Epidemiology, ventilation and outcomes of acute respiratory failure in immunocompromised patients from 103 intensive care units in 26 countries were discussed at ISICEM this week.
Efraim III, a large multinational retrospective observational study, examined the epidemiology, management, and outcomes of acute hypoxaemic respiratory failure (ARF) in immunocompromised patients admitted to ICUs. The study addresses an important clinical gap, as contemporary large-scale data in this population have been limited despite significant advances in oncology, transplantation, and immunosuppressive therapies that have increased both survival and ICU admissions among immunocompromised individuals.
The study included 9854 adult patients from 103 ICUs across 26 countries between 2017 and 2023. Patients had ARF and a form of immunodeficiency, most commonly haematological malignancy (48.3%) or solid tumours (38.7%). The median age was 64 years, and the cohort was predominantly male (60%). The primary objective was to determine 30-day mortality and identify predictors of mortality and intubation using multivariable statistical models.
ARF in this population is complex, often resulting from multiple overlapping causes. Infection was the leading cause, accounting for 62% of cases, with bacterial infections being the most frequent (41.7%), followed by viral (20.9%) and invasive fungal infections (8.2%). Non-infectious causes included disease-specific lung infiltrates, cardiogenic pulmonary oedema, and pulmonary embolism. In 15.1% of patients, no cause of ARF could be identified despite extensive diagnostic investigations, and more than half of the patients had multiple contributing causes.
Management strategies varied, reflecting real-world practice. Initial oxygenation approaches included conventional oxygen therapy (52.0%), high-flow nasal oxygen (HFNO; 20.3%), non-invasive ventilation (9.3%), and immediate intubation (20.4%). Diagnostic procedures such as bronchoscopy and imaging were widely used, although protocols were not standardised across centres.
The study found a high mortality rate: 37.1% at ICU discharge, 47.3% at 30 days, and 54.8% at 90 days. Several independent predictors of increased 30-day mortality were identified. These included older age, greater comorbidity burden, higher frailty scores, delayed ICU admission after hospitalisation, higher respiratory rate, and coma at ICU admission. Clinical factors such as invasive fungal infection, disease-specific infiltrates, and particularly an unidentified cause of ARF were strongly associated with worse outcomes. The need for organ support, including vasoactive drugs and renal replacement therapy, was also linked to higher mortality.
Several factors were associated with improved survival. These included having undergone a solid organ transplant, having systemic vasculitis or connective tissue disease, a higher PaO₂/FiO₂ ratio and cardiogenic pulmonary oedema as the cause of ARF. The use of HFNO therapy was associated with reduced mortality in the primary analysis, although this finding was not consistent in sensitivity analyses.
Intubation was required in 56% of patients. The strongest predictors of intubation were the need for vasoactive drugs, renal replacement therapy, coma at ICU admission, and invasive fungal infection. Higher oxygenation (PaO₂/FiO₂ ratio), cardiogenic pulmonary oedema, and certain underlying diseases such as acute myeloid leukaemia were associated with a lower likelihood of intubation. The risk of intubation increased with worsening hypoxaemia regardless of the initial oxygenation strategy.
The study provides several important clinical insights. First, it suggests that outcomes for immunocompromised patients with ARF may have improved compared with earlier studies, likely due to advances in supportive care and earlier ICU admission. Mortality among intubated patients was lower than previously reported, challenging the perception that invasive mechanical ventilation is futile in this population.
Second, the findings highlight the critical importance of early and accurate diagnosis. The strong association between unidentified ARF causes and mortality underscores the need for improved diagnostic strategies. Emerging technologies such as metagenomics and transcriptomics may help identify elusive pathogens or immune-mediated conditions in the future.
Third, the study emphasises modifiable factors that could improve outcomes. Delays in ICU admission were associated with worse survival, suggesting that earlier referral and escalation of care may be beneficial. Similarly, the potential role of HFNO therapy and the need for targeted prevention and early treatment of fungal infections warrant further investigation.
Overall, this study identifies key predictors of mortality and intubation in immunocompromised patients with ARF. It highlights the importance of frailty, comorbidities, timely ICU admission, and accurate diagnosis, while suggesting that outcomes may be improving and that intubation should not be considered futile. The findings provide a foundation for improving clinical decision-making and guiding future research in this high-risk population.
Source: The Lancet
Image Credit: iStock
