Background and objective: Drug-resistant pathogen (DRP) risk stratification is important for choosing a treatment strategy for community-onset pneumonia. Evidence for benefits of non-antipseudomonal β-lactam plus macrolide combination therapy (BLM) on mortality is limited in patients at low DRP risk. Risk factors for mortality remain to be clarified.
Methods: Post hoc analysis using a prospective multicentre study cohort of community-onset pneumonia was performed to assess 30-day differences in mortality between non-antipseudomonal β-lactam monotherapy (BL) and BLM groups. Logistic regression analysis was performed to assess the therapeutic effect and risk factors for mortality in patients at low DRP risk.
Results: In total, 594 patients with community-onset pneumonia at low DRP risk (369 BL and 225 BLM) were analysed. The 30-day mortality in BL and BLM was 13.8% and 1.8%, respectively (P < 0.001). Multivariate analysis showed that BLM reduced the 30-day mortality (adjusted odds ratio: 0.28, 95% CI: 0.09-0.87) compared with BL. Independent prognostic factors for 30-day mortality included arterial partial pressure of carbon dioxide (PaCO2 ) > 50 mm Hg, white blood cell count < 4000/mm3 , non-ambulatory status, albumin < 3.0 g/dL, haematocrit < 30%, age ≥ 80 years, respiratory rate > 25/min and body temperature < 36°C.
Conclusion: In patients with community-onset pneumonia at low DRP risk, BLM treatment reduced 30-day mortality compared with BL. Independent risk factors for mortality are potential confounding factors when assessing antibiotic effects in randomized clinical trials.
Keywords: algorithms; antibiotics; community-acquired pneumonia; drug resistance, bacterial; pneumonia.
© 2017 Asian Pacific Society of Respirology.