Inflammatory biomarkers and prediction for intensive care unit admission in severe community-acquired pneumonia.

OBJECTIVE: Increased inflammatory response is related to severity and outcome in community-acquired pneumonia, but the role of inflammatory biomarkers in deciding intensive care unit admission is unknown. We assessed the relationship between inflammatory response, prediction for intensive care unit admission, delayed intensive care unit admission, and outcome in patients with community-acquired pneumonia.
DESIGN: Prospective clinical study.
SETTING: Intensive care units of two university hospitals. PATIENTS: We included 627 ward and 58 intensive care unit patients with community-acquired pneumonia, 36 with direct and 22 with delayed intensive care unit admission.
INTERVENTIONS: Serum levels of C-reactive protein, procalcitonin, tumor necrosis factor-α, interleukin-1, interleukin-6, interleukin-8, and interleukin-10 at admission.
MEASUREMENTS AND MAIN RESULTS: We assessed the prediction for intensive care unit admission of biomarkers and the Infectious Diseases Society of America/American Thoracic Society guidelines minor criteria for severe community-acquired pneumonia. Procalcitonin (p=.001), C-reactive protein (p=.005), tumor necrosis factor-α (p=.042), and interleukin-6 (p=.003) levels were higher in intensive care unit-admitted patients; however, the Infectious Diseases Society of America/American Thoracic Society guidelines minor severity criteria predicted better intensive care unit admission (odds ratio, 12.03; 95% confidence interval, 5.13-28.20; p<.001). No patient with severe community-acquired pneumonia by three or more minor severity criteria and procalcitonin levels below the optimal cutoff (0.35 ng/mL) needed intensive care unit admission compared with 14 (23%) with levels above the cutoff (p=.032). In patients initially admitted to wards, procalcitonin (p=.012) and C-reactive protein (p=.039) were higher in those 22 patients subsequently transferred to the intensive care unit after adjusting for age, comorbidities, and Pneumonia Severity Index risk class. Despite initially admitted to wards, 14 (64%) patients with delayed intensive care unit admission had already criteria for severe community-acquired pneumonia at admission compared with 73 (12%) ward patients (p<.001).
CONCLUSION: Inflammatory biomarkers identified patients needing intensive care unit admission, including those with delayed intensive care unit admission. Patients with severe community-acquired pneumonia by minor criteria and low levels of procalcitonin may be safely admitted to wards. Correctly applying the Infectious Diseases Society of America/American Thoracic Society guidelines would reduce substantially delayed intensive care unit admission.