Respiratory support strategies in interstitial lung disease with acute respiratory failure: a real-world ICU study

Acute exacerbation of interstitial lung disease (AE-ILD) complicated by acute respiratory failure (ARF) is associated with high short-term mortality. Evidence describing clinical profiles and observed outcomes across sustained respiratory-support modalities, together with factors associated with 28-day mortality, remains limited. To describe the clinical characteristics and observed outcomes of patients with AE-ILD receiving different sustained respiratory-support modalities and to identify factors associated with 28-day mortality. We conducted a single-center retrospective cohort study of 192 consecutive adults with AE-ILD and ARF admitted to a respiratory intensive care unit between January 2018 and December 2022. Patients were classified according to the initial respiratory-support modality sustained for at least 72 h: noninvasive respiratory support (NIRS, including HFNC and NIV) or invasive mechanical ventilation (IMV). The outcome was all-cause mortality within 28 days after initiation of respiratory support. Factors associated with mortality were evaluated in the full analytic cohort using multivariable Firth logistic regression. Multicollinearity in the final five-variable model was assessed using variance inflation factors, and the association between sustained IMV requirement and the dichotomized low PaO 2 /FiO 2 ratio (P/F) indicator was assessed using Spearman correlation. Among 192 patients (median age, 71 years; 58.3% male), 160 received sustained NIRS and 32 received sustained IMV. The IMV group had more severe baseline hypoxemia and higher observed 28-day mortality than the NIRS group (90.6% vs. 52.5%, P < 0.001); this comparison was descriptive because respiratory-support selection was confounded by illness severity. Mortality did not differ significantly between idiopathic pulmonary fibrosis (IPF) and non-IPF subgroups (61.0% vs. 57.9%, P = 0.805). In full-cohort multivariable Firth regression, elevated white blood cell count (adjusted OR, 2.28; 95% CI, 1.10–4.82), low hemoglobin (adjusted OR, 4.04; 95% CI, 1.80–9.71), elevated procalcitonin (adjusted OR, 5.08; 95% CI, 2.38–11.33), reduced P/F (adjusted OR, 3.53; 95% CI, 1.73–7.41), and sustained IMV requirement (adjusted OR, 5.38; 95% CI, 1.68–24.18) were associated with 28-day mortality. All variance inflation factors in the final five-variable model were < 2, and the correlation between sustained IMV requirement and low P/F was weak (Spearman r = 0.145, P = 0.044). Among patients with AE-ILD and ARF who survived and received sustained respiratory support for at least 72 h, sustained IMV requirement and four admission clinical variables were associated with 28-day mortality. These findings may support closer monitoring, frequent reassessment, and early multidisciplinary discussion of prognosis and treatment goals. Not applicable.