Short‐term outcomes in infants with mild neonatal encephalopathy: a retrospective, observational study

In this study, we found that 18 % of infants with mild NE had abnormal short-term outcomes such as abnormal brain MRI findings, abnormal neurological examination findings at discharge, or seizures. The Thompson score at admission may be useful for prediction of abnormal short-term outcomes in infants with mild NE.

Unlike moderate or severe cases, infants with mild NE have been considered to have a good prognosis [11, 12]. However, accumulating evidence suggests that neonates with mild NE are also at risk for neurodevelopmental disorders [13‐16]. One reason for these inconsistent reports may be a difference in timing of evaluation, in that the earlier studies defined mild NE by serial neurological examination during the first week after birth whereas the therapeutic hypothermia trials defined the severity of NE based on evaluation within 6 h of birth [25]. Moreover, the timing and methods used to assess the prognosis varied in the earlier studies. In the present study, we found that a considerable number of infants with mild NE had abnormal short-term outcomes, which is consistent with previous studies. Most of the infants with abnormal short-term outcomes in this study had abnormal brain MRI findings; furthermore, there may have been infants with poor prognosis who went undetected because brain MRI was not performed in all cases. Although we did not assess neurodevelopmental outcomes at older ages, brain MRI abnormalities obtained within 4 weeks after birth are reported to be associated with neurodevelopmental impairments [22, 23]. In addition, we must consider not only problems that occur in early childhood but also those that arise in school-aged and adolescent children. van Kooij et al. reported that MRI abnormalities found in the neonatal period are also related to intelligence quotient (IQ) scores obtained in school-aged children and in children with special educational needs, and that even among infants with normal or mild lesions on neonatal MRI, 23.1 % (3/13) had an IQ score ≤ 85 [26]. Furthermore, neuropsychological problems, such as poor attention/executive function, poor memory performance, and behavioral difficulties, have been identified in studies that have followed children with NE through to school age [27]. The long-term neuropsychological outcome of mild NE is controversial; some papers report that even mild NE can cause psychological problems [28, 29] while others report no concern if they are not severe [30]. Although the long-term prognosis of mild NE is not fully understood, we should be aware that the risk of poor neurological prognosis in children with NE of mild severity is not negligible. Therefore, it is important to continue to follow up infants with mild NE in the long term and accumulate data on their long-term outcomes.

Although the Sarnat score is the method most widely used to test for NE, the Thompson score at admission may also be a useful predictor of the short-term outcome in infants with mild NE. Introduced by Sarnat et al. in 1976, the Sarnat score requires evaluation of many items and is somewhat complicated to use [11]. The Thompson score was originally introduced as a simple and effective scoring system for predicting neurodevelopmental outcomes in neonates with hypoxic-ischemic encephalopathy before the era of therapeutic hypothermia. It is a numeric scoring system that requires no equipment or specific training to use. Scores ≥ 15 have been associated with cerebral palsy or severe developmental retardation [19]. Another study reported that the median Thompson score was 4 in mild to moderate encephalopathy and 9 in severe encephalopathy, which is consistent with our data for mild encephalopathy [31]. The Thompson score also has a high predictive value for outcomes in infants treated with therapeutic hypothermia [32]. Although we found that a Thompson score of 4 was a highly sensitive and highly specific predictor of an abnormal short-term outcome, the value was lower than that used by earlier researchers. Compared with previous studies of the predictive ability of the Thompson score [19, 32], we set a milder definition of abnormal outcomes. Therefore, it is possible that our cutoff score was lower than that in previous studies. A recent prospective study by Chalak et al. found that the total Sarnat score recorded in the first 6 h after birth can predict disability at 18–22 months in infants with mild NE [33]. We propose inclusion of this simple scoring system, the Thompson score, into the prognostic assessment of mild NE, which is often treated in institutions without a neonatal neurological specialist.

We could not reach a conclusion regarding the effectiveness of therapeutic hypothermia for mild NE. While there are some papers suggesting that therapeutic hypothermia is effective in mild NE [34, 35], a recent meta-analysis could not confirm its effectiveness [36]. Although therapeutic hypothermia is often already provided for infants with mild NE, prospective trials are needed to confirm its efficacy in these infants. Our present findings suggest that poor prognosis is unlikely in children with milder NE who have a lower Thompson score, such as 1 or 2, but that it is necessary to screen high-risk infants with mild NE in the trials. We hope that our findings will be useful in future research.

The limitations of this study include its retrospective design and lack of long-term follow-up in most cases. However, some infants had brain MRI abnormalities at discharge, which is associated with poor long-term prognosis. Therefore, we assume that mild encephalopathy does not always have a good prognosis.