Late versus early surgical correction for congenital diaphragmatic hernia in newborn infants

Abstract

Background

Congenital diaphragmatic hernia, although rare (1 per 2‐4,000 births), is associated with high mortality and cost. Opinion regarding the timing of surgical repair has gradually shifted from emergent repair to a policy of stabilization using a variety of ventilatory strategies prior to operation. Whether delayed surgery is beneficial remains controversial.

Objectives

To summarize the available data regarding whether surgical repair in the first 24 hours after birth rather than later than 24 hours of age improves survival to hospital discharge in infants with congenital diaphragmatic hernia who are symptomatic at or immediately after birth.

Search methods

Search of MEDLINE (1966 to Sept 2003), EMBASE (1978 to Oct 2003) and the Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library, Issue 3, 2003); citations search, and contact with experts in the field to locate other published and unpublished studies.

This search was updated in 2009.

Selection criteria

Studies were eligible for inclusion if they were randomized or quasi‐randomized trials that addressed infants with CDH who were symptomatic at or shortly after birth, comparing early (< 24 hours) vs late (> 24 hours) surgical intervention, and evaluated mortality as the primary outcome.

Data collection and analysis

Data were collected regarding study methods and outcomes including mortality, need for ECMO and duration of ventilation, both from the study reports and from personal communication with investigators. Analysis was performed in accordance with the standards of the Cochrane Neonatal Review Group.

Main results

Two trials met the pre‐specified inclusion criteria for this review. Both were small trials (total n<90) and neither showed any significant difference between groups in mortality. Meta‐analysis was not performed because of significant clinical heterogeneity between the trials.

Authors' conclusions

There is no clear evidence which favors delayed (when stabilized) as compared with immediate (within 24 hours of birth) timing of surgical repair of congenital diaphragmatic hernia, but a substantial advantage to either one cannot be ruled out. A large, multicenter randomized trial would be needed to answer this question.

Plain language summary

Late versus early surgical correction for congenital diaphragmatic hernia in newborn infants

No clear evidence about when to perform surgery to correct congenital diaphragmatic hernia. Congenital diaphragmatic hernia is a rare but often fatal condition. It occurs when a newborn baby's diaphragm has a defect in it that allows abdominal organs (such as the stomach or liver) to enter the chest and displace the lung and heart. Surgery can correct the defect, but damage to the lung may be so severe that the baby still cannot survive. It has been thought that correcting the defect was so urgent that emergency surgery should be performed within the first 24 hours following birth, but more recent thinking suggests that a period of stabilization before surgery could help the lung develop. Only two trials have been done, and these provide no clear evidence to support delayed surgery over emergency surgery.

Background

Description of the condition

Congenital diaphragmatic hernia (CDH) occurs in approximately 1 in every 2,000 to 4,000 newborn infants, and was generally considered a fatal condition through the early 20th century. Despite many advances in the medical and surgical care of infants, the mortality for this condition remains quite high. CDH also ranks among the most costly of correctable conditions, with an estimated cost per new case of US$250,000, and an overall estimated yearly cost of US$264,000,000 in the United States (Metkus 1995).

Description of the intervention

In 1940, Ladd and Gross demonstrated that repair could be successfully undertaken in an infant (Ladd 1940) Their earliest reports were encouraging, demonstrating a marked improvement in survival compared to the previous approach of watchful waiting. They justified urgent surgery by noting that these infants had a large amount of gas in the bowel and thorax, which was thought to cause lung compression and respiratory distress resulting in the need for urgent correction. A policy of early intervention evolved so that by 1970 CDH was considered one of the true surgical emergencies in the newborn (Holder 1979; Guzzetta 1989).

How the intervention might work

Surgery has been shown to be associated with deterioration in lung compliance, which may be ameliorated by preoperative stabilization. Several reports in the 1980s suggested that survival was no different with delayed repair and recommended that a period of stabilization prior to operation was warranted (Cartlidge 1986; Langer 1988; Sakai 1987). Many pediatric surgeons are currently employing a strategy of delay for some period of time prior to correction of the defect in order to achieve a minimal level of ventilator support, to document absence of pulmonary hypertension, to show improvement in pulmonary compliance, or to show that lung radiographic appearance is improved prior to repair (Tracy 1994; Ryan 1995; West 1992; Breaux 1991). Alternative ventilatory strategies such as permissive hypercapnia, nitric oxide administration, and liquid ventilation have been used in an effort to stabilize these infants prior to surgery or as rescue therapy following surgery. Extracorporeal membrane oxygenation (ECMO) has been used as an alternative mode of ventilatory support in infants with severe hypoxemia and respiratory failure in the preoperative, intra‐operative and post‐operative period (Lally 1992; Reickert 1996). Whether there is any interaction between the use of ECMO or other ventilatory strategies and the timing of surgery is not known.

Why it is important to do this review

CDH remains one of the most difficult neonatal conditions to treat and is associated with high mortality and high cost. We plan to summarize the available data from randomized trials regarding whether surgical repair later than 24 hours of age rather than in the first 24 hours after birth improves survival to hospital discharge in infants with congenital diaphragmatic hernia who are symptomatic at or immediately after birth.

Objectives

The objective of this review is to summarize the available data from randomized trials regarding whether surgical repair later than 24 hours of age rather than in the first 24 hours after birth improves survival to hospital discharge in infants with congenital diaphragmatic hernia who are symptomatic at or immediately after birth.

Methods

Criteria for considering studies for this review

Types of studies

Randomized or quasi‐randomized trials, which compared a policy of late versus early surgical correction of CDH. Blinding of the investigators to the groups after allocation is highly unlikely and was not required. A minimum follow up of 80% of patients was required. Studies using assignment by convenience or by investigator preference and studies using historical controls were excluded.

Types of participants

Infants who developed signs and symptoms such as respiratory distress and hypoxemia, which were attributed to CDH, within the first 12 hours of life.

Types of interventions

Allocation to surgical correction of the hernia either late (at > 24 hours of age) or early (at ≤ 24 hours of age), regardless of whether infants actually received the surgical intervention within the time period allocated.

Types of outcome measures

Primary outcomes

The primary outcome measurement was mortality.

Secondary outcomes

Secondary outcome measurements that were sought included need for and duration of ECMO, duration of ventilatory support in survivors, pulmonary function, and developmental outcome.

Search methods for identification of studies

See Collaborative Review Group Search Strategy.

Electronic searches

MEDLINE (1966 to September 2003), EMBASE (1978 to October 2003), the Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library, Issue 3, 2003) were searched using the search terms "congenital diaphragmatic hernia" and "surg*".

In 2009, we updated the search as follows: MEDLINE (search via PubMed), CINAHL, EMBASE and CENTRAL (The Cochrane Library) were searched from 2003 to 2009. Search terms: congenital diaphragmatic hernia and surg*. Limits: human, newborn infant and clinical trial. No language restrictions were applied.

Searching other resources

Reference lists for each identified article were reviewed to locate other potentially relevant articles. A Science Citations Index search was performed on four important early papers (Langer 1988; Cartlidge 1986; Hazebroek 1988; Sakai 1987) to locate other potentially relevant studies. Members of the CDH Study Group (Clark 1998) and other acknowledged experts in CDH management were contacted to locate other trials, unpublished trials and technical reports which were not found in published databases.

Studies published in all languages were included in the review.

Data collection and analysis

The standard method of conducting a systematic review, as described in the Cochrane Collaboration Handbook was used for this review.

Selection of studies

Each study was independently reviewed by two reviewers, one with expertise in the clinical management of CDH, and the other with expertise in study design. Reviewers were not blinded to authorship, journal or results at the time of the review.

Data extraction and management

Two review authors separately extracted, assessed and coded all data for each study using a form that was designed specifically for this review. Any standard error of the mean was replaced by the corresponding standard deviation. For each study, final data was entered into RevMan by one review author and then checked by a second review author. Any disagreements were addressed and resolved by consensus.

Assessment of risk of bias in included studies

Studies which met the inclusion criteria for the review were evaluated for the presence of allocation concealment in order to ameliorate selection bias. Differences in the conditions under which surgery was performed, and in the indications and/or use of diagnostic and/or therapeutic co‐interventions were sought from the trial reports. When information could not be obtained from the reports, it was obtained directly from the investigators when possible. The trials were assessed for differences in withdrawals from the study, and for differences in how the outcome assessments were made between the groups.

For the updated review in 2009, the following information was extracted for the Risk of Bias Table:

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

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

‐ adequate (any truly random process e.g. random number table; computer random number generator); 

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

‐ unclear. 

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

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

‐ adequate (e.g. telephone or central randomization; consecutively numbered sealed opaque envelopes); 

‐ inadequate (open random allocation; unsealed or non‐opaque envelopes, alternation; date of birth); 

‐ unclear. 

(3) Blinding (checking for possible performance bias). Was knowledge of the allocated intervention adequately prevented during the study? At study entry? At the time of outcome assessment?

For each included study, we catagorized 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 classes of outcomes. We catagorized the methods as: 

‐ adequate, inadequate or unclear for participants; 

‐ adequate, inadequate or unclear for personnel; 

‐ adequate, inadequate or unclear for outcome assessors. 

In some situations there may be partial blinding e.g. where outcomes are self‐reported by unblinded participants but they are recorded by blinded personnel without knowledge of group assignment.  Where needed “partial” was added to the list of options for assessing quality of blinding. 

(4) 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 randomized 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 catagorized the methods as: 

‐ adequate (< 20% missing data); 

‐ inadequate (≥ 20% missing data): 

‐ unclear. 

(5) 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: 

‐ adequate (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); 

‐ inadequate (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 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);

‐ unclear. 

(6) 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:

‐ yes; no; or unclear.  

If needed, we planned to explore the impact of the level of bias through undertaking sensitivity analyses.

Measures of treatment effect

Statistical analyses were performed using Review Manager software. Categorical data (such as mortality) were analyzed using relative risk (RR), risk difference (RD) and the number needed to treat (NNT). Continuous data (such as length of hospital stay) were analyzed using weighted mean difference (WMD). The 95% Confidence interval (CI) was reported on all estimates.

Assessment of heterogeneity

If meta‐analysis was performed, we planned to estimate the treatment effects of individual trials and examine heterogeneity between trials by inspecting the forest plots and quantifying the impact of heterogeneity using the I² statistic. If we detected statistical heterogeneity, we planned to explore the possible causes (for example, differences in study quality, participants, intervention regimens, or outcome assessments) using post hoc subgroup analyses.

We did not perform meta‐analysis of the two identified trials due to obvious clinical heterogeneity in study design.

Data synthesis

If meta‐analysis was performed, we planned on using Review Manager software (RevMan 5) supplied by the Cochrane Collaboration. For estimates of typical relative risk and risk difference, we planned to use the Mantel‐Haenszel method. For measured quantities, we planned to use the inverse variance method. All meta‐analyses were to be done using the fixed effect model.

Results

Description of studies

See Table of included studies.

Twenty‐six studies were identified, of which 24 were excluded because they did not use a randomized or quasi‐randomized trial design. Two trials were included in the analysis. Both trials were randomized, but neither report specified how the randomization was accomplished. The end point for both trials was mortality; however, one study (Nio 1994) evaluated survival to hospital discharge (at about 1 to 2 months of age) and the other (de la Hunt 1996) evaluated survival at 6 months.

Nio 1994 studied a total of 32 patients; 14 patients underwent repair within 6 hours of the onset of symptoms (mean age at operation, 10.1 hours) and 18 patients underwent repair at over 96 hours of life. The study was performed in two institutions and the authors provided no information about preoperative or postoperative management, including pharmacotherapy and ventilator management other than the frequency of use of ECMO. The number of patients in the study was too small to detect significant differences.

de la Hunt 1996 also studied a small number of patients, with a total of 54 infants enrolled in the study from two institutions. There were 26 infants randomized to repair within 4 hours of arrival (all but one repaired at less than 11 hours of age; one infant's surgery was delayed to 296 hours pending genetic evaluation) and 28 infants who had repair after 24 hours. As with the Nio 1994 study, there was no standardized preoperative or postoperative management described in the report.

Neither report offers specifics about the postnatal management of these patients with the exception of using ECMO in 75% of the patients in the Nio trial (Nio 1994) and 4% in the de la Hunt report (de la Hunt 1996). In neither study were the criteria for ECMO outlined. Other aspects of care that could significantly alter outcome such as the ventilator strategies and use of adjunctive therapies were not controlled for or specified, although de la Hunt 1996 stated that they were to be the same for both groups.

Risk of bias in included studies

In both included studies, allocation was stated to be random but the randomization scheme was not described in the published report. Personal communication with the authors of the Nio study (Nio 1994) revealed that sealed envelopes were used, with randomization after consent was obtained. Whether allocation was effectively concealed is not clear in either of these reports. Although all patients were accounted for, two patients assigned to early surgical intervention in the Nio study acutely deteriorated and were dropped from the study (Nio 1994). Information on the outcome of those two patients is being requested from the investigators and will be included in a future update of this review. As expected, neither patients (families) nor clinicians were blinded to group assignment. Data are insufficient to determine the balance of prognostic factors in the two groups in either study. More patients received ECMO therapy in the delayed repair group in the Nio study (Nio 1994); other co‐interventions are not well described.

Effects of interventions

Mortality (Outcome 1.1): 
 Mortality was evaluated differently in the two studies, Nio evaluating deaths prior to discharge (which occurred at 1 to 2 months of age for most infants) and de la Hunt evaluating deaths prior to 6 months of age. In the de la Hunt study (de la Hunt 1996), 43% of the late surgical group and 54% of the early surgical group had died by 6 months (RR for death with delayed surgery 0.80, 95%CI 0.46,1.39; RD ‐0.11, 95%CI ‐0.38,0.16). In the Nio study (Nio 1994), death prior to discharge occurred with almost equal frequency with late and early surgical intervention respectively (RR for death with delayed surgery 1.11, 95%CI 0.32,3.81, RD 0.03, 95%CI ‐0.29,0.35) .

Length of stay of survivors (Outcome 1.2): 
 In the de la Hunt study (de la Hunt 1996), the survivors in the late repair group remained hospitalized for a geometric mean of 32.7 days, compared to 20.7 days for those in the early repair group (SDs not provided.) In the Nio study, surviving patients were discharged at a mean of 63.5 (SD 63.5) days in the late repair group and 49.6 (30.4) days in the early repair group (mean difference 13.9, 95%CI ‐25.9, 53.7).

Use of extracorporeal membrane oxygenation: 
 In the Nio study (Nio 1994), ECMO was used in 89% of those in the delayed surgery group and 67% of those in the early surgery group. Criteria for ECMO use for both groups was an Oxygen Index (mean airway pressure x FIO2/PaO2 x100) > 40 for more than 2 hours. In the de la Hunt study, two infants in the late surgery group and none in the early surgery group received ECMO (criteria not stated).

Duration of ventilatory support in survivors: 
 In the de la Hunt study, survivors in the late surgery group required more days of ventilation than did those in the early surgery group (geometric mean of 10.7 vs. 5.8 days), but the difference was not significant. Nio did not report duration of ventilatory support.

Pulmonary function, developmental outcome: 
 Neither pulmonary function nor developmental outcome were evaluated in these two studies.

Evaluation of heterogeneity: 
 While "early" surgery occurred well before 24 hours of age (means of 6 and 10 hours), the timing of "late" surgery was different in the two studies. In the Nio study, the infants in the delayed surgery arm underwent surgery at a mean of 173 hours of age; in the de la Hunt study, late surgery occurred at a mean of 48 hours of age. In addition, the outcome variable was survival to discharge in the Nio study and survival to 6 months in the de la Hunt study. Because of these clinical differences in the study, meta‐analysis with statement of summary estimates was not felt to be appropriate.

Discussion

After the first successful repair of a diaphragmatic hernia in a newborn, the predominant medical opinion expressed was that an emergent operation was required to alleviate compression of the patient's "good" lung. The earliest reports by Ladd and Gross on correction of CDH in the neonate were encouraging, demonstrating a marked improvement in survival compared to the previous approach of watchful waiting (Ladd 1940). This prompted most surgeons to correct the defect as early as possible with the hope that removing the bowel and/or liver from the chest would allow the lungs to expand. The majority of pediatric surgeons treated this as a true emergency and the patients were rushed to the operating room a very short time after diagnosis or arrival at the tertiary hospital. Subsequently, Sakai and colleagues (Sakai 1987), and others, demonstrated that early operation actually worsened pulmonary compliance, which sparked a broader interest in delaying operation until the infant was stable. However, what has been reported as "delayed" operation has not been well defined. One concept is to stabilize for a short period of time and therefore avoid an emergency operation. This approach would have the patient undergo operation within a day or two after birth. An alternative approach would be to delay surgical repair until physiological changes occur, such as improved pulmonary compliance, decreased pulmonary hypertension, or others. With this approach, operation would be delayed for days or weeks as necessary.

The two studies that met criteria for this review actually address two different questions. The Nio study evaluated the potential advantage of a lengthy delay before surgical repair, allowing a physiologic change in the patient's condition and indeed, the average age at operation in the "late" group in that study was 173 hours (Nio 1994). In comparison, the "delayed" patients in the study by de la Hunt and colleagues underwent repair at an average age of 48 hours (de la Hunt 1996). The de la Hunt trial primarily focused on a brief period of stabilization compared to emergent operation while the Nio trial allowed for a lengthy period of physiologic stability before operation. Another important difference in the two studies was the availability or actual use of ECMO, with only 4% of the patients in the de la Hunt trial receiving ECMO compared with 75% of patients in the Nio trial.

Current practice has evolved towards avoiding emergent operation and allowing some period of stabilization prior to repair. Data published from a registry of over 1,000 patients with CDH show that most pediatric surgeons are currently employing a strategy of delaying correction of the defect for some period of time, with 62% of patients undergoing operation beyond 24 hours of age (Clark 1998). It remains unclear, however, what constitutes an appropriate period of stabilization. Various authors have recommended that patients achieve a minimal level of ventilator support, that pulmonary hypertension be absent, that there be improvement in pulmonary compliance, or the lung radiographic appearance be improved prior to repair (Nakayama 1991; Tracy 1994).

The widespread availability of ECMO during the time when delayed repair became more popular led to the use of this intervention to stabilize critically ill infants with CDH. Once the infant is on extracorporeal support, the timing of operative intervention remains controversial. Some initial reports demonstrated a high complication rate associated with repair on ECMO, with bleeding being a significant issue (Lally 1992; Wilson 1992). These studies were not controlled trials. Subsequently, refinements in operative technique such as the use of fibrin glue or aminocaproic acid, diaphragmatic patching, and abdominal wall patching have been associated with lower complication rates. The ideal timing for repair of CDH on ECMO remains unclear.

Authors' conclusions

Implications for practice.

No clear evidence favors emergent (immediate) vs. urgent (within 24 hours) vs. delayed (after stabilization) repair of congenital diaphragmatic hernia.

Implications for research.

A large, controlled, randomized trial of this question would be useful. Currently, there is not widespread agreement on a standardized postnatal management strategy. It would be desirable to have some level of agreement on postnatal management prior to a trial. The relative rarity of CDH and the small number of patients seen in each institution will make this type of trial difficult. Given this situation, it will be necessary to select specific issues, such as permissive hypercapnia or "gentle" ventilation as well as indication for ECMO, for future study.

What's new

Date	Event	Description
3 February 2010	New search has been performed	This review updates the existing review "Late versus early surgical correction for congenital diaphragmatic hernia in newborn infants" published in the Cochrane Database of Systematic Reviews (Moyer 2002). Updated search found no new trials. No changes to conclusions.

History

Protocol first published: Issue 3, 1999
 Review first published: Issue 3, 2000

Date	Event	Description
28 October 2008	Amended	Converted to new review format.
30 October 2003	New search has been performed	This is an update of the existing review of "Late versus early surgical correction for congenital diaphragmatic hernia in newborn infants", published in The Cochrane Library, Issue 3, 2002 (Moyer 2002). No new trials were located in the search dated September 2003, and as a result, no substantive changes were made in the review. There is no change to the conclusion that no clear evidence favors delayed versus immediate timing for surgical repair of congenital diaphragmatic hernia.

Data and analyses

Comparison 1. Late vs early surgery.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Mortality	2		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
2 Length of stay in survivors	2		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

1.1. Analysis.

Comparison 1 Late vs early surgery, Outcome 1 Mortality.

1.2. Analysis.

Comparison 1 Late vs early surgery, Outcome 2 Length of stay in survivors.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

de la Hunt 1996.

Methods	Open, 2 center randomized trial. Masking of allocation: No Masking of intervention: No Complete follow‐up: Yes
Participants	54 newborns with congenital diaphragmatic hernia. 26 infants were randomized to early surgery, 28 to delayed surgery. No patients were excluded. Severity of symptoms was not well described.
Interventions	Infants either underwent surgical repair within the first 4 hours of admission or after 24 hours when they were considered stable.
Outcomes	Primary outcome: Survival to 6 months. Secondary outcomes: duration of hospitalization.
Notes	Average age at repair in the delayed group was 48 hours; 1 patient in early group was delayed for 11 days for genetic evaluation. 2 infants (4%) received ECMO.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	High risk	Masking of allocation: No
Blinding? All outcomes	High risk	Masking of intervention: No
Incomplete outcome data addressed? All outcomes	Low risk	Complete follow‐up: Yes

Nio 1994.

Methods	Open, 2 center randomized trial. Masking of allocation: No Masking of intervention: No Complete follow‐up: Yes
Participants	32 newborns with congenital diaphragmatic hernia. Infants less than 34 weeks, 2 kilograms or older than 12 hours were excluded. 14 infants were randomized to immediate repair, 18 to delayed repair. All patients developed respiratory distress within 3 hours of birth.
Interventions	Infants either underwent surgical repair within the first 6 hours of admission or after 96 hours when there was no evidence of pulmonary hypertension
Outcomes	Primary outcome: Survival to hospital discharge. Secondary outcomes: Use of ECMO, procedure used, duration of hospitalization for survivors, and complications.
Notes	2 of the initial 14 patients in the early group were placed on ECMO before repair, and were excluded. Outcome data being requested from investigators. 75% of included patients received ECMO.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	High risk	Masking of allocation: No
Blinding? All outcomes	High risk	Masking of intervention: No
Incomplete outcome data addressed? All outcomes	Low risk	Complete follow‐up: Yes

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Adolph 1995	Observational, non‐randomized, retrospective comparison of surgical strategies.
Al‐Hathal 1998	Observational retrospective review of consecutive patients, no control group
Boloker 2002	Observational review of consecutive patients compared to published survival rates in other institutions.
Breaux 1991	Observational, non‐randomized study using historical controls.
Cartlidge 1986	Observational, non‐randomized study using historical controls.
Charlton 1991	Observational, non‐randomized study.
Chu 2000	Observational study, very few infants with early surgery.
Davenport 1992	Observational study of survivors of delayed repair.
Desfrere 2000	Observational study of combined intervention of change in ventilatory management and change in surgical timing.
Goh 1992	Observational, non‐randomized comparison of surgical strategies.
Hazebroek 1988	Observational study ‐ case series.
Hodson 2000	Observational study, very small number of infants with delayed surgery.
Kamata 1998	Observational study using historical controls at three centers.
Lally 1992	Observational, non‐randomized, retrospective study.
Langer 1988	Observational, non‐randomized, study using historical controls.
Nakayama 1991	Observational non‐randomized study comparing surgical strategies.
Okuyama 2002	Observational non‐randomized retrospective study using historical controls.
Reickert 1996	Observational, non‐randomized, retrospective study using historical controls.
Reyes 1998	Observational study, case series.
Sakai 1987	Observational study describing lung mechanics.
Sigalet 1995	Observational, non‐randomized, retrospective study.
Stranak 1999	Observational comparison of survivors and non‐survivors in a case series.
Wilson 1992	Observational, non‐randomized study.
Wung 1995	Observational, non‐randomized, retrospective study.

Contributions of authors

Virginia A Moyer (VAM), Fernando R Moya, Dick Tibboel, Paul D Losty, Masahiro Nagaya and Kevin P Lally wrote the original review and updated the review in 2002.

The recent update (December 2009) was conducted centrally by the Cochrane Neonatal Review Group staff (Yolanda Montagne, Diane Haughton, and Roger Soll).  This update was reviewed and approved by VAM.

Declarations of interest

None

New search for studies and content updated (no change to conclusions)