|Home | About | Journals | Submit | Contact Us | Français|
Contributors: HTS had the idea for the study. He discussed the idea and study design with all coauthors. BN, HL, LP, and SPJ retrieved and validated the data, and HTS and LP did the statistical analysis. HTS wrote the first draft, and all authors contributed to the final paper. HTS is the guarantor.
Infantile hypertrophic pyloric stenosis affects 0.5 to 3.0 per 1000 live births.1,2 The aetiology is unclear, but family history and sex distribution seem important and might indicate a genetic role.3 However, a recently reported fall in the incidence of infantile hypertrophic pyloric stenosis suggests that environmental factors are also important.4 During the last decade a similar reduction in the proportion of pregnant women who smoke was reported in Denmark,5 raising the possibility that maternal smoking is a risk factor for infantile hypertrophic pyloric stenosis. This hypothesis is supported by the reported ecological association between infantile hypertrophic pyloric stenosis and sudden infant death syndrome.1 Maternal smoking is also a risk factor for sudden infant death syndrome,1 and we therefore examined whether maternal smoking was a risk factor for infantile hypertrophic pyloric stenosis in a population based Danish cohort.
We conducted the study between 1 January 1991 and 31 December 2000 in the Danish county of North Jutland and included data on all women who delivered a live infant after 28 weeks' gestation. The county birth registry contains information on all births since 1 January 1991. The main data comprise maternal age, self reported smoking status at the first visit to the midwife, birth order, gestational age, length and weight of neonates at birth, civil status, and civil registry numbers for both mother and child.
We identified all cases of infantile hypertrophic pyloric stenosis from the county hospital discharge registry, which contains data on all discharges from the hospitals since 1977, the civil registry number, dates of admission and discharge, surgical procedures performed, and up to 20 diagnoses classified according to the international classification of diseases (ICD-8 until the end of 1993 and then ICD-10). The codes for infantile hypertrophic pyloric stenosis were 750.19 (ICD-8) and Q40.0 (ICD-10). The civil registry numbers were used to link the records in both registries.
We used SAS version 8.02 for logistic regression. We estimated the risk of infantile hypertrophic pyloric stenosis among infants born to smoking mothers and non-smoking mothers adjusted for maternal age, civil status, birth order, and sex of the child.
There were 57996 births during the study, and 16725 (28.8%) mothers smoked. We identified 78 cases of infantile hypertrophic pyloric stenosis, yielding a cumulative incidence of 1.3%. Thirty five cases (0.2%) occurred among maternal smokers compared with 43 (0.1%) among maternal non-smokers, giving a relative risk of 2.0. The adjusted odds ratio was also 2.0 (95% confidence interval 1.3 to 3.1, table). The correlation between the proportion of smokers and annual incidence of infantile hypertrophic pyloric stenosis per birth year was 0.65 (Spearman's τ=0.65, P=0.04). The highest incidence was 0.22% in 1991 (33% were smokers), the lowest 0.05% in 1997 (28% smokers).
Our data suggest that maternal smoking is a risk factor for infantile hypertrophic pyloric stenosis. We cannot determine whether the association is caused by smoking during pregnancy or postnatally—that is, through breast milk or passive smoking. We were able to adjust for some risk factors, but confounding by unknown factors associated with smoking cannot be ruled out.
The strengths of our study are the population based design and the prospective registration of smoking status and infantile hypertrophic pyloric stenosis. In addition, the data were complete and collected independently of the hypothesis. The incidence corresponds closely with that previously reported.5 Any misclassification of smoking status and infantile hypertrophic pyloric stenosis will bias the risk estimates towards unity.
Funding: Western Danish Research Forum for Health Sciences.
Competing interests: None declared.