Respiratory management of critically ill pneumocystis pneumonia patients: a multicenter retrospective study

Pneumocystis jirovecii pneumonia (PjP) is a rising cause of acute respiratory failure in immunocompromised patients, often requiring Intensive Care Unit (ICU) admission. However, optimal ventilatory strategies remain unclear.

For the present study, we conducted an ancillary analysis of the PRONOCYSTIS study, a large multicenter cohort of PjP patients. Patients admitted to the ICUs were compared according to initial respiratory management (High-Flow Nasal Cannula (HFNC), standard Oxygen (SO) or Non-Invasive Ventilation (NIV). A propensity score adjustment [inverse probability of treatment weighting (IPTW) analysis] was implemented to account for potential confounders. The primary outcome was intubation rate. Univariable and multivariable Cox regressions were also used to assess variables associated with survival.

Over the study period, 248 patients with PjP were included in the present analysis. Of those, 70 were treated by HFNC while 118 and 60 received SO and NIV, respectively. HFNC patients had a decreased intubation rate (28.6% versus 45.0% in NIV and 55.4% in SO patients; p = 0.003). When assessing the impact of respiratory management on intubation by IPTW, HFNC remained an independent protective factor (weighted Hazard Ratio (HR) 0.41 (95% CI 0.24-0.69); p < 0.001). While, NIV was not associated with intubation (HR 0.62 (95% CI 0.37-1.02); p = 0.056). Through adjusted survival analysis, long-term corticosteroids treatment (aHR 4.03 (95% CI 2.01-8.08); p < 0.001), Solid tumor (aHR 3.37 (95% CI 1.45-7.86); p = 0.005) and the Sequential Organ Failure Assessment score (aHR 1.24 (95% CI 1.15-1.35); p < 0.001) were found to be independent predictor for death. Initial respiratory support was not associated with survival either in the Cox multivariable analysis or in the IPTW analysis.

Through this multicenter observational study of severe PjP patients, although oxygenation strategy was not associated with D90 survival, HFNC support appeared to be associated with a lower intubation rate. Further prospective studies are warranted to refine respiratory management in critically ill PjP patients.