Answers are grounded in this paper and cite their sources.
Acute respiratory distress syndrome (ARDS) is a highly heterogeneous condition associated with high mortality. Although the modified shock index (MSI) predicts clinical outcomes, standard static evaluations fail to capture dynamic physiological responses across distinct subphenotypes. This study aimed to identify clinical subphenotypes and evaluate the dynamic time-varying associations between the MSI and hospital mortality. In this retrospective cohort study, we used data from MIMIC-IV and eICU-CRD data, and independently applied consensus K-means clustering in each cohort to identify ARDS subphenotypes. Generalized additive models were used to characterize the dynamic time-varying associations between MSI and hospital mortality. We identified three reproducible subphenotypes among 1,938 patients in the MIMIC-IV cohort and 5,177 in the eICU-CRD cohort. Subphenotype 1 was characterized by severe multiorgan dysfunction and the poorest prognosis. Subphenotype 2 presented with preserved hemodynamics and the highest survival probability, whereas Subphenotype 3 featured older age, alkalosis, and intermediate prognosis. Dynamic modeling revealed an lower-risk MSI range of 0.7 to 1.1 for the overall population. Temporal risk profiles varied significantly across subphenotypes. Subphenotype 1 demonstrated a lower-risk MSI interval of 0.5 to 1.1, with mortality risk escalating rapidly at higher exposures. Subphenotype 2 showed a time-varying U-shaped profile but lacked a consistent lower-risk interval across two database. Subphenotype 3 exhibited a J-shaped trajectory with a lower-risk interval of 0.7 to 1.1. Across all groups, sustained exposure to MSI values higher than 1.5 was consistently associated with elevated mortality. The dynamic time-varying association between MSI and hospital mortality differed across clinical ARDS subphenotypes.These findings support interpretation of continuous MSI measurements within the broader clinical subphenotypes of each patient for risk stratification, rather than use of a single MSI threshold as a universal hemodynamic treatment target.
No takes yet. Share an insight, caveat, or question.
Yuan et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: