Inter-Observer Agreement on the Diagnosis of Neurocardiogenic Injury Following Aneurysmal Subarachnoid Hemorrhage

Abstract

Neurocardiogenic injury results from increased sympathetic nervous system activation following acute brain injury. No diagnostic criteria for neurocardiogenic injury exist, and agreement has not been tested. We investigated the agreement by neurointensivists for the presence of neurocardiogenic injury on routine cardiac studies. Six neurointensivists rated 100 consecutive cases of aneurysmal subarachnoid hemorrhage (aSAH) for the presence of neurocardiogenic injury. A fixed-panel design was employed for the agreement among the whole cohort, as well as stratified by modified Fisher Scale (mFs), Hunt and Hess grade, gender, and the presence of elevated cardiac enzymes. Overall percent agreement, paired agreement, and agreement above change (Fleiss' Kappa) were calculated. Overall percent agreement between groups was compared using Chi square tests. Six raters completed the survey for a total 600 responses. Overall percent agreement was 79.3 %, and agreement among cases at least one rater thought had neurocardiogenic injury was 66.5 % (paired agreement). Fleiss' Kappa was 0.66 (95 % CI, 0.1-0.71; p < 0.0001), indicating substantial agreement above chance. Similarly, on subgroup analysis, significant agreement beyond chance was seen in all groups (p < 0.001). Overall percent agreement was significantly better among mFs 3-4 compared to mFs ≤ 2 (81.3 vs. 63.6 %; p = 0.018) and among cases with positive cTI (96.9 vs. 70.1 %; p ≤ 0.001). Overall, we demonstrated substantial agreement for the presence of neurocardiogenic injury on early cardiac studies following aSAH. However, inter-observer variability increased when evaluating patients without the objective finding of elevated cTI and among those with lower clinical and radiographic grades.