Maternal Risk Factors and Perinatal Characteristics for Hirschsprung Disease
BACKGROUND AND OBJECTIVES: Hirschsprung disease (HSCR) is a congenital defect of the enteric nervous system characterized by a lack of ganglion cells in the distal hindgut. The aim of this study was to assess the birth prevalence, perinatal characteristics, and maternal risk factors in HSCR patients in Sweden.
METHODS: This was a nationwide, population-based, case-control study of all children born in Sweden between 1982 and 2012 and registered in the Swedish Medical Birth Register. Cases were identified in the Swedish National Patient Register and data on potential maternal risk factors and patient characteristics were collected from the Swedish National Patient Register and the Swedish Medical Birth Register. Five age- and sex-matched controls were randomly selected for each case. The association between studied risk factors and HSCR was analyzed using conditional logistic regression to calculate the odds ratio (OR) and 95% confidence interval (CI).
RESULTS: The study population comprised 600 HSCR cases and 3000 controls with a male-to-female ratio of 3.7:1. The birth prevalence of HSCR was 1.91/10 000. Maternal obesity was associated with an increased risk for the child to have HSCR (OR 1.74; CI 1.25–2.44). Children with HSCR were born at an earlier gestational age (OR 1.60; CI 1.18–2.17) than control children. Associated malformations were identified in 34.5% of the cases.
CONCLUSIONS: This study showed that the Swedish birth prevalence of HSCR was 1.91/10 000. Children with HSCR disease were born at a lower gestational age than controls. Maternal obesity may increase the risk for the child to have HSCR.
- CI —
- confidence interval
- HSCR —
- Hirschsprung disease
- ICD —
- International Classification of Diseases
- MBR —
- Swedish Medical Birth Register
- NPR —
- National Patient Register
- OR —
- odds ratio
- SGA —
- small for gestational age
What’s Known on This Subject:
Hirschsprung disease (HSCR) is a multifactorial disease. The incidence varies from 1 per 2000 to 1 per 12 000. Being firstborn may decrease the risk of having the disease.
What This Study Adds:
The incidence of HSCR is 1.91 per 10 000 live newborns in Sweden. Maternal obesity increases the risk for having a child with HSCR, and there is increased risk for children with HSCR to be born prematurely.
Hirschsprung disease (HSCR) is a congenital defect of the enteric nervous system characterized by a lack of ganglion cells in the distal hindgut. Motility disturbances in the distal colon usually lead to neonatal intestinal obstruction. The majority of the patients undergo surgical treatment during the first year of life.
The incidence of HSCR has been assessed in both demographic and epidemiologic studies and varies from 1 in 2000 to 1 in 12 000 live births.1 Best et al recently reported a slightly increasing prevalence of HSCR in Europe.2
HSCR is known to be a multifactorial disease caused by both genetic and environmental factors. Mutations in >10 genes have been associated with HSCR, particularly the RET gene, in which 15% to 20% of patients with isolated HSCR have mutations.3 Familial occurrence, male predominance, and the pattern of associated malformations, encountered in 4% to 35% of the cases, also imply a genetic etiology.4 Furthermore, patients with Down syndrome (trisomy 21) have a 100-fold risk for HSCR compared with the normal population.5 The importance of environmental factors is not well known. In a small study of patients with Down syndrome and HSCR, extensive coffee drinking and maternal fever during the first trimester increased the risk for HSCR.6 There have been speculations about the role of hypothyroidism, vitamin A deficiency, and maternal intake of Ibumetin or mycophenolate during pregnancy, but none of these associations have been confirmed.7⇓⇓⇓–11
The purpose of this study was to investigate the maternal risk factors and perinatal characteristics of HSCR in Sweden.
This was a nationwide, population-based case-control study. The study base includes all neonates born in Sweden during the observational period January 1, 1982, to December 31, 2012, and registered in the Swedish Medical Birth Register (MBR). The study outcome involved HSCR and the study exposures being assessed through linkage with the Swedish National Patient Register (NPR) for both cases and their mothers. All residents in Sweden are assigned a unique 10-digit personal identification number after birth or immigration, which enables linkage among different national registers.
The NPR contains prospectively collected information on all hospital admissions in Sweden. The register is maintained by the Swedish National Board of Health and Welfare. It was initiated in 1964 and covers all hospitals in Sweden since 1987. The data include sex, age, geographic data, surgical procedures, primary and secondary diagnoses, and dates of admission and discharge. The International Classification of Diseases (ICD) has been modified over the years: ICD-8 in 1969–1986, ICD-9 in 1987–1996, and ICD-10 since 1997. Since 2001, data on outpatient specialist care have also been included in the register. The latest validation of the register showed that the diagnoses are valid in 85% to 95% of the cases.12
The MBR contains data on all pregnancies and deliveries in Sweden since 1973. The Swedish National Board of Health and Welfare administers the register. Correlation between register data, corresponding original medical records, and the validity of the study exposures has been shown to be excellent.13 The data are collected prospectively from antenatal care clinics, obstetric clinics, and maternity wards and include maternal age and parity, maternal weight and height, maternal smoking and diseases, duration of pregnancy, single or multiple birth, parity, and birth weight.
Cases and Controls
All cases with an ICD code for HSCR in the Swedish NPR (ICD-8 751.39; ICD-9 751D; ICD-10 Q431) during the study period were identified (n = 816) to confirm that the subjects had HSCR and were not misclassified by mistake. For instance, we wanted to avoid including neonates with suspected HSCR admitted for rectal suction biopsies in cases in which the biopsies turned out to be negative or patients admitted only to a hospital without pediatric surgery. Each case had to satisfy 1 of the following inclusion criteria:
HSCR as the main diagnosis and a surgical intervention number specific for HSCR;
admission to a pediatric surgical center at least twice, with a hospital stay of at least 4 days at least once, and HSCR as the main diagnosis for both hospital stays; and/or
One long admission to a pediatric surgical center once and >1 outpatient visit to a pediatric surgical center with HSCR as the main diagnosis.
Using these criteria, 216 patients were excluded, ending up with 600 HSCR cases; because there were 10 siblings among the cases, 590 mothers were identified. For each case, 5 controls matched for birth year and sex and without a history of HSCR (n = 3000) were randomly sampled from the study base by incidence density sampling by using the RANUNI function in SAS 9.4 (SAS Institute, Cary, NC). A flowchart is shown in Fig 1.
Definition and Categorization of Study Variables
The birth prevalence of HSCR was assessed by dividing the number of HSCR cases that met the inclusion criteria by the number of live births in Sweden in 1987–2012, according to the Swedish National Board of Health and Welfare. The birth prevalence was also assessed separately for males and females. We chose the study period from 1987 instead of 1982 because the NPR became national in 1987.
Maternal Risk Factors
Exposure data on maternal age, maternal smoking, parity, and maternal BMI were obtained from the MBR. Maternal age was categorized into 5 groups: <20 years, 20 to 24 years, 25 to 29 years, 30 to 35 years, and >35 years. Data on maternal smoking were defined as smoking at the time of registration at the antenatal care clinic at gestational weeks 10 to 12 and were categorized into nonsmoker, <10 cigarettes daily, and ≥10 cigarettes daily. Birth order was categorized into 3 groups; first, second, and third or greater. Maternal BMI was measured in kg/m2 and calculated from the periods 1982 to 1989 and 1992 onward because of changes in the register. Maternal BMI was registered at the first antenatal care clinic visit. BMI was categorized into 4 groups: <18.5 (underweight), 18.5 to 24.9 (normal weight), 25.0 to 29.9 (overweight), and ≥30.0 (obese) according to the World Health Organization classification.14
The exposure data on maternal diseases were based on occurrences of the following ICD codes: ICD-8: 242, 244, 245, 250, 340.99, 563,00-563,10; ICD-9: 242, 244, 245, 250, 340A-341×, 555A-X, 556; ICD-10: E03, E05, E06, E10, E11, E12, E13, E89, G35.9, K50-51.
Exposure data on delivery mode, gestational age, birth weight, and neonatal mortality were obtained from the MBR. Delivery mode was categorized concerning caesarean and vaginal delivery. Gestational age was categorized into 2 groups: <37 gestational weeks (preterm) and ≥37 gestational weeks (term). Birth weight for gestational age was categorized into 2 groups according to Marsál et al: small for gestational age (SGA) or not, based on growth curves.15 For the analysis of weight for gestational age (SGA), twins were excluded. Data on preeclampsia were collected from both the MBR and the NPR, based on the following ICD codes: ICD 8: 637.00-637.04, 637.09-637.10, 637.99; ICD 9: 642 E-G; ICD-10: O14.0-O14.1, O14.9, O11.9. Neonatal mortality was collected from the MBR.
Congenital malformations and chromosomal abnormalities comprised the following diagnoses in either the NPR or the MBR: ICD-8: 750.00-759.99; ICD-9: 740-759×; ICD-10: Q00-Q99. The following diagnoses were excluded: Meckel diverticulum (751, 751A, Q430), undescended testis (752.1, 752F, Q531-532, Q539), and urachus remnant (753H, Q644) due to a diagnostic bias. The malformations were categorized into the following categories: gastrointestinal, cardiovascular, musculoskeletal, urogenital, ophthalmic, central nervous system, and other malformations. The gastrointestinal malformations were further analyzed, and cases with only 1 diagnosis from an outpatient visit or 1 admission to a hospital without pediatric surgery were excluded. Data on newborns with Down syndrome during 1999–2012 were collected from the Swedish National Board of Health and Welfare to assess the cumulative incidence of HSCR among patients with Down syndrome.
The associations between HSCR and perinatal characteristics were examined by using univariable logistic regression models with preterm delivery, SGA, and mode of delivery as the outcomes and HSCR as an explanatory variable, presented with odds ratio (OR) estimates and 95% confidence intervals (CIs). Possible risk factors for HSCR (maternal age, maternal BMI, maternal smoking, and parity) were analyzed by using a conditional logistic regression procedure (clogit in R) stratifying over the matched pairs. The results were presented with OR and 95% CI. All odds of HSCR were evaluated by using a univariable approach. A multivariable analysis was considered if the univariable analysis showed that it was appropriate. All statistics are performed in R software version 2.38 (http://CRAN.R-project.org/package=survival).16
The Regional Ethics Review Board at Karolinska Institutet, Stockholm, approved the study.
The birth prevalence of HSCR in Sweden was 1.91 in 10 000 births between 1987 and 2012. The birth prevalence by sex and the total birth prevalence over the years are shown in Fig 2.
Maternal Risk Factors
The maternal risk factors are summarized in Table 1, and the most prominent finding was the increased risk for obese females to have children with HSCR. There appears to be a linear increase over BMI categories; the ORs are not significant for underweight and overweight, but this is evident from the effect sizes. The same gradient effect is shown in male cases. An analysis using a multivariate model did not change the results. A subanalysis of infant sex is shown in Table 2.
Of the 600 HSCR cases, 466 were boys, resulting in a male-to-female ratio of 3.7:1. There were 19 discordant twins among the cases and 77 among the controls (OR 1.24, 95% CI 0.75–2.07). One of the cases did not survive the first month of life, and 6 controls died (OR 0.83, 95% 0.10–6.92). Data on gestational age, birth weight, and delivery mode are summarized in Table 3. The data presented are not causative factors but could instead be associated with HSCR. The gestational age of cases and controls is shown in Fig 3. The odds of HSCR were significantly greater in preterm compared with term deliveries in girls (OR 2.69, 95% CI: 1.38–5.26) and in boys (OR: 1.41, 95% CI: 1.00–1.99). The odds of HSCR for SGA were as follows: girls, OR 1.59 (95% CI: 0.51–4.97); boys, OR 1.22 (95% CI: 0.62–2.37). Preeclampsia occurred as frequently in mothers of cases as in those of controls (24 and 95, respectively).
Among the HSCR cases, 207 (34.5%) individuals had at least 1 other congenital malformation, including chromosomal anomalies. Excluding the cases with only chromosomal anomalies, 191 of the cases had associated malformations (Fig 4). Altogether, 59 (9.8%) of the cases had Down syndrome and 18 (3%) had other chromosomal anomalies. Between 1999 and 2012, 2203 children with Down syndrome were born in Sweden. Of these, 25 also had HSCR, which gives a birth prevalence of 1.1%.
This is the first population-based case-control study including 600 cases of HSCR. The study shows that the birth prevalence of HSCR was 1 in 5000 in Sweden between 1987 and 2012. We found an association between HSCR in the child and maternal obesity during the first trimester. Also mothers pregnant with their third child or greater showed an increased risk of having a child with HSCR, and patients with HSCR tended to be born at a lower gestational age than healthy control children.
Sweden has well-recognized population-based medical registers that make it possible to perform powerful studies like this one. All data are prospectively collected, thus avoiding the risk for recall bias. The controls were randomly selected from the study base, thereby decreasing the risk of selection bias. Our study assesses a large number of HSCR cases with cross-linked data on exposures with the same definition in cases as in controls. This methodology decreases the risk of differential misclassification of exposure. However, the study also has limitations because there was no histopathology register available for linkage to confirm the HSCR diagnosis. This means that it was necessary to base the diagnosis on the ICD code, which was the reason for using additional inclusion criteria to increase the specificity of the study. Among the excluded cases, we found that the majority had only been admitted once during the neonatal period without having surgery, or admitted to a hospital without pediatric surgery services. Another possible limitation is the risk for type II errors due to the limited sample size. Data on HSCR that occurred in stillbirth or termination for fetal anomaly were not possible to retrieve from the registers and may be a limitation of the study.
Our study showed that 1.91 of 10 000 or 1 in 5000 live births involved HSCR in Sweden from 1987 to 2012. The birth prevalence did not change significantly over the study period. Earlier studies have reported an incidence of HSCR of 1 in 5000, although this varies from 1 in 2000 to 1 in 12 000 live births.1 To calculate the incidence, the population at risk is needed. Because the population at risk for HSCR, which includes the number of conceptions that reach the gestational age when the defect occurs, is unknown, it is impossible to calculate the incidence. Therefore, Mason et al recommended calculation of birth prevalence instead. The previously published incidence data are comparable to our birth prevalence data.17
Our results indicate that obesity may be a maternal risk factor or at least a risk indicator for the child to have HSCR. It is well known that prepregnancy obesity is associated with several birth defects, but this has not been shown specifically for HSCR.18⇓⇓–21 Two known risk factors for birth defects that may confound this association are blood folate levels and prepregnancy diabetes. The blood folate level among the cases’ mothers is not available from the registers and is therefore unknown. The rate of diabetes among the mothers was analyzed, and no differences could be demonstrated between cases and controls. However, there is 1 uncommon syndrome, small left colon syndrome, that occurs particularly in infants of diabetic mothers and with an unknown etiology.22 The syndrome can simulate HSCR in neonates within the first 24 to 48 hours of life, but in our data, there was no overrepresentation of maternal diabetes, which makes this an unlikely explanation.
Untreated hypothyroidism, which may lead to obesity in the mothers, could also be considered as a possible confounder. A case report describing congenital hypothyroidism associated with HSCR speculates about the importance of thyroxin levels and the development of HSCR.7 Another speculation may be that obese mothers have vitamin deficiencies due to nutritional habits. In mice, vitamin A deficiency increased the penetrance and severity of aganglionosis in an experimental model of HSCR.8 Another possible confounder could be that obese mothers may use medical treatments that may increase the risk for HSCR in the embryo. It has been suggested in animal models that maternal intake of ibuprofen or mycophenolate may cause enteric nervous system malformations with an HSCR-like pathology, which calls for studies on maternal intake of drugs during pregnancy and the associated risk for HSCR.9,11
In our study, we found that a parity of ≥3 children increases the risk of having a child with HSCR. On subanalysis for child sex, boys show the same pattern, but not girls, perhaps because of the small sample size and lack of power. This result is similar to what Ryan et al described, namely, that being firstborn decreased the risk for having HSCR.23 Goldberg showed an association between maternal age and the risk for HSCR in the offspring, which could not be confirmed by Best et al or Russel et al.2,24,25
No correlation could be found between the risk of having a child with HSCR and maternal diseases. Our group previously reported a family with autosomal dominantly inherited HSCR associated with multiple sclerosis. This family had a novel EDNRB gene mutation that could potentially play a role in the development of the diseases.26 However, in this study, maternal multiple sclerosis was not associated with HSCR in the offspring.
We also showed that HSCR patients were born at a lower gestational week than controls, thus confirming earlier studies.23 The prevalence of HSCR in premature infants in a systemic review extending from 2000 to 2013 was 4% to 19.4% (overall prevalence, 14%).27 Preterm birth may have many causes, for example, preeclampsia or congenital malformations. Maternal preeclampsia was not overrepresented among the cases’ mothers in this study. Another possible marker for a problematic pregnancy, caesarean delivery, was not statistically different between the groups. Downey et al studied preterm patients with HSCR and concluded that they had more associated anomalies than term patients with HSCR.28 The rate of associated malformations reported in this study (34.5%) is fairly high compared with other studies. This could be explained by our broad definition of malformation: at least 1 diagnosis in any of the registers, which could cause an overrepresentation of the patients with associated malformations. Because a high frequency of gastrointestinal malformations was found, these malformations were further investigated. The birth prevalence of congenital malformation in the Swedish general birth population in 2012 was 1.5%; however, that might be underestimated due to lack of insufficient validity of that specific register.29
This study shows a stable birth prevalence of HSCR of 1.91 of 10 000 in Sweden. Maternal obesity and parity may be a risk factor for the child to develop HSCR and affected children were born at a lower gestational age than control subjects.
- Accepted April 20, 2016.
- Address correspondence to Anna Löf Granström, MD, Division of Pediatric Surgery, Astrid Lindgren Children’s Hospital, Q3:03, Karolinska University Hospital, Solna, SE-17176 Stockholm. E-mail:
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: This study was supported by the Foundation Frimurare Barnhuset, Her Royal Highness Crown Princess Lovisa Foundation, and the Sällskapet Barnavård Foundation.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- Moore SW
- Fu M,
- Sato Y,
- Lyons-Warren A, et al
- R Core Team
- Correa A,
- Marcinkevage J
- Goldberg EL
- ↵Socialstyrelsen. Fosterskador och kromosomavvikelser 2012 [Birth defects and chromosomal abnormalities 2012]. Stockholm, Sweden: Sveriges officiella statistik, Hälso- och sjukvård; 2013
- Copyright © 2016 by the American Academy of Pediatrics