Introduction Published prevalence rates of congenital diaphragmatic hernia (CDH) vary. This study aims to describe the epidemiology of CDH using data from high-quality, population-based registers belonging to the European Surveillance of Congenital Anomalies (EUROCAT).
Methods Cases of CDH delivered between 1980 and 2009 notified to 31 EUROCAT registers formed the population-based case series. Prevalence over time was estimated using multilevel Poisson regression, and heterogeneity between registers was evaluated from the random component of the intercept.
Results There were 3373 CDH cases reported among 12 155 491 registered births. Of 3131 singleton cases, 353 (10.4%) were associated with a chromosomal anomaly, genetic syndrome or microdeletion, 784 (28.2%) were associated with other major structural anomalies. The male to female ratio of CDH cases overall was 1:0.69. Total prevalence was 2.3 (95% CI 2.2 to 2.4) per 10 000 births and 1.6 (95% CI 1.6 to 1.7) for isolated CDH cases. There was a small but significant increase (relative risk (per year)=1.01, 95% credible interval 1.00–1.01; p=0.030) in the prevalence of total CDH over time but there was no significant increase for isolated cases (ie, CDH cases that did not occur with any other congenital anomaly). There was significant variation in total and isolated CDH prevalence between registers. The proportion of cases that survived to 1 week was 69.3% (1392 cases) for total CDH cases and 72.7% (1107) for isolated cases.
Conclusions This large population-based study found an increase in total CDH prevalence over time. CDH prevalence also varied significantly according to geographical location. No significant association was found with maternal age.
- congenital diaphragmatic hernia
- Maternal age
- Population based
Statistics from Altmetric.com
What is already known on this topic
Previous population-based studies have found total prevalence rates of congenital diaphragmatic hernia (CDH) between 2.4 and 4.1 per 10 000 births. Studies that have investigated CDH prevalence over time have found no significant trends. Significant associations with maternal age have not been seen.
What this study adds
This study provides a prevalence estimate for congenital diaphragmatic hernia based on the largest number of cases and denominator population ever studied. Total prevalence is lower than previously reported and a slight increase over time was observed. No association was found with maternal age.
Congenital diaphragmatic hernia (CDH) is a malformation of the diaphragm which can lead to a hole remaining in the diaphragm allowing the abdominal viscera to protrude into the chest cavity.1 The diaphragm normally starts developing during the fourth week of gestation and closes during the eighth week. However, in cases of CDH the diaphragm fails to close completely during this period.2
In previous population-based register studies, estimates of total prevalence rates of CDH range between 2.4 and 4.1 per 10 000 births in Atlanta (1968–1999) and in the UK (1995–2000), respectively.3 ,4 Studies investigating CDH prevalence over time observed increasing and decreasing trends, but none were found to be statistically significant.3 ,5–8
The causes of CDH are unknown, but CDH is more common in chromosomal anomalies such as trisomy 18 (Edwards syndrome), trisomy 13 (Patau's syndrome) and trisomy 21 (Down's syndrome)6 ,8–10 and may therefore have a genetic predisposition.11 CDH is often observed in patients who are diagnosed with Fryns’ syndrome, Pallister–Killian, Wolf–Hirschhorn or Cornelia De Lange. Despite CDH occurring alongside several genetic syndromes, there is limited evidence to suggest any significant level of familial clustering. Up to 2% of cases may have some form of familial association.12 ,13 A meta-analysis undertaken to examine factors associated with mortality in CDH cases14 estimated that 39.3% of all CDH cases occurred with other major structural anomalies.
A recent population-based study in the UK estimated the survival rate for CDH to be 64.6% at 1 week, 58.4% at 1 year and 57.1% at 20 years.15
Several studies have investigated the influence of maternal age on CDH prevalence.5–8 Two suggested that there may be a slight increased risk for older mothers (aged >35)6 ,8 and one suggested an increased risk in women under 20.5 However, none of these associations reached statistical significance.
The aim of this study is to describe the epidemiology of CDH, in particular, associated anomalies, pregnancy outcomes, maternal age, sex distribution and prevalence rates over time and according to geographical location, using high-quality, population-based European register data.
The European Surveillance of Congenital Anomalies (EUROCAT)16 is a collaborative network of population-based congenital anomaly registers. Forty-three registers in 20 countries use multiple sources to collect data on anomalies occurring in spontaneous fetal losses at ≥20 weeks’ gestation, termination of pregnancy for fetal anomaly (TOPFA) following prenatal diagnosis at any gestation and live births. EUROCAT surveys approximately 1.6 million births in Europe, representing almost 30% of the European birth population.17 Cases are coded using the WHO International Classification of Disease V.9 or 10 (ICD 9 or ICD 10). Minor anomalies are excluded; further details regarding data collection are available on the EUROCAT website.18 ,19
All cases of CDH (ICD 9: 756.6 or ICD 10: Q790) with a delivery date between 1 January 1980 and 31 December 2009, notified to the 31 registers that participated in this study, form the population-based case series. The remaining registers could not contribute their data in time. Denominator (live births and stillbirths), and denominator data by maternal age, were provided by EUROCAT.17
Inclusion and exclusion criteria
Cases occurring in multiple pregnancies or those with a chromosomal anomaly, genetic syndrome, microdeletion or teratogenic syndrome were excluded from complex statistical analysis due to their differing aetiologies. All other cases were included and were coded as isolated cases or cases with associated structural anomalies. Cases with associated structural anomalies that were directly related to CDH (eg, lung hypoplasia or intestinal malrotation) were classed as isolated CDH.
Descriptive statistics were produced for associated congenital anomalies (by group and subtype), pregnancy outcome (categorised as live birth, stillbirth or TOPFA). Maternal age (categorised as <20, 20–24, 25–29, 30–34 and ≥35 years), 1-week survival (classed as alive or died), gestational age at delivery (weeks) and birth weight (grams). Gestational age data and birth weight data were missing for 6.94% and 4.4% of cases, respectively. These cases were excluded from analysis of these variables. Total prevalence rates for CDH in each register were calculated as the number of cases (whether ending in fetal death, stillbirth, TOPFA or live birth) per 10 000 total births; 95% CIs were derived from the binomial distribution.
The relative risk (RR) of CDH per year increase in year of birth was estimated using a multilevel Poisson regression. The EUROCAT registers survey distinct geographical areas and contributed data for different time periods, therefore a simple analysis of prevalence over time could have underestimated SEs and introduced confounding.20 The number of cases per year were nested within register and modelled with a random intercept. The offset was equal to the log of denominator, and year as a continuous predictor. Heterogeneity between registers was evaluated from the random component of the intercept. Inter-regional differences in trends were tested through the incorporation of a random slope, and variation terms were added to check for compliance with the Poisson distribution. However, neither was necessary.
To investigate associations between maternal age at delivery and CDH prevalence, all models were refitted to include age. Five age groups were categorised: <20, 20–24, 25–29, 30–34 and ≥35 years. There were 113 cases with no maternal age data, which were excluded from this analysis. The registers in Northern England, Valencia Region and Reunion had more than 5% of their denominator data uncategorised for several years and so were excluded completely from maternal age analysis. Similarly, Tuscany, South Portugal, Saxony Anhalt and Thames Valley were excluded for the years 1995, 1993–1994 and 2007, 1990 and 1990–1994, respectively. Interactions between study period and maternal age were explored by adding cross-product terms.
The RR (per year) of 1-week survival was also estimated using multilevel Poisson models, with a random intercept and with log of the number of CDH cases as the offset. This analysis was restricted to live born cases only. There were 13.3% of cases with missing 1-week survival data and these cases were therefore excluded from all analyses of survival.
Multilevel model parameters were estimated using a random walk (Metropolis–Hastings) Markov Chain Monte Carlo (MCMC) algorithm. Assuming diffuse uniform priors, the procedure was run for a burn-in sample of 5000 observations, and an analysis sample of 1 000 000 thinned by 10 (numbers guided by Raftery–Lewis calculations21). Ninety-five percent credible intervals (CrIs) were obtained from the posterior distribution for each parameter.
All analyses were carried out on total births (cases occurring in live births, fetal losses or TOPFA), except for analysis of 1-week survival, gestational age at delivery and birth weight, which was carried out on live births only.
Statistical analyses were performed using Stata V.11 (descriptive analysis) and MLwiN 2.14 (multilevel analysis). p<0.05 was considered statistically significant.
Figure 1 shows the flow of cases through the study. A total of 3373 cases of CDH were reported to the 31 EUROCAT registers. Of these, 139 (4.1%) occurred in multiple births (133 in twins and six in triplets) and 103 (3.1%) had unknown number of fetuses and were excluded from the analysis. Seven (5.3%) twins had a chromosomal anomaly and one had a genetic syndrome. Forty-one (30.8%) twins and one (16.7%) set of triplets had associated structural anomalies; 84 (63.2%) twins and five (83.3%) triplets had isolated CDH. There were two examples of twins concordant for CDH. Two cases occurred with a teratogenic syndrome and were also excluded.
Associated structural anomalies
Of the 3129 singleton cases, 226 (7.2%) were associated with a chromosomal anomaly, including trisomy 18 in 132 cases (4.2%), trisomy 13 in 36 cases (1.1%) and trisomy 21 in 28 cases (0.9%). A further 110 (3.5%) cases occurred as part of genetic syndromes, including 17 (0.5%) with Fryns’ syndrome. Seventeen (0.54%) cases occurred with a microdeletion.
These 353 cases were excluded from further analysis, leaving 2776 singleton cases. Of these, 784 (28.2%) occurred with one or more associated major structural anomalies (figure 1). Many cases occurred with more than one associated congenital anomaly (and therefore contributed to each relevant subtype in table 1). Cardiac anomalies were most commonly associated with CDH, occurring in 398 cases (14.3%). Urinary anomalies occurred in 154 (5.5%) cases, limb anomalies in 152 (5.5%) cases and nervous system anomalies in 134 (4.8%) cases (table 1). There were 1992 (71.8% of 2776) isolated CDH cases.
Gestational age and birth weight
Of the 2317 live born children, 2216 (95.6%) had known gestational age at delivery. The median gestational age at delivery was 39 weeks (IQR 37–40) for all cases and 39 (IQR 37–40) for isolated CDH cases.
Of the live born cases, the mean birth weight was 2960 g (95% CI 2931 to 2990, n=2156) for all cases and 3217 g (95% CI 3190 to 3243, n= 1631) for isolated cases. Of the term cases (gestational age ≥ 37 weeks, n=1506), the mean birth weight was 3173 (95% CI 3149 to 3198, n=1653) for all cases and 3217 (95% CI 3191 to 3243, n=1282) for isolated cases.
Pregnancy outcomes, sex distribution and antenatal detection
Of the 2776 total cases, 2317 (83.4%) resulted in a live birth, 360 (13.0%) in a TOPFA and 99 (3.6%) in a stillbirth. The proportion of cases resulting in a live birth decreased over time from 86.4% in 1980–1984 to 82.5% in 2005–2009, but the trend was not significant (test for trend: p= 0.240). The proportion of cases resulting in a TOPFA increased significantly over time, from 4.6% in 1980–1984 to 14.4% in 2005–2009 (test for trend: p=0.003).
Of the isolated cases, 1767 (88.7%) resulted in a live birth, 177 (8.9%) resulted in a TOPFA and 48 (2.4%) were stillbirths. The proportion of isolated cases resulting in a live birth decreased significantly over time, from 95.2% in 1980–1984 to 87.2% in 2005–2009 (test for trend: p=0.004). The proportion of cases resulting in a TOPFA increased significantly over time, from 1.6% in 1980–1984 to 10.4% in 2005–2009 (test for trend: p<0.001). Stillbirths and TOPFA occurred more frequently in pregnancies involving CDH associated with other structural anomalies compared with isolated CDH cases.
Where sex was known (2684 (96.7%) cases), the M:F ratio for all cases was 1:0.69 and 1:0.64 for isolated cases.
Data on antenatal diagnosis were available in 2337 (84%) cases. Antenatal diagnosis of any anomaly was made in 1148 (49.1%) CDH cases. While rates increased from 32.9% in 1980–1984 to 43.3% in 2005–2009, there was no significant trend in detection over time (p=0.936). Antenatal diagnosis was made in 759 (45.3%) of the isolated cases, with an increase in detection rates over time, from 6.0% in 1980–1984 to 52% in 2005–2009 (test for trend: p<0.001). TOPFA occurred in 23.3% of all isolated cases that were identified antenatally.
Between 1980 and 2009, there were 2776 cases among 12 155 491 registered births, giving a total prevalence rate of 2.30 (95% CI 2.2 to 2.4) per 10 000 births (figure 2).
There was significant variation between registers (p<0.001); total prevalence ranged from 1.39 (95% CI 1.1 to 1.7) in Tuscany to 4.10 (95% CI 3.0 to 5.5) per 10 000 births in Malta (figure 2, table 2). However, the Maltese rate was based on only 46 cases in a birth population of approximately 110 000. Strasbourg, Paris, Wales and Northern England are larger registers and also have higher prevalence rates of 3.4 (95% CI 2.8 to 4.1), 3.1 (95% CI 2.8 to 3.5), 3.1 (95% CI 2.6 to 3.7) and 3.2 (95% CI 2.6 to 3.9) per 10 000 births respectively.
The total prevalence rate for isolated CDH cases for all registers was 1.6 (95% CI 1.6 to 1.7) per 10 000 births. The total prevalence of isolated CDH varied significantly between registers (p<0.001), Valencia Region having the highest rate (3.1 per 10 000 (95% CI 2.0 to 4.0)) and Cork & Kerry having the lowest rate (0.7 per 10 000 births (95% CI 0.3 to 1.4)).
The isolated CDH live birth prevalence was 1.5 (95% CI 1.4 to 1.5) per 10 000 live births for all registers.
Trends in prevalence over time
After accounting for variation between registers, the multilevel analysis identified a small, but significant increase in total prevalence for all CDH cases over time (RR (per year)=1.01, 95% CrI 1.00 to 1.01; p=0.030). This ranged from a modelled prevalence of 2.04–2.39 per 10 000 in 1980–1984 to 2004–2009.
For isolated CDH cases, no significant increase over time was observed (RR (per year)=1.00, 95% CrI 0.99 to 1.01; p=0.359), with modelled prevalence ranging from 1.86 to 1.90 per 10 000 in 1980–1984 to 2004–2009.
There was a suggestion of a U-shaped association between CDH prevalence at maternal age at delivery (table 3). However, after accounting for heterogeneity between registers and year of delivery, there were no statistically significant associations with maternal age (table 3) (p=0.390).
Among isolated cases, prevalence appeared to increase over increasing maternal age categories (table 3). After accounting for variation in registers and year of delivery, this was evident in the effect sizes of the RRs. However, there were no statistically significant associations in any maternal age categories (p=0.302).
There was a small but significant increase in prevalence over time after adjustment for maternal age (RR=1.01 (1, 1.01); p=0.006). There was no evidence of an interaction between maternal age and year. For isolated cases, there were no significant associations with maternal age in any category.
Survival at 1 week
One week survival status was known for 86.7% (2008) of live born cases. Of these, 69.3% (1392) of all cases and 72.7% (1107) of isolated cases survived the first week of life. Over time, the risk of death in the first week decreased significantly by 3% per year for all cases (RR=0.97, 95% CrI 0.93 to 0.99; p=0.023) and 7% per year for isolated cases (RR= 0.93, 95% CrI 0.88 to 0.98; p=0.003).
Using data from 31 population-based European registers over a 29-year period, including over 12 million births, we found CDH prevalence rates of 2.3 and 1.6 per 10 000 births for all cases and isolated cases, respectively. This study found evidence that total prevalence rates have increased over time. We also found differences in prevalence rates between registers. We found no significant association with maternal age, although prevalence increased linearly with maternal age for all cases of CDH and in a U-shape for isolated CDH cases.
The main strength of this study is that it used data derived from established, high-quality, population-based congenital anomaly registers. The registers use standard methods to identify cases and multiple sources of notification to ensure a high case ascertainment. A further strength is our use of multilevel methods, which provide more accurate SE estimates for nested data than classical statistical approaches. Furthermore, this method limits potential for confounding due to registers contributing data from different time points.
It was not possible to account for differences in case ascertainment between registers or over time. It is therefore possible that variation between registers and the increase in prevalence over time is due to ascertainment. The study was also unable to differentiate between the different types of CDH which have differing aetiologies.
A further limitation is that potential exposures or risk factors are not included in the routine data collected by all registers. However, given that knowledge of causal factors for CDH is limited, it is unclear what exposures would be informative. Previous studies have described prevalence by ethnic group,3 ,6–8 with limited evidence that prevalence of isolated CDH may be lower in some ethnic groups. This could not be investigated in this study as ethnicity data were incomplete.
The total prevalence rate reported by this study (2.3 per 10 000 births) is slightly lower than that reported in previous studies.3 ,5 ,6 ,22 ,23 Differences in overall prevalence rates may result from variations in the inclusion and exclusion criteria. Other studies included all multiple births and/or all chromosomal anomalies in their prevalence calculations.4 ,9 ,22 ,24 ,25 Our isolated CDH prevalence rate of 1.6 per 10 000 is similar to other published figures which ranged from 1.34 to 2.6 per 10 000 births.3–10 ,22–26 Perhaps isolated CDH is a less heterogeneous group than non-isolated CDH and therefore ascertainment between studies may have been more consistent. Using multilevel analysis, we identified a small, but significant increase in CDH prevalence over time. For isolated CDH cases, no significant increase was observed. Although this is the first time a significant trend over time has been observed, the effect is only very small and it is quite possible that it is due to increased ascertainment rather than a true increase in prevalence. Increased survival rates due to improved perinatal emergency care may have contributed to improved ascertainment of CDH, if perinatal deaths are incorrectly attributed to asphyxia when there is no autopsy.
Although there was significant variation in total prevalence observed between registers for all cases and for isolated cases, one of the study limitations is that variation between registers may be explained in part by differences in case ascertainment. Several of the registers with highest prevalence rates such as Malta, Paris, Strasbourg and Northern England have been long established (since the 1980s). However, geographical variation in prevalence may reflect a true difference, for example in exposures between regions.
We found that cases of CDH were associated with a chromosomal or genetic syndrome or microdeletion in approximately 11% of cases, which is comparable to other studies.5 ,7 ,8 ,22 ,24 ,25 The identification of trisomy 18 as the most common chromosomal anomaly is also consistent with previous studies5 ,7 ,8 ,22 ,24 ,25 although the proportion was lower in our study. Trisomy 18 accounted for 37% of all chromosomal anomalies, but figures from other studies have reported it as high as 60%.5 ,6
An association of CDH with Fryns’ syndrome has been reported in several studies,6 ,7 ,10 ,22 ,23 many at higher rates than that found in our study. CDH forms a component of the diagnosis of Fryns’ syndrome and is diagnosed clinically; this may lead to an artificially high number of CDH cases being classed as Fryns’ and CDH.11 Fryn's may be underrepresented in this study as it is not a specific syndrome diagnosis listed within the EUROCAT guidance. It is also more difficult to diagnose syndromes when a TOPFA has been undertaken.
The prevalence of congenital anomalies among cases was much higher than we would expect in the general population. For example, we identified 14.3% and 4.8% of cases with cardiac anomalies and nervous system anomalies whereas in the general population we would expect 79.9 and 25.1 per 10 000 total birth, respectively.27 However, our rates of associated congenital anomalies are generally comparable to those reported elsewhere. For example, when comparing our rates of associated cardiac anomalies with studies that used similar inclusion criteria, the 14.8% of cases observed in our study lies within the middle reported rates in the literature.5–7 ,9Similarly, our rates of association with central nervous system, limb and urinary anomalies (4.8–5.5%) also broadly match those reported in the published literature.3 ,4 ,7 ,8 ,10
We identified a U-shaped trend in the prevalence of all CDH (including those with associated structural anomalies), and a linear increase in the prevalence of isolated CDH over increasing maternal age categories, but neither reached statistical significance. This is comparable to studies that found no association or observed a slight, non-significant increase in prevalence in older age groups or a non-significant risk in those under 20 years of age.5–8 Robert et al6 propose that the slightly elevated, non-significant risk of CDH for mothers aged over 30 may be due to the mistaken inclusion of infants with chromosomal anomalies given the well established association between maternal age and chromosomal anomalies.28
The sex ratios calculated in this study of 1:0.69 for total cases and 1:0.64 for isolated cases are comparable to the highest ratios previously reported in the literature (1:0.70 for total cases and 1:0.63 for isolated cases).3 ,6–8 Differences in prevalence by sex were not calculated in this paper, as the denominator data were not available. Previous research has reported a pooled estimate of the RRs of CDH for male cases compared with female cases of 1.34 (95% CI 1.25 to 1.43).29
We found that cases occurring with other major structural congenital anomalies were more likely to have an anomaly diagnosed antenatally than isolated cases. This confirms the findings of several other studies.3 ,8 ,22 ,23 ,30 We also identified an increase over time in the proportion of isolated cases diagnosed antenatally, but not in cases with other anomalies. This suggests that the antenatal diagnosis rates of CDH specifically have increased. Each country has different policies regarding timing and frequency of antenatal routine ultrasound screening.9 ,31 As a result, antenatal detection rates are likely to vary significantly between registers.
We found a higher survival rate among isolated cases than identified in some studies,7 ,8 ,23 ,24 ,30 but these studies reported survival at discharge (which may be beyond 1 week) or at 1 year. A recent study by Tennant et al15 found similar survival rates to this study, with a 65% survival rate at 1 week.
This large population-based study found no evidence of an increase in the prevalence of isolated CDH over time. A statistically significant but marginal increase in total CDH prevalence was observed. CDH prevalence varied significantly according to geographical location, a finding which requires further investigation. No association was found with maternal age for all CDH cases or for isolated cases.
Contributors MRM undertook the analysis and interpretation of the data, drafted the manuscript and approved the final manuscript as submitted. KEB participated in the analysis and interpretation of data, critically reviewed and revised the manuscript and approved the final manuscript as submitted. JR conceived the project and participated in the analysis and interpretation of data, critically reviewed and revised the manuscript and approved the final manuscript as submitted. DW made substantial contribution to data acquisition and analysis, interpretation of the data, revising the article and final approval to be published. RG, M-CA, LA, HdW, JB, FB, EC, BD, ESD, EG, MG, MH, BK, KK, AL-B, MO, PB, BM, CM, VN, AQ-L, HR, AR, CR, AS, RT, DT, WW and CM made substantial contribution to data acquisition, interpretation of the data, revising the article and final approval to be published.
Funding The North of England register is funded by the UK Healthcare Quality Improvement Partnership. This research was undertaken by MM (supervised by JR and KB) in partial fulfilment of the requirements for the MSc in Public Health and Health Services Research at Newcastle University (degree awarded December 2012). No external funding was secured for this study. Malformation Monitoring Centre Saxony-Anhalt is funded by the Ministry of Labour and Social Affairs Federal State of Saxony-Anhalt.
Competing interests None.
Patient consent No.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.