risk settings for occupational TB exposure.

Baseline tuberculin positivity was 36.2% (600 of 1,658), 15.5% (143 of 922) among HCWs born in the United States, and 48.5% (182 of 375) among HCWs not born in the United States. There were 36 tuberculin conversions during 2,754 observation-years (rate, 1.3 per 100 person-years). For HCWs born in the United States, the risk for tuberculin conversion was greater in high-risk occupational settings compared with low-risk settings (OR 5.7; CI95, 1.7–19.2; P < .01). HCWs not born in the United States and those employed at the Office of the Chief Medical Examiner (OCME) were at high risk for baseline tuberculin positivity (OR, 3.2; CI95,1.7–5.8; P < .001); OCME HCWs (OR 4.7; CI95, 2.3–9.4; P < .001), those of Asian ethnicity (OR 4.3; CI95,1.4–13.5; P < .01), and older HCWs (OR, 1.0; CI95,1.0–1.1; p < .05) were at a higher risk for conversion.

Although the prevalence of tuberculin positivity decreased after the peak of the recent TB epidemic in New York City, the conversion rate among HCWs in high-risk occupational settings for TB exposure was still greater than that among HCWs in low-risk settings. Continued surveillance of occupational TB infection is needed, especially among high-risk HCWs.

To determine the frequency with which methicillin-resistant Staphylococcus aureus (MRSA) is spread from colonized or infected patients to their household and community contacts.

Retrospective cohort study.

University hospital.

Household and community contacts of MRSA-colonized or -infected patients for whom MRSA screening cultures were performed.

MRSA was isolated from 25 (14.5%) of 172 individuals. Among the contacts of index patients who had at least one MRSA-colonized contact, those with close contact to the index patient were 7.5 times more likely to be colonized (53% vs 7%; 95% confidence interval, 1.1 to 50.3; P = .002). An analysis of antimicrobial susceptibility and DNA fingerprint patterns suggested person-to-person spread.

MRSA colonization occurs frequently among household and community contacts of patients with nosocomially acquired MRSA, suggesting that transmission of nosocomially acquired MRSA outside of the healthcare setting may be a substantial source of MRSA colonization and infection in the community.

To estimate the frequency of and risk factors for surgical-site infections (SSIs) in Bolivia, and to study the performance of the National Nosocomial Infections Surveillance (NNIS) System risk index in a developing country.

A prospective study with patient follow-up until the 30th postoperative day.

A general surgical ward of a public hospital in Santa Cruz, Bolivia.

Patients admitted to the ward between July 1998 and June 1999 on whom surgical procedures were performed.

Follow-up was complete for 91.5% of 376 surgical procedures. The overall SSI rate was 12%. Thirty-four (75.6%) of the 45 SSIs were culture positive. A logistic regression model retained an American Society of Anesthesiologists score of more than 1 (odds ratio [OR], 1.87), a not-clean wound class (OR, 2.28), a procedure duration of more than 1 hour (OR, 1.81), and drain (OR, 1.98) as independent risk factors for SSI. There was no significant association between the NNIS System risk index and SSI rates. However, a “local” risk index constructed with the above cut-off points showed a linear trend with SSI (P < .001) and a relative risk of 3.18 for risk class 3 versus a class of less than 3.

SSIs cause considerable morbidity in Santa Cruz. Appropriate nosocomial infection surveillance and control should be introduced. The NNIS System risk index did not discriminate between patients at low and high risk for SSI in this hospital setting, but a risk score based on local cutoff points performed substantially better.

Antianaerobic antibiotic therapy promotes persistent high-density growth of vancomycin-resistant enterococci (VRE) in the stool of colonized patients. We tested the hypothesis that antibiotic regimens with potent antianaerobic activity promote overgrowth of coexisting antibiotic-resistant, gram-negative bacilli in the stool of VRE-colonized patients.

Eight-month prospective study examining the effect of antibiotic therapy on the stool density of gram-negative bacilli resistant to ceftazidime, ciprofloxacin, or piperacillin/tazobactam.

A Department of Veterans Affairs medical center including an acute care hospital and nursing home.

All VRE-colonized patients with at least 3 stool samples available for analysis.

One-hundred forty stool samples were obtained from 37 study patients. Forty-nine (61%) of 80 stool samples obtained during therapy with an antianaerobic regimen were positive for an antibiotic-resistant, gram-negative bacillus, whereas only 14 (23%) of 60 samples obtained 4 or more weeks after completion of such therapy were positive (P < .001). Twenty-four (65%) of the 37 patients had one or more stool cultures positive for a gram-negative bacillus resistant to ciprofloxacin, ceftazidime, or piperacillin/tazobactam. The density of these organisms was higher during therapy with antianaerobic regimens than in the absence of such therapy for at least 2 weeks (mean ± standard deviation, 5.6 ± 1.4 and 3.9 ± 0.71 log10 organisms/g; P < .001).

Limiting the use of antianaerobic antibiotics in VRE-colonized patients may reduce the density of colonization with coexisting antibiotic-resistant, gram-negative bacilli.

To assess the compliance of a group of Nigerian dentists with standard infection control practices.

A confidential self-administered questionnaire survey was conducted among dentists engaged in active clinical practices in public hospitals in Lagos, Nigeria (n = 185).

One hundred forty-six questionnaires were returned (response rate, 78.9%). Most (70.6%) of the dentists always wore gloves when treating patients, whereas 29.4% sometimes did. Regarding facemasks, 45.9% always wore them, 52.7% sometimes wore them, and 1.4% never wore them. Protective eye wear was always worn by 4.8% of the dentists, sometimes worn by 52.7%, and never worn by 42.5%. Approximately half (50.7%) of the respondents had received hepatitis B vaccination. Sterilization was performed using a combination of methods, including auto-claving (84.1%), boiling (19.3%), dry heat (17.5%), and chemicals (29.7%). Nonavailability of materials was the major reason for noncompliance with infection control practices.

Nigerian dentists need continuous education regarding infection control. Also, Nigerian hospitals urgently need adequate funding for up-to-date and functional equipment and materials.

To investigate the molecular epidemiology of systemic nosocomial infections caused by Enterobacter cloacae.

Neonatal intensive care unit (NICU) of a tertiary-care university hospital.

Forty-two high-risk neonates with systemic infections caused by E. cloacae.

From 1995 to 1997, the variables associated with death in these patients were evaluated. The molecular epidemiology of the strains responsible for the systemic infections, and 14 unrelated strains, was studied using plasmid analysis and pulsed-field gel electrophoresis (PFGE).

The overall mortality rate for infection caused by E. cloacae was 34%, whereas the crude mortality rate during the study period was 8.12% (P < .001). Gestational age (preterm neonates) and birth weight (small for gestational age) were not associated with a higher risk of death. Insertion of a venous catheter by dissection of a peripheral vein was the only invasive procedure related to death (P = .016) in this study. A molecular analysis showed that three outbreaks, each occurring in a different year, were caused by strains with distinctive DNA profiles. Only one outbreak was identified by the infection control service, in the NICU. Plasmid analysis and PFGE showed similar ability to discriminate control strains from the E. cloacae strains isolated from the neonates.

Systemic infections caused by E. cloacae in our NICU were associated with a high mortality rate and occurred as small, unrecognized outbreaks. These results may not be generalizable because the data were from a single center (Infect Control Hosp Epidemiol 2003;24:490-494).

To test the hypothesis that patients colonized with vancomycin-resistant Enterococcus (VRE) have a higher frequency of colonization or infection with other nosocomial pathogens than do patients who are not colonized with VRE.

A rectal swab culture survey was conducted to determine the point-prevalence of stool colonization with ceftazidime-resistant gram-negative bacilli in hospitalized patients with or without VRE stool colonization. For a 6-month period, the frequency of Clostridium difficile diarrhea and isolation of antibiotic-resistant (ie, ceftazidime-, piperacillin/tazobactam-, levofloxacin-, or trimethoprim/sulfamethoxazole-resistant) gram-negative bacilli, methicillin-resistant Staphylococcus aureus (MRSA), and non-albicans Candida species from clinical specimens other than stool was examined.

A Department of Veterans Affairs medical center.

All patients hospitalized in the acute care facility and one nursing home unit during a 1-week period in February 2001.

VRE-colonized patients had a higher point-prevalence of rectal colonization with ceftazidime-resistant gram-negative bacilli than did patients not colonized with VRE (17% vs 4%; P = .026). During a 6-month period, the VRE-colonized patients were more likely to have Clostridium difficile–associated diarrhea (26% vs 2%; P = .001), MRSA infection (17% vs 4%; P = .017), or colonization or infection with gram-negative bacilli resistant to 4 different antibiotics.

VRE-colonized patients in our institution have a higher frequency of colonization or infection with other nosocomial pathogens than do patients who are not colonized with VRE. This suggests that isolation measures implemented to control VRE could help limit the dissemination of other, coexisting pathogens.

To identify the source of an outbreak of acute hepatitis C virus (HCV) infection among 3 patients occurring within 8 weeks of hospitalization in the same ward of a Florida hospital during November 1998.

A retrospective cohort study was conducted among 41 patients hospitalized between November 11 and 19, 1998. Patients' blood was tested for antibodies to HCV, and HCV RNA-positive samples were genotyped and sequenced.

Of the 41 patients, 24 (59%) participated in the study. HCV genotype 1b infections were found in 5 patients. Three of 4 patients who received saline flushes from a multidose saline vial on November 16 had acute HCV infection, whereas none of the 9 patients who did not receive saline flushes had HCV infection (P = .01). No other significant exposures were identified. The HCV sequence was available for 1 case of acute HCV and differed by a single nucleotide (0.3%) from that of the indeterminate case.

This outbreak of HCV probably occurred when a multidose saline vial was contaminated with blood from an HCV-infected patient. Hospitals should emphasize adherence to standard procedures to prevent blood-borne infections. In addition, the use of single-dose vials or prefilled saline syringes might further reduce the risk for nosocomial transmission of blood-borne pathogens.

Defining risk factors for central venous catheter (CVC)-associated bloodstream infections (BSIs) is critical to establishing prevention measures, especially for factors such as nurse staffing and antimicrobial-impregnated CVCs.

We prospectively monitored CVCs, nurse staffing, and patient-related variables for CVC-associated BSIs among adults admitted to eight ICUs during 2 years.

A total of 240 CVC-associated BSIs (2.8%) were identified among 4,535 patients, representing 8,593 CVCs. Antimicrobial-impregnated CVCs reduced the risk for CVC-associated BSI only among patients whose CVC was used to administer total parenteral nutrition (TPN, 2.6 CVC-associated BSIs per 1,000 CVC-days vs no TPN, 7.5 CVC-associated BSIs per 1,000 CVC-days; P = .006). Among patients not receiving TPN, there was an increase in the risk of CVC-associated BSI in patients cared for by “float” nurses for more than 60% of the duration of the CVC. In multivariable analysis, risk factors for CVC-associated BSIs were the use of TPN in non-antimicrobial-impregnated CVCs (P = .0001), patient cared for by a float nurse for more than 60% of CVC-days (P = .0019), no antibiotics administered to the patient within 48 hours of insertion (P = .0001), and patient unarousable for 70% or more of the duration of the CVC (P = .0001). Peripherally inserted central catheters (PICCs) were associated with a lower risk for CVC-associated BSI (P = .0001).

Antimicrobial-impregnated CVCs reduced the risk of CVC-associated BSI by 66% in patients receiving TPN. Limiting the use of float nurses for ICU patients with CVCs and the use of PICCs may also reduce the risk of CVC-associated BSI.

To investigate the clinical, microbiological, and epidemiologic features of three outbreaks caused by Klebsiella during 3 years.

Neonatal intensive care unit of a university hospital.

Thirty affected neonates.

Data were collected through chart reviews and conversations with physicians. Screening samples were obtained from the staff, the neonates, and the environment. Antibiogram typing and arbitrarily primed polymerase chain reaction-based fingerprinting were used to type the strains.

The first outbreak had 13 K. pneumoniae strains isolated. The second outbreak had 10 K. oxytoca strains isolated. The third outbreak had 20 K. pneumoniae strains isolated. More than half of the patients had low birth weights, were premature, and underwent mechanical ventilation and intravenous catheterization. Approximately three-fourths of the patients died. The isolates tested were completely susceptible to meropenem, cefoxitin, and ciprofloxacin and were resistant to cephalothin. More than half of these strains were resistant to many beta-lactam antibiotics, amikacin, and trimethoprim/sul-famethoxazole. Typing procedures yielded 3 antibiotypes and 3 genotypes among the isolates of the first outbreak, 3 antibiotypes with 1 subtype and 2 genotypes with 1 subtype in the second outbreak, and 2 antibiotypes and 2 genotypes in the third outbreak.

Klebsiella outbreaks mainly affected premature neonates with intravenous catheters, mechanical ventilation, or both. The high mortality rate (76.7%) was notable. Resistance to multiple antibiotics, but mainly to broad-spectrum beta-lactam antibiotics, was observed, particularly in K. pneumoniae isolates. Molecular typing indicated that the three outbreaks were not related to one other (Infect Control Hosp Epidemiol 2003;24:495-500).

To describe the relative contribution of and risk factors for both community-acquired and nosocomial methicillin-resistant Staphylococcus aureus (MRSA) infections.

Retrospective cohort study.

270-bed, tertiary-care children's hospital.

All MRSA-infected children from whom MRSA was recovered between October 1, 1999, and September 30, 2001.

Demographic, clinical, and risk factor data were abstracted from medical records. Categorical variables were analyzed using the chi-square or Fisher's exact test and continuous variables were analyzed using the Mann-Whitney test.

Of the 62 patients with new MRSA infection, 37 had community-acquired MRSA and 25 had nosocomial MRSA. Most community-acquired MRSA infections were of the skin and soft tissue, the middle ear, and the lower respiratory tract. Nosocomial MRSA infections occurred in the lower respiratory tract, the skin and soft tissue, and the blood. Risk factors for infection, including underlying medical illness, prior hospitalization, and prior surgery, were similar for patients with community-acquired MRSA and nosocomial MRSA. History of central venous catheterization and previous endotracheal intubation was more common in patients with nosocomial MRSA. Only 3 patients with community-acquired MRSA had no identifiable risk factor other than recent antibiotic use. Resistance for clindamycin, erythromycin, and levofloxacin was similar between strains of community-acquired MRSA and nosocomial MRSA.

Similarities in patient risk factors and resistance patterns of isolates of both community-acquired and nosocomial MRSA suggest healthcare acquisition of most MRSA. Thus, classifying MRSA as either community acquired or nosocomial underestimates the amount of healthcare-associated MRSA.