Key Point

Reducing the risk of ventilator-associated infection was the main focus of a symposium held at the annual meeting of the American College of Chest Physicians.

MONTREAL—Investigators have been scrutinizing all aspects of infection in the critically ill to improve the management and outcomes of such infections, particularly those involving the ventilator. During a symposium at the annual meeting of the American College of Chest Physicians, experts in this field discussed their most recent findings.¹

In the multicenter Assessment of Local Antimicrobial Resistance Measures (ALARM) study, Lee E. Morrow, MD, and colleagues identified five predictors of mortality from baseline to 72 hours among 398 patients with ventilator-associated pneumonia. The predictors included an increasing APACHE II score, no response to therapy, the presence of risk factors for multidrug-resistant pathogens, the failure of clinicians to use bronchoalveolar lavage in pneumonia diagnosis, and inability to de-escalate antibiotic therapy.

"For each 1-point increase in the APACHE score, the relative risk [of death] increased by about 5%," related Dr. Morrow, Assistant Professor of Medicine in the Division of Pulmonary and Critical Care at Creighton University in Omaha. The other four mortality predictors were associated with relative mortality risks of 1.40, 2.19, 2.32, and 2.85, respectively. The relative risk was 10.75 for patients whose antibiotic therapy was escalated compared to those whose antibiotics were de-escalated.

The ALARM study validates the recent ATS recommendation to take into account risk factors for multidrug-resistant pathogens when selecting empiric therapy for ventilator-associated pneumonia, stated Dr. Morrow. The study also supports invasive diagnostic methods (ie, bronchoalveolar lavage) for suspected ventilator-associated pneumonia and the de-escalation of antibiotic therapy for the infection when possible, he added.

The ALARM findings suggest that "de-escalation is occurring to some extent," remarked Marin H. Kollef, MD, who discussed management patterns in ventilator-associated pneumonia. In ALARM, a mean duration of antibiotic therapy of 11.8 days was indicative of de-escalation. Also, a reduction in the number of antibiotics prescribed was reported for 22.7% of the patients; however, antibiotic therapy usually was not changed, and it was occasionally escalated.

Overall mortality in ALARM was 25.1%. Death occurred in 16.9% of patients whose antibiotic therapy was de-escalated, in 31.6% of those whose antibiotic regimens were unchanged, and in 42.4% of those whose antibiotics were escalated, observed Dr. Kollef, Professor of Medicine in the Division of Pulmonary and Critical Care Medicine at Washington University in St. Louis.

TRACHEAL COLONIZATION IN THE NEWLY INTUBATED

Microbial colonization of the trachea occurs in nearly two thirds of mechanically ventilated patients within 24 hours of intubation, and the proportion rises from there, suggest preliminary data reported by Lakshmi Durairaj, MBBS. Tracheal colonization was found to precede ventilator-associated pneumonia in several studies; pneumonia occurred in 10% to 20% of the colonized patients, pointed out Dr. Durairaj, Assistant Professor in the Department of Internal Medicine at the University of Iowa in Iowa City.

She shared the results of quantitative cultures of tracheal aspirates from 29 intubated adults with a mean age of 56.

Nineteen of the patients were receiving antibiotics on the first day of intubation. The proportion of patients with tracheal colonization on days 1 through 4 was 63%, 72%, 76%, and 93%, respectively.

"None of the group that was initially colonized ... went on to acquire Candida in their aspirates," Dr. Durairaj related. Among those who were initially free of pathogens, she continued, about 40% went on to acquire Candida.

There was no significant change in the rate of colonization with Candida or bacteria over the four-day observation period. "By day 4, these two groups were mutually exclusive," said Dr. Durairaj.

Interestingly, there was a tendency for patients who had no colonization on day 1 to remain uncolonized, while microbial counts tended to rise in those who were colonized initially. There were no significant differences between these two groups in length of ICU stay or mortality.

COLONIZATION PREDICTS INFECTIOUS ETIOLOGY

A strong relationship exists between respiratory and gastrointestinal tract colonization and the microbial etiology of subsequent infection, suggested a study by Effrosyni D. Manali, MD, and coworkers. Thus, weekly surveillance of respiratory and gastrointestinal tract specimens "could be helpful in predicting microbial etiology of infection and guiding appropriate empiric treatment in the critically ill," Dr. Manali suggested.

She and her colleagues retrospectively reviewed data for 148 patients admitted to the ICU during a one-year period. In that time, there were 55 documented infections—38 bacteremias, 10 ventilator-associated pneumonias, four soft-tissue infections, and three intra-abdominal infections—in the study population.

"For us, Acinetobacter is a very big problem," stressed Dr. Manali, who is with the Attikon University Hospital in Athens, Greece. Acinetobacter was the pathogen most often isolated in patients with infections, followed by Klebsiella and Pseudomonas. "Methicillin- resistant Staphylococcus aureus is still a very small problem for us," Dr. Manali noted.

Colonization of the respiratory tract with Acinetobacter was predictive of ventilator-associated pneumonia etiology with 75% sensitivity and 100% specificity; gastrointestinal tract colonization with Klebsiella was 57% sensitive and 92% specific for prediction of bacteremia etiology. "Fecal or bronchial colonization predicted etiology in all intra-abdominal or soft-tissue infections, respectively," added Dr. Manali.

Pathogens isolated from patients with ventilator-associated pneumonia correlated with bronchial or stool colonizers in 88% of cases. In isolates from catheter-related infections, gram-negative organisms had previously colonized bronchial secretions or stool. "Empiric antibiotic treatment based on results of colonization tests permitted 90% adequacy in treatment of [ventilator-associated pneumonia] and 80% in primary bacteremia treatment," Dr. Manali reported.

PROBIOTIC PROPHYLAXIS

In his second talk at the symposium, Dr. Morrow discussed preliminary study findings suggesting that administration of the probiotic agent Lactobacillus may prevent ventilator-associated pneumonia. Among 40 mechanically ventilated adults studied so far, Lactobacillus appeared to maintain a nonpathogenic flora and was associated with a trend toward lower rates of ventilator-associated pneumonia, Dr. Morrow explained.

The study participants were randomized to placebo or 109 colony-forming units of Lactobacillus delivered to the stomach and oropharynx every 12 hours.

From baseline to 72 hours, Lactobacillus was significantly more effective than placebo at preserving the normal oral flora, and there were nonsignificant trends in the Lactobacillus group toward less pathogenic colonization of the mouth and stomach.

Likewise, differences did not reach significance between the Lactobacillus and placebo groups in rates of clinically diagnosed and microbiologically confirmed ventilator-associated pneumonia (26% vs 45% and 11% vs 33%, respectively) or mortality (0% vs 10%).

Based on these preliminary data, Dr. Morrow concluded that probiotic therapy with Lactobacillus is safe and appears to favorably alter the microbiotic flora in the critically ill. "It may provide a novel, inexpensive, and nonantibiotic opportunity for [ventilator-associated pneumonia] prophylaxis," he suggested.

—Timothy Begany

Reference
1. Restrepo MI, Kollef MH, Morrow LE, et al. Ventilator-associated infections. Presented at: annual meeting of the American College of Chest Physicians; October 31, 2005; Montreal, Quebec.

Return to table of contents