Phenotypes of Functional Decline or Recovery in Sepsis ICU Survivors: Insights From a 1-Year Follow-Up Multicenter Cohort Analysis.
Summary
In a 21-ICU prospective cohort (n=220), four discharge phenotypes of sepsis-associated PICS were identified spanning none, mild, moderate, and severe multi-domain impairments. Mild physical/cognitive deficits improved by 3 months, but moderate and severe phenotypes showed persistent disability at 12 months; severe phenotype had ongoing depression and declining survival.
Key Findings
- Four discharge phenotypes identified: no PICS (n=62), mild (physical and cognitive; n=55), moderate (all domains; n=53), severe (all domains; n=50).
- Mild phenotype improved by 3 months; moderate and severe phenotypes showed persistent disability over 12 months.
- Psychiatric PICS improved in moderate/severe groups, but depression persisted at 12 months in the severe group.
- All groups had persistently reduced QoL and low employment (0–50%), with continuously decreasing survival in the severe group.
Clinical Implications
Use discharge assessments (Barthel, SMQ, HADS, IES-R, EQ-5D-5L, frailty, grip/MRC) to assign patients to phenotypes and prioritize intensive, multi-domain rehabilitation and mental health support for moderate/severe PICS; monitor severe phenotype closely for mortality risk.
Why It Matters
This phenotyping provides a pragmatic framework to tailor post-ICU follow-up and rehabilitation resources based on early discharge assessments, potentially improving long-term outcomes for sepsis survivors.
Limitations
- Moderate sample size (n=220) with potential selection bias to survivors
- Cluster phenotypes are data-driven and require external validation for generalizability
Future Directions
Validate phenotypes in larger, international cohorts and test phenotype-guided rehabilitation and mental health interventions in pragmatic trials to improve survivorship outcomes.
Study Information
- Study Type
- Cohort
- Research Domain
- Prognosis
- Evidence Level
- III - Prospective multicenter cohort identifying prognostic phenotypes with 1-year outcomes.
- Study Design
- OTHER