Maayke Hunfeld

32 Chapter 2 In a recent retrospective study (n=41) EEG background suppression (<10 μV) was associated with poor outcome (16). There are important limitations. First, many studies were retrospective (16, 21-24, 26) with often a small cohort (median n=34, interquartile range (IQR) 28-56) and all studies were single center. Second, clinicians were not blinded for the EEG results which might have influenced the decision to withdraw treatment (21-26). Third, most studies had a short-term follow-up (ranging from hospital/PICU discharge in eight studies up to 3 years in one study) with crude outcome scores (i.e. Pediatric Cerebral Performance Category (PCPC)). These studies labeled different PCPC scores as good outcome; PCPC 1 and 2 where defined as good outcome in some studies, other studies used a PCPC score of 1, 2 and 3 as good outcome. When outcome was nonsurvival, cause of death was not specified (did patient die due to withdrawal of life sustaining therapies based on a poor neurological prognosis, or due to refractory circulatory failure?). No neuropsychological assessments were performed. Fourth, different studies used different EEG classifications, restricting comparability. Fifth, timing of the EEG was not always fully specified. This should be taken into account as studies in adults after CA suggest that EEG patterns at 12 hours after ROSC have the best predictive value for good outcome and patterns at 24 hours post ROSC have the best predictive value for poor outcome (27). Finally, although a normal EEG is associated with a good outcome, there were children with a normal EEG and poor outcome and vice versa (22, 25). Transcranial Doppler A retrospective study by Lin et al (n=17) showed that TCD can serve as a prognostic tool in children post CA receiving hypothermia (28) (Supplemental Table 4, Supplemental Digital Content 2, http://links.lww.com/PCC/B352) . A normal mean flow velocity of the middle cerebral artery (MCA) in the rewarming phase led to a better outcome compared with a low peak flow velocity (p=0.009). A normal pulsatility index (PI) value led to a significantly better outcome than a high PI value in the hypothermia and rewarming phases (p=0.002 and 0.003, respectively). However, this was a retrospective study with a small cohort. Also, autoregulation is often impaired in children after CPR which might influence TCD values (29). Brain MRI We included four retrospective studies and one observational exploratory study on brainMRI made within 2 weeks after pediatric CA in relation to outcome (Supplemental Table 5, Supplemental Digital Content 2, http://links.lww.com/PCC/B352) .