Hyperbaric oxygen for term newborns with hypoxic ischaemic encephalopathy

Abstract

This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:

The objective of this review was to determine the effects of therapeutic HBO (with or without hypothermia) on death and long‐term neurodevelopmental disabilities in term newborn infants with HIE. Furthermore, the short‐term medical effects and side effects of HBO were studied. Primary comparisons:

1. HBO versus supportive care;

2. HBO versus hypothermia therapy;

3. HBO plus hypothermia versus hypothermia alone;

4. HBO plus erythropoietin versus erythropoietin alone.

Subgroup analyses were planned on the basis of:

1. severity of HIE (mild, moderate, severe) (Sarnat 1976; Finer 1981);

2. the therapeutic time window, by timing of commencement of the intervention (< 6 hours versus 6 hours to 7 days versus >7 days);

3. the peak pressure of HBO (<2 ATA vs. >2ATA);

4. the number of courses of HBO (single course versus multiple courses).

Background

Description of the condition

Perinatal asphyxia is caused by lack of oxygen (hypoxia) or lack of perfusion (ischaemia), or both, to various organs of the fetus or the newborn. Perinatal asphyxia can occur in utero, during labor and delivery, or in the postnatal period. Estimates of the incidence of perinatal asphyxia vary from one to eight per 1000 live births (van Handel 2007). The World Health Organization has estimated that perinatal asphyxia is the cause of 23% of mortality in the neonatal period and 8% of global child mortality at the age of less than five years (Bryce 2005).

In the neonate, hypoxic ischaemic encephalopathy (HIE) is characterized by clinical and laboratory evidence of acute or subacute brain injury due to perinatal asphyxia. The cause of HIE is not always obvious; HIE is not a single disease entity but a condition resulting from diverse causes that manifest as signs of brain injury (Higgins 2006). Moderate or severe HIE caused by perinatal asphyxia affects 0.5 to 1 per 1000 live births (Jacobs 2013). HIE is one of the main causes of disabilities in infants born at term. Fifteen per cent to 20% of affected newborns will die during the postnatal period and an additional 25% will sustain permanent clinical deficits (Ferriero 2004). Deficits include functional motor deficits (cerebral palsy, which is a non‐progressive motor or postural disorder originating in early life), cognitive deficits (mental retardation or subnormal intellectual function resulting in impaired language skills, learning, executive functions or social ability), and auditory impairment (Gonzalez 2006; Jiang 2008a). Historically, treatment has been limited to supportive intensive care (Roka 2008). Recently, trials of therapeutic mild hypothermia treatment in infants with moderate to severe HIE has proven effective in reducing death and serious developmental disability (Jacobs 2013; Kirpalani 2007).

Description of the intervention

Hyperbaric oxygen (HBO) therapy is defined by the Undersea and Hyperbaric Medical Society as a treatment in which a patient is placed in a treatment chamber and intermittently breathes 100% oxygen while the treatment chamber is pressurized to a pressure greater than sea level (a pressure greater than 1 atmosphere absolute (ATA), where 1 ATA = 750 mm Hg = 0.1 MPa). The roots of HBO therapy can be traced to over three centuries ago, when Henshaw built the first hyperbaric chamber (Gill 2004). HBO is regarded as the only treatment for decompression sickness and arterial gas embolism and has been approved by the Undersea and Hyperbaric Medical Society for a variety of other medical conditions (Gill 2004).

HBO therapy has been identified as a promising therapy for ischaemic injury to the central nervous system. In experimental animal models of neonatal hypoxic ischaemic brain injury, HBO therapy has been demonstrated to improve neurological outcome (Matchett 2009). In 1963, HBO therapy was applied in resuscitation of the newborn with perinatal asphyxia (Hutchison 1963). Further exploratory studies have been performed since that time. There is a concern that excessive oxygen may cause retinopathy of prematurity or bronchopulmonary dysplasia, leading some investigators to recommend against the use of HBO in neonates (Yang 2008). However, hyperbaric oxygen has been used to treat newborns with neonatal HIE in clinical studies in China. The time window of HBO is still controversial. In clinical studies, HBO is usually initiated within one to seven days after birth, administered one to three times per day at 0.15 to 0.17 MPa for 60 to 120 minutes, and continued for one to four courses of treatment (Liu 2006).

How the intervention might work

The mechanisms of HBO therapy are not completely understood. HBO therapy has been reported to improve brain injuries in different cerebral regions, including the cortex (Wang 2009a), white matter (Wang 2007a), and hippocampus (Liu 2007; Bai 2008; Wang 2008a), after hypoxia ischaemia. Hypoxia during asphyxia leads to impairment of mitochondrial function, energy failure, accumulation of purine derivatives, and increased generation of reactive oxygen species. After ischaemia, the resultant hypoxia leads to a vicious cycle of events including reduced energy metabolism, brain edema and elevated intracranial pressure that can ultimately result in cell death (Chang 1999; Littlejohns 2005; Stiefel 2005; Van Putten 2005). The application of HBO is based on the theory that inhalation of oxygen at increased atmospheric pressure might produce a marked elevation of oxygen partial pressure in arterial blood and thus improve oxygen tension in the hypoxic brain (Nighoghossian 1997; Calvert 2007).

Recent studies have demonstrated that neuronal death occurs in two phases following hypoxic ischaemic insult. In the first phase, there may be immediate 'primary neuronal death' that is related to cellular hypoxia with exhaustion of the cellular high energy stores (primary energy failure) immediately after the insult (Jacobs 2013). Hyperbaric oxygen can produce a marked elevation of oxygen partial pressure in the arterial blood and thus improve the oxygen tension in the hypoxic brain.

After a latent period of at least six hours, the secondary phase of 'delayed neuronal death' begins. The mechanisms of delayed neuronal death include hyperemia, cytotoxic edema, mitochondrial failure, accumulation of excitotoxins, cellular apoptosis, nitric oxide synthesis, free radical damage, and cytotoxic actions of activated microglia. In several animal models, HBO therapy has been demonstrated to reduce these processes. In a rat brain injury model, HBO treatment significantly increased brain tissue partial pressure of oxygen (PO₂) after injury and restored the mitochondrial redox potential (a measure of mitochondrial function) by four hours (Daugherty 2004). In seven‐day old rat pups subjected to unilateral carotid artery ligation, HBO therapy restored the levels of adenosine triphosphate (ATP) and phosphocreatine and increased the utilization of energy, ultimately leading to a reduction in brain injury (Calvert 2007a). Hyperbaric oxygen preconditioning provides brain protection against hypoxic ischaemic (HI) insult via inhibition of neuronal apoptosis pathways (Li 2008). In a middle cerebral artery occlusion rat model, HBO therapy prevented apoptosis and promoted neurological function and the opening of the mitochondrial ATP‐sensitive potassium channel (Lou 2006), inhibited neutrophil infiltration in the injured brain, decreased inflammation (Atochin 2000; Miljkovic‐Lolic 2003), reduced basal lamina degradation, and preserved the integrity of the blood‐brain barrier after cerebral ischaemia (Veltkamp 2006). In a rat traumatic brain injury model, the application of HBO during the early phase significantly diminished intracranial pressure elevation and decreased the mortality level (Rogatsky 2005). In neonatal rats with intrauterine hypoxic ischaemic brain damage, early HBO treatment can increase synaptic transmission efficiency, improve central nervous electrophysiological conduction velocity, and reduce neuronal death (Chen 2009).

In a clinical trial, Zhou and colleagues investigated the roles of HBO in antioxidant capacity in neonates with HIE (Zhou 2008). They found that the serum superoxide dismutase (SOD) level increased and serum levels of malondialdehyde (MDA), nitric oxide (NO) and nitric oxide synthetase (NOS) decreased significantly after HBO therapy. The antioxidant capacity increases with increasing HBO pressure in neonates with HIE.

Advances in the understanding of stem cell biology may lead to promising approaches for rescue therapy in developing brains (Ferriero 2002). HBO promotes the proliferation of neural stem cells in hypoxic ischaemic brain damaged (HIBD) neonatal rats (Wang 2007). HBO also promotes cortical migration and differentiation of endogenous neural stem cells in neonatal rats with HIBD (Wang 2009a). HBO can promote the differentiation of implanted human neural stem cells into neurons in neonatal rats following HIBD (Bai 2008).

A potential reason for the failure of developing new therapy for HIE in newborns is that animal studies involving a mature nervous system are extrapolated to the neonatal brain and further translated to clinical trials (Ferriero 2002). In this assumption, the therapeutic time window is limited to a few hours (such as six hours) after the event. In fact, it is clear that neuropathological changes evolve over weeks for the developing nervous system. There is a wide therapeutic window of opportunity in the developing nervous system (Ferriero 2002). Delayed administration (96 hours after the insult) of HBO treatment still reduces HIBD in neonatal rats. With increasing courses of HBO treatment, the inhibition of apoptosis and neuronal protection gradually increased (Wang 2009).

HBO treatment plays a role in regulating genes and the protein expression of neurons that might be neuroprotective. The genes and proteins regulated by HBO include factors associated with stress responses, transport, neurotransmission, signal transduction, and transcription factors (Chen 2009b). In addition, HBO therapy leads to activation of ion channels (Mrsic‐Pelcic 2004), up‐regulation of SOD (Freiberger 2006) and decreased caspases (Li 2008; Chen 2009a), suppression of p38 mitogen activated protein kinase (Yamashita 2009), and activation of Wnt signaling (Wang 2007), which might affect neuroprotection.

Why it is important to do this review

There is a growing body of research in the use of HBO for hypoxic ischaemic injury. On the one hand, the protective effects of HBO for the treatment of hypoxic ischaemic injury have been demonstrated most extensively in experimental animal models and in some clinical trials. On the other hand, the following potential side effects of HBO have been reported; HBO might lead to the formation of oxygen radical species (Narkowicz 1993) resulting in consumption of antioxidants ( Kot 2003; Bader 2007) and reduction in antioxidant enzyme activity (Benedetti 2004), ultimately causing lipid peroxidation (Barth 2008; Benedetti 2004;Muth 2004) and DNA damage (Muth 2004; Groger 2005; Hauser 2006).

Complications and side effects of HBO treatment include barotrauma to the ear, round window blowout, 'sinus squeeze', visual refractive changes, numb fingers, dental problems, claustrophobia, seizures, and pulmonary oxygen toxicity, although these effects are either very rare or are only temporary (Phillips 2013). HBO treatment seems to be the proverbial 'double‐edged sword' in cerebral hypoxic ischaemic insults. In clinical studies, the therapeutic time window, pressure, timing, dosing intervals, and courses of HBO treatment in HIE vary amongst trials.Therefore, it is important to review the available evidence to evaluate the effectiveness and safety of HBO treatment for neonates with HIE and to determine the optimal therapeutic parameters of HBO treatment.

A systematic review of HBO for neonatal HIE performed in 2006 reported that treatment with hyperbaric oxygen possibly reduces mortality and neurological sequelae in term neonates with HIE (Liu 2006). However, there are several limitations to this paper. First, the inconsistent diagnostic criteria for HIE may lead to heterogeneity among the included studies. Second, the included trials may be of poor quality since the design and methods of the trials are not clear; this needs to be clarified by phone or mail. Third, for the included 20 trials the severity of HIE, exposure time to HBO treatment, and other baseline characteristics were not consistent. Finally, no trials of negative results were found, which means publication bias is possible or the searching strategy should be improved. Therefore, it is necessary to perform this systematic review, searching for all the randomized controlled trials of HBO therapy for neonates with HIE, to summarize currently available evidence. Additional randomized controlled clinical trials have now been completed.

Objectives

The objective of this review was to determine the effects of therapeutic HBO (with or without hypothermia) on death and long‐term neurodevelopmental disabilities in term newborn infants with HIE. Furthermore, the short‐term medical effects and side effects of HBO were studied. Primary comparisons:

1. HBO versus supportive care;

2. HBO versus hypothermia therapy;

3. HBO plus hypothermia versus hypothermia alone;

4. HBO plus erythropoietin versus erythropoietin alone.

Subgroup analyses were planned on the basis of:

1. severity of HIE (mild, moderate, severe) (Sarnat 1976; Finer 1981);

2. the therapeutic time window, by timing of commencement of the intervention (< 6 hours versus 6 hours to 7 days versus >7 days);

3. the peak pressure of HBO (<2 ATA vs. >2ATA);

4. the number of courses of HBO (single course versus multiple courses).

Methods

Criteria for considering studies for this review

Types of studies

Randomised controlled trials (RCTs) and quasi‐randomised trials, cluster trials were included. Studies completed but unpublished and studies reported only as abstracts were included. No language restrictions was applied.

Types of participants

Inclusion criteria

1. Newborn infants: term neonates (> 37 weeks) with postnatal age of 28 days or less.

2. Evidence of peripartum asphyxia, with each enrolled infant satisfying at least one of the following criteria:

   1. Apgar score of five or less at five or 10 minutes;

   2. Mechanical ventilation or resuscitation at 10 minutes;

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

3. Evidence of encephalopathy according to Sarnat staging (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 reflex;

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

4. No major congenital abnormalities or syndromes recognizable at birth.

Types of interventions

Included trials investigated all forms of HBO therapy, regardless of initial time, duration, frequency and pressure of treatment. HBO treatment was given alone or in combination with another medical treatment such as hypothermia. HBO intervention was compared with no HBO treatment at normothermia or hypothermia.

Types of outcome measures

Primary outcomes

1. Infant mortality (death at 28 days and at 12 months).

2. Long‐term (> 18 months) major neurodevelopmental disabilities among all participants or survivors (cerebral palsy, developmental delay (Bayley or Griffith assessment more than 2 SD below the mean), or intellectual impairment (IQ more than 2 SD below mean), blindness (vision < 6/60 in both eyes), sensorineural deafness requiring amplification).

Secondary outcomes

1. Potential adverse effects of HBO during treatment period and after the treatment period:

   1. Traumatic injury to the ears, nasal sinuses, or lungs from the compression or expansion of gas pressure (barotrauma);

   2. Seizures (post‐treatment);

   3. Retinopathy (retrolental fibroplasia): acute retrolental fibroplasia (any stage of retrolental fibroplasia during the weeks after birth, observed by direct or indirect ophthalmoscopic examination), and severe retrolental fibroplasia (Stage 3 or greater);

   4. Bronchopulmonary dysplasia: oxygen dependency (oxygen > 21% or positive pressure, or both) at 28 days postnatal age with or without compatible clinical and radiographic changes.

2. Short‐term medical effects of HBO (early indicators of a neurodevelopmental outcome after HBO therapy):

   1. Thompson neurological scores (Thompson 1997) or Sarnat scores (Sarnat 1976; Finer 1981);

   2. Neonatal Behavioral Neurological Assessment (NBNA) score (Bao 1993);

   3. Severity of electroencephalogram (EEG) abnormality:

      1. Severe: isoelectric or burst‐suppression pattern,

      2. Moderate: low voltage or discontinuous background,

      3. Mild: electrographic seizures, dysmaturity;

   4. Incidence and severity of seizures (and number of anticonvulsants);

   5. Basal ganglia, thalami, posterior limb of internal capsule or white matter injury, parasagittal neuronal necrosis on magnetic resonance imaging (MRI) (at day 7);

   6. Days to full sucking feeds.

Search methods for identification of studies

We used the criteria and standard methods of Cochrane and Cochrane Neonatal (see the Cochrane Neonatal search strategy for specialized register).

Electronic searches

We used the standard search strategy of Cochrane Neonatal. We did not apply any language restrictions. We used the following search terms: (Asphyxia OR Hypoxic Ischemic Encephalopathy OR hypoxic isch* OR cerebral hypoxic isch* OR hypoxia brain OR encephalopathy OR cerebral ischemia OR cerebral hypoxia OR perinatal asphyxia OR perinatal hypoxic ischemic encephalopathy OR perinatal hypoxic ischemia encephalopathy) AND (Hyperbaric Oxygenation OR Hyperbaric oxygen OR HBO OR HBOT), plus limiters for RCTs and neonates (see Appendix 1 for the full search strategies which we amended for specific databases). We searched the following databases:

1. Cochrane Central Register of Controlled Trials (CENTRAL 2017, Issue 1) in The Cochrane Library; MEDLINE via PubMed (1966 to 3 February 2017); EMBASE (1980 to 3 February 2017); and CINAHL (1982 to 3 February 2017).

2. ISI Web of Science (1969 to 2016).

3. DORCTHIM (Database of Randomised Controlled Trials in Hyperbaric Medicine) at www.hboevidence.com (from inception to 2016). DORCTHIM was compiled from an unfocused search of PubMed using "hyperbaric oxygenation" as a MESH term, along with handsearching of primarily hyperbaric journals since first publication and checking references in identified RCTs.

4. Five major Mainland Chinese academic literature databases using keywords in Chinese: CNKI (China National Knowledge Infrastructure) (1979 to 2010), VIP (Wei Pu Information) (1989 to 2017), Wang Fang Data (1980 to 2017), CMCI (Chinese Medical Citation Index) (1994 to 2017), CBM (Chinese Biologic Medical database) (1978 to 2017).

5. The China Hyperbaric Oxygen Medicine Information Center (1980 to 2017).

6. Clinical trials registries for ongoing or recently completed trials (clinicaltrials.gov; the World Health Organization’s International Trials Registry and Platform www.whoint/ictrp/search/en/, and the ISRCTN Registry).

Searching other resources

We handsearched selected journals and conference proceedings and contacted known experts in the field to identify additional published or unpublished trials. We also searched the reference lists of any articles selected for inclusion in this review in order to identify additional relevant articles

Data collection and analysis

Selection of studies

We assessed all published articles identified as potentially relevant by the literature search. Tao Xiong and Jing Zhao read abstracts retrieved from the search independently to identify all trials that met the inclusion criteria. If needed, full text articles were to be retrieved and reviewed. We planned to resolve differences in opinion by a third review author (Dezhi Mu) and discussion among the review authors. If the details of the primary trials were not clear, the trial authors were to be contacted for clarification.

Data extraction and management

We designed a form to extract data. Two review authors (Tao Xiong, Hongju Chen) independently extracted, assessed, and coded all available data for each study using a specially designed data extraction form. If it was necessary, additional information and clarification of published data was requested from the authors of individual trials. Review Manager software (RevMan 5.0) was used to enter all the data, by Tao Xiong; Jing Zhao checked it.

Assessment of risk of bias in included studies

Two review authors (initials) independently assessed the risk of bias (low, high, or unclear) of all included trials using the Cochrane ‘Risk of bias’ tool (Higgins 2011) for the following domains:

• Sequence generation (selection bias)

• Allocation concealment (selection bias)

• Blinding of participants and personnel (performance bias)

• Blinding of outcome assessment (detection bias)

• Incomplete outcome data (attrition bias)

• Selective reporting (reporting bias)

• Any other bias

Any disagreements were resolved by discussion or by a third assessor. See Appendix 2 for a more detailed description of risk of bias for each domain

Measures of treatment effect

We calculated relative risk ratio (RR) with its 95% confidence interval (CI) for dichotomous data. We calculated the risk difference (RD) with 95% CI and the number needed to treat (NNT). We expressed continuous data as mean difference (MD) and the 95% confidence interval (CI).

Unit of analysis issues

As detailed in Section 16.3 of the Cochrane Handbook for Systematic Reviews of Interventions, particular biases in cluster randomised trials include: (i) recruitment bias; (ii) baseline imbalance; (iii) loss of clusters; (iv) incorrect analysis; and (v) comparability with individually randomized trials. One way to avoid unit‐of‐analysis errors in cluster‐randomised trials is to conduct the analysis at the same level as the allocation, using a summary measurement from each cluster. In the case of a neonatal intensive care unit (NICU)‐cluster trial, we looked for evidence of adjustments at the level of each neonate as well as each allocated neonatal unit. Alternatively, we might use statistical methods based on a 'multilevel model', a 'variance components analysis', or 'generalized estimating equations (GEEs)' to determine whether the method used was appropriate. Effect estimates and their standard errors from correct analyses of cluster‐randomised trials might be meta‐analysed using the generic inverse‐variance method in RevMan (Higgins 2011).

Crossover trials were excluded.

Dealing with missing data

We obtained data from the primary investigator, as feasible, for unpublished trials or when published data were incomplete. If this approach was unsuccessful, analyses were restricted to available data. Evaluation of important numerical data such as screened, eligible, and randomized patients as well as intention‐to‐treat (ITT) and per‐protocol (PP) populations were carefully performed. Dropouts, missing at follow‐up, and withdrawn study participants were investigated. Issues of last observation‐carried‐forward (LOCF), ITT and PP were critically appraised and compared to the specification of primary outcome parameters and the power calculation. We performed sensitivity analyses to assess how the overall results were affected with and without the inclusion of studies with significant dropout rates. For a particular outcome, if less than 70% of patients allocated to the treatments were reported on at the end of the trial, the data were not used as they were considered to be too prone to bias.

Assessment of heterogeneity

We used a fixed‐effect model if there was no evidence of significant heterogeneity between studies. The Chi² test (if P ≤0.10, substantial or considerable heterogeneity was present) was employed to determine whether there was statistically significant heterogeneity. The degree of statistical heterogeneity was assessed by examining the I² statistic (if I² ≥ 50%, substantial or considerable heterogeneity was present). Trials were explored to investigate possible explanations for heterogeneity. If heterogeneity was identified among a group of studies, we checked the data and again explore the reasons for heterogeneity. When there was heterogeneity that cannot readily be explained, we might divide the studies into subgroups if they had an appropriate basis.

Assessment of reporting biases

Overcoming, detecting, and correcting for publication bias was still problematic. Publication bias was tested using funnel plots, or other corrective analytical methods, depending on the number of clinical trials included in the systematic review. The funnel plot should be seen as a generic means of displaying small‐study effects. Asymmetry could be due to publication bias or to a relationship between trial size and effect size. Therefore, true heterogeneity in intervention effects is just one cause of funnel plot asymmetry; we will not pay too much attention to its asymmetry (Egger 1997;Higgins 2011).

Data synthesis

If more than one eligible trial was identified and there was sufficient homogeneity among the studies with respect to participants and reported outcomes, statistical analyses were performed using the standard methods of the Neonatal Review Group, using the RevMan software with the fixed‐effect model for meta‐analysis. Categorical data were presented as relative risk (RR) with 95% confidence interval (CI). The weighted mean difference (WMD) with 95% CI was used for outcomes measured on a continuous scale. The number needed to treat (NNT) was presented, as appropriate.

Quality of evidence

We used the GRADE approach, as outlined in the GRADE Handbook (Schünemann 2013), to assess the quality of evidence for the following (clinically relevant) outcomes: 1) Death (during the neonatal period and infancy) or major disability, 2) Mortality (during the neonatal period and infancy), and 3) Major disability.

Two authors independently assessed the quality of the evidence for each of the outcomes above. We considered evidence from randomized controlled trials as high quality but downgraded the evidence one level for serious (or two levels for very serious) limitations based upon the following: design (risk of bias), consistency across studies, directness of the evidence, precision of estimates and presence of publication bias. We used the GRADEpro GDT Guideline Development Tool to create a ‘Summary of findings’ table to report the quality of the evidence.

The GRADE approach results in an assessment of the quality of a body of evidence in one of four grades:

1. High: We are very confident that the true effect lies close to that of the estimate of the effect.

2. Moderate: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.

3. Low: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect.

4. Very low: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Subgroup analysis and investigation of heterogeneity

We performed subgroup analysis on the basis of the following issues:

1. Severity of HIE (mild, moderate, severe) (Sarnat 1976; Finer 1981);

2. The therapeutic time window, timing of commencement of intervention (< 6 hours versus 6 hours to 7 days versus >7 days);

3. The peak pressure of HBO (< 2 ATA versus > 2ATA);

4. The number of courses of HBO (single course versus multiple courses).

Sensitivity analysis

Sensitivity analyses were performed based on missing data and study quality. In the case of missing data, we employed sensitivity analyses using different approaches to imputing missing data.

If appropriate, we conducted sensitivity analysis by study quality based on the presence or absence of a reliable random allocation method, concealment of allocation, and blinding of participants or outcome assessors.

Appendices

Appendix 1. Standard search methodology

PubMed: ((infant, newborn[MeSH] OR newborn OR neonate OR neonatal OR premature OR low birth weight OR VLBW OR LBW or infan* or neonat*) AND (randomized controlled trial [pt] OR controlled clinical trial [pt] OR randomized [tiab] OR placebo [tiab] OR drug therapy [sh] OR randomly [tiab] OR trial [tiab] OR groups [tiab]) NOT (animals [mh] NOT humans [mh]))

Embase: (infant, newborn or newborn or neonate or neonatal or premature or very low birth weight or low birth weight or VLBW or LBW or Newborn or infan* or neonat*) AND (human not animal) AND (randomized controlled trial or controlled clinical trial or randomized or placebo or clinical trials as topic or randomly or trial or clinical trial)

CINAHL: (infant, newborn OR newborn OR neonate OR neonatal OR premature OR low birth weight OR VLBW OR LBW or Newborn or infan* or neonat*) AND (randomized controlled trial OR controlled clinical trial OR randomized OR placebo OR clinical trials as topic OR randomly OR trial OR PT clinical trial)

Cochrane Library: (infant or newborn or neonate or neonatal or premature or preterm or very low birth weight or low birth weight or VLBW or LBW)

Appendix 2. Risk of bias tool

We used the standard methods of Cochrane and Cochrane Neonatal to assess the methodological quality (to meet the validity criteria) of the trials. For each trial, we sought information regarding the method of randomisation, and the blinding and reporting of all outcomes of all the infants enrolled in the trial. We assessed each criterion as low, high, or unclear risk. Two review authors separately assessed each study. We resolved any disagreement by discussion. We added this information to the table Characteristics of included studies. We evaluated the following issues and entered the findings into the risk of bias table:

1. Sequence generation (checking for possible selection bias). Was the allocation sequence adequately generated?

For each included study, we categorized the method used to generate the allocation sequence as:

a. Low risk (any truly random process e.g. random number table; computer random number generator);

b. High risk (any non‐random process e.g. odd or even date of birth; hospital or clinic record number);

c. Unclear risk.

2. Allocation concealment (checking for possible selection bias). Was allocation adequately concealed?

For each included study, we categorized the method used to conceal the allocation sequence as:

a. Low risk (e.g. telephone or central randomisation; consecutively numbered sealed opaque envelopes);

b. High risk (open random allocation; unsealed or non‐opaque envelopes, alternation; date of birth);

c. Unclear risk

3. Blinding of participants and personnel (checking for possible performance bias). Was knowledge of the allocated intervention adequately prevented during the study?

For each included study, we categorized the methods used to blind study participants and personnel from knowledge of which intervention a participant received. Blinding was assessed separately for different outcomes or class of outcomes. We categorized the methods as:

a. Low risk, high risk or unclear risk for participants;

b. Low risk, high risk or unclear risk for personnel;

4. Blinding of outcome assessment (checking for possible detection bias). Was knowledge of the allocated intervention adequately prevented at the time of outcome assessment?

For each included study, we categorized the methods used to blind outcome assessment. Blinding was assessed separately for different outcomes or class of outcomes. We categorized the methods as:

a. Low risk for outcome assessors.

b. High risk for outcome assessors.

c. Unclear risk for outcome assessors.

5. Incomplete outcome data (checking for possible attrition bias through withdrawals, dropouts, protocol deviations). Were incomplete outcome data adequately addressed?

For each included study and for each outcome, we described the completeness of data including attrition and exclusions from the analysis. We noted whether attrition and exclusions were reported, the numbers included in the analysis at each stage (compared with the total randomised participants), reasons for attrition or exclusion where reported, and whether missing data were balanced across groups or were related to outcomes. Where sufficient information was reported or supplied by the trial authors, we re‐included missing data in the analyses. We categorized the methods as:

a. Low risk (< 20% missing data);

b. High risk (≥ 20% missing data);

c. Unclear risk.

6. Selective reporting bias. Are reports of the study free of suggestion of selective outcome reporting?

For each included study, we described how we investigated the possibility of selective outcome reporting bias and what we found. We assessed the methods as:

a. Low risk (where it is clear that all of the study's pre‐specified outcomes and all expected outcomes of interest to the review have been reported);

b. High risk (where not all the study's pre‐specified outcomes have been reported; one or more reported primary outcomes were not pre‐specified outcomes of interest and are reported incompletely and so cannot be used; study fails to include results of a key outcome that would have been expected to have been reported);

c. Unclear risk.

7. Other sources of bias. Was the study apparently free of other problems that could put it at a high risk of bias?

For each included study, we described any important concerns we had about other possible sources of bias (for example, whether there was a potential source of bias related to the specific study design or whether the trial was stopped early due to some data‐dependent process). We assessed whether each study was free of other problems that could put it at risk of bias as:

a. Low risk;

b. High risk;

c. Unclear risk

If needed, we explored the impact of the level of bias through undertaking sensitivity analyses.

What's new

Date	Event	Description
8 March 2019	Amended	The Cochrane Neonatal Editorial team has decided to withdraw this protocol, as it has not made adequate progress towards a completed review.

Contributions of authors

Tao Xiong is responsible for all aspects of the review, contributes to the design, development and revise.

Hongju Chen is responsible for data collection, screening search results, and organizing retrieval of papers.

Jing Zhao is responsible for search.

Wenbin Dong and Yi Qu provide support for the part of background.

Honghao Li is responsible for screening retrieved papers against inclusion criteria and appraising quality of papers.

Taixiang Wu works as a co‐supervisor at both protocol and review stages.

Dezhi Mu is responsible for revise of the protocol and supervises whole review progresses.

Sources of support

Internal sources

• Department of Pediatrics, West China Second University Hospital, Sichuan University, Chengdu, China.

• Department of Neonatology, Affiliated Hospital of Luzhou Medical College,Luzhou, China.

External sources

• Chinese Cochrane Center, West China Hospital of Sichuan University, China.

• Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Department of Health and Human Services, USA.

  Editorial support of the Cochrane Neonatal Review Group has been funded with Federal funds from the Eunice Kennedy Shriver National Institute of Child Health and Human Development National Institutes of Health, Department of Health and Human Services, USA, under Contract No. HHSN275201600005C

Declarations of interest

None

Notes

The Cochrane Neonatal Editorial team has decided to withdraw this protocol, as it has not made adequate progress towards a completed review.

Withdrawn from publication for reasons stated in the review