RESUMO
BACKGROUND: The role of respiratory coinfections at diagnosis of Pneumocystis jirovecii pneumonia (PcP) on clinical impact has been underestimated. METHODS: A retrospective observational study was conducted January 2011 to April 2019 to evaluate respiratory coinfections at diagnosis of PcP patients in 2 tertiary care hospitals. Coinfection was defined by identification of pathogens from P. jirovecii-positive samples. RESULTS: Of 7882 respiratory samples tested for P. jirovecii during the 8-year study, 328 patients with diagnosis of PcP were included. Mean age was 56.7 (SD 14.9) years, 193 (58.8%) were male, 74 (22.6%) had positive HIV serology, 125 (38.1%) had viral coinfection, 76 (23.2%) bacterial coinfection, and 90-day mortality was 25.3%. In the overall population, 90-day mortality was independently associated with solid tumor underlying disease (odds ratio [OR], 11.8; 95% confidence interval [CI], 1.90-78.0; P = .008), sepsis-related organ failure assessment score (SOFA) at admission (OR, 1.62; 95% CI, 1.34-2.05; P< .001), and cytomegalovirus (CMV) respiratory coinfection (OR, 3.44; 95% CI, 1.24-2.90; P = .02). Among HIV-negative patients, respiratory CMV coinfection was associated with worse prognosis, especially when treated with adjunctive corticosteroid therapy. CONCLUSIONS: Respiratory CMV coinfection at PcP diagnosis was independently associated with increased 90-day mortality, specifically in HIV-negative patients.