EDITOR—The pooled outcomes used in Lucassen et al's systematic review of treatments for infantile colic are not reported with their level of heterogeneity, but the three trials dealing with soy formula milk seem to be heterogeneous as the confidence intervals do not overlap.1 It may therefore be more useful to look at the differences between the three trials rather than pool the results. Excluding the trial of lower methodological quality does not seem to eliminate the heterogeneity.

The trial showing the large effect size (that by Campbell2) is reported as having no babies who were partially or totally breast fed, while in the trial with the least effect (that by Evans et al3) all babies were breast fed. I wonder if this might explain the difference in outcomes.

The same reservation about pooling of the results would apply to the two trials of hypoallergenic formula milk. Although there seems to be no statistical heterogeneity between them, in that by Hill et al two thirds of the infants were breast fed4 and in the other (by Forsyth) none were breast fed.5 Pooling results from diet trials with such diverse levels of breast feeding may not make clinical sense; surely results from breast fed and bottle fed infants would be better analysed separately.

EDITOR—I am concerned that the literature analysis by Lucassen et al may be used to make recommendations for the care of any infant reported to be colicky or to cry persistently.1Indeed, the advice to start a one week trial of hypoallergenic formula is …