*Department of Clinical Sciences, Unit for Social Epidemiology
†Department of Clinical Sciences, Unit for Diabetes and Celiac Disease, Lund University, Malmo, Sweden.
Address correspondence and reprint requests to Carl J. Wingren, MD, Unit for Social Epidemiology CRC, Jan Waldenströmsgata 35, Skåne University Hospital, SE-20502 Malmö, Sweden (e-mail: email@example.com).
Received 1 June, 2012
Accepted 16 July, 2012
Supplemental digital content is available for this article. Direct URL citations appear in the printed text, and links to the digital files are provided in the HTML text of this article on the journal's Web site (www.jpgn.org).
This work was financially supported by the Centre for Economic Demography at Lund University, the Swedish Council for Working Life and Social Research (FAS) (Dnr 2010-0402, PI Juan Merlo), and the Swedish Research Council (VR) (Dnr K2011-69X-15377-07-6, PI Juan Merlo).
The authors report no conflicts of interest.
Congenital anomalies and chromosomal aberrations have previously been associated with celiac disease (CD) (1–8). Especially well documented is the increased risk of CD in populations with Down or Turner syndrome (2,7); however, less is known about the risk of CD in children having other congenital anomalies (6). We set out to investigate the association between congenital anomalies and CD before the age of 6 years, using population-based registers in Sweden.
We included all of the singleton children (792,401) born in Sweden between 1987 and 1993, identified using the Swedish Medical Birth Registry. Statistics Sweden and the Centre for Epidemiology at the Swedish National Board of Health and Welfare performed the record linkage using the personal identification number assigned to each person residing in Sweden. Before the data were delivered to us, these personal identifiers were encrypted by the Swedish authorities to ensure the anonymity of the subjects. The study was approved by the regional ethics review board in southern Sweden.
Assessment of Variables
We identified cases with CD before 6 years of age in the Swedish National Inpatient Registry, using the Swedish version of the International Classification of Diseases (ICD), 9th (579.0; ICD-9) or 10th (K90.0; ICD-10) editions (9).
We categorized congenital anomalies according to chapter XIV, “congenital anomalies,” in the Swedish version of ICD-9 and considered each sections defined by the first 3 digits (740–759) as separate entities (9). We identified cases of anomalies using the Swedish Medical Birth Registry as well as the Swedish National Inpatient Registry, in which we considered diagnoses of anomalies before the child was 3 months of age.
We defined being born small for gestational age (SGA) as a birth weight (grams) below 2 standard deviations from the expected mean birth weight for that gestational age (days). As described previously (10), the mean birth weight for each gestational day was calculated using a sex-specific formula previously derived from ultrasonic intrauterine measurements of healthy fetuses in uncomplicated pregnancies (10). We used the non-SGA children as reference in the analysis.
We included maternal age at time of birth of the child according to the following categories: younger than 20 years; 20 to 24 years; 25 to 29 years; 30 to 34 years; 35 to 39 years; and older than 39 years.
Statistical and Epidemiological Analysis
We investigate the association between having a congenital anomaly and CD before the age 6 years, using simple Cox regression models. The sections of congenital anomalies (ICD-9 codes 740–759) associated with CD were also investigated in strata of chromosomal abnormalities (ICD-9 code 758) and SGA. We also applied sibling design models (Cox regression models stratified by maternal identification number) to investigate the association between anomalies and CD in siblings discordant for the congenital anomaly. To some extent, a sibling design model adjusts for shared genetics and shared environmental exposures such as maternal CD (11,12).
We estimated hazard ratios (HR) and 95% confidence intervals (CI) using Cox regression models. We performed the analyses using SPSS version 20.0.0 (SPSS Inc, Chicago, IL).
A total of 792,401 children were included, of whom 0.37% (2859) were diagnosed as having CD before 6 years of age. In Table 1, we present the rates of the congenital anomalies in the population according to CD status.
We present the results of the simple Cox regression analyses modeling the association between congenital anomalies and CD in Table 2. We observed that the risk of CD was higher in those children affected by anomalies of the ear, face, and neck (ICD-9 code 744); heart (ICD-9 codes 745 and 746); the digestive tract below the level of the stomach (ICD-9 code 751); and chromosomes (ICD-9 code 758) (Table 2). Adjusting the models for maternal age had only a slight effect on the estimates and did not affect the conclusions made (results not shown).
We tested the associations of the anomalies conclusively associated with CD by applying sibling designs, as well as by stratifying the analyses according to the presence or not of a chromosomal abnormality and by SGA status (Table 3). In the sibling design model, the association between CD and anomalies of the heart (ICD-9 code 745 [bulbus cordis and septal closure]; HR 3.27, 95% CI 1.38–7.72) and anomalies of chromosomes (ICD-9 code 758; HR 8.75, 95% CI 2.61–29.3) persisted. Conversely, anomalies of the ear, face, and neck (ICD-9 code 744); other anomalies of the heart (ICD-9 code 746); and other anomalies of the digestive tract (ICD-9 code 751) indicated positive associations with CD in the sibling analyses, however inconclusively (Table 3).
We also analyzed the associations between congenital anomalies and CD separately in boys and girls (see supplemental Table 4, http://links.lww.com/MPG/A158) because girls have an increased risk of CD compared with boys (13). Generally, it appears that anomalies are more important as risk factors for CD in boys compared with girls. In fact, the associations between congenital anomalies of the ear, face, and neck (ICD-9 code 744); other anomalies of the heart (ICD-9 code 746); and anomalies of the lower digestive tract (ICD-9 code 751) and CD are inconclusive in girls, although the point estimates indicate a positive effect (see supplemental Table 4, http://links.lww.com/MPG/A158).
The sections in ICD-9 chapter XIV “congenital anomalies” conclusively associated with CD in the simple Cox regression models (Table 2) were further subdivided into the 4-digit level, (the fourth digit is a letter in the Swedish version). We investigated each of these malformation diagnoses in a simple Cox regression model for association with CD before 6 years of age (see supplemental Table 5, http://links.lww.com/MPG/A159).
We observed that congenital anomalies of the heart, the digestive tract, ear, face, and neck, as well as chromosomal anomalies, were conclusively associated with CD before 6 years of age in conventional population-based Cox regression models; however, in sibling design models, only anomalies of the heart (ICD-9 code 745: bulbus cordis and septal closure) and chromosomal anomalies (ICD-9 code 758) remained conclusively associated with CD. Our results support previous observations that congenital anomalies in the child, especially concerning the heart and chromosomes, are associated with childhood CD (2,6–8).
Furthermore, the sex of the child seemed to condition the association between congenital anomalies and CD because a malformation appears to be a stronger risk factor for CD in boys compared with girls (see supplemental tables, http://links.lww.com/MPG/A158, http://links.lww.com/MPG/A159). This is in line with previous findings that SGA and low socioeconomic position seem to be more important as risk factors for CD in boys (1,14).
Our study has several strengths because it encompasses a large number of children and uses population-based registries with prospectively collected information. Also, we applied sibling designs, which are more appropriate when investigating causal effects (11,12); however, the study also experiences some limitations, such as using the national inpatient registry to identify cases of CD. This approach may identify those cases having more overt disease; we are not able to determine the criteria applied to set the diagnosis of CD. Also, we have no information about breast-feeding duration and timing of gluten introduction. Despite a large population of children, our analyses experience low statistical power. Another limitation is the approach of obtaining cases with congenital anomalies using the medical birth registry in combination with the national inpatient registry, to some extent discussed elsewhere (15).
Several explanations for an association between congenital anomalies and CD exist. Previously, it has been suggested that maternal CD increases the risk of anomalies in the offspring through mechanisms of malnutrition (16,17); however, results have been contradictory (18,19). In any case, if so, the association between anomalies in the child and later CD could be caused by maternal CD because first-degree relatives of patients with CD have an increased risk of CD (20). In our study this explanation seems less likely in the case of cardiac anomalies (ICD-9 code 745) because the risk of CD associated with the malformation persists in the sibling design model that by design adjusts for maternal CD (Table 3). Besides, congenital anomalies may condition other exposures such as a shorter duration of breast-feeding or infections during childhood that may mediate the increased risk of CD (6,21). Moreover, intrauterine infections are recognized for the potential to induce anomalies in the growing fetus, and we could speculate on intrauterine infections in CD risk (22); however, since a large proportion of cases with CD in fact are undiagnosed, we must consider diagnostic bias as an explanation for our results (23). Screening for CD may be more liberal in children experiencing a congenital anomaly; alternatively, the malformation may mimic symptoms of CD (eg, failure to thrive, diarrhea) or even require a gastroscopy during the diagnostic workup.
In conclusion, we observed an increased risk of CD by the age of 6 years in children having a congenital anomaly in either of the cardiac system, the ear, face, and neck, or in parts of the digestive system (below the level of the stomach), as well as in children experiencing chromosomal anomalies; however, whether congenital anomalies are in the causal pathway for CD is still unknown.
1. Wingren CJ, Agardh D, Merlo J. Revisiting the risk of celiac disease in children born small for gestational age: a sibling design perspective. Scand J Gastroenterol
2. Mortensen KH, Cleemann L, Hjerrild BE, et al. Increased prevalence of autoimmunity in Turner syndrome—influence of age. Clin Exp Immunol
3. Bettendorf M, Doerr HG, Hauffa BP, et al. Prevalence of autoantibodies associated with thyroid and celiac disease in Ullrich-Turner syndrome in relation to adult height after growth hormone treatment. J Pediatr Endocrinol Metab
4. Digilio MC, Giannotti A, Castro M, et al. Screening for celiac disease in patients with deletion 22q11.2 (DiGeorge/velo-cardio-facial syndrome). Am J Med Genet A
5. Giannotti A, Tiberio G, Castro M, et al. Coeliac disease in Williams syndrome. J Med Genet
6. Congdon PJ, Fiddler GI, Littlewood JM, et al. Coeliac disease associated with congenital heart disease. Arch Dis Child
7. Nowak TV, Ghishan FK, Schulze-Delrieu K. Celiac sprue in Down's syndrome: considerations on a pathogenetic link. Am J Gastroenterol
8. Hilhorst MI, Brink M, Wauters EA, et al. Down syndrome and coeliac disease: five new cases with a review of the literature. Eur J Pediatr
9. Klassifikation av sjukdomar 1987. Systematisk förteckning (in Swedish) [Classification of diseases 1987, Systematic listing]. Stockholm, Sweden: The National Board of Health and Welfare; 1986.
10. Marsal K, Persson PH, Larsen T, et al. Intrauterine growth curves based on ultrasonically estimated foetal weights. Acta Paediatr
11. Donovan SJ, Susser E. Commentary: advent of sibling designs. Int J Epidemiol
12. Strully K, Mishra G. Theoretical underpinning for the use of siblings studies in life course epidemiology. In: Lawlor DA, Gira D, eds. Family Matters Designing, Analysing and Understanding Family-based Studies in Life-Course Epidemiology
. New York: Oxford University Press; 2009:39–56.
13. Ivarsson A, Persson LA, Nystrom L, et al. The Swedish coeliac disease epidemic with a prevailing twofold higher risk in girls compared to boys may reflect gender specific risk factors. Eur J Epidemiol
14. Wingren CJ, Bjorck S, Lynch KF, et al. Coeliac disease in children: a social epidemiological study in Sweden. Acta Paediatr
15. Registration of Congenital Malformations in the Swedish Health Registers
. Stockholm, Sweden: The National Board of Health and Welfare, Centre for Epidemiology; 2004.
16. Hozyasz KK. Coeliac disease and birth defects in offspring. Gut
17. Arakeri G, Arali V, Brennan PA. Cleft lip and palate: an adverse pregnancy outcome due to undiagnosed maternal and paternal coeliac disease. Med Hypotheses
18. de Almeida RC, Lima BO, Castro LC, et al. Maternal celiac disease: improbable risk factor for neural tube defect. Eur J Gastroenterol Hepatol
19. Dickey W, Stewart F, Nelson J, et al. Screening for coeliac disease as a possible maternal risk factor for neural tube defect. Clin Genet
20. Dube C, Rostom A, Sy R, et al. The prevalence of celiac disease in average-risk and at-risk Western European populations: a systematic review. Gastroenterology
21. Sandberg-Bennich S, Dahlquist G, Kallen B. Coeliac disease is associated with intrauterine growth and neonatal infections. Acta Paediatr
22. Gilbert-Barness E. Teratogenic causes of malformations. Ann Clin Lab Sci
23. Myleus A, Ivarsson A, Webb C, et al. Celiac disease revealed in 3% of Swedish 12-year-olds born during an epidemic. J Pediatr Gastroenterol Nutr