Denise M. Murphy

Mechanical Ventilation with or without 7-Day Circuit Changes: A Randomized Controlled Trial

Nosocomial pneumonia is the leading cause of death among all hospital-acquired infections [1]. The estimated incidence of nosocomial pneumonia in intensive care units ranges from 10% to 65%; most studies [2-6] show case fatality rates of more than 20%. Ventilator-associated pneumonia specifically refers to nosocomial pneumonia that develops in a mechanically ventilated patient and that was not present at the time of airway intubation [7]. Various clinical risk factors have been associated with an increased incidence of ventilator-associated pneumonia, either because they predispose the patient to bacterial colonization of the oropharynx and stomach (for example, the administration of antacids or histamine-2-receptor antagonists) or because they facilitate aspiration of contaminated contents from these sites (for example, supine positioning) [1, 2, 8, 9]. Craven and colleagues [10] first showed that the frequency of ventilator circuit changes also influences the incidence of ventilator-associated pneumonia. They found that changing circuits every 24 rather than every 48 hours was independently associated with the occurrence of nosocomial pneumonia [10]. This association has been attributed to increased manipulation of the patient, the endotracheal tube, and the ventilator circuit, which results in increased aspiration of contaminated tubing condensate or upper airway secretions [10, 11]. More recently, several groups of investigators have found that ventilator circuits can be used safely for more than 48 hours without increasing the incidence of nosocomial pneumonia [12-16]. However, because of limitations in the design of these studies and the small number of patients prospectively examined, the Centers for Disease Control and Prevention has given no clear recommendation for the maximum length of time that ventilator circuits can safely be left in place during prolonged mechanical ventilation [17]. This has resulted in the development of ambiguous guidelines about the frequency with which ventilator circuits should be changed [18, 19] and in a call for well-designed investigations to resolve this issue [20]. We did a randomized, controlled trial to compare the effect and cost-efficacy of routine and no routine ventilator circuit changes in patients having prolonged mechanical ventilation. Our main goals were to determine 1) the incidence and outcome of ventilator-associated pneumonia in patients receiving scheduled ventilator circuit changes and 2) whether this incidence was increased in patients whose ventilator circuits remained unchanged. Methods Study Location and Patients The study was conducted at two university-affiliated teaching hospitals: Barnes Hospital (900 beds) and Jewish Hospital (450 beds). During a 7-month period (June 1994 to December 1994), all patients receiving mechanical ventilation in the intensive care units of these hospitals (surgical, trauma, medical, cardiothoracic, and neurosurgical units at Barnes Hospital; surgical, medical, and cardiothoracic units at Jewish Hospital) were potentially eligible for this investigation. Patients were entered into the trial if they were older than 18 years and had received mechanical ventilation for more than 5 days. Mechanical ventilation for more than 5 days was predetermined, on the basis of our previous experience at these institutions [2, 21], to be necessary so that a more homogeneous cohort of patients requiring prolonged mechanical ventilation could be accrued. Patients were excluded if they were likely to be extubated within 24 hours of randomization, if they had transferred from other hospitals and had already received mechanical ventilation for more than 24 hours, if they had had lung transplantation, or if they had active hemoptysis. Barnes Hospital and Jewish Hospital share the same respiratory therapy and infection control departments. The study was approved by the Washington University School of Medicine Human Studies Committee and the Institutional Review Board of Jewish Hospital. Both waived the requirement for informed consent because this study was a quality assessment of two low-risk practices already in clinical use. Study Design Patients were randomly assigned to receive no routine ventilator circuit changes or circuit changes every 7 days within 24 hours of meeting eligibility criteria. A schedule of changing ventilator circuits every 7 days was selected on the basis of available clinical data [12-16] and our survey of 16 regional medical centers (DM Baker. Unpublished communication). Stratification according to hospital site was done before randomization to control for differences in patient populations and health care personnel. Randomization within each hospital was done using opaque, sealed envelopes, which were opened at the time each patient was enrolled in the study. For the purposes of this investigation, ventilator circuits were defined to include gas delivery tubing, humidifier water reservoirs, water traps, and medication delivery devices (such as metered-dose inhaler chambers or adapters). Ventilator circuits could be changed at any time, at the discretion of individual care providers (physicians, nurses, and respiratory therapists), secondary to a mechanical failure of the ventilator circuit (such as an air leak) or visible soil (such as that resulting from hemoptysis or aspirated emesis). Scheduled ventilator circuit changes were done during the evening or night shifts to minimize the identification of individual patient group assignments to blinded investigators. All nonscheduled circuit changes were done when an appropriate indication for the circuit change (that is, a mechanical defect or soil) was identified. Patients transferred to the operating room for a surgical procedure (such as tracheotomy) or to diagnostic radiology received the same mechanical ventilator and circuit when they returned to the intensive care unit. The ventilators used for this study included Siemens Servo 900C (Siemens-Elema Ventilator Systems, Schaumburg, Illinois), Puritan-Bennett 700 Series (Puritan-Bennett Corporation, Carlsbad, California), and Bird 8400 Series ventilators (Bird Products Corporation, Palm Springs, California). All ventilators were equipped with wick-type humidifiers (Concha Therm III Plus, Hudson Respiratory Care, Inc., Temecula, California) filled with sterile irrigation water. All ventilator circuits were disposable (Hudson Respiratory Care, Inc., model 1613) and equipped with Y connectors. Each ventilator circuit had an attached trap for the collection of tubing condensate (Marquest Medical Products, Inc., Englewood, Colorado). As per our standard procedure, all ventilator circuits were monitored at least every 2 hours and water traps were emptied when full. Data Collection For all study patients, the following characteristics were prospectively recorded by one of the investigators: age, sex, diagnosis at hospital admission, indication for mechanical ventilation, Premorbid Lifestyle score, the ratio of arterial blood oxygen tension to the concentration of inspired oxygen (Pao 2: Fio 2), severity of illness based on APACHE II (Acute Physiology and Chronic Health Evaluation [22]) scores, the Organ System Failure Index, and the occurrence of a witnessed aspiration event. Specific processes of medical care examined to assess risk factors for ventilator-associated pneumonia were the administration of antacids or histamine-2-receptor antagonists, pharmacologic aerosol treatments during mechanical ventilation (such as bronchodilators, antibiotics, and mucolytics), fiberoptic bronchoscopy, surgical tracheostomy, and the number of ventilator circuit changes done and the indications for those changes (scheduled according to the study protocol, soil, or mechanical defect). Two of the investigators made daily rounds in the intensive care units of each hospital to identify eligible patients. Patients entered into the study were prospectively followed for the occurrence of ventilator-associated pneumonia until they were successfully weaned from mechanical ventilation, were discharged from the hospital, or died. All patients suspected by these investigators of having ventilator-associated pneumonia were prospectively and independently reviewed by another investigator who was blinded to the patients treatment group assignment. The diagnosis of ventilator-associated pneumonia was strictly based on the predetermined criteria described below. Patients could not be entered into the study more than once during the same hospitalization, and only the first episode of ventilator-associated pneumonia was evaluated. In addition to the occurrence of ventilator-associated pneumonia, secondary outcomes assessed included the length of hospitalization, the duration of mechanical ventilation, hospital mortality, and mortality directly attributed to ventilator-associated pneumonia. All study variables were recorded in data collection books maintained at each of the participating hospitals. Definitions All definitions were selected prospectively as part of the original study design. The Premorbid Lifestyle score was used as previously defined [23]: Zero indicated that the patient was employed without restriction; 1 indicated that the patient was independent, fully ambulatory, not employed, or employed with restriction; 2 indicated that the patient had restricted activities, could live alone and get out of the house to do basic necessities, or had severely limited exercise ability; 3 indicated that the patient was housebound, could not get out of the house unassisted, could not live alone, or could not do heavy chores; and 4 indicated that the patient was bed- or chairbound. We calculated APACHE II scores on the basis of clinical data available from the 24-hour period before study enrollment (day 5 of mechanical ventilation). The Organ System Failure Index was modified from that used by Rubin and coworkers [24]. One point was given for acquire

Results of a comprehensive infection control program for reducing surgical-site infections in coronary artery bypass surgery.

OBJECTIVE To evaluate the efficacy of a comprehensive infection control program on the reduction of surgical-site infections (SSIs) following coronary artery bypass graft (CABG) surgery. DESIGN Prospective cohort study. SETTING 1,000-bed tertiary-care hospital. PATIENTS Persons undergoing CABG with or without concomitant valve surgery from April 1991 through December 1994. INTERVENTIONS Prospective surveillance, quarterly reporting of SSI rates, chlorhexidene showers, discontinuation of shaving, administration of antibiotic prophylaxis in the holding area, elimination of ice baths for cooling of cardioplegia solution, limitation of operating room traffic, minimization of flash sterilization, and elimination of postoperative tap-water wound bathing for 96 hours. Logistic regression models were fitted to assess infection rates over time, adjusting for severity of illness, surgeon, patient characteristics, and type of surgery. RESULTS 2,231 procedures were performed. A reduction in infection rates was noted at all sites. The rate of deep chest infections decreased from 2.6% in 1991 to 1.6% in 1994. Over the same period, the rate of leg infections decreased from 6.8% to 2.7%, and of all SSI from 12.4% to 8.9%. The adjusted odds ratio (OR) for all SSIs for the end of 1994 compared to December 31, 1991, was 0.37 (95% confidence interval [CI95], 0.22-0.63). For deep chest and mediastinal infections, the adjusted OR comparing the same period was 0.69 (CI95, 0.28-1.71). CONCLUSIONS We observed significant reductions in SSI rates of deep and superficial sites in CABG surgery following implementation of a comprehensive infection control program. These differences remained significant when adjusted for potential confounding covariables.

Results of a comprehensive infection control program for reducing surgical-site infections in coronary artery bypass surgery: further data from the authors.

We welcome Dr. Lee’s contribution to the discussion of the issues raised in our article1 and value his demonstrated perspicacity and clarity. The leg wounds of the patients in our study were closed by subcuticular closure with Dexon suture in the majority of cases during the study years. When time constraints dictated, or in particularly obese patients, there was a preference for skin staples. There was no systematic change in surgical techniques with regard to leg-wound closure during the study. It is, however, likely that greater attention was paid to issues such as wound irrigation and wound drainage during the study, as individual surgical assistants responsible for wound closure were receiving feedback regarding their specific surgical-site infection rates. To detect possible surgical-site infections, we used a prospective cardiothoracic (CT) surgery database and an infection control (IC) database. A coordinator reviewed all cases and all outcomes. The coordinators and IC staff remained the same throughout the study period. IC staff made rounds on the ward and the intensive care unit 2 or 3 times per week. The CT nurses and medical staff called the IC department for every suspicious wound. IC staff reviewed the hospital chart and diagnosis of suspected cases. We monitored antibiotic use, as well as microbiology culture results. The practice of all cardiac surgeons at Barnes Hospital is to see patients 1 month postoperatively, and greater than 90% are seen here at least once in followup. Infection control nurses at other hospitals call us if they identify an infection thought to be due to an operation in Barnes Hospital. The primary outcome in this study was total surgicalsite infections. This decreased significantly. We also looked at two subgroups: deep chest infections, as defined in our article, and combined deep and superficial incisional leg infections, referred to in the article as “leg infections.” Most of the improvement in infection rates occurred in leg infections (Table 1). We did not show a statistically significant decrease in deep chest infections; however, our sample size is too small for the study to have the power to detect a clinically important change in this outcome. Superficial incisional chest infection rates increased for the first 3 years and decreased to a rate above the initial rate in the last year of the study. Our term deep chest infections is not standard terminology, but it is unambiguously defined in the “Methods” section of the article in terms of the Centers for Disease Control and Prevention categories outlined by Dr. Lee. It includes mediastinal and sternal infections, but excludes cutaneous and subcutaneous infections of the sternal incision. We have one reservation about the validity of the analysis and have addressed this as described below. In general, any regression model assumes that the outcome of interest for a particular individual is independent of all other individuals in the study sample. Based on this assumption, one calculates the errors of the estimates, and thus one generates a probability value for hypothesis testing. However, in a situation like the one in our study, the independence of the events cannot be taken for granted. Is it possible that when one individual gets a surgical-site infection others are at higher risk? This certainly would be biologically plausible. Events that occur over a period of time often have a tendency toward related outcomes for events that are in close temporal proximity. This is timeseries auto-correlation. If it were present in a series of cases such as we present, it would invalidate the hypothesis testing in the logistic regression model. We have tested for the presence of auto-correlation in the residual values for the models described in the article. In this case, none was found. However, it is possible that the model could be sensitive to small degrees of auto-correlation that were not statistically significant.