Objectives: Assessment of whether admission plasma troponin I level is associated with mortality or requirement for vasoactive drugs in pediatric intensive care.
Design: Retrospective cohort study.
Setting: Single centre, tertiary referral general PICU, without a cardiac surgical program.
Patients: Three hundred and nineteen patients 0-18 years old in two cohorts. Cohort 1 was admitted between January 2009 and September 2012 and the cohort 2 between April 2014 and April 2015.
Interventions: None.
Measurements and main results: Plasma troponin I was measured in patients in cohort 1 only if the attending physician ordered the test due to clinical concern regarding myocardial injury. The second cohort had plasma troponin I routinely measured at admission. The primary outcome was death during PICU admission, and the secondary outcome was maximum inotrope requirement during PICU stay, measured by Vasoactive Inotrope Score. Plasma troponin I was a discriminator for mortality in both cohorts (area under the receiver-operating characteristic curve of 0.73 and 0.86 in cohorts 1 and 2, respectively). In an adjusted analysis using Cox regression, accounting for Pediatric Index of Mortality 2 risk of mortality and age, elevated plasma troponin I was significantly associated with death in both cohorts (hazard ratio, 4.99; p = 0.033; hazard ratio, 10.5; p = 0.026 in cohorts 1 and 2, respectively). Elevated plasma troponin I was only associated with increased Vasoactive Inotrope Score following multivariate analysis in the cohort 2.
Conclusions: Detectable plasma troponin I at admission to PICU is independently associated with death. The utility of troponin I as a stratification biomarker requires further evaluation.