Timing of onset of persistent critical illness in Japan: a nationwide registry study.
retrospective_cohort · Level III
Where this comes from
- Record sourced from PubMed, PMID 40688165.
- Also identified by DOI 10.1016/j.lanwpc.2025.101632 and PMC identifier 12274935.
- Licence recorded as CC BY.
- The licence permits redistribution, so the abstract is shown in full and the full text is available from the publisher.
Abstract
Persistent critical illness (PerCI) occurs in patients whose ongoing ICU stay is no longer primarily driven by their presenting illness, but rather by new clinical instability. Studies in Western countries have found PerCI onset at 5-11 days, but timing in Japan or other Asian countries is unknown. Using the Japanese Intensive Care Patient Database (JIPAD), PerCI onset was assessed by comparing the predictive ability of antecedent characteristics components (demographics, comorbidities) and acute illness components (physiological status at ICU admission) for in-hospital mortality, analyzed through multivariable logistic regression models among patients still in ICU from day 1 to day 28. The day when the Area Under the Receiver Operating Characteristic curve (AUROC) of antecedent characteristics exceeded that of acute illness was defined as the onset of PerCI. We studied 285,567 patients from 101 ICUs in Japan between April 2015 and March 2023. Overall in-hospital mortality was 8.2%. The AUROCs of antecedent characteristics and acute illness components for in-hospital mortality was 0.752 and 0.920 on day 1, 0.669 and 0.779 on day 7, 0.668 and 0.743 on day 14 and 0.667 and 0.700 on day 28, respectively. Through day 28, the AUROC of acute illness component remained consistently higher than that of antecedent characteristics component, with no observed onset of PerCI. Unlike findings from Western countries, PerCI onset was not observed in ICU in Japan. This suggests that variation in end-of-life decision making and the critical care delivery system may influence the onset of PerCI. None.