Results

Colic occurred in 4.7% of all infants and was, with statistical significance, more frequent in younger infants and infants of smoking mothers (table 1). After adjustment for infant age, smoking was only associated with a higher prevalence of colic among completely formula fed children. Obstetric history and demographic and socioeconomic background did not explain these findings. Analyses that used another outcome, crying for more than three hours on only one day during the preceding week, yielded similar results (data not shown).

Analyses restricted to infants aged 1 month, the group with the highest overall prevalence of colic, yielded somewhat larger differences between groups. Odds ratios (OR) (95% confidence intervals, CI) for smoking less than 15 cigarettes per day and for 15 and over, were 2.01 (1.18 to 3.42) and 2.85 (1.41 to 5.74), respectively, compared to infants of non-smoking mothers. However, compared with infants of non-smoking mothers, infants of smoking mothers that were breastfed had a less increased risk than completely formula fed infants: ORs are 1.61 (0.77 to 3.36) and 2.52 (1.50 to 4.25), respectively.

Conclusion

Our study shows that maternal smoking and colic are associated but that breastfeeding weakens this association rather than reinforcing it.

It is unlikely that selective non-response has affected our findings because of the very high response rate. Some parents may not report their smoking, however, because of the warnings against it. This would give an underestimation of the strength of the association. However, another Dutch study that used mailed questionnaires yielded a similar prevalence of maternal smoking in infancy.7

Whether the association that we found can be explained by maternal smoking causing colic, needs additional study. Infantile crying may also have caused more maternal smoking, instead of the reverse. Furthermore, mothers who smoke during the infancy of their offspring may also have smoked during pregnancy, which also harms infant health,5 and might cause colic. Finally, smoking mothers may be more sensitive for crying, thus only reporting more crying, or may be less responsive so that their baby cries for longer.

Our findings regarding maternal smoking and breastfeeding confirm those of Said and coworkers,3 in a much larger sample. In contrast, Matheson and Rivrud only found an association with crying among breastfed infants, but their study concerned a mailed questionnaire survey four months postpartum (78% response) and they defined colic less stringently.2 Finally, Haggart and Giblin reported an overall lack of association, but a reanalysis of their data yields an OR (95% CI) for the association of parental smoking with colic of 1.66 (0.69 to 4.03),4 which fits with our findings.

If maternal smoking really causes colic, this may result in simple, though difficult to implement,5 ,7 advice to prevent colic: stop smoking. However, further research is needed to assess whether this association is indeed causal.