Djamel Mokart, Michael Darmon, Achille Kouatchet, Peter Pickkers, Marcio Soares, Jordi Rello, Philippe R Bauer, Andry van de Louw, Virginie Lemiale, Fabio Silvio Taccone, Ignacio Martin-Loeches, Katerina Rusinova, Laveena Munshi, Massimo Antonelli, Peter Schellongowski, Andreas Barratt-Due, Miia Valkonen, Precious Pearl Landburg, Ramin Brandt Bukan, Frédéric Pène, Victoria Metaxa, Gaston Burghi, Colombe Saillard, Lene B Nielsen, Emmanuel Canet, Naike Bigé, Fabrice Bruneel, Alexandre Demoule, Laurent Chow-Chine, Elie Azoulay, Efraim Investigators and the Nine-I Study Group
Immune dysfunction strongly influences ARF patterns and prognosis. Immune-guided diagnostic strategies could improve early etiological identification and outcomes in high-risk patients with hematological malignancies.
BACKGROUND: Diagnosing acute respiratory failure (ARF) in patients with hematologic malignancies is complex and prognostically critical. We hypothesized that immune dysfunction driven by disease and treatments determines ARF etiologies and outcomes.
METHODS: This post-hoc analysis of a prospective multicenter cohort included patients with hematological malignancies admitted to ICU with hypoxemic ARF. Immune dysfunctions were classified as phagocytic, humoral, or cellular. Associations between malignancies, immune profiles, treatments, and ARF causes were assessed by multivariate logistic regression. Unsupervised hierarchical clustering identified immune-phenotypic subgroups.
RESULTS: Among 708 patients, the most frequent malignancies were acute myeloid leukemia (AML, 29.5%), non-Hodgkin lymphoma (25.6%), and myeloma (21.3%). Immune dysfunctions were frequent: humoral (71.5%), cellular (57.9%), and phagocytic (55.5%). Main ARF etiologies were bacterial pneumonia (40.4%), viral infections (18.8%), and invasive fungal infections (16.8%), while 12.7% remained undiagnosed. Phagocytic dysfunction was associated with bacterial (OR 1.60, p = 0.026) and infiltrative causes (OR 2.17, p = 0.007). Humoral dysfunction was linked to viral infections (OR 1.83, p = 0.014), and cellular dysfunction to viral infections (OR 1.70, p = 0.01) and Pneumocystis jirovecii pneumonia (PJP) (OR 2.33, p = 0.03). AML was associated with infiltrative ARF (OR 2.63, p = 0.009) but negatively with influenza (OR 0.34, p = 0.002) and PJP (OR 0.15, p = 0.009). Lymphomas were linked to pleural effusion (OR 4.31, p < 0.001), airway obstruction (OR 4.30, p = 0.009), and aspiration pneumonia (OR 2.58, p = 0.042). Corticosteroids correlated with undiagnosed ARF (OR 2.07, p = 0.03), and neutropenia with Gram-negative (OR 2.42, p < 0.001) and fungal infections (OR 1.73, p = 0.050). Cluster 1 (myeloid infiltration/uncertainty) had 53% mortality, Cluster 2 (mixed immunosuppression) 49%, and Cluster 3 (chronic lymphoid diseases) 39%.
CONCLUSION: Immune dysfunction strongly influences ARF patterns and prognosis. Immune-guided diagnostic strategies could improve early etiological identification and outcomes in high-risk patients with hematological malignancies.