Fetal surgery for congenital diaphragmatic hernia.

Abstract

The article of Harrison et al1,2 is a major contribution to the literature despite its disappointing findings. It represents the gold standard in methodology in assessment of new health care procedures and technology.Fetal tracheal occlusion initially with clips and subsequently with intraluminal occluders was applied to stimulate lung development with a high level of technical success in 11 of 24 fetuses with left-sided congenital diaphragmatic hernia (CDH) in a randomized fashion. The fetoscopic surgery represents a technical tour de force and the skills learned in this program will surely benefit many other areas of pediatric surgery. Unfortunately, the authors were unable to solve the problem of tocolysis, with the entire tracheal occlusion group suffering preterm labor. It is interesting to speculate whether the outcome would be better in the tracheal occlusion group if the tocolysis had been more effective.The authors evaluated tracheal occlusion in a randomized, controlled trial (RCT). The study population was designed very sensibly to be a high-risk cohort of infants with liver herniation into the chest (liver-up) and pulmonary hypoplasia (low head-lung ratio). The exclusion of right-sided CDH made the study population homogenous, eliminating the possible confounding effect of large right-sided hernias, which are more difficult to manage. Despite the selection of a cohort of infants expected to have a poor outcome, the survival in the intervention and control groups were not significantly different, with 73% and 77% survival at 90 days, respectively. Only 1 patient in the control group required extracorporeal membrane oxygenation (ECMO).The trial was unable to demonstrate improved outcome because of the greater-than-expected survival in the control group. CDH management in the modern era demonstrated by Stolar and coworkers,3 Wilson and coworkers,4 Kays and coworkers,5 and Reickert et al6 consists initially of gentle ventilation with permissive hypercapnia and hypoxia. If homeostasis cannot be maintained with gentle ventilation, ECMO is used to support gas exchange until the pulmonary hypertension has resolved. CDH repair can either be deferred until the infant is off ECMO or performed on ECMO by using adapted surgical technique and antithrombolytic agents (ie, Amicar, Aprotinin) (see ref. 7 and ref. 8 and references therein). These measures have been effective in improving the outcome in infants with CDH, compared with the time period used to calculate the expected survival in the control group. Once again, this demonstrates the limitations in using historical controls to demonstrate improved outcome. It is a tribute to the authors’ scientific principals that they did not stop evaluating the technique once they had developed its feasibility but continued the evaluation with this well-designed RCT.The infants who underwent tracheal occlusion all seemed to have adequately patent tracheas during the period reported in the study. The possibility of tracheal stenosis at the site of instrumentation should definitely be considered in following these children in the future. It is likely that tracheal stenosis would be seen if this technology had become widespread and disseminated beyond one highly specialized center. Subsequent management of tracheal stenosis could be difficult.9Does fetal surgery have a future? Obviously this trial means that we will not yet be seeing fetal surgery for CDH. We await the findings of the neural tube defect trial with great interest. Even if this study fails to show an improved outcome, the field of fetal surgery remains an important area of study, because there are several other conditions with a poor prognosis in neonatal life in which the concept of a fetal cure is attractive, such as severe Shones’ complex and Ebsteins’ anomaly. However, fetal open-heart surgery remains some way off in the future.10,11Any additional evaluation of fetal surgery should follow the principal of evaluation demonstrated by Harrison et al1 to truly determine whether a treatment effect is present. In addition, any future study of fetal surgery should include long-term evaluation of quality-of-life data for the study patients, comparing treated and control patients well into adulthood. We should also reconsider whether conventional RCT design is the most appropriate in fields such as this, where the volume is small, the rate of change is high, and the case mix is heterogenous. The model of “tracker” studies proposed by Lilford’s group may well be a more appropriate tool than a conventional RCT in these areas (see ref. 12 and references therein).The immediate aim for researchers in the field of fetal surgery must be to solve the problem of preterm labor. Better tocolysis would seem to be the most important objective in making fetal surgery a viable proposition in the future.