Management and Prognosis of Community-Acquired Pneumonia in Adults

Abstract

Community-acquired pneumonia (CAP) is a common cause of morbidity and mortality worldwide. In this thesis, different aspects of both management and prognosis of adults admitted with CAP have been addressed in separate parts. In the first part, we evaluated empirical antibiotic therapy and statins as adjunctive therapy. Since the optimal empirical antibiotic strategy for patients with CAP admitted to non-ICU wards was unknown, we have performed a multi-center cluster randomized cross-over trial in 7 Dutch hospitals comparing three empirical antibiotic strategies (CAP-START study). Among 2,283 patients with a working diagnosis of CAP, the absolute risk differences on 90-day mortality were 1.9% (-0.6%; 4.4%) and -0.6% (-2.8%; 1.9%) for beta-lactam/ macrolide combination and fluoroquinolone monotherapy, respectively, as compared to beta-lactam monotherapy, demonstrating non-inferiority of beta-lactam monotherapy by the pre-specified 3% non-inferiority margin. There were no major differences in secondary outcomes or costs. Furthermore, in a post-hoc analysis of this trial, macrolide use was associated with hospital-acquired cardiac complications (HR 1.61, 95% CI 1.14-2.26), which, although we cannot fully exclude observational bias, adds to the changing risk-benefit assessment of macrolides as empirical treatment in this patient population. Given these observations, the lack of superiority, and the threat of increasing antimicrobial resistance, beta-lactam/ macrolide combination therapy and fluoroquinolone monotherapy should not be universally advised as empirical therapy for CAP patients admitted to non-ICU wards. Next, we have evaluated the role of statins as adjunctive therapy in CAP in a systematic review; at this moment, there is a lack of evidence to support this role. In part two, concerning the prognosis of CAP, we have evaluated the prediction of mortality and cardiac complications using data from the CAP-START study cohort. First, we studied the additional value of commonly available ‘biomarkers’ (i.e. the neutrophil/ lymphocyte ratio (NLR) and admission glucose levels) to existing prediction systems. In 1,549 patients with radiologically confirmed CAP, the NLR was associated with mortality, OR 1.19 (95% CI 1.02-1.38) per ten units increase, but did not improve prediction systems with non-significant AUC differences when added to pneumonia-severity-index (PSI) (0.752 vs 0.761, p=0.10) and CURB-65 (confusion, uremia, respiratory rate, blood pressure, age ≥65 years) score (0.698 vs 0.709, p=0.246). Admission glucose level has a significant non-linear association with 30-day, but its addition to models with PSI and CURB-65 score also hardly improved model discrimination (AUC change for PSI: from 0.747 to 0.764, p=0.156 and for CURB-65 score: from 0.697 to 0.722, p=0.082). Next, in the 2,107 patients with a working diagnosis of CAP, admitted without a cardiac event on presentation, the incidence of hospital-acquired cardiac events was 6.9%. The hospital-mortality rate of these patients was 9.6%, and the overall hospital-mortality rate was 3.1%, of which 14.4% (95% CI 6.9-22.0%) was attributable to hospital-acquired cardiac events. Ages, comorbid cardiac disease, diabetes, increasing levels of blood urea nitrogen and C-reactive protein on admission were independent risk factors. Recognition of patients at high risk for these complications, might aid in developing preventive strategies to improve outcome.