Towards explaining the Swedish epidemic of celiac disease : an epidemiological approach

Sammanfattning: Background: Celiac disease occurs worldwide in approximately 1% of the population, whereof the majority of cases are undiagnosed. Sweden experienced an epidemic (1984-1996) of clinically detected celiac disease in children below 2 years of age, partly attributed to changes in infant feeding. Whether the epidemic constituted a change in disease occurrence and/or a shift in the proportion of diagnosed cases remains unknown. Moreover, the cause of the epidemic is not fully understood.Objective: To increase the knowledge regarding the occurrence of celiac disease in Sweden, with focus on the epidemic period and thereafter, as well as the etiology of celiac disease in general, by investigating the Swedish epidemic and its potential causes.Methods: We performed a two-phased cross-sectional multicenter screening study investigating the total prevalence, including both clinically- and screening-detected cases, of celiac disease in 2 birth cohorts of 12-year-olds (n=13 279): 1 of the epidemic period (1993) and 1 of the post-epidemic period (1997). The screening strategy entailed serological markers analyses, with subsequent small intestinal biopsy when values were positive. Diagnosis was ascertained in clinical cases detected prior to screening. Infant feeding practices in the cohorts were ascertained via questionnaires. An ecological approach combined with an incident case-referent study (475 cases, 950 referents) performed during the epidemic were used for investigating environmental- and lifestyle factors other than infant feeding. Exposure information was obtained via register data, a questionnaire, and child health clinic records. All studies utilized the National Swedish Childhood Celiac Disease Register.Results: The total prevalences of celiac disease were 2.9% and 2.2% for the 1993 and 1997 cohorts, respectively, with 2/3 cases unrecognized prior to screening. Children born in 1997 had a significantly lower celiac disease prevalence compared to those born in 1993 (prevalence ratio, 0.75; 95% confidence interval [CI], 0.60-0.93). The cohorts differed in infant feeding; more specifically in the proportion of infants introduced to dietary gluten in small amounts during ongoing breastfeeding. Of the environmental and lifestyle factors investigated, no additional changes over time coincided with the epidemic. Early vaccinations within the Swedish program were not risk factors for celiac disease. Early infections (≥3 parental-reported episodes) were associated with increased risk for celiac disease (adjusted odds ratio [OR] 1.5; 95% CI, 1.1-2.0), a risk that increased synergistically if, in addition to having ≥3 infectious episodes, the child was introduced to gluten in large amounts, compared to small or medium amounts, after breastfeeding was discontinued (OR 5.6; 95% CI, 3.1-10). Early infections probably made a minor contribution to the Swedish epidemic through the synergistic effect with gluten, which changed concurrently. In total, approximately 48% of the epidemic could be explained by infant feeding and early infections.Conclusion: Celiac disease is both unexpectedly prevalent and mainly undiagnosed in Swedish children. Although the cause of the epidemic is still not fully understood, the significant difference in prevalence between the 2 cohorts indicates that the epidemic constituted a change in disease occurrence, and importantly, corroborates that celiac disease can be avoided in some children, at least up to 12 years of age. Our findings suggest that infant feeding and early infections, but not early vaccinations, have a causal role in the celiac disease etiology and that the infant feeding practice – gradually introducing gluten-containing foods from 4 months of age, preferably during ongoing breastfeeding – is favorable.