Types of studies

We included all randomised controlled trials (RCTs) and quasi‐RCTs that compare cooling plus xenon versus cooling alone.

Types of participants

1. Newborn infants at 35 weeks’ gestation or greater with:

   1. evidence of peripartum asphyxia, with each enrolled infant satisfying at least one of the following criteria.

      1. Apgar score ≤ 5 at 10 minutes.

      2. Mechanical ventilation or resuscitation at 10 minutes.

      3. Cord pH < 7.1, or arterial pH < 7.1, or base deficit ≥ 12 within 60 minutes of birth.

   2. evidence of encephalopathy according to Sarnat staging (Finer 1981; Sarnat 1976).

      1. Stage 1 (mild): hyperalertness, hyperreflexia, dilated pupils, tachycardia, absence of seizures.

      2. Stage 2 (moderate): lethargy, hyperreflexia, miosis, bradycardia, seizures, hypotonia with weak suck, and Moro.

      3. Stage 3 (severe): stupor, flaccidity, small to mid‐position pupils that react poorly to light, decreased stretch reflexes, hypothermia, and absent Moro.

   3. induced therapeutic hypothermia treatment (whole body or selective head cooling to 32°C to 34°C) initiated within six hours after birth; and

   4. no major congenital abnormalities recognisable at birth.

Types of interventions

Inhaled xenon (irrespective of timing and concentrations used) as an adjuvant to therapeutic hypothermia versus therapeutic hypothermia alone, based on the following prespecified definitions.

1. Therapeutic hypothermia.

   1. Standard therapeutic hypothermia (whole body or selective head cooling to 32°C to 34°C initiated within six hours after birth and continued for 72 hours before slow rewarming.

2. Xenon administration.

   1. Irrespective of timing (starting age and duration) and concentrations used.

Types of outcome measures

Electronic searches

We used the criteria and standard methods of Cochrane and the Cochrane Neonatal Review Group. We undertook a comprehensive search of the following electronic sources.

We used MeSH terms and keywords to search the following.

1. MEDLINE (1966 to 01 August 2017).

2. Embase (1966 to 01 August 2017).

3. Cochrane Library (01 August 2017; 2017, Issue 8).

We used keywords (to retrieve e‐publications and items not indexed in MEDLINE).

1. PubMed (1966 to 01 August 2017).

Others.

1. Conference proceedings of the Perinatal Society of Australia and New Zealand (from 2005 to 01 August 2017).

2. Conference proceedings of the Pediatric Academic Societies (from 2000 to 01 August 2017).

We have presented in Appendix 1 the full search strategies for each database. We screened the reference lists of any cited articles.

Searching other resources

We searched clinical trial registries for ongoing and recently completed trials (e.g. ClinicalTrials.gov (clinicaltrials.gov), World Health Organization International Trials Registry and Platform, International Standard Randomised Controlled Trial Number (ISRCTN) registry (www.isrctn.com/)).

Selection of studies

Two review authors (CR and JC) independently searched and identified eligible trials that met the inclusion criteria using Covidence, which is an online screening and data extraction tool used for Cochrane Reviews (Covidence 2017). First, we screened titles and abstracts to identify potentially relevant citations; then we retrieved the full text of all potentially relevant articles. We independently assessed the eligibility of studies in accordance with the specified inclusion criteria. We reviewed studies for relevance based on study design and types of participants, interventions, and outcome measures. We resolved disagreements by discussion and, if necessary, by consultation with a third review author (PD).

We have provided details of studies excluded from the review in the Characteristics of excluded studies table, along with reasons for exclusion. We contacted trial authors if details of primary trials were unclear.

Data extraction and management

Two review authors (CR and JC) separately extracted, assessed, and coded all data for each study using an online screening and data extraction tool for Cochrane Reviews (Covidence 2017). One review author (CR) checked exported data using Review Manager 5 (RevMan 5) software (Review Manager 2014). A third review author addressed disagreements.

Assessment of risk of bias in included studies

Two review authors (CR and JC) independently assessed the risk of bias (as low, high, or unclear) of all included trials using the Cochrane 'Risk of bias' tool (Higgins 2011) for the following domains.

1. Sequence generation (selection bias).

2. Allocation concealment (selection bias).

3. Blinding of participants and personnel (performance bias).

4. Blinding of outcome assessment (detection bias).

5. Incomplete outcome data (attrition bias).

6. Selective reporting (reporting bias).

7. Any other bias.

We resolved disagreements by discussion or by consultation with a third review author. See Appendix 2 for a detailed description of risk of bias for each domain.

Measures of treatment effect

We performed statistical analyses using standard methods of the Cochrane Neonatal Review Group. We analysed results of studies using RevMan 5 (Review Manager 2014), and we presented results as risk ratios (RRs), risk differences (RDs), number needed to treat for an additional beneficial outcome (NNTB), or number needed to treat for an additional harmful outcome (NNTH) for categorical variables. We used mean differences (MDs) for continuous variables. We reported 95% confidence intervals (95% CIs) for all estimates.

Unit of analysis issues

We included all RCTs and quasi‐RCTs. We took into account the level at which randomisation occurred, including cross‐over trials, cluster‐randomised trials, and multiple observations for the same outcome.

Dealing with missing data

We planned to request additional data from the authors of each trial if data on important outcomes were missing or needed clarification. When data were still missing, we planned to examine the effects of losses by performing sensitivity analysis. We performed these analyses by intention‐to‐treat.

Assessment of heterogeneity

We assessed statistical heterogeneity by examining the I² statistic ‐ a quantity that describes the proportion of variation in point estimates that is due to variability across studies rather than to sampling error (Higgins 2011). We applied I² statistic cutoffs and labels for heterogeneity as follows.

1. Less than 25%: no heterogeneity.

2. 25% to 49%: heterogeneity.

3. 50% to 74%: moderate heterogeneity.

4. ≥ 75%: high heterogeneity.

We considered statistical heterogeneity to be substantial when the I² statistic value was greater than 50%. In addition, we employed the Chi² test of homogeneity to determine the strength of evidence that heterogeneity is genuine. We explored clinical variation across studies by comparing the distribution of important participant factors among trials and trial factors (randomisation concealment, blinding of outcome assessment, loss to follow‐up, treatment type, and co‐interventions). We considered a threshold P value less than 0.1 as an indicator of important heterogeneity (genuine variation in effect sizes).

Assessment of reporting biases

We planned to use funnel plots to assess small‐study effects. Owing to several possible explanations for funnel plot asymmetry, we planned to interpret results carefully (Sterne 2011).

Data synthesis

We planned to perform statistical analyses using standard methods of the Cochrane Neonatal Review Group. We used RevMan 5 software with the fixed‐effect model for meta‐analysis (Review Manager 2014). We used standardised mean differences (SMDs) to combine trials that measured the same outcome using different methods. We used weighted mean differences (WMDs) with 95% CIs for outcomes measured on a continuous scale.

Subgroup analysis and investigation of heterogeneity

We carried out the following subgroup analyses.

1. Severity of HIE.

   1. Based on Sarnat score (Finer 1981; Sarnat 1976).

      1. Mild versus moderate/severe.

   2. Based on EEG or aEEG at baseline.

      1. Mild (electrographic seizures, dysmaturity) versus moderate/severe (low voltage or discontinuous background/isoelectric or burst‐suppression pattern).

2. Xenon administration.

   1. Concentration: < 30% versus ≥ 30%.

   2. Starting age: < six hours versus ≥ six hours after insult.

   3. Duration: < 12 hours versus ≥ 12 hours.

3. Gestational age.

   1. Late preterm (35 0/7 through 36 6/7 gestational weeks) versus term infants (≥ 37 0/7 gestational weeks).

4. Quality of outcome assessment.

   1. High quality (≥ 18 months with formal psychological testing and review by developmental paediatrician for diagnosis of cerebral palsy (CP)) versus lower quality.

Sensitivity analysis

We conducted sensitivity analyses to explore the effects of methodological quality of trials and to ascertain whether studies at high risk of bias overestimate the effects of treatment.