WASHINGTON — Five clinical findings at initial presentation in patients with severe, community-acquired pneumonia identified those who failed to respond to the first 3 days of antibiotic therapy in a study with 260 patients.
Prompt assessment of these five clinical flags may potentially distinguish patients who require close monitoring from those who face little risk from an early hospital discharge and oral antibiotic therapy, Martine Hoogewerf, M.D., said at the annual Interscience Conference on Antimicrobial Agents and Chemotherapy.
But the prognostic value of these five findings requires validation in a larger group of patients before they can be used in routine practice, cautioned Dr. Hoogewerf, an infectious diseases specialist at the University Medical Center in Utrecht, the Netherlands.
The five findings were confusion, a gram-negative infection, an arterial pH of less than 7.35, an arterial oxygen pressure (PO2) of less than 60 mm Hg, and a diagnosis of heart failure.
In a multivariate analysis of potential predictors of early treatment failure, these five parameters were associated with statistically significant differences in response. The first four factors identified patients with an increased risk of failure, while the presence of heart failure identified patients with a reduced risk.
Two possible explanations for the “protective” role of heart failure were that some patients with heart failure are mistakenly diagnosed with pneumonia, and that when they actually have pneumonia, it more often has a viral etiology, Dr. Hoogewerf said. Among the patients in the study, 68% of those with heart failure had no identified etiologic pathogen; among the patients without heart failure, 45% had no identified etiologic pathogen.
Dr. Hoogewerf and her associates converted their findings on predictive features into a scoring system. Based on the relative predictive value of these five clinical parameters, they created a scoring scheme that health care workers might use when first evaluating patients with severe, community-acquired pneumonia.
The system scores an arterial pH of less than 7.35 as 3 points, confusion as 2 points, a gram-negative infection as 2 points, an arterial PO2 of less than 60 mm Hg as 2 points, and heart failure as -2 points.
Among the 260 patients in the study, 61% had a score of less than 2 points, and 18% of these patients had treatment failure after 3 days of antibiotic therapy. Another 35% of patients had a score of 2-4 points, and they had a treatment failure rate of 47%. The remaining 4% of patients had a score of 5 points or higher, and their failure rate was 80%, Dr. Hoogewerf said at the conference, sponsored by the American Society for Microbiology.
The study involved adults with community-acquired pneumonia seen at seven hospitals and medical centers in the Netherlands. All patients had either a pneumonia severity index score of more than 90 at their initial evaluation or met the criteria of the American Thoracic Society for severe pneumonia. All patients were started on antibiotic therapy and reassessed 3 days later.
A total of 80 patients (31%) had not responded to treatment after 3 days. The clinical importance of treatment failure at this time was documented by additional follow-up of all patients. The average hospital length of stay for patients who had early treatment success was 9.6 days, compared with an average length of stay of 13.4 days for the early treatment failures. The difference of 3.8 days was statistically significant.
In addition, patients with early treatment success had a 4.4% mortality rate by 28 days after the start of treatment. Patients who were classified as early treatment failures had a 12% mortality rate 28 days after treatment started.
The multivariate analysis of baseline parameters showed that patients with an arterial pH of less than 7.35 had a 4.46-fold increased risk of treatment failure, compared with patients whose arterial pH was at or above this cutoff. Patients who were confused had a 3.79-fold increased risk of failure, compared with those who were not confused.
Patients with a gram-negative infection had a 3.15-fold increased risk of treatment failure, compared with patients without an identified gram-negative infection. Patients with an arterial PO2 of less than 60 mm Hg had a 77% increased risk of failure, compared with patients whose arterial PO2 was at or above this cutoff.
Patients with heart failure had a 68% reduced risk of treatment failure, compared with patients with no heart failure.
KEVIN FOLEY, RESEARCH