Etiologies and Outcomes of Acute Respiratory Failure in Patients Treated with Immune Checkpoint Inhibitors.

Background: Indications for immune checkpoint inhibitors (ICIs) are expanding across cancer types and stages. While this carries promise for improved survival, increasing use has raised concern for respiratory complications in patients with multi-morbidities. Pulmonary complications such as checkpoint inhibitor pneumonitis (CIP), a recognized immune-related ICI toxicity, can lead to severe acute respiratory failure (ARF). However, not all ARF following ICI exposure is attributable to CIP, and while different causes imply distinct management and prognostic implications, the types of complications leading to ARF after ICI remain poorly understood. Methods: We conducted a retrospective cohort study of ICI-treated adults who developed ARF, defined as ≥ 24 hours of invasive or noninvasive ventilatory support within 2 years of first ICI receipt, treated at five hospitals in a single health system (2017-2024). ARF etiology was determined by manual chart review, and suspected CIP cases underwent multidisciplinary adjudication. We evaluated associations between CIP versus non-CIP ARF and 90-day all-cause mortality using multivariable Cox proportional hazards models. Results: Among 197 ICI-treated patients with ARF, the median age was 66 years, and the most frequent cancer type was non-small cell lung cancer. Among these patients, 29 (15%) had CIP, and 168 (85%) had other etiologies of ARF (e.g., pneumonia, progressive cancer, aspiration, fluid overload). In the first 24 hours, CIP patients exhibited more severe hypoxemia but less multiorgan dysfunction than non-CIP patients. Ninety-day mortality was 66% for CIP and 71% for non-CIP ARF. In adjusted analyses, CIP vs. non-CIP ARF was not significantly associated with 90-day mortality (adjusted hazard ratio [aHR] = 0.66; 95% CI: 0.40-1.08) but was associated with decreased in-hospital mortality (aHR = 0.43; 95% CI: 0.21-0.88). Among ARF survivors, 13% restarted ICI therapy: 5% of adjudicated CIP vs 17% of adjudicated non-CIP (p = 0.28); only 7% of suspected but non-adjudicated CIP cases restarted ICI. Conclusions: In this cohort of ICI-treated patients who developed acute respiratory failure, most ARF was not attributable to CIP. While CIP and non-CIP ARF had similar 90-day mortality, CIP-related respiratory failure was associated with lower in-hospital mortality and lower likelihood of ICI continuation, underscoring the importance of accurate etiologic attribution in this vulnerable group.