Identification and validation of chronic critical illness subphenotypes using urea-creatinine ratio trajectories.

BACKGROUND & AIMS: Chronic critical illness (CCI) is common in intensive care units (ICUs) and is highly heterogeneous. We utilized urea-creatinine ratio (UCR) trajectories to develop and validate subphenotypes and reveal the heterogeneous treatment effects of medical nutrition therapy.

METHODS

This retrospective study included all CCI patients from three public critical care databases. Group-based trajectory modeling (GBTM) was applied to the UCR for subphenotype development and validation. A multivariate Cox regression model was used to assess the independent associations of these subphenotypes with mortality. Generalized estimation equations were used to reveal the potential heterogeneity in medical nutrition therapy between survivors and nonsurvivors in different subphenotypes.

RESULTS

A total of 4047 CCI patients were divided into three subphenotypes on the basis of their UCR trajectories. Stable low subphenotype had a low UCR with a slight upwards trend, the youngest age, and the fewest comorbidities. Intermediate subphenotype was characterized by a medium UCR. Stable high subphenotype had a high UCR with a rapid increase and the highest ICU mortality (14.72 %), hospital mortality (36.20 %) and 28-day mortality (39.26 %) (p < 0.05). In the multivariate Cox regression with Stable low subphenotype as control, Stable high subphenotype had increased risks of hospitalization mortality (HR: 2.74; 95 % CI: 2.01-3.72; P < 0.001), 28-day mortality (HR: 3.20; 95 % CI: 2.36-4.34; P < 0.001) and ICU mortality (HR: 2.78; 95 % CI: 1.71-4.52; P < 0.001). In CCI patients, the dose of nutritional intake within 5 days after CCI diagnosis were greater in the survival group (P < 0.05), especially survivors in Intermediate and Stable high subphenotypes.

CONCLUSION

We demonstrated that longitudinal UCR trajectories during the first 10 days of ICU admission serve as robust biomarkers for predicting CCI into three distinct subphenotypes to understand patient heterogeneity. For CCI patients, higher doses of enteral nutrition and protein intake after CCI diagnosis may improve prognosis, especially for patients in the subphenotypes with the higher baseline UCR values showing an upward trend and poorer prognosis. A prospective study is needed to validate these findings, inform practice and guide future research on personalized care.