Outcomes associated with community- and
healthcare-acquired pneumonia
By John Gever
Patients presenting with pneumonia -- or what appears to be pneumonia -- are among the most common serious situations in most hospital. According to Sharon Ngan, M.D., pulmonary and critical care resident at New York Methodist Hospital in Brooklyn, N.Y., "The number of patients hospitalized with CAP [community-acquired pneumonia] is higher than myocardial infarction [or] stroke." But data presented by Ngan at CHEST 2006, the annual meeting of the American College of Chest Physicians, indicates that CAP may be overdiagnosed at hospital admission.
Among those patients correctly diagnosed, blood testing for C-reactive protein (CRP) may help direct patients -- especially those with the severest illness -- to the most effective treatment, according to research presented by Lusine Melik-Adamyan, critical care resident at New York Hospital Queens in Queens, N.Y.
The meeting was held Oct. 21-26 in Salt Lake City, Utah. Ngan and Melik-Adamyan spoke at a session devoted to community- and healthcare-associated pneumonia diagnosis and treatment.
Ngan led a study at her institution, an urban teaching hospital, delving into the accuracy of initial CAP diagnoses. Her group reviewed records of 113 patients admitted to the hospital's general wards for CAP in a four-month period in early 2006. The goal was to determine how frequently the diagnosis actually conformed to official guidelines for the American Thoracic Society's official definition of pneumonia. Under that definition, pneumonia is a combination of two respiratory symptoms such as cough, fever, dyspnea, sputum production or pleuritic chest pain, and opacity on a chest x-ray as documented in a formal radiology report.
In the study's 113-patient sample, fewer than half (54) met these criteria for a pneumonia diagnosis. Moreover, of those, only 3 were eventually confirmed with a positive microbiological culture, Ngan said.
Of the 59 patients whose diagnosis failed to meet the official definition, 17 (15 percent) were found to have congestive heart failure, 5 (4 percent) had chronic obstructive pulmonary disease (COPD), 3 (3 percent) had acute bronchitis and 34 (30 percent) had a non-respiratory illness such as cellulitis, sepsis or urinary tract infection.
The mean duration of hospitalization in this sample was similar to national averages for CAP reported in the literature. Ngan said that the patients in her sample with longer hospital stays were most likely to have pneumonia confirmed by a positive blood culture.
Her conclusion: "CAP might be overdiagnosed at initial evaluation," she said, indicating that physicians have an extremely high clinical suspicion of pneumonia in patients with breathing difficulties.
"It demonstrates a need for appropriate diagnosis and less inappropriate use of antibiotics," Ngan said, adding that better management of confounding illnesses such as heart failure and COPD would help as well.
Meanwhile, Melik-Adamyan presented results from a prospective study of CRP as a prognostic marker in CAP and its cousin, healthcare-associated pneumonia (HCAP). The latter resembles nosocomial pneumonia in clinical course and/or microbiological genotype, but the onset of HCAP occurs outside the hospital setting.
The study was inspired by observations that CRP is elevated in a range of inflammatory conditions. Further, some earlier research had indicated that high levels of CRP predicted fatal outcomes in lower respiratory tract infections, although this had not been reliably confirmed.
Her group included 25 consecutive patients with severe CAP or HCAP over a four-month period ending in March 2006. Patients were evaluated at admission and 48 hours later for blood CRP levels and Clinical Pulmonary Infection Score (CPIS). Most patients were elderly (mean age of 80), and a large majority were sufficiently ill to require mechanical ventilation. Eighty percent were determined to have CAP as opposed to HCAP.
Thirteen of the patients eventually died. The study's key finding was that the 0-48 hour change in CRP level was significantly correlated with fatal outcome: patients who died showed a mean CRP increase of 5.5 mg/dl, whereas CRP levels declined by a mean of 4.7 mg/dl in the survivors (p < 0.01). CRP levels declined in only two of the 13 fatalities, and these both showed extremely high CRP levels at admission, Melik-Adamyan said.
The study examined a range of other factors as well, including age, gender, CPIS score at admission, ventilation and tracheostomy status, and presence of multilobar illness. Only the last was significantly correlated with outcome, Melik-Adamyan said, which was expected based on earlier studies.
Melik-Adamyan said the clinical implication is that CRP testing can help "decide whether or not further, more invasive, diagnostic procedures are needed and whether therapeutic interventions should be maintained or modified." In particular, more aggressive therapy may be required in patients with markedly increasing CRP. Conversely, declining CRP could be an indication that antibiotic therapy is no longer needed.