Temperature management and outcomes after cardiac arrest: a retrospective registry study

We examined whether the association between targeted temperature management (TTM) and neurological outcomes differed by cardiopulmonary resuscitation (CPR) duration. This retrospective Taiwanese registry study compared low-temperature TTM (33–34 °C) with heterogeneous active temperature management. Outcomes were in-hospital mortality and unfavorable neurological outcome (Cerebral Performance Category [CPC] 3–5). Adjusted models included initial shockable rhythm. Analyses included propensity matching, continuous CPR-duration modeling and 50 treatment-stratified multiple imputations. The ten-minute subgroup threshold was selected post hoc. Among 433 patients, 342 received low-temperature TTM and 91 received comparator treatment; adjusted complete-case models included 397 patients. Good neurological recovery (CPC 1–2) occurred in 27.2% versus 27.5%. The adjusted neurological association was nonsignificant overall ( p = 0.235), but differed by CPR duration (interaction p = 0.003; multiplicity-adjusted p = 0.061). For CPR lasting ten minutes or less, the adjusted odds ratio (OR) was 0.08 (95% confidence interval [CI] 0.01–0.44); no significant association was observed after longer CPR (OR 1.19, 95% CI 0.56–2.51). The interaction persisted after multiple imputation in 558 patients ( p = 0.007). Lower in-hospital mortality was associated with low-temperature TTM (OR 0.50, 95% CI 0.28–0.89), with sensitivity to model specification. Low-temperature TTM may be associated with better neurological outcomes in patients whose CPR duration was ten minutes or less, though this finding requires external validation. Trial registration ClinicalTrials.gov, NCT03578328; registered 6 July 2018.