CHICAGO – Use of the Clinical Pulmonary Infection Score to guide therapy for ventilator-associated pneumonia in critically ill patients has had mixed results, and a recent study has raised further questions about whether CPIS can accurately determine resolution of the infection in critically injured trauma ICU patients.
"Using CPIS to determine the appropriate duration of antimicrobial therapy in trauma patients could potentially be harmful by unnecessarily prolonging exposure to antibiotics," Dr. Nancy A. Parks said at the annual meeting of the American Association for the Surgery of Trauma. It could also lead to undertreatment, resulting in relapse, she noted.
Dr. Nancy A. Parks
Conventional diagnosis of ventilator-assisted pneumonia (VAP) has been based on a CPIS of 6 or higher, but her 6-year study of 1,028 critically ill patients diagnosed with VAP showed resolution of VAP in patients with higher CPIS scores.
Dr. Parks, of the Elvis Presley Memorial Trauma Center in Memphis, cited previously published work that questioned the ability of the CPIS to help differentiate the infection of VAP from posttraumatic systemic inflammatory response syndrome (J. Trauma 2006;60:523-8). While this study explored the efficacy of CPIS for initiating antibiotic therapy, it left unanswered the question of whether CPIS could help determine when to end therapy.
Dr. Parks and her coinvestigators answered that in the negative.
Their protocol to confirm VAP involved using bronchoscopy with bronchoalveolar lavage (BAL) in patients with a clinical suspicion of VAP based on three of four elements in the CPIS: body temperature, white cell count, purulent secretions, or new or changing infiltrate on chest x-ray. The investigators started patients on antibiotics empirically, then adjusted treatment if BAL effluent was 105 colony-forming units (CFU) per milliliter or higher. The investigators considered VAP resolved if repeat BAL showed a reading of 103 CFU/mL or less on day 4 of therapy. If repeat BAL was done within 2 weeks, they considered any findings greater than 105 CFU/mL as a recurrence.
The study population, which had an average Injury Severity Score of 31 and an average base deficit of 4.4, was composed primarily of blunt trauma patients, Dr. Parks reported. Overall mortality was 9.4%. Slightly more than half of the patients had community-acquired pathogens (50% methicillin-sensitive Staphylococcus aureus [MSSA], the remainder split between streptococcus and haemophilus), and slightly fewer had hospital-acquired infections (including about 19% methicillin-resistant Staphylococcus aureus [MRSA] and 27% pseudomonas). The recurrence rate was 1% overall.
A CPIS score of 6 or greater is considered positive for VAP, but this study showed resolution of VAP at CPIS scores above that. "The average CPIS on discontinuation of antibiotics was 6.9 in the community-acquired group and 6.3 in the hospital-acquired group – both well above our clinical cutoff of 6," Dr. Parks said.
"If we had used CPIS to guide treatment, we would have seen a sensitivity of 69% in our community-acquired group and 72% in our hospital-acquired group," Dr. Parks said. "Specificity would have been 51% and 53%, respectively. Our positive predictive value was 59% and 61%, and negative predictive value 62% and 66%. If we were to use CPIS to guide therapy, 59% of our patients would have had antibiotics continued inappropriately."
"This disparity – that CPIS may be potentially beneficial and useful in medical patients but not in trauma patients – is very important," said Dr. Lena Napolitano, a discussant at the meeting. "It will have important implications, particularly as we strive nationally to modify the definition of VAP and optimize duration of microbial therapy for VAP, hoping to move toward more short-term antimicrobial therapy and hopefully not lead to greater prevalence of recurrent pneumonia," said Dr. Napolitano, of the University of Michigan, Ann Arbor.
Dr. Parks had no disclosures, and the study received no outside funding.