Prognostic performance of gray-white matter ratio in adult out-of-hospital cardiac arrest patients after receiving extracorporeal cardiopulmonary resuscitation

Takashi Hongo,Hiromichi Naito,Michitaka Nasu,Tetsuya Yumoto,Yoshinori Kosaki,Takashi Yorifuji,Toru Hifumi,Akihiko Inoue,Tetsuya Sakamoto,Yasuhiro Kuroda,Atsunori Nakao,SAVE-J Ⅱ Study Group
DOI: https://doi.org/10.1016/j.resuscitation.2024.110351
IF: 6.251
Resuscitation
Abstract:Background: Gray-to-white matter ratio (GWR), measured by computed tomography (CT), is commonly used to predict poor neurological outcomes after out-of-hospital cardiac arrest (OHCA). The prognostic performance of GWR in OHCA patients receiving extracorporeal cardiopulmonary resuscitation (ECPR) is not known. Methods: This study is a secondary analysis of data from the SAVE-J II registry, a retrospective, multicenter study. Participants were divided into four groups according to average GWR (aGWR) values ranging from 1.00 to 1.39, separated by 0.1 intervals. The aGWR values were calculated for bilateral basal ganglia, centrum semiovale, and high convexity obtained by head CT within 24 h after ECPR. Primary outcome was poor neurological outcomes at 30-day. Results: In total, 1,146 OHCA patients treated with ECPR were included in our analysis. Overall, participants with lower aGWR more likely had poor neurological outcomes, aGWR 1.00-1.09 (94.6%), aGWR 1.10-1-19 (87.8%), aGWR 1.20-1.29 (78.5%), and aGWR 1.30-1.39 (70.3%). Multivariable logistic regression showed that lower aGWR was associated with poor neurological outcome at 30-day, aGWR 1.30-1.39: reference, aGWR 1.00-1.09: adjusted odds ratio (aOR) 10.01 (95% confidence interval (CI) [3.58-27.99]), aGWR 1.10-1.19: aOR 4.83 (95% CI [2.31-10.12]), aGWR 1.20-1.29: aOR 2.16 (95% CI [1.02-4.55]). Receiver operating characteristic curve analysis revealed that the prognostic performance of aGWR had an area under the curve of 0.628, 95% CI [0.59-0.66]). The aGWR threshold of 1.005 for predicting poor neurological outcome reached 100% specificity with 0.1% sensitivity. Conclusion: Early neuro-prognostication depending on GWR may not be sufficient after ECPR and requires a multimodal approach.
What problem does this paper attempt to address?