Article Text
PDF
Abstract
Purpose This study aims to describe short-term outcomes of live-born infants with congenital diaphragmatic hernia (CDH) and to identify prognostic factors associated with early mortality.
Design A prospective population cohort study was undertaken between April 2009 and September 2010, collecting data on live-born infants with CDH from all 28 paediatric surgical centres in the UK and Ireland using an established surgical surveillance system. Management and outcomes are described. Prognostic factors associated with death before surgery are explored.
Results Two hundred and nineteen live-born infants with CDH were reported within the data collection period. There were 1.5 times more boys than girls (n=133, 61%). Thirty-five infants (16%) died without an operation. This adverse outcome was associated with female sex (adjusted OR (aOR) 3.96, 95% CI 1.66 to 9.47), prenatal diagnosis (aOR 4.99, 95% CI 1.31 to 18.98), and the need for physiological support in the form of inotropes (aOR 9.96, 95% CI 1.19 to 83.25) or pulmonary vasodilators (aOR 4.09, 95% CI 1.53 to 10.93). Significant variation in practice existed among centres, and some therapies potentially detrimental to infant outcomes were used, including pulmonary surfactant in 45 antenatally diagnosed infants (34%). Utilisation of extracorporeal membrane oxygenation was very low compared with published international studies (n=9/219, 4%). Postoperative 30-day survival was 98% for 182 infants with CDH who were adequately physiologically stabilised and underwent surgery.
Conclusion This is the first British Isles population-based study reporting outcome metrics for infants born with CDH. 16% of babies did not survive to undergo surgery. Factors associated with poor outcome included female sex and prenatal diagnosis. Early postoperative survival in those who underwent surgical repair was excellent.
- neonatology
- outcomes research
- paediatric surgery
- congenital diaphragmatic hernia
- baps-cass
Statistics from Altmetric.com
What is already known on this topic?
Congenital diaphragmatic hernia (CDH) is a rare disease with an incidence of 1 in 3000 births.
Newborns with prenatally diagnosed CDH have more severe birth defects and worse survival outcomes.
What this study adds?
Evidence of clinical variation in prenatal and postnatal management of CDH across centres in the UK and Ireland, including deviation from ‘best practice’ strategies.
Evidence that live-born female infants with CDH are more likely to die before surgery than male infants.
Introduction
‘Congenital diaphragmatic hernia’ (CDH) describes a spectrum of abnormalities ranging from small muscular defects to major aberrations in thoracic development affecting the lung and pulmonary vasculature. Recent decades have seen major advances in fetal medical and neonatal care, many of which have had an impact on outcomes for these infants.1 Researchers and clinicians continue to strive to improve outcomes in infants born with CDH as survival of live-born infants has remained at 75%–85% for the last two decades.2–4
A strong evidence base for most of the management practices used in CDH is conspicuously lacking,5 6 although where evidence of benefit or harm exists it has been used (in part) to inform European management guidelines.7 We describe management practices at the time of data collection for infants born in the UK and Ireland. Variation in outcomes among centres and departure from evidence-based therapy(s) are reported. Prognostic factors linked to death before surgery are also explored.
Methods
Case identification
A prospective population cohort study was undertaken, collecting data from all 28 specialist paediatric surgical centres in the UK and Ireland over an 18-month period from April 2009 to September 2010. The surveillance methodology, using the British Association of Paediatric Surgeons Congenital Anomaly Surveillance System (BAPS-CASS), has been previously described.8 In brief, designated paediatric surgeons at UK and Ireland surgical centre(s) are sent a monthly reporting card throughout the duration of the study, and they then return the card with the number of cases seen in the preceding month or respond that they have not seen any cases.
Case definition
Infants eligible for study inclusion were those who were presented to specialist paediatric surgical centres within the data collection period meeting the following case definition: any live-born infant with CDH, defined as a developmental defect of the diaphragm present at birth allowing herniation of abdominal contents into the chest.
Data collection
After cases were reported to BAPS-CASS, data were collected from centres using a standardised data collection form and then double-entered into a customised database. Duplicate cases were identified and removed. Missing data were sought from reporting centres and clinicians by email/telephone contact.
Statistical analyses
Statistical analyses were undertaken using Stata V.13. Median and IQR values are presented for continuous data and differences analysed using Mann-Whitney tests. χ2 or Fisher’s exact tests were used to assess differences in proportions between groups. Univariable and multivariable logistic regression analyses were performed to identify factors associated with death before surgery. Potential risk factors were determined a priori from the literature. These included being small for gestational age, defined as being less than the 10th percentile for birth weight for gestational age using the British normative data from the LMS growth study.9
Risk factors with a significant OR (P<0.10) were included in a multivariable model in a stepwise forward method in order of their statistical significance. Covariates that improved the fit of the model on likelihood ratio testing were retained. Factors with a P value <0.05 within the final model were considered statistically significantly associated with death before surgery.
Results
Two hundred and nineteen live-born infants with CDH were identified. During the 18-month surveillance period, there were 1.3 million live births in the UK and Ireland, giving an estimated live-birth prevalence of 1.7 per 10 000 (95% CI 1.5 to 1.9), or 1 in 5880. The number of infants reported by individual centres ranged from 1 to 18 (IQR 4–12). Sixty-one per cent of infants were male. Demographic information and characteristics of those presenting antenatally and postnatally are shown in table 1.
Demographics and antenatal management of infants with CDH in the cohort
Antenatal findings and management
One hundred and thirty-four infants had CDH detected antenatally (61%). There was no statistical difference between the proportion of male (48/133, 63%) and female (50/86, 58%) whose CDH was detected antenatally (P=0.467). Where diagnosis of CDH was established antenatally, 78% of parents received prenatal counselling involving paediatric surgeons.
Postnatal management
Table 2 summarises postnatal management of the study cohort. Thirty-three antenatally diagnosed newborns (25%) were transferred to another hospital after delivery compared with 64 babies diagnosed postnatally (76%).
Postnatal medical management in infants born with CDH
Fifty-three (25%) of all infants were administered surfactant after delivery, 29 of these were born at term (17% of infants born at term). Forty-five antenatally diagnosed infants (34%) received surfactant; 24 of these were born at term (26% of infants born at term following antenatal diagnosis). Seventy-eight babies received pulmonary vasodilator therapy(s) for pulmonary hypertension after birth (36%); 75 of these (96%) had inhaled nitric oxide.
All 134 antenatally diagnosed infants with CDH and 64/85 (75%) of those postnatally diagnosed were ventilated. Some infants received more than one mode of ventilation: 193 (97%) underwent conventional mechanical ventilation and 87 (44%) high frequency oscillation. Extracorporeal membrane oxygenation (ECMO) use was very low among the cohort (nine infants, 4% of the cohort).
Sixty-one per cent of CDH newborns received inotropes (n=134), with a statistically significant difference in inotrope use between those diagnosed antenatally and postnatally (77% and 36%, P<0.001).
Surgical management
One hundred and eighty-two infants underwent surgical repair of their diaphragmatic hernia (83%). Thirty-five infants (16%) died before surgery. One infant was lost to follow-up before surgery. Details of surgical management are summarised in table 3. Sixty-five per cent of infants had primary repair of the diaphragmatic defect (n=118). Thirty-five per cent (n=64) underwent repair with a patch, 50% and 16% in antenatally and postnatally diagnosed infants, respectively (P<0.001). There was no statistical difference in the proportions of female and male newborns who underwent patch diaphragm reconstruction (23/63 (37%) vs 41/119 (34%), P=0.782).
Surgical management of infants in the cohort
Early outcomes
Of the 182 infants who underwent surgical repair, 4 died within 30 days of surgery, giving a 30-day postoperative survival rate of 98%. Two infants had a recurrence of their CDH within 30 days of surgery. Twelve infants developed chylothorax following surgery (7%).
Factors associated with mortality before surgical correction of CDH
The relationship between a number of factors and death without surgery was assessed with univariable logistic regression (table 4). In addition to those factors shown in table 4, the relationship between use of surfactant and death before surgery was assessed by univariable logistic regression analysis. The unadjusted OR for the use of surfactant in all infants was 3.35 (1.57–7.14). In those born at <37 weeks’ gestation, the OR was 0.80 (0.19–3.42), and for babies born at ≥37 weeks the OR was 6.59 (2.60–16.71). As the impact of surfactant utilisation appeared so different in term and preterm babies, this was not added to the multivariable model.
Univariable logistic regression to assess factors associated with death before surgery
The use of pulmonary vasodilators after birth, inotrope use, prenatal and postnatal diagnosis, female sex and gestational age at birth all met the criteria for inclusion in the multivariable analysis and were entered into the model in that order. The first four of these factors were retained within the model after likelihood ratio testing, and all were statistically significant in the final model. The results of the multivariable analysis are summarised in table 5.
Multivariable logistic regression to assess factors associated with death before surgery
Discussion
A population-based assessment of incidence, management and early postoperative outcomes of CDH has been undertaken for the first time in the UK and Ireland. Consistent with the findings of several recent studies, we observed that there were 1.5 times as many male live-born infants with CDH as female (61% vs 39%).10 11
Sixteen per cent of infants died before surgery. Factors associated with poor outcome were identified. Two factors, the use of inotropes and vasodilators after birth, are hallmarks of severe physiological instability. The prognostic significance of diagnosis before birth was also strikingly apparent within this cohort, with antenatally diagnosed infants almost five times more likely to die before surgery than those detected postnatally, a reflection of the greater anatomical severity of CDH in antenatally diagnosed fetuses. Twenty-four per cent of newborns with prenatal diagnosis died without having an operation, in comparison with 4% of CDH cases who had defects detected in the postnatal period. It is also noteworthy that a larger proportion of antenatally diagnosed babies with CDH had inotropes administered before surgery (77% vs 36%, P<0.001), emphasising their greater likelihood of physiological instability. In addition, more underwent patch repair (50% vs 16%, P<0.001).
We found that female infants were more likely to die without undergoing surgery than male infants (26% and 10%, respectively). This association with gender was statistically significant even after adjusting for prenatal diagnosis and the two markers of physiological instability. A large study from the CDH study group likewise identified that female infants were more likely to die before hospital discharge.10 It could be postulated that female sex is associated with more severe pulmonary hypoplasia, and hence girls are less likely to be stable enough to undergo surgery. It is noteworthy though that girls were just as likely to have CDH defects detected antenatally as boys (P=0.48). Our study results showed no difference(s) in the number of girls and boys having patch repair (P=0.782), although the defect sizes of those who died before surgery remain unknown.
There is some emerging evidence that there may be sex-related differences in the genetics of isolated CDH, with girls carrying more de novo coding variants; however, the correlation between genetic aetiology and CDH severity is as yet unclear.12 The reasons for the higher proportion of girls dying before operative repair of CDH merit further investigation.
While 16% of infants died without undergoing an operation, it is useful to report that only a small number of early postoperative deaths were identified in the study cohort. This could be a product of related phenomena: case selection of infants who are physiologically stable and appropriate surgical candidates and excellent postoperative care in individual centres. It is also necessary to assess outcomes beyond 30 days to provide meaningful data on longer term postoperative survival.
The strength of this study is that it represents, for the first time, a comprehensive population-based assessment of early outcomes in CDH across individual surgical centres, with all paediatric surgery centres in the UK and Ireland contributing data to the study. The live-birth prevalence of 1.7 per 10 000 is somewhat lower than that estimated during the years 2009–2010 from geographical regions covered by the British Association of Congenital Anomalies Register(s) (2.66 (95% CI 2.24 to 3.15))13; the exact reasons for this difference in ascertainment are unclear. While the data set reported here includes only those cases reported to paediatric surgeons, established ascertainment methodology, meticulous chasing of incomplete data and querying of inconsistent responses mean that we are confident that the data set presents a true picture of surgically treated CDH cases for the period in question. However, as this study collected data from surgical centres only, it is possible that some infants who were born alive died outside of these centres, and surgeons were not informed of this. The number of these infants is likely to be very small as antenatally diagnosed CDH deliveries are planned routinely to take place in surgical centres and the rate of non-survival in postnatally diagnosed infants is low. Cross-validation with the congenital anomalies registers existent at the time of the study may have helped to identify such cases in the areas covered by these registers.
The relatively small number of infants presenting with CDH at each specialist centre is noteworthy (median of 5 newborns per centre annually) and precluded any meaningful outcome comparison between units. The influence on outcome of birth in a low-volume centre has been widely debated, with studies from Canadian Pediatric Surgery Network (CAPSNet) and Scandinavia suggesting improved survival in hospitals managing more than five or six CDH cases a year.14 15 Controversially however a recent large study from the USA was not able to demonstrate any clear relationship between CDH mortality and management in high-volume versus low-volume centres (OR 1.03, 95% CI 0.86 to 1.23).3 The small number of cases managed in individual surgical centres across the UK and Ireland does, however, bring into sharp focus a crucial need for collaborative research platforms through organisations such as BAPS-CASS, CAPSNet, CDH study group, the National Congenital Anomaly and Rare Disease Registration Service and the CDH European Consortium to collectively study, analyse and report large data sets to inform ‘best practice’ for rare diseases.
Variation in practice in prenatal and postnatal CDH care across individual centres in the UK and Ireland was readily apparent from this study. It is notable that 22% of parents who had an antenatal diagnosis of CDH did not receive counselling from a paediatric surgeon before the birth of their baby. This is an area where improvements must be made to facilitate multidisciplinary antenatal counselling and reflects recommendations issued by the Mothers and Babies : Reducing Risk Through Audits And Confidential Enquiries across the UK (MBBRACE) UK 2014 Perinatal Confidential Enquiry into CDH.16 Counselling by surgeons alongside obstetricians and neonatologists affords the opportunity for knowledge exchange and information sharing with regard to the prognosis for the fetus, including the site and mode of delivery, newborn management and long-term outcome(s). There is strong evidence that counselling by paediatric surgeons before birth reduces levels of parental anxiety, and parents value this interaction.17–20
Current practice regarding the use of exogenous surfactant therapy as part of early postnatal CDH management is inconsistent. In this study, approximately one-third of all newborns with antenatally detected CDH received surfactant. There was some evidence that this may increase the risk of death before surgery in term infants (OR 6.59, 95% CI 2.60 to 16.71). It is likely though that term babies who received surfactant may have been more unstable, leading neonatologists to use it as a perceived ‘salvage’ therapy. Although the quality of the evidence remains poor, studies have emerged suggesting the use of surfactant in infants with CDH may adversely affect survival and be linked with the development of chronic lung disease.21 22 Current European guidelines advise against the use of this costly and potentially harmful intervention in infants born with CDH either at or before term.7
The very limited use of ECMO in infants with CDH across the UK and Ireland in this study cohort is noteworthy. Five centres currently provide this facility and infants are transferred to these units for ECMO where appropriate. Only 4% of the entire CDH cohort received ECMO. This is in marked contrast to contemporary study series from ECMO centres across the USA, where ECMO is deployed for over half of all infants with CDH.23 The CDH International Study Group has also reported a wide variation in usage of ECMO, noting that on average 30% of infants received this therapy particularly in high-volume centres.1 The low utilisation of ECMO in the UK and Ireland reported here probably reflects a more judicious use of this high-cost healthcare intervention with little evidence of clinical efficacy.5 7
Conclusions
This study shows variation in the management of CDH in the UK and Ireland before, during and after birth. Mortality remains high for this rare condition, with some 16% of all live-born infants with CDH not surviving to undergo surgical repair. Published outcomes should be clearly stratified by criteria of disease severity, including antenatal diagnosis and size of diaphragm defect. The small number of babies with CDH managed in individual surgical centres further highlights the need for neonatal networks and collaborative multidisciplinary research to generate robust outcome data and facilitate interventional studies to improve outcomes.
Acknowledgments
We acknowledge the support, dedication and hard work of all those who contributed to this study. Sean Marven, Elizabeth Draper, Paul Johnson, Peter Brocklehurst, David Howe and Judith Rankin were involved in the original set-up and design of the study.
References
Footnotes
Contributors PDL, MK and JJK designed and coordinated the study, along with the BAPS-CASS collaborators. A-ML and KJB analysed the data, and A-ML wrote the manuscript, which was reviewed and edited by all other authors.
Funding This project was funded by Action Medical Research.
Competing interests None declared.
Patient consent Detail has been removed from this case description/these case descriptions to ensure anonymity. The editors and reviewers have seen the detailed information available and are satisfied that the information backs up the case the authors are making.
Ethics approval The study received ethics committee approval from the London Research Ethics Committee: Ref 09/H0718/10.
Provenance and peer review Not commissioned; externally peer reviewed.
Collaborators Sean Marven, Elizabeth Draper, Paul Johnson, Peter Brocklehurst, David Howe, Judith Rankin, Adil Aslam, Wajid Jawaid, David Wilkinson, Andrew Robb, Anthony Lander, Tariq Burki, Laura Coates, Eleri Cusick, Julia Fishman, Nicholas Madden, Boma Adikibi, Gordon MacKinlay, Atul Sabharwal, Celia Larcombe, Joe Curry, Sanja Besarovic, Khaled Ashour, Paul Johnson, Helen Carnaghan, Mark Davenport, Ian Sugarman, Shawqui Nour, Thomas Tsang, Anu Paul, Brian Davies, Danielle McLaughlin, Prem Puri, Robin Abel, Christopher Driver, Aneis Mahomed, Evelyn Ervine, William McCallion, Simon Phelps, Ross Craigie, Antonino Morabito, Michael De La Hunt, Gareth Hosie, Karen Lloyd, Rhoda Taylor, Melanie Drewett, Bruce Okoye, Catherine Richards, Juliette King and Simon Huddart.