A Therapeutic Strategy for All Pneumonia Patients: A 3-Year Prospective Multicenter- Cohort Study Using Risk Factors for Multidrug Resistant Pathogens To Select Initial Empiric Therapy.

Background Empiric therapy of pneumonia is currently based on the site of acquisition (community or hospital), but could be chosen, based on risk factors for multidrug-resistant (MDR) pathogens, independent of site of acquisition. Methods We prospectively applied a therapeutic algorithm based on MDR risks, in a multicenter cohort study of 1089 patients with 656 community-acquired pneumonia (CAP), 238 healthcare-associated pneumonia (HCAP), 140 hospital-acquired pneumonia (HAP), or 55 ventilator-associated pneumonia (VAP). Results Approximately 83% of patients were treated according to the algorithm, with 4.3% receiving inappropriate therapy. The frequency of MDR pathogens varied, respectively, with VAP (50.9%), HAP (27.9%), HCAP (10.9%), and CAP (5.2%). Those with 2 MDR risks had MDR pathogens more often than those with 0-1 MDR risk (25.8% vs 5.3%, P < .001). The 30-day mortality rates were as follows: VAP (18.2%), HAP (13.6%), HCAP (6.7%), and CAP (4.7%), and were lower in patients with 0-1 MDR risks than in those with 2 MDR risks (4.5% vs 12.5%, P < .001). In multivariate logistic regression analysis, 5 risk factors (advanced age, hematocrit <30%, malnutrition, dehydration, and chronic liver disease), as well as hypotension and inappropriate therapy were significantly correlated with 30-day mortality, whereas the classification of pneumonia type (VAP, HAP, HCAP, CAP) was not. Conclusions Individual MDR risk factors can be used in a unified algorithm to guide and simplify empiric therapy for all pneumonia patients, and were more important than the classification of site of pneumonia acquisition in determining 30-day mortality. Clinical Trials Registration JMA-IIA00146. We applied a single algorithm to all forms of pneumonia, which was followed in 82.5% of a cohort of 1089 patients. Only 4.3% received inappropriate therapy; in multivariate analysis, site of pneumonia acquisition was not predictive of 30-day mortality.