To evaluate the dynamics of bacterial contamination of healthcare workers' (HCWs) hands during neonatal care.

The 20-bed neonatal unit of a large acute care teaching hospital in Geneva, Switzerland.

Structured observation sessions were conducted. A sequence of care began when the HCW performed hand hygiene and ended when the activity changed or hand hygiene was performed again. Alcohol-based handrub was the standard procedure for hand hygiene. An imprint of the five fingertips of the dominant hand was obtained before and after hand hygiene and at the end of a sequence of care. Regression methods were used to model the final bacterial count according to the type and duration of care and the use of gloves.

One hundred forty-nine sequences of care were observed. Commensal skin flora comprised 72.4% of all culture-positive specimens (n = 360). Other microorganisms identified were Enterobacteriaceae (n = 55, 13.8%); Staphylococcus aureus (n = 10, 2.5%); and fungi (n = 7,1.8%). Skin contact, respiratory care, and diaper change were independently associated with an increased bacterial count; the use of gloves did not fully protect HCWs' hands from bacterial contamination.

These data confirm that hands become progressively contaminated with commensal flora and potential pathogens during neonatal care, and identify activities at higher risk for hand contamination. They also reinforce the need for hand hygiene after a sequence of care, before starting a different task, and after glove removal.

To obtain accurate data regarding the handwashing behavior and patterns of visits to patients by healthcare workers (HCWs).

All visits by HCWs to selected patient rooms were recorded for 3 days and 2 nights. Additionally, 5 nurses were observed for 1 day each and 2 nurses were observed for 1 night each. Nurses were observed for their entire shifts and all of their activities were recorded.

A general medical ward in a tertiary-care hospital.

Convenience samples of HCWs and patients.

Patients were visited every 25 minutes on average. Monitoring rooms and observing nurses resulted in similar rates of patient visits. The highest level of risk was contact with body fluids in 11% of visits and skin in 40% of visits. The overall rate of handwashing was 46%; however, the rate was higher for visits involving contact with body fluids (81%) and skin (61%). Nurses returned immediately to the same patient 45% of the time. The rate of handwashing was higher for the last of a series of visits to a patient's room (53% vs 30%, P < .0001).

Nurses adjusted their handwashing rates in accordance with the risk level of each visit. Monitoring patient rooms and observing nurses yielded similar estimates of patient visits and proportions of visits involving contact with skin or body fluids. Education programs about hand hygiene may be more effective if patterns of care and levels of risk are incorporated into recommendations.

The Fulkerson scale lists 15 typical clinical activities ranked according to their risk of hand contamination.

To investigate how often healthcare workers (HCWs) apply antiseptic handrubs after each of the specified activities ranked in the Fulkerson scale.

Observational study.

University Hospital of Heidelberg, a 1,600-bed teaching hospital.

Using a structured protocol, 41 observers recorded whether HCWs rubbed their hands in 1 of 15 given situations.

Handrubs were used in 1,115 (52.2%) of 2,138 observations. Comparatively high rates of handrub use were observed after contact with sterile materials (39.6%) and after contact with excretions (90% to 97%). Nurses used handrubs significantly more frequently than did physicians after contact with the inanimate environment, but physicians applied handrubs significantly more frequently after contact with body secretions.

This study showed better compliance with hand hygiene than that reported by many previous studies, but compliance was still inadequate and varied significantly by type of HCW, type of activity, and location in the hospital.

The Centers for Disease Control and Prevention has published a new guideline on hand hygiene promoting the use of the alcohol-based handrub, but the technique was not addressed. The goal of this study was to evaluate the influence of technique on the efficacy of the alcohol-based handrub.

Healthcare workers (HCWs) attending a course in hospital epidemiology.

A fluorescent dye was added to a hand antiseptic, and hands were checked under ultraviolet light after antiseptic cleansing. Data regarding the numbers of predefined fluorescent areas on the skin were collected in addition to demographic data such as age, gender, job description, and job experience. Results of the visualization test were compared with the data from microbiological samples before and after the procedure by the hand plate technique.

Sixty HCWs were tested, 63% of whom had worked in infection control for more than 10 years. Sixty-six percent of all participants still had detectable bacteria after antisepsis. The mean log10 CFU reduction was 2.0 (range, 0–3.85). Twenty-five percent of all HCWs achieved less than 1.1 log10 CFU. Staphylococcus aureus was isolated from 13% (one of them being methicillin resistant) and gram-negative bacilli from 6.7%. After using the alcohol handrub, one subject still remained positive for S. aureus. Years of experience was the single most important factor predicting antimicrobial efficacy.

Technique is of crucial importance in hand antisepsis. Major deficiencies were detected among even highly trained HCWs. Training should be provided before switching from handwashing to the alcohol handrub.

From April to June 2001, an outbreak of extended-spectrum beta-lactamase (ESBL)–producing Klebsiella pneumoniae infections was investigated in our neonatal intensive care unit.

Cultures of the gastrointestinal tracts of patients, the hands of healthcare workers (HCWs), and the environment were performed to detect potential reservoirs for ESBL-producing K. pneumoniae. Strains of K. pneumoniae were typed by pulsed-field gel electrophoresis using Xbal. A case–control study was performed to determine risk factors for acquisition of the outbreak clone (clone A); cases were infants infected or colonized with clone A and controls (3 per case) were infants with negative surveillance cultures.

During the study period, 19 case-infants, of whom 13 were detected by surveillance cultures, harbored clone A. The overall attack rate for the outbreak strain was 45%; 9 of 19 infants presented with invasive disease (n = 6) or developed invasive disease (n = 3) after colonization was detected. Clone A was found on the hands of 2 HCWs, 1 of whom wore artificial nails, and on the designated stethoscope of a case-infant. Multiple logistic regression analysis revealed that length of stay per day (odds ratio [OR], 1.05; 95% confidence interval [CI95], 1.02 to 1.09) and exposure to the HCW wearing artificial fingernails (OR, 7.87; CI95, 1.75 to 35.36) were associated with infection or colonization with clone A.

Short, well-groomed, natural nails should be mandatory for HCWs with direct patient contact.

To determine the frequency of conjunctival colonization, identify the colonizing flora, and correlate culture results with physical findings in infants in a NICU

Surveillance study.

Level III NICU of a large university teaching hospital.

All infants admitted for longer than 24 hours during a 26-week period.

Weekly bacterial conjunctival cultures were performed for all infants. The conjunctival appearance at the time of culture was recorded. The frequency, identity, and correlation of culture results with physical findings were determined.

One thousand ninety-one cultures were performed for 319 infants: 133 (42%) had no positive cultures and 186 (58%) had at least one positive culture. Culture analysis revealed that 411 (38%) were positive and yielded 494 isolates comprising more than 18 bacterial species. Bacteria most commonly isolated included coagulase-negative Staphylococcus (CoNS) (75%), viridans group streptococci (8.7%), Staphylococcus aureus (3.8%), Enterococcus species (2.6%), and Serratia marcescens (2.4%). The frequency of non-CoNS isolates increased significantly during the first 6 weeks of patient hospital stay (6% [1 to 3 weeks] to 12% [4 to 6 weeks]; P = .01), with an increasing trend to 15 weeks (18%). Correlation of bacteriologic results with physical findings demonstrated that infants with non-CoNS isolates exhibited conjunctival edema, erythema, or exudates more frequently than did infants with CoNS alone (30% vs 13%; P = .0001).

Conjunctival colonization was common among infants in a NICU. Prolonged hospitalization predisposes to colonization with potentially pathogenic organisms. Physical findings were more likely in patients with non-CoNS conjunctival isolates.

To determine the cause of an outbreak of Klebsiella pneumoniae bloodstream infections (BSIs) among neonates in a high-risk nursery and to institute control measures.

During the on-site investigation, a cohort study to identify risk factors for K. pneumoniae BSI, a point-prevalence study to assess K. pneumoniae colonization, a maternal cohort study to determine maternal K. pneumoniae colonization, and an observational study to evaluate healthcare worker (HCW) compliance with infection control practices were conducted.

Neonates in a 40-bed high-risk nursery in a 700-bed university hospital in Cali, Colombia.

Cohorting of neonates colonized with K. pneumoniae.

The overall K. pneumoniae BSI attack rate was 10 of 105 (9.5%). In the retrospective cohort study, the number of blood transfusions (OR, 3.1 per transfusion; P = .02; CI95,1.4–9.7) and intravenous injections (OR, 1.2 per injection; P = .04; CI95, 1.0–1.5) were independently associated with K. pneumoniae BSI. The overall prevalence of K. pneumoniae colonization was 61% among neonates and 7% among mothers. During the HCW assessment, suboptimal intravenous therapy practices were observed. A cohorting intervention resulted in a significant reduction in K. pneumoniae colonization (12% vs 61%; RR, 0.19; P < .001). During the intervention period, no K. pneumoniae BSIs occurred.

This investigation suggested that the outbreak probably occurred due to widespread colonization and suboptimal infection control and intravenous therapy practices. Cohorting successfully reduced the overall prevalence of K. pneumoniae colonization and, along with improved infection control practices, probably prevented K. pneumoniae BSIs.

To assess the rate of bacterial contamination of intravenous administration sets at their rubber injection ports and matching infusates.

Cultures of injection ports and infusate during 26 visits to 4 hospitals.

Four public general pediatric hospitals in Mexico City with substandard care practices.

Hospitalized pediatric patients receiving intravenous solutions.

Overall, 176 of 251 injection ports were contaminated (70.1%; 95% confidence interval [CI95], 64.5% to 75.8%), 35 (13.9%; CI95, 9.7% to 18.2%) with gram-negative rods, primarily of the tribe Klebsielleae. Cultures of infusates were positive in 17 cases (6.8%, CI95, 3.7% to 9.9%), 5 of which grew gram-negative rods (2%; CI95, 0.6% to 4.6%). In 3 cases (1.2%), the same species with gram-negative rods was found in the infusates and on the injection ports. During one visit, 8 clustered cases of injection port contamination with a clonal Enterobacter cloacae were found; this agent was also found in the blood culture, intravenous fluid, and parenteral nutrition of one patient. Inadequate chlorination of tap water, a potential risk factor, was recorded during 22 visits (84.6%).

These data suggest that external contamination of the intravenous administration set could play a role in infusate contamination.

To identify the source of an epidemic of Burkholderia cepacia bloodstream infections during 7 years (411 episodes in 361 patients).

Outbreak investigation.

A 250-bed university hospital in Beirut, Lebanon.

Matched case-control and retrospective cohort studies, and microbiological surveillance and polymerase chain reaction–restriction fragment length ascertainment were employed. Special media and filtration techniques were used to isolate organisms from water and diluted alcohol solutions.

In a group of 50 randomly selected case-matched patients from 1999, the positive blood cultures were concomitant with fever in 98%, intravenous phlebitis in 44%, and recurrent bacteremia in 20%. Fever disappeared approximately 6 hours after intravenous catheter removal. Polymerase chain reaction–restriction fragment length polymorphism revealed strain homogeneity in patient, water, and alcohol isolates. Contaminated tap water had been used to dilute alcohol for skin antisepsis and for decontamination of the caps of heparin vials. Only sporadic cases directly attributable to breach of protocol were reported after single-use alcohol swabs were substituted.

This is potentially the largest single-source nosocomial bloodstream infection outbreak ever reported, and the first report of an alcohol skin antiseptic contaminated by tap water as a source for nosocomial bacteremia.

To determine the optimal time interval for the routine replacement of intravenous administration sets when crystalloids or parenteral nutrition are administered via a central or peripheral catheter in an acute care setting.

Systematic review of all randomized or systematically allocated controlled trials addressing the frequency of replacing intravenous administration sets.

The Cochrane Controlled Trials Register (June 2001) and the Ovid databases (Medline, CINAHL, and CancerLit-July 2001) were searched. Bibliographies, relevant conference proceedings, and any product information were also checked for references.

Eighteen studies were selected for review. The 12 included studies were separated into 3 intravenous administration set change comparisons; 24 hours versus 48 hours or more; 48 hours versus 72 hours or more; and 72 hours versus 96 hours or more. There was good evidence that changing intravenous administration sets every 72 hours or more does not increase the risk of infusate-related bloodstream infection (BSI) in patients with central or peripheral catheters and a fair level of evidence that it does not increase the risk of catheter-related BSI. There were insufficient data regarding the incidence of BSI among patients receiving parenteral nutrition, particularly lipid-containing parenteral nutrition.

It appears that intravenous administration sets containing crystalloids can be changed in patients with central or peripheral catheters every 72 hours or more without increasing the risk of BSI. However, it is not possible to conclude that intravenous administration sets containing parenteral nutrition, particularly lipid-containing parenteral nutrition, can be changed at this interval.

Nosocomial infections are an important public health problem in many developing countries, particularly in the intensive care unit (ICU) setting. No previous data are available on the incidence of device-associated nosocomial infections in different types of ICUs in Argentina.

We performed a prospective nosocomial infection surveillance study during the first year of an infection control program in six Argentinean ICUs. Nosocomial infections were identified using the Centers for Disease Control and Prevention National Nosocomial Infections Surveillance System definitions, and site-specific nosocomial infection rates were calculated.

The rate of catheter-associated bloodstream infections in medical-surgical ICUs was 30.3 per 1,000 device days; it was 14.2 per 1,000 device-days in coronary care units (CCUs). The rate of ventilator-associated pneumonia in medical-surgical ICUs was 46.3 per 1,000 device-days; it was 45.5 per 1,000 device-days in CCUs. The rate of symptomatic catheter-associated urinary tract infections in medical-surgical ICUs was 18.5 per 1,000 device-days; it was 12.1 per 1,000 device-days in CCUs.

The high rate of nosocomial infections in Argentinean ICUs found during our surveillance suggests that ongoing targeted surveillance and implementation of proven infection control strategies is needed in developing countries such as Argentina.

To determine the cause of acute illness on August 30, 2000, among patients at an outpatient dialysis center (center A).

We performed a cohort study of all patients receiving dialysis on August 30, 2000; reviewed dialysis procedures; and analyzed dialysis water samples using microbiologic and chemical assays.

Dialysis center (center A).

A case-patient was defined as a patient who developed chills within 5 hours after starting hemodialysis at center A on August 30, 2000.

Sixteen (36%) of 44 patients at center A met the case definition. All case-patients were hospitalized; 2 died. Besides chills, 15 (94%) of the case-patients experienced nausea; 12 (75%), vomiting; and 4 (25%), fever. Illness was more frequent on the second than the first dialysis shift (16 of 20 vs 0 of 24, P < .001); no other risk factors were identified. The center's water treatment system had received inadequate maintenance and disinfection and a sulfurous odor was noted during sampling of the water from the reverse osmosis (RO) unit. The water had elevated bacterial counts. Volatile sulfur-containing compounds (ie, methanethiol, carbon disulfide, dimethyldisulfide, and sulfur dioxide) were detected by gas chromatography and mass spectrometry in 8 of 12 water samples from the RO unit and in 0 of 28 samples from other areas (P < .001). Results of tests for heavy metals and chloramines were within normal limits.

Parenteral exposure to volatile sulfur-containing compounds, produced under anaerobic conditions in the RO unit, could have caused the outbreak. This investigation demonstrates the importance of appropriate disinfection and maintenance of water treatment systems in hemodialysis centers.