Lung ultrasound and mortality in a cardiogenic shock population: A prospective registry‐based analysis

Aims: Lung ultrasound (LUS) is a widely used technique to assess de-aeration in critically ill patients with respiratory failure. There is paucity of data on LUS in cardiogenic shock (CS). We sought to evaluate the epidemiology of lung congestion and its relation with outcome. Methods and results: The Altshock-2 registry is a multicentre, prospective, observational registry including all-comer CS patients. The LUS protocol included the examination of four zones using dichotomous assessment of lung congestion severity: ≤50% or >50%. LUS was performed at admission and at 24 h. Univariate and multivariate logistic regression analyses were performed. Overall, 185 patients (mean age 64.2 ± 13.5 years; 25.9% female) had a LUS at admission. A total of 128 patients (69.2%) had ≥50% of the investigated lung field with B-lines. At univariate Cox regression analysis, B-lines ≥50% at 24 h were significantly associated with increased 30-day mortality (hazard ratio [HR] 4.705; 95% confidence interval [CI] 2.329–9.508) and the reduction of B-lines during 24 h was associated with lower 30-day mortality (HR 0.739; 95% CI 0.571–0.956; p = 0.021). Results were confirmed at multivariate analysis after adjustment for significant covariates: B-lines ≥50% at 24 h (HR 2.23; 95% CI 1.042–8.654; p = 0.041) and the reduction in B-lines from baseline to 24 h (HR 0.815; 95% CI 0.415–1.132; p = 0.039). The sensitivity analysis, excluding patients with cardiac arrest, led to significantly increased accuracy in outcome prediction. Conclusion: Assessment and monitoring of lung congestion with LUS over the first 24 h in patients with CS allow to further stratify clinical outcomes with higher accuracy when added to SCAI classification, especially when excluding patients with cardiac arrest at CS presentation.