Abstract
The purpose of this study was to determine the impact of a scheduled change of antibiotic classes, used for the empiric treatment of suspected gram-negative bacterial infections, on the incidence of ventilator-associated pneumonia and nosocomial bacteremia. Six hundred eighty patients undergoing cardiac surgery were evaluated. During a 6-mo period (i.e., the before-period), our traditional practice of prescribing a third generation cephalosporin (ceftazidime) for the empiric treatment of suspected gram-negative bacterial infections was continued. This was followed by a 6-mo period (i.e., the after-period) during which a quinolone (ciprofloxacin) was used in place of the third-generation cephalosporin. The incidence of ventilator-associated pneumonia was significantly decreased in the after-period (n = 327) compared with the before-period (n = 353) (6.7 versus 11.6%; p = 0.028). This was primarily due to a significant reduction in the incidence of ventilator-associated pneumonia attributed to antibiotic-resistant gram-negative bacteria (0.9 versus 4.0%; p = 0.013). Similarly, we observed a lower incidence of bacteremia attributed to antibiotic-resistant gram-negative bacteria in the after-period compared with the before-period (0.3 versus 1.7%; p = 0.125). These data suggest that a scheduled change of antibiotic classes can reduce the incidence of ventilator-associated pneumonia attributed to antibiotic-resistant gram-negative bacteria.
Hospital-acquired pneumonia and bacteremia are the leading causes of death from nosocomial infection in critically ill patients (1-4). The prior administration of antibiotics, particularly broad-spectrum antibiotics such as the third-generation cephalosporins, has recently become recognized as an important risk factor for both hospital-acquired pneumonia and bacteremia caused by antibiotic-resistant gram-negative bacteria (5-9). Antimicrobial agents have become one of the most frequently prescribed classes of drugs in hospitalized patients (10). Intensive care units, partly because of the density of antimicrobial usage in these confined areas, are recognized as important locations for the emergence and spread of infections caused by antibiotic-resistant microorganisms (5-9).
In an attempt to reverse the trend of increasing antimicrobial resistance among hospital-acquired infections, several antimicrobial control strategies have been advocated that directly restrict the use of antibiotics by physicians or offer practice guidelines for the administration of these drugs (11). As part of our hospital's ongoing quality improvement program, we performed a prospective before-after study in patients undergoing cardiac surgery. The main goal of our study was to determine whether the incidence of ventilator-associated pneumonia or nosocomial bacteremia attributed to antibiotic-resistant gram-negative bacteria could be reduced using a scheduled change of antibiotic classes.
The study was conducted at a university-affiliated teaching hospital: Barnes-Jewish Hospital (900 beds). During a 12-mo period (August 1995 through August 1996), all patients undergoing cardiac surgery were potentially eligible for this investigation. Patients were excluded if they were younger than 18 yr of age and if they were undergoing heart transplantation. The study was approved by the Washington University School of Medicine Human Studies Committee, and the requirement for informed consent was waived.
We employed a before–after study design to test the hypothesis that a scheduled change of antibiotic classes, used for the empiric treatment of suspected gram-negative bacterial infections, can reduce the incidence of nosocomial pneumonia or bacteremia attributed to antibiotic-resistant gram-negative bacteria. Patients in the before-period received a third generation cephalosporin (ceftazidime) for the empiric treatment of suspected gram-negative bacterial infections. Third generation cephalosporins have been well documented to represent the major antibiotic class utilized for the empiric treatment of gram-negative bacterial infections at Barnes-Jewish Hospital prior to the performance of this investigation (12, 13). Patients in the after-period received a quinolone (ciprofloxacin) for the same indication. Empiric administration of vancomycin and/or an aminoglycoside could also be prescribed to patients in both study periods according to the clinical impression of the treating physicians. The need for empiric antibiotics and the duration of antibiotic treatment were solely determined by the physicians caring for the study patients.
One of the investigators (J.V.) made daily rounds with the cardiac surgery patient care physician team to insure optimal understanding and compliance with the study protocol. Additionally, monthly in-services were given to the nursing staff and physician housestaff teams working in the cardiothoracic intensive care unit by a nurse from the Department of Infection Control (D.M.). These in-services were targeted at maintaining a uniform practice of infection control procedures throughout both study periods.
For all study patients, the following characteristics were prospectively recorded by one of the investigators: age, sex, race, Premorbid Lifestyle Score (14), prophylactic antibiotic administration, serum albumin (g/L), the duration of surgery from the first skin incision until closure, the time spent on cardiopulmonary bypass, the aortic cross-clamp time, the ratio of arterial blood oxygen tension to the concentration of inspired oxygen (PaO2 /Fi O2 ) at the time of ICU admission, severity of illness based on APACHE II (Acute Physiology and Chronic Health Evaluation) scores (15), the presence of presurgical conditions, including congestive heart failure requiring medical therapy with diuretics and/or vasodilators, chronic obstructive pulmonary disease requiring medical therapy with inhaled bronchodilators or corticosteroids, underlying malignancy, and immunosuppression, the specific type of cardiac surgery performed, and whether or not the surgery was performed on an emergent basis. Specific processes of medical care examined after cardiac surgery included the use of hemodialysis, intra-aortic balloon counterpulsation, extracorporeal membrane oxygenation, presence of a ventricular assist device, reintubation, use of a nasogastric tube, presence of a tracheostomy, administration of antacids, histamine-2-receptor antagonists, sucralfate, or aerosol therapy (including bronchodilators, mucolytics, and antibiotics), fiberoptic bronchoscopy, positioning of the head of the bed, use of chemical paralysis, postoperative administration of antibiotics, the clinical indications for postoperative antibiotics, and the duration of pulmonary artery catheter (or other central vein catheter) placement.
One of the investigators made daily rounds on all study patients recording relevant data from the medical records, bedside flow sheets, and the hospital's main frame computer for reports of microbiologic studies (sputum Gram's stains and sputum, blood, pleural fluid, and lower respiratory tract cultures) and bacterial antibiotic sensitivity profiles. All chest radiographs were prospectively reviewed by one of the investigators (M.H.K.), and the computerized radiographic reports were also reviewed (24 to 48 h later). Patients were evaluated for nosocomial pneumonia during the period of mechanical ventilatory support and for 48 h after extubation. This was based on our previous experience, which suggested that nosocomial pneumonia occurring without mechanical ventilation was uncommon in this patient population (13, 16).
All definitions were selected prospectively as part of the original study design. Nosocomial bloodstream infections were defined according to criteria established by the Centers for Disease Control and Prevention (17). The diagnostic criteria for ventilator-associated pneumonia were those established by the American College of Chest Physicians (18). Ventilator-associated pneumonia was considered to be present when a new or progressive roentgenographic infiltrate developed in conjunction with one of the following: radiographic evidence of pulmonary abscess formation (i.e., cavitation within preexisting pulmonary infiltrates); histologic evidence of pneumonia in lung tissue; a positive blood or pleural fluid culture; the presence of a positive quantitative culture of a sample of secretions from the lower respiratory tract; or two of the following clinical criteria in the absence of an alternative explanation for the pulmonary infiltrates: fever, leukocytosis, and purulent tracheal aspirate. Blood and pleural fluid cultures could not be related to another source and both had to be obtained within 48 h before or after the clinical suspicion of ventilator-associated pneumonia. Microorganisms recovered from blood or pleural fluid cultures also had to be identical to the organisms recovered from cultures of respiratory secretions.
A new radiographic infiltrate was prospectively defined as one occurring more than 48 h after the start of mechanical ventilation or within 48 h of extubation. Radiographic infiltrates occurring within 48 h of intubation were attributed to other disease processes, including perioperative aspiration (19). Therefore, only episodes of ventilator- associated pneumonia classified as late-onset according to a recent consensus statement were evaluated (18). Fever was defined as an increase in the core temperature of 1° C or more and a core temperature of more than 38.3° C. Leukocytosis was defined as a 25% increase in circulating leukocytes from baseline and a leukocyte count of more than 10 × 103/mm3 (10 × 109/L). Tracheal aspirates were considered purulent if Gram's stain showed more than 25 neutrophils per high power field.
We routinely performed minibronchoalveolar lavage (mini-BAL) in patients with suspected ventilator-associated pneumonia to obtain lower respiratory specimens for culture and Gram's stain (20). The accuracy of mini-BAL for establishing a microbiologic diagnosis of ventilator-associated pneumonia has been previously established (20, 21). Microbiologically confirmed cases of ventilator-associated pneumonia required the isolation of bacteria in significant quantity from mini-BAL samples (⩾ 103 colony forming units/mL) (20, 21). All patients were also prospectively screened for possible alternative causes for fever and radiographic chest densities as suggested by other investigators (19).
Multiorgan dysfunction was defined as derangements of three or more organ systems using the criteria of Rubin and coworkers (22). The definitions used for the systemic inflammatory response syndrome (SIRS), sepsis, severe sepsis, and septic shock, were those proposed by the American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference (23). Postoperative indications for antibiotic administration were prospectively classified as either empiric treatment or infection-directed treatment. Empiric treatment was considered to be present when antibiotics were prescribed for postoperative fever without a specific localized source of infection. Infection-directed treatment was defined as the administration of antibiotics for a specific clinically localized source of infection (e.g., pneumonia, urinary tract, wound, bloodstream) as documented in the patient's medical record. Clinically localized sources of infection, excluding bloodstream infections, did not require microbiologic confirmation by Gram's stain or culture in order to be classified as infection-directed treatment.
Antibiotic-resistant gram-negative bacteria were defined as being resistant to at least one class of antibiotics typically used in the treatment of gram-negative bacterial infections. These antibiotic classes included the aminoglycosides (gentamicin, tobramycin, amikacin), third-generation cephalosporins (ceftazidime), extended-spectrum penicillins (pipercillin), quinolones (ciprofloxacin), and the carbapenems (imipenem). Antibiotic-resistant gram-positive bacteria were defined as staphylococci and enterococci resistant to oxacillin or vancomycin.
All comparisons were unpaired and all tests of significance were two-tailed. Continuous variables were compared using Student's t test for normally distributed variables and Wilcoxon's rank-sum test for non-normally distributed variables. The chi-square test for Fisher's exact test were used to compare categorical variables. The primary data analysis compared patients from the before-period to patients from the after-period.
Multiple logistic regression analysis was used to determine the independent risk factors for ventilator-associated pneumonia, nosocomial bloodstream infections, and hospital mortality (24). A stepwise approach was used to enter new terms into the logistic regression models where 0.05 was set as the limit for the acceptance or removal of new terms. Results of the logistic regression analyses are reported as adjusted odds ratios (AORs) with 95% confidence intervals (CIs). Relative risks (RRs) and their 95% CIs were calculated using standard methods. Values are expressed as the mean ± SD (continuous variables) or as a percentage of the group from which they were derived (categorical variables). All p values were two-tailed and p ⩽ 0.05 was considered statistically significant.
A total of 680 consecutive patients were prospectively evaluated (Tables 1 and 2). The mean age of the patients was 63.8 ± 12.9 yr (range, 18 to 90 yr). Four hundred forty-one (64.9%) patients were men and 239 (35.1%) were women. The mean APACHE II Score of the entire study cohort was 12.8 ± 4.4 (range, 1.0 to 38.0). The surgical procedures performed on these patients included 555 (81.6%) coronary artery bypass operations (33 with an accompanying valve replacement), 83 (12.2%) valve operations, 29 (4.3%) maze procedures, and 13 (1.9%) miscellaneous surgeries (three aortic arch repairs, four atrial septal defect closures, four pericardiectomies, and two left ventricle stab wound repairs).
| Characteristic | Before-Period (n = 353) | After-Period (n = 327) | p Value | |||
|---|---|---|---|---|---|---|
| Age, yr* | 63.0 ± 12.9 | 64.7 ± 12.9 | 0.077 | |||
| Sex | ||||||
| Male, n (%) | 242 (68.6) | 199 (60.9) | 0.036 | |||
| Female, n (%) | 111 (31.4) | 128 (39.1) | ||||
| Race | ||||||
| White, n (%) | 315 (89.2) | 282 (86.3) | 0.477 | |||
| Black, n (%) | 35 (9.9) | 41 (12.5) | ||||
| Other, n (%) | 3 (0.9) | 4 (1.2) | ||||
| Premorbid Lifestyle Score, n (%) | ||||||
| 0 | 130 (36.8) | 83 (25.4) | < 0.001 | |||
| 1 | 207 (58.7) | 206 (63.0) | ||||
| 2 | 12 (3.4) | 31 (9.5) | ||||
| 3 | 4 (1.1) | 5 (1.5) | ||||
| 4 | 0 (0.0) | 2 (0.6) | ||||
| Congestive heart failure, n (%) | 62 (17.6) | 91 (27.8) | < 0.001 | |||
| Chronic obstructive pulmonary disease, n (%) | 18 (5.1) | 25 (7.6) | 0.173 | |||
| Underlying malignancy, n (%) | 16 (4.5) | 23 (7.0) | 0.161 | |||
| Immunosuppressed, n (%) | 7 (2.0) | 5 (1.5) | 0.653 | |||
| Albumin, g/L* | 39.4 ± 6.1 | 39.7 ± 4.4 | 0.417 | |||
| PaO2 /Fi O2 * | 269 ± 106 | 283 ± 102 | 0.072 | |||
| APACHE II Score* | 12.5 ± 4.2 | 13.1 ± 4.6 | 0.177 | |||
| Surgery type | ||||||
| CABG, n (%) | 286 (81.0) | 269 (82.3) | 0.383 | |||
| Valve, n (%) | 41 (11.6) | 42 (12.8) | ||||
| Other, n (%) | 26 (7.4) | 16 (4.9) | ||||
| Emergent surgery, n (%) | 35 (9.9) | 39 (11.9) | 0.400 | |||
| Duration of surgery, h* | 5.6 ± 1.7 | 5.4 ± 1.9 | 0.011 | |||
| Cardiopulmonary bypass time, h* | 2.5 ± 0.9 | 2.5 ± 0.9 | 0.689 | |||
| Aortic cross-clamp time, h* | 1.4 ± 0.6 | 1.5 ± 0.6 | 0.266 |
| Characteristic | Before-Period (n = 353) | After-Period (n = 327) | p Value | |||
|---|---|---|---|---|---|---|
| Dialysis, n (%) | 7 (2.0) | 9 (2.8) | 0.508 | |||
| Reintubation, n (%) | 24 (6.8) | 17 (5.2) | 0.381 | |||
| Tracheostomy, n (%) | 14 (4.0) | 8 (2.5) | 0.263 | |||
| Intra-aortic balloon counterpulsation, n (%) | 68 (19.3) | 48 (14.7) | 0.112 | |||
| ECMO, n (%) | 1 (0.3) | 0 (0.0) | > 0.99 | |||
| Ventricular assist device, n (%) | 7 (2.0) | 7 (2.1) | 0.885 | |||
| Nasogastric tube, n (%) | 345 (97.7) | 320 (97.9) | 0.911 | |||
| Antacids, n (%) | 24 (6.8) | 16 (4.9) | 0.291 | |||
| Histamine type-2 antagonists, n (%) | 62 (17.2) | 66 (20.2) | 0.383 | |||
| Sucralfate, n (%) | 325 (92.1) | 293 (89.6) | 0.264 | |||
| Aerosol administration, n (%) | 47 (13.3) | 44 (13.5) | 0.957 | |||
| Witnessed aspiration, n (%) | 6 (1.7) | 1 (0.3) | 0.125 | |||
| Bronchoscopy, n (%) | 23 (6.5) | 7 (2.1) | 0.006 | |||
| Semi-recumbent head positioning, n (%) | 328 (92.9) | 313 (95.7) | 0.117 | |||
| Chemical paralysis, n (%) | 66 (18.7) | 33 (10.1) | 0.001 | |||
| Duration of Pa catheterization, d* | 2.1 ± 1.7 | 2.0 ± 1.9 | 0.233 | |||
| Duration of central vein catheterization, d*,† | 2.1 ± 4.7 | 2.3 ± 6.5 | 0.269 |
The clinical outcomes evaluated for the two study groups are shown in Table 3. The rates of ventilator-associated pneumonia (relative risk, 0.58; 95% CI, 0.35 to 0.95; p = 0.028) and ventilator-associated pneumonia attributed to antibiotic-resistant gram-negative bacteria (relative risk, 0.23; 95% CI, 0.07 to 0.80; p = 0.013) were significantly lower during the after-period compared with the before-period (Figure 1). The duration of mechanical ventilation prior to the onset of ventilator-associated pneumonia was similar for both study periods (after-period, 9.9 ± 12.8 d; before-period, 6.0 ± 3.5 d; p = 0.518). The distribution of the microorganisms recovered from the lower airways of patients with ventilator-associated pneumonia are shown in Table 4.
| Outcome | Before-Period (n = 353) | After-Period (n = 327) | p Value | |||
|---|---|---|---|---|---|---|
| VAP, n (%) | 41 (11.6) | 22 (6.7) | 0.028 | |||
| Microbiologically confirmed VAP, n (%) | 34 (9.6) | 18 (5.5) | 0.043 | |||
| VAP caused by antibiotic-resistant GNB, n (%) | 14 (4.0) | 3 (0.9) | 0.013 | |||
| Bloodstream infection, n (%) | 13 (3.7) | 11 (3.4) | 0.822 | |||
| Bloodstream infection caused by antibiotic-resistant GNB, n (%) | 6 (1.7) | 1 (0.3) | 0.125 | |||
| Sepsis, n (%) | 47 (13.3) | 30 (9.2) | 0.089 | |||
| Severe sepsis, n (%) | 16 (4.5) | 8 (2.5) | 0.141 | |||
| Septic shock, n (%) | 4 (1.1) | 2 (0.6) | 0.687 | |||
| Duration of mechanical ventilation, d* | 2.5 ± 4.8 | 2.4 ± 4.6 | 0.799 | |||
| Length of intensive care, d* | 3.7 ± 5.0 | 3.5 ± 4.7 | 0.729 | |||
| Hospital length of stay, d* | 11.8 ± 8.7 | 13.1 ± 11.4 | 0.280 | |||
| Acquired organ system derangements, n* | 0.9 ± 1.1 | 0.8 ± 1.2 | 0.708 | |||
| Hospital mortality, n (%) | 18 (5.1) | 26 (8.0) | 0.131 | |||
| Hospital mortality attributed to VAP caused by antibiotic-resistant GNB, n (%) | 6 (1.7) | 2 (0.6) | 0.289 |
Fig. 1. Occurrence rates of ventilator-associated pneumonia (VAP) in the Before-Period compared with those in the After-Period for all etiologies (closed columns) and for pneumonia attributed to antibiotic-resistant gram-negative bacteria (ARGNB) (open columns).
[More] [Minimize]| Before-Period (n = 41) | After-Period (n = 22) | ||||||
|---|---|---|---|---|---|---|---|
| Pseudomonas aeruginosa | 7 | Pseudomonas aeruginosa | 3 | ||||
| Serratia marcescens | 7 | Enterobacter species | 3 | ||||
| Enterobacter species | 4 | OSSA | 3 | ||||
| OSSA | 4 | Klebsiella species | 2 | ||||
| Cytomegalovirus† | 3 | Stenotrophomonas maltophilia | 2 | ||||
| Aspergillus species | 3 | Hemophilus influenzae | 2 | ||||
| Citrobacter freundii | 2 | Acinetobacter baumannii | 1 | ||||
| Morganella morganii | 2 | Proteus mirabilis | 1 | ||||
| Klebsiella species | 2 | Escherichia coli | 1 | ||||
| Acinetobacter baumannii | 2 | ORSA | 1 | ||||
| Proteus mirabilis | 2 | Alcaligenes xylosoxidans | 1 | ||||
| Stenotrophomonas maltophilia | 2 | Streptococcus pneumoniae | 1 | ||||
| Herpes simplex virus† | 2 | Herpes simplex virus† | 1 | ||||
| Hemophilus influenza | 1 | ||||||
| Escherichia coli | 1 | ||||||
| ORSA | 1 | ||||||
| Candida species | 1 | ||||||
Eleven patients (seven in the before-period, four in the after-period) meeting the study criteria for ventilator-associated pneumonia, and with no other explanation for their fever and radiographic infiltrates, had negative mini-BAL culture results. All 11 of these patients had been started on broad spectrum antibiotics at least 12 h prior to obtaining the mini-BAL samples. The rate of ventilator-associated pneumonia, which was microbiologically confirmed using mini-BAL, was significantly lower during the after-period than during the before-period (relative risk, 0.57; 95% CI, 0.33 to 0.99; p = 0.043) (Table 3).
The incidence of bloodstream infections was similar for patients in the after-period compared with patients in the before-period (relative risk, 0.91; 95% CI, 0.42 to 2.01; p = 0.822) (Figure 2). The incidence of bloodstream infections caused by antibiotic-resistant gram-negative bacteria was less in the after-period than in the before-period, although this difference did not reach statistical significance (relative risk, 0.18; 95% CI, 0.02 to 1.49; p = 0.125). The average duration of intensive care prior to the development of a nosocomial bloodstream infection was similar for both study periods (after-period, 9.3 ± 8.9 d; before-period, 11.2 ± 14.0 d; p = 0.693). The distribution of pathogens recovered from the blood cultures of patients with a nosocomial bloodstream infection are shown in Table 5.
Fig. 2. Occurrence rates of bloodstream infections in the Before-Period compared with those in After-Period for all etiologies (closed columns) and for bacteremia attributed to antibiotic-resistant gram-negative bacteria (ARGNB) (open columns).
[More] [Minimize]| Before-Period (n = 13) | After-Period (n = 11) | ||||||
|---|---|---|---|---|---|---|---|
| OSSA | 3 | Candida species | 5 | ||||
| Candida species | 3 | Enterococcus species | 4 | ||||
| ORSA | 2 | OSSA | 1 | ||||
| Pseudomonas aeruginosa | 2 | Pseudomonas aeruginosa | 1 | ||||
| Enterobacter species | 2 | Enterobacter species | 1 | ||||
| Serratia marcescens | 1 | Serratia marcescens | 1 | ||||
| Klebsiella species | 1 | Klebsiella species | 1 | ||||
| Proteus mirabilis | 1 | ||||||
| Acinetobacter baumannii | 1 | ||||||
| Enterococcus species | 1 | ||||||
Among the 41 gram-negative bacterial isolates associated with either ventilator-associated pneumonia or bacteremia in the before-period, 20 (48.8%) were classified as antibiotic-resistant. This was significantly greater than the number of gram-negative isolates associated with infection that were classified as antibiotic-resistant in the after-period (four of 20 bacterial isolates [20.0%]; p = 0.05). Thirteen (31.7%) of the gram-negative isolates associated with infection in the before-period were resistant to ceftazidime and five (12.2%) were resistant to ciprofloxacin. Two (10%) of the gram-negative isolates associated with infection in the after-period were resistant to ceftazidime (p = 0.111 compared with the before-period) and two (10%) were resistant to ciprofloxacin (p = 0.801 compared with the before-period).
Univariate analysis identified 23 potential risk factors for ventilator-associated pneumonia (Table 6). Multiple logistic regression analysis demonstrated that four variables were independently associated with the development of ventilator-associated pneumonia (Table 7). Multiple logistic regression analysis also identified the duration of pulmonary artery catheterization in 1-d increments (adjusted odds ratio, 1.28; 95% CI, 1.17 to 1.39; p < 0.001), prior administration of empiric antibiotics (adjusted odds ratio, 1.72; 95% CI, 1.43 to 2.06; p < 0.001), duration of surgery in 1-h increments (adjusted odds ratio, 1.34; 95% CI, 1.21 to 1.48; p = 0.004), and the duration of central vein catheterization in 1-d increments (adjusted odds ratio, 1.05; 95% CI, 1.03 to 1.08; p = 0.016) as independent risk factors for the development of a nosocomial bloodstream infection.
| Variable | Ventilator-associated Pneumonia | p Value | ||||
|---|---|---|---|---|---|---|
| Present (n = 63) | Absent (n = 617) | |||||
| Age, yr* | 66.8 ± 12.9 | 63.5 ± 12.9 | 0.056 | |||
| Albumin, g/L* | 37.6 ± 4.4 | 39.8 ± 5.4 | < 0.001 | |||
| APACHE II Score* | 15.9 ± 4.8 | 12.4 ± 4.3 | < 0.001 | |||
| Duration of surgery, h* | 6.9 ± 2.3 | 5.4 ± 1.7 | < 0.001 | |||
| Cardiopulmonary bypass time, h* | 3.0 ± 1.3 | 2.4 ± 0.8 | < 0.001 | |||
| Aortic cross-clamp time, h* | 1.7 ± 0.9 | 1.4 ± 0.6 | 0.008 | |||
| Duration of mechanical ventilation, d* | 7.4 ± 8.2† | 1.5 ± 1.4 | < 0.001 | |||
| Congestive heart failure, n (%) | 20 (31.8) | 133 (21.6) | 0.065 | |||
| Chronic obstructive pulmonary disease, n (%) | 8 (12.7) | 35 (5.7) | 0.029 | |||
| Requiring hemodialysis, n (%) | 6 (9.5) | 10 (1.6) | 0.001 | |||
| Reintubation, n (%) | 24 (38.1) | 17 (2.8) | < 0.001 | |||
| Tracheostomy, n (%) | 18 (28.6) | 4 (0.6) | < 0.001 | |||
| Intra-aortic balloon counterpulsation, n (%) | 30 (47.6) | 86 (13.9) | < 0.001 | |||
| Ventricular assist device, n (%) | 4 (6.4) | 10 (1.6) | 0.033 | |||
| Antacids, n (%) | 8 (12.7) | 32 (5.2) | 0.016 | |||
| Histamine type-2 antagonists, n (%) | 21 (33.3) | 107 (17.3) | 0.002 | |||
| Aerosol administration, n (%) | 22 (34.9) | 69 (11.2) | < 0.001 | |||
| Witnessed aspiration, n (%) | 5 (7.9) | 2 (0.3) | < 0.001 | |||
| Bronchoscopy, n (%) | 16 (25.4) | 14 (2.3) | < 0.001 | |||
| Chemical paralysis, n (%) | 23 (36.5) | 76 (12.3) | < 0.001 | |||
| Use of vasopressors/inotropic agents n (%) | 60 (95.2) | 427 (69.2) | < 0.001 | |||
| Empiric postoperative antibiotics, n (%) | 51 (81.0) | 207 (33.5) | < 0.001 | |||
| Patient assignment to Before Period, n (%) | 41 (65.1) | 312 (50.6) | 0.028 | |||
| Variable | Adjusted Odds Ratio | 95% CI | p Value | |||
|---|---|---|---|---|---|---|
| Duration of mechanical ventilation†: (1-day increments) | 1.60 | 1.48–1.73 | < 0.001 | |||
| Prior administration of empiric antibiotics | 2.20 | 1.73–2.79 | < 0.001 | |||
| Patient assignment to Before Period | 4.14 | 2.39–7.18 | 0.009 | |||
| Reintubation | 3.87 | 2.07–7.24 | 0.028 |
Perioperative antibiotic prophylaxis was administered to 670 (98.5%) patients. Ten patients did not receive antibiotic prophylaxis. Antibiotic prophylaxis consisted of a first-generation cephalosporin in 464 (69.3%) patients, vancomycin in 112 (16.7%) patients, and vancomycin plus a first generation cephalosporin in 94 (14.0%) patients. Prophylactic antibiotics were initially administered preoperatively in 659 (98.4%) patients, intraoperatively in five (0.7%) patients, and postoperatively in six (0.9%) patients. The average duration of prophylactic antibiotic administration was 2.0 ± 2.3 d. Patients developing a bloodstream infection (37.5%) were statistically more likely to receive vancomycin plus a first-generation cephalosporin compared with patients without a bloodstream infection (13.2%) (relative risk, 2.86; 95% CI, 1.65 to 5.0; p < 0.001). No other significant associations were found between antibiotic prophylaxis and the occurrence of ventilator-associated pneumonia or nosocomial bloodstream infections (p > 0.20).
Overall, 296 (43.5%) patients received postoperative antibiotics in addition to their perioperative antibiotic prophylaxis (Table 8). The rate of exposure to postoperative antibiotics was similar for both the before-period and the after-period (44.5 versus 42.5%; p = 0.605). Among patients receiving postoperative antibiotics, 183 (61.8%) received empiric therapy and 113 (38.2%) received infection-directed therapy. The indications for postoperative antibiotics were similar in two study groups (p = 0.463). Patients in the before-period were significantly more likely to receive ceftazidime and less likely to receive ciprofloxacin than were patients in the after-period. Patients in the before-period were also more likely to receive vancomycin or an aminoglycoside than were patients in the after-period; however, these differences were not statistically significant.
| Antibiotic Exposure | Before-Period (n = 353) | After-Period (n = 327) | p Value | |||
|---|---|---|---|---|---|---|
| Third-generation cephalosporins | ||||||
| Patients exposed, n (%) | 70 (19.8) | 24 (7.3) | < 0.001 | |||
| Days of exposure, entire cohort* | 1.1 ± 3.0 | 0.4 ± 2.3 | < 0.001 | |||
| Days of exposure, patients prescribed drug* | 5.7 ± 4.5 | 5.7 ± 6.4 | 0.289 | |||
| Ciprofloxacin | ||||||
| Patients exposed, n (%): | 11 (3.1) | 61 (18.7) | < 0.001 | |||
| Days of exposure, entire cohort* | 0.2 ± 1.8 | 1.0 ± 2.9 | < 0.001 | |||
| Days of exposure, patients prescribed drug* | 8.0 ± 7.0 | 5.2 ± 4.8 | 0.497 | |||
| Vancomycin | ||||||
| Patients exposed, n (%) | 151 (42.8) | 128 (39.1) | 0.336 | |||
| Days of exposure, entire cohort* | 1.8 ± 3.6 | 1.4 ± 2.7 | 0.247 | |||
| Days of exposure, patients prescribed drug* | 4.2 ± 4.4 | 3.6 ± 3.4 | 0.407 | |||
| Aminoglycosides | ||||||
| Patients exposed, n (%) | 43 (12.2) | 31 (9.5) | 0.258 | |||
| Days of exposure, entire cohort* | 0.4 ± 1.5 | 0.3 ± 2.0 | 0.273 | |||
| Days of exposure, patients prescribed drug* | 3.0 ± 3.5 | 3.6 ± 5.5 | 0.617 |
The mortality rate of patients in the before-period was not significantly different from the mortality rate of patients in the after-period (relative risk, 0.64; 95% CI, 0.36 to 1.15; p = 0.131) (Table 3). The mortality rate of patients with ventilator-associated pneumonia was significantly greater than the mortality rate of patients without ventilator-associated pneumonia (27.0 versus 4.4%; p < 0.001). Similarly, patients with a nosocomial bloodstream infection had a significantly greater mortality rate than did patients without a nosocomial bloodstream infection (37.5 versus 5.3%; p < 0.001). The mortality rate attributed to ventilator-associated pneumonia caused by antibiotic-resistant gram-negative bacteria was not significantly different between the study periods (relative risk of before-period compared with after-period, 2.78; 95% CI, 0.56 to 13.67; p = 0.289) (Table 3). Using a Poisson approximation, we estimate that 1,028 patients would have needed to be studied to see a statistically significant reduction in mortality attributed to a decrease in the incidence of ventilator-associated pneumonia caused by antibiotic-resistant gram-negative bacteria.
Multiple logistic regression analysis identified female sex (adjusted odds ratio, 3.16; 95% CI, 1.88 to 5.30; p = 0.026), chemical paralysis (adjusted odds ratio, 7.86; 95% CI, 4.64 to 13.32; p < 0.001), APACHE II Score in 1-point increments (adjusted odds ratio, 1.19; 95% CI, 1.13 to 1.26; p < 0.001), multiorgan dysfunction (adjusted odds ratio, 3.09; 95% CI, 2.55 to 3.75; p < 0.001), and the aortic cross-clamp time in 1-h increments (adjusted odds ratio, 2.06; 95% CI, 1.50 to 2.83; p = 0.024) as independent risk factors for hospital mortality.
In this preliminary investigation, we demonstrated that a scheduled change of antibiotic classes, used for the empiric treatment of suspected gram-negative bacterial infections, can significantly decrease the incidence of ventilator-associated pneumonia caused by antibiotic-resistant gram-negative bacteria. We also found a trend favoring a lower incidence of bacteremia attributed to antibiotic-resistant gram-negative bacteria among patients in the study group assigned to receive a scheduled change of antibiotic classes. These results were obtained without increasing the overall use of postoperative antibiotics. However, we observed no significant difference in hospital mortality, lengths of stay, or the development of multiorgan dysfunction. This may be due in part to the relatively small number of patients we examined whose hospital mortality could be attributed to ventilator-associated pneumonia caused by antibiotic-resistant gram-negative bacteria. Other interventional studies have similarly demonstrated decreases in infection rates without significant changes in patient mortality because of insufficient numbers of study patients (i.e., Type II error) (25).
Previous antibiotic exposure has become recognized as an important risk factor for the development of ventilator-associated pneumonia, primarily pneumonia caused by antibiotic-resistant bacteria (7, 16, 26). Fagon and coworkers (6) demonstrated that patients receiving prior antimicrobial therapy had a greater incidence of ventilator-associated pneumonia caused by Pseudomonas or Acinetobacter species than did patients without previous antimicrobial therapy. Similarly, Rello and coworkers (27) demonstrated that the use of prior antibiotics predisposed to ventilator-associated pneumonia attributed to oxacillin-resistant Staphylococcus aureus. Prior antibiotic exposure appears to increase the risk of ventilator-associated pneumonia by facilitating patient colonization with antibiotic-resistant pathogens (4, 16). Colonization of the lower respiratory tract by multiresistant organisms such as Pseudomonas aeruginosa or oxacillin-resistant staphylococci has previously been shown to be closely correlated with the subsequent development of overt pneumonia (28-30). Therefore, it is biologically plausible that efforts directed at reducing such colonization could decrease the occurrence rates of ventilator-associated pneumonia caused by antibiotic-resistant bacteria.
Prior antibiotic administration is also recognized as an important risk factor for the development of nosocomial bloodstream infections. Chow and coworkers (5) showed that patients who received a third-generation cephalosporin were significantly more likely to develop bacteremia with multiresistant Enterobacter species than were patients without such an exposure. Prior antibiotic therapy has also been demonstrated to be an independent predictor of hospital mortality among hospitalized patients with bloodstream infections (31). In part, this may be due to the development of secondary bacteremia attributed to either partially treated or antibiotic-resistant pathogens (i.e., superinfections).
The use of broad-spectrum antibiotics, particularly third-generation cephalosporins, has greatly proliferated in hospitals during the last decade (5). In our study, 33.5% of patients without ventilator-associated pneumonia received empiric antibiotics. This was presumably due to physician concerns about not treating an underlying nosocomial infection. Several clinical investigations have directly implicated the increased use of broad-spectrum antibiotics as an explanation for the increasing emergence of antibiotic-resistant infections (5, 13, 27, 32). Additionally, many gram-negative bacteria now possess extended-spectrum beta-lactamases, which usually confer resistance to all beta-lactam antibiotics except for imipenem (32). This emerging resistance to our most commonly prescribed antibiotics has contributed to recent calls for the development of improved strategies for the use of all antimicrobials in order to reverse this pattern (33, 34). In an attempt to accomplish this goal, several antibiotic utilization programs have been examined, including restricted hospital formularies, formalized antibiotic guidelines, scheduled antibiotic rotation, and computerized information systems (11).
To date, widespread acceptance and utilization of any single antibiotic control strategy has not occurred, and the administration of these drugs is still usually performed without considering the risk of superinfection (12, 33). In part, this may be due to a lack of recognition of the importance of this problem at institutions without specific identifiable outbreaks of antibiotic-resistant infections. Nevertheless, several investigations have demonstrated the potential benefits of antibiotic control programs (35, 36). Meyer and coworkers (32) showed that restricting the use of ceftazidime allowed them to control an outbreak of ceftazidime-resistant Klebsiella pneumonia. Gerding and colleagues (37) employed a scheduled rotation of amikacin and gentamicin when resistance to gentamicin by Pseudomonas aeruginosa reached high levels. This strategy reduced the incidence of gentamicin resistance and allowed it to be reintroduced for the treatment of gram-negative bacterial infections. Most recently, Pestotnik and coworkers (35) used a computer-assisted decision support system to decrease the total pharmacy acquisition costs of antibiotics, improve preoperative antibiotic utilization, and stabilize antimicrobial resistance patterns despite increased overall use of antibiotics at their institution.
There are several potential explanations for the results we observed. First, reducing the use of third-generation cephalosporins may have decreased the occurrence of ineffective antimicrobial therapy for infections due to bacteria resistant to this class of antibiotics (32). Second, the use of a quinolone class may have prevented the emergence of infections not previously suppressed by third-generation cephalosporins (38). This mechanism would be most important in patients colonized with gram-negative bacteria resistant to the third-generation cephalosporins but sensitive to the quinolones. Lastly, we cannot exclude the possibility that the decrease in the overall rate of ventilator-associated pneumonia was due to factors other than the change in antibiotic classes. This is suggested by our finding that reintubation and the duration of mechanical ventilation were independently associated with the development of ventilator-associated pneumonia (Table 7).
Several limitations of our study should be acknowledged. First, the relatively small sample size may not have sufficient statistical power to identify all important risk factors for the outcomes examined, or to identify all significant outcome differences between the study groups. Our data suggest that a larger study may have identified significant benefits from a scheduled change in antibiotic classes on the incidences of bacteremia and hospital mortality attributed to antibiotic- resistant gram-negative bacteria. Second, we examined only one change of antibiotic classes. Studies of longer duration are needed to determine the influence of scheduled changing of antibiotic classes on the incidence of nosocomial infections, antibiotic resistance patterns, and patient outcomes. Lastly, we did not perform routine surveillance cultures to determine the impact of a scheduled change in antibiotic classes on patient colonization with antibiotic-resistant bacteria. Therefore, we could not determine if the reduction in ventilator-associated pneumonia due to antibiotic-resistant gram-negative bacteria was associated with a decrease in colonization because of these pathogens.
In summary, these preliminary data suggest that a scheduled change of antibiotic classes can decrease the incidence of ventilator-associated pneumonia attributed to antibiotic-resistant gram-negative bacteria. Additionally, we demonstrated the importance of prior empiric antibiotic administration as a risk factor for the development of ventilator-associated pneumonia and nosocomial bacteremia. Our results also indicate that clinical efforts directed at reducing the durations of mechanical ventilation and intravascular catheter use should help to reduce the incidence of these infections. Future studies are needed to validate our results and to determine whether scheduled changes or rotation of antibiotic classes is an effective strategy for reducing nosocomial infections attributed to antibiotic-resistant bacteria. Until such studies are performed, clinicians should be aware of the hazards of bacterial superinfection as a cause of morbidity in their patients and the role of prior antibiotic exposure in the occurrence of this complication.
The writers wish to thank Daniel P. Schuster, M.D., for his review of the manuscript.
Supported in part by a grant from the American Lung Association of Eastern Missouri.
| 1. | Vincent J. L., Bihari D. J., Suter P. M., Bruining H. A., White J., Nicolas-Chanoin M. H., Wolff M., Spencer R. C., Hemmer M.The prevalence of nosocomial infection in intensive care units in Europe. Results of the European Prevalence of Infection in Intensive Care (EPIC) Study. EPIC International Advisory Committee. J.A.M.A.2741995639644 |
| 2. | Brun-Buisson C., Doyon F., Carlet J., Dellamonica P., Govin F., Lepoutre A., Mercier J. C., Offenstadt G., Regnier B.Incidence, risk factors, and outcome of severe sepsis and septic shock in adults: a multicenter prospective study in intensive care units. J.A.M.A.2741995968974 |
| 3. | Wenzel R. P., Thompson R. L., Landry S. M., Russell B. S., Miller P. J., Ponce de Leon S., Miller G. B.Hospital-acquired infections in intensive care unit patients: an overview with emphasis on epidemics. Infect. Control41983371375 |
| 4. | Craven D. E., Steger K. A.Epidemiology of nosocomial pneumonia: new perspectives on an old disease. Chest10819951S16S |
| 5. | Chow J. W., Fine M. J., Shlaes D. M., Quinn J. P., Hooper D. C., Johnson M. P., Ramphal R., Wagener M. M., Miyashira D. K., Yu V. L.Enterobacter bacteremia: clinical features and emergence of antibiotic resistance during therapy. Ann. Intern. Med.1151991585590 |
| 6. | Fagon J. Y., Chastre J., Domart Y., Troulliet J. L., Pierre J., Darne C., Gibert C.Nosocomial pneumonia in patients receiving continuous mechanical ventilation: prospective analysis of 52 episodes with use of a protected specimen brush and quantitative culture technique. Am. Rev. Respir. Dis.1391989877884 |
| 7. | Kollef M. H., Silver P., Murphy D. M., Trovillion E.The effect of late-onset ventilator-associated pneumonia in determining patient mortality. Chest108199516551662 |
| 8. | Fagon J. Y., Chastre J., Hance A. J., Montravers P., Novara A., Gibert C.Nosocomial pneumonia in ventilated patients: a cohort study evaluating attributable mortality and hospital stay. Am. J. Med.941993281288 |
| 9. | Rello J., Ausina V., Ricart M., Castella J., Prats G.Impact of previous antimicrobial therapy on the etiology and outcome of ventilator-associated pneumonia. Chest104199312301235 |
| 10. | Huth T. S., Burke J. P.Infections and antibiotic use in a community hospital, 1971–1990. Infect. Control Hosp. Epidemiol.121991525534 |
| 11. | McGowen J. E.Do intensive hospital antibiotic control programs prevent the spread of antibiotic resistance? Infect. Control Hosp. Epidemiol.151994478483 |
| 12. | Dunagan W. C., Woodward R. S., Medoff G., Gray J. L., Casabar E., Lawrenz C., Spitznagel E., Smith M. D.Antibiotic misuse in two clinical situations: positive blood cultures and administration of aminoglycosides. Rev. Infect. Dis.131991405412 |
| 13. | Kollef M. H., Wragge T., Pasque C.Determinants of mortality and multiorgan dysfunction in cardiac surgery patients requiring prolonged mechanical ventilation. Chest107199513951401 |
| 14. | Menzies R., Gibbons W., Goldberg P.Determinants of weaning and surviving among patients with COPD who require mechanical ventilation for acute respiratory failure. Chest951989398405 |
| 15. | Knaus W. A., Draper E. A., Wagner D. P., Zimmerman J. E.APACHE II: a severity of disease classification system. Crit. Care Med.131985818829 |
| 16. | Kollef M. H.Ventilator-associated pneumonia: a multivariate analysis. J.A.M.A.270199319651970 |
| 17. | Garner J. S., Jarvis W. R., Emori T. G., Horan T. C., Hughes J. M.CDC definitions for nosocomial infections. Am. J. Infect. Control161988128140 |
| 18. | Pingleton S. K., Fagon J. Y., Leeper K. V.Patient selection for clinical investigation of ventilator-associated pneumonia: criteria for evaluating diagnostic techniques. Chest1021992553S556S |
| 19. | Marquette C. H., Georges H., Wallet F., Ramon P., Saulnier F., Neviere R., Mathieu D., Rime A., Tonnel A. B.Diagnostic efficiency of endotracheal aspirates with quantitative bacterial cultures in intubated patients with suspected pneumonia: comparison with the protected specimen brush. Am. J. Respir. Crit. Care Med.1481993138144 |
| 20. | Kollef M. H., Bock K. R., Richards R. D., Hearns M. L.The safety and diagnostic accuracy of minibronchoalveolar lavage in patients with suspected ventilator-associated pneumonia. Ann. Intern. Med.1221995743748 |
| 21. | Papazian L., Thomas P., Garbe L., Guignon I., Thirion X., Charrel J., Bollet C., Fuentes P., Gouin F.Bronchoscopic or blind sampling techniques for the diagnosis of ventilator-associated pneumonia. Am. J. Respir. Crit. Care Med.152199519821991 |
| 22. | Rubin D. B., Wiener-Kronish J. P., Murray J. F., Green D. R., Turner J., Luce J. M., Montgomery A. B., Marks J. D., Matthay M. A.Elevated von Willebrand factor antigen is an early phase predictor of acute lung injury in nonpulmonary sepsis syndrome. J. Clin. Invest.861990474480 |
| 23. | American College of Chest Physicians/Society of Critical Care Medicine Consensus CommitteeDefinitions for sepsis and organ failures and guidelines for the use of innovative therapies in sepsis. Chest101199216581662 |
| 24. | SAS/STAT User's Guide. 1990. SAS Institute, Inc., Cary, NC. 1071–1126. |
| 25. | Valles J., Artigas A., Rello J., Bonsoms N., Fontanals D., Blanch L., Fernandez R., Baigorri F., Mestre J.Continuous aspiration of subglottic secretions in preventing ventilator-associated pneumonia. Ann. Intern. Med.1221995179186 |
| 26. | Celis R., Torres A., Gatell J. M., Almela M., Rodriguez-Roisin R., Agusti-Vidal A.Nosocomial pneumonia: a multivariate analysis of risk and prognosis. Chest931988318324 |
| 27. | Rello J., Torres A., Ricart M., Valles J., Gonzalez J., Artigas A., Rodriguez-Roisin R.Ventilator-associated pneumonia by Staphylococcus aureus: comparison of methicillin-resistant and methicillin-sensitive episodes. Am. J. Respir. Crit. Care Med.150199415451549 |
| 28. | Johanson W. G., Pierce A. K., Sandford J. P., Thomas G. D.Nosocomial respiratory infections with gram-negative bacilli: the significance of colonization of the respiratory tract. Ann. Intern. Med.771972701706 |
| 29. | de Latorre F. J., Pont T., Ferrer A., Rossello J., Palomar M., Planas M.Pattern of tracheal colonization during mechanical ventilation. Am. J. Respir. Crit. Care Med.152199510281033 |
| 30. | Sheagran J. N.Staphylococcus aureus: the persistent pathogen. N. Engl. J. Med.310198414371442 |
| 31. | Pittet B., Thievent B., Wenzel R. P., Li N., Auckenthaler R., Suter P. M.Bedside prediction of mortality from bacteremic sepsis: a dynamic analysis of ICU patients. Am. J. Respir. Crit. Care Med.1531996684693 |
| 32. | Meyer K. S., Urban C., Eagan J. A., Berger B. J., Rahal J. J.Nosocomial outbreak of Klebsiella infection resistant to late-generation cephalosporins. Ann. Intern. Med.1191993353358 |
| 33. | Goldmann D. A., Weinstein R. A., Wenzel R. P., Tablan O. C., Duma R. J., Gaynes R. P., Schlosser J., Martone W. J.Strategies to prevent and control the emergence and spread of antimicrobial- resistant microorganisms in hospitals. J.A.M.A.2751996234240 |
| 34. | Wunderink R. G.Mortality and ventilator-associated pneumonia: the best antibiotics may be the least antibiotics. Chest1041993993995 |
| 35. | Pestotnik S. L., Classen D. C., Evans R. S., Burke J. P.Implementing antibiotic practice guidelines through computer-assisted decision support: clinical and financial outcomes. Ann. Intern. Med.1241996884890 |
| 36. | McGowan J. E., Gerding D. N.Does antibiotic restriction prevent resistance? New Horizons41996370376 |
| 37. | Gerding D. N., Larson T. A., Hughes R. A., Weiler M., Shanholtzer C., Peterson L. R.Aminoglycoside resistance and aminoglycoside usage: ten years of experience in one hospital. Antimicrob. Agents Chemother.35199112841290 |
| 38. | Potgieter P. D., Hammond J. M.Prophylactic use of the new quinolones for prevention of nosocomial infection in the intensive care unit. Drugs4919958691 |
