Maayke Hunfeld

159 MRI and prognosis Discussion This is, to the best of our knowledge, the first study that shows that a normal brain MRI (without evidence of post-hypoxic brain injury) within 1 week after pediatric OHCA is 100% predictive for good neurological outcome 2 years post-OHCA. Conversely, the presence of extensive brain injury (defined as ≥ 50% of the cortex/white matter or in 4 or more of the defined brain regions) is 100% predictive for poor neurological outcome (both including and excluding death) or death at hospital discharge and 2 years post-OHCA. The finding that a normal brain MRI is predictive for good outcome 2 years post- OHCA is in line with previous work from Oualha et al. (11). In their study brain MRI was performed in 20 children within one week after cardiac arrest. All 5 children with a normal brain MRI had good outcome (outcome was determined beyond 24 months, but the exact follow-up interval per patient has not been specified). However, they included both IHCA and OHCA children. In a retrospective cohort by Fink et al., including 28 children with CA, all 16 children with a normal brain MRI (median 6 days post-CA, IQR 4-11 days) survived until hospital discharge. Of those, ten children had no change in Glasgow Outcome Score (GOS) at discharge (compared to their GOS pre- arrest), but whether the outcome was good or poor is not described (12). Differences between our study and Fink’s cohort makes comparison difficult: 1.) In Fink’s cohort children with both IHCA and OHCA were included versus only children with OHCA in our study. 2.) Median time of brain MRI was 6 days in Fink’s cohort. This implies that a number of brain MRIs were performed beyond 1 week post-CA. After week pseudo- normalization of DWI can occur (20) 3.) In Fink’s study, outcome was determined at hospital discharge versus up to 2 years in ours. We have shown in this study and in previous work that outcome at discharge can improve over time (21). This underlines the importance that follow-up should not end at hospital discharge. This is the first study describing that the presence of extensive brain injury has a 100% PPV for a poor neurological outcome and for death at discharge and 2 years post-OHCA, without the use of other OHCA variables. Other studies did show an association between the presence of multiple affected brain areas on MRI after CA and poor outcome and death (mainly at hospital discharge) (11, 12, 14, 22), but PPV was not reported. The presence of focal injury on brain MRI, whether or not involving deep gray matter, has a PPV of only around 50% both and 2 years post-OHCA, reflecting that around 50% of the children with focal brain injury has poor neurological outcome or dies. This means that purely based on MRI with focal brain injury, it is impossible to predict neurological outcome accurately enough to support management decisions, especially WLST. Other variables (such as pre-arrest, intra-arrest and post-arrest 5