Emily Shepherd, Rehana A Salam, Philippa Middleton, Shanshan Han, Maria Makrides, Sarah McIntyre, Nadia Badawi, Caroline A Crowther

Cerebral palsy is an umbrella term that encompasses disorders of movement and posture attributed to non-progressive disturbances occurring in the developing foetal or infant brain. As there are diverse risk factors and aetiologies, no one strategy will prevent cerebral palsy. Therefore, there is a need to systematically consider all potentially relevant interventions for prevention. PrimaryTo summarise the evidence from Cochrane Systematic Reviews regarding effects of neonatal interventions for preventing cerebral palsy (reducing cerebral palsy risk).SecondaryTo summarise the evidence from Cochrane Systematic Reviews regarding effects of neonatal interventions that may increase cerebral palsy risk. We searched the Cochrane Database of Systematic Reviews (27 November 2016) for reviews of neonatal interventions reporting on cerebral palsy. Two review authors assessed reviews for inclusion, extracted data, and assessed review quality (using AMSTAR and ROBIS) and quality of the evidence (using the GRADE approach). Reviews were organised by topic; findings were summarised in text and were tabulated. Interventions were categorised as effective (high-quality evidence of effectiveness); possibly effective (moderate-quality evidence of effectiveness); ineffective (high-quality evidence of harm); probably ineffective (moderate-quality evidence of harm or lack of effectiveness); and no conclusions possible (low- to very low-quality evidence). Forty-three Cochrane Reviews were included. A further 102 reviews pre-specified the outcome cerebral palsy, but none of the included randomised controlled trials (RCTs) reported this outcome. Included reviews were generally of high quality and had low risk of bias, as determined by AMSTAR and ROBIS. These reviews involved 454 RCTs; data for cerebral palsy were available from 96 (21%) RCTs involving 15,885 children. Review authors considered interventions for neonates with perinatal asphyxia or with evidence of neonatal encephalopathy (3); interventions for neonates born preterm and/or at low or very low birthweight (33); and interventions for other specific groups of 'at risk' neonates (7). Quality of evidence (GRADE) ranged from very low to high.Interventions for neonates with perinatal asphyxia or with evidence of neonatal encephalopathyEffective interventions: high-quality evidence of effectivenessResearchers found a reduction in cerebral palsy following therapeutic hypothermia versus standard care for newborns with hypoxic ischaemic encephalopathy (risk ratio (RR) 0.66, 95% confidence interval (CI) 0.54 to 0.82; seven trials; 881 children).No conclusions possible: very low-quality evidenceOne review observed no clear differences in cerebral palsy following therapeutic hypothermia versus standard care.Interventions for neonates born preterm and/or at low or very low birthweightPossibly effective interventions: moderate-quality evidence of effectivenessResearchers found a reduction in cerebral palsy with prophylactic methylxanthines (caffeine) versus placebo for endotracheal extubation in preterm infants (RR 0.54, 95% CI 0.32 to 0.92; one trial; 644 children).Probably ineffective interventions: moderate-quality evidence of harmResearchers reported an increase in cerebral palsy (RR 1.45, 95% CI 1.06 to 1.98; 12 trials; 1452 children) and cerebral palsy in assessed survivors (RR 1.50, 95% CI 1.13 to 2.00; 12 trials; 959 children) following early (at less than eight days of age) postnatal corticosteroids versus placebo or no treatment for preventing chronic lung disease in preterm infants.Probably ineffective interventions: moderate-quality evidence of lack of effectivenessTrial results showed no clear differences in cerebral palsy following ethamsylate versus placebo for prevention of morbidity and mortality in preterm or very low birthweight infants (RR 1.13, 95% CI 0.64 to 2.00; three trials, 532 children); volume expansion versus no treatment (RR 0.76, 95% CI 0.48 to 1.20; one trial; 604 children); gelatin versus fresh frozen plasma (RR 0.94, 95% CI 0.52 to 1.69; one trial, 399 children) for prevention of morbidity and mortality in very preterm infants; prophylactic indomethacin versus placebo for preventing mortality and morbidity in preterm infants (RR 1.04, 95% CI 0.77 to 1.40; four trials; 1372 children); synthetic surfactant versus placebo for respiratory distress syndrome in preterm infants (RR 0.76, 95% CI 0.55 to 1.05; five trials; 1557 children); or prophylactic phototherapy versus standard care (starting phototherapy when serum bilirubin reached a pre-specified level) for preventing jaundice in preterm or low birthweight infants (RR 0.96, 95% CI 0.50 to 1.85; two trials; 756 children).No conclusions possible: low- to very low-quality evidenceNo clear differences in cerebral palsy were observed with interventions assessed in 21 reviews.Interventions for other specific groups of 'at risk' neonatesNo conclusions possible: low- to very low-quality evidenceReview authors observed no clear differences in cerebral palsy with interventions assessed in five reviews. This overview summarises evidence from Cochrane Systematic Reviews regarding effects of neonatal interventions on cerebral palsy, and can be used by researchers, funding bodies, policy makers, clinicians, and consumers to aid decision-making and evidence translation. To formally assess other benefits and/or harms of included interventions, including impact on risk factors for cerebral palsy, review of the included Reviews is recommended.Therapeutic hypothermia versus standard care for newborns with hypoxic ischaemic encephalopathy can prevent cerebral palsy, and prophylactic methylxanthines (caffeine) versus placebo for endotracheal extubation in preterm infants may reduce cerebral palsy risk. Early (at less than eight days of age) postnatal corticosteroids versus placebo or no treatment for preventing chronic lung disease in preterm infants may increase cerebral palsy risk.Cerebral palsy is rarely identified at birth, has diverse risk factors and aetiologies, and is diagnosed in approximately one in 500 children. To date, only a small proportion of Cochrane Systematic Reviews assessing neonatal interventions have been able to report on this outcome. There is an urgent need for long-term follow-up of RCTs of such interventions addressing risk factors for cerebral palsy (through strategies such as data linkage with registries) and for consideration of the use of relatively new interim assessments (including the General Movements Assessment). Such RCTs must be rigorous in their design and must aim for consistency in cerebral palsy outcome measurement and reporting to facilitate pooling of data and thus to maximise research efforts focused on prevention.