The ALT SHOCK criteria for respiratory failure in cardiogenic shock

Cardiogenic shock (CS) is frequently complicated by acute respiratory failure (ARF), yet no standardized definitions of ARF have been applied in this context. The Berlin definition for acute respiratory distress syndrome (ARDS) has shown predictive validity in critically ill patients but has not been tested in CS. This study aimed to evaluate whether hypoxemia severity, as defined by the Berlin criteria, predicts outcomes in a large, prospective registry of CS patients with ARF.

Patients requiring positive pressure mechanical ventilation- either non-invasive or invasive - within the AltShock-2 registry (NCT04295252) were included. Demographic and clinical features were analyzed with a focus on Horowitz index along with the other ARDS Berline criteria adapted to the CS setting. Multivariable logistic regression and Cox proportional hazards models were used to identify predictors in-hospital mortality.

Among 1056 patients enrolled from March 2020 to December 2024, 526 met inclusion criteria. Mortality differed by ARF severity: 29.9% (mild), 31.7% (moderate), and 40.5% (severe) (p = 0.043). According to the Horowitz index, 10.6% had severe, 58.9% moderate, and 30.4% mild ARF. Severe ARF was associated with higher prevalence of smoking, hypertension, atrial fibrillation, elevated lactate and PaCO₂, and greater lung congestion, while vasopressor-inotropic scores were similar across groups. Noninvasive ventilation was applied in 213 patients (40.4%), with 22 progressing to invasive ventilation within 24 hours; severe ARF patients were more likely to receive IMV (81.2% vs. 59.9% vs. 49.2%; p = 0.026). V-A ECMO was required in 7.6%, that were more likely to have severe ARF than the other group (p = 0.001). No differences in Horowitz ratio distribution amongst SCAI classes were noted, nor among those who experienced cardiac arrest at presentation (p < 0.05). Severe ARF was more frequent in ischemic vs non-ischemic CS (p = 0.002) and, among ischemic, in STEMI patients (p = 0.026). Severe hypoxemia independently predicted mortality at admission [HR 2.55 (1.16–3.87)] and at 24 h [HR 2.63 (1.98–4.86)], alongside lung congestion severity.

According to our results, severe hypoxemia acts as an outcome modifier and the integration of PaO2/FiO2 ratio should be considered as an additional parameter in the SCAI classification.