The association of erythromycin and infantile hypertrophic pyloric stenosis: causal or coincidental?

The safety profile of erythromycin is notable for the frequent occurrence of intolerable gastrointestinal effects. One of the more serious of these is infantile hypertrophic pyloric stenosis (IHPS). A recent cluster of IHPS cases prompted an epidemiological investigation which identified oral erythromycin chemoprophylaxis of pertussis as the major risk factor. Evidence suggests an association between early postnatal erythromycin exposure and IHPS. There is no substantive evidence of a risk associated with prenatal exposure, with the single published case-control study to date producing negative findings. The epidemiological investigations of the association with early postnatal exposure have reported significantly elevated odds ratios but have a variety of methodological limitations that prevent definitive conclusions being made. Nevertheless, the concordance of findings across studies increases the strength of evidence favouring an association. The prominent gastrokinetic properties of erythromycin have been postulated as the mechanism behind this phenomenon. A comprehensive assessment of this potential adverse effect should consider its biological plausibility in light of known gastrointestinal physiology, its modulation by erythromycin, and the known pathophysiology of IHPS. Gastrointestinal motor activity in the fasted mammal consists of three phases, phase III being large amplitude contractions called migrating motor complexes (MMC) that can be initiated by motilin and erythromycin. The gastrokinetic effects of erythromycin are variable and complex and include effects on the timing, duration, amplitude and distribution of MMCs. It has been speculated that the motilinomimetic effects of erythromycin on antral smooth muscle function, such as the MMC, may mediate the effect via work hypertrophy. Although intuitively plausible and consistent with hypertrophic obstructive changes similar to IHPS observed in hyperplastic rat ileum after artificially induced mechanical obstruction, there is no direct evidence of this phenomenon. Further complicating the association is the limitations of our knowledge about the pathophysiology of IHPS, including numerous genetic abnormalities, increased parietal cell mass, and gastric hyperacidity. The implications of the reported findings with erythromycin on the benefit-risk profiles of newer macrolides and azalides must be considered. The available data on the comparative gastrokinetic properties of macrolides are significant for the potent gastrokinetic properties and its acid degradation products, the marked variation in gastrokinetic properties associated with macrolide ring size, and the requirement for specific glycosidic linkages at the C-3 and C-5 carbons of the macrolide ring. The variation in gastrokinetic properties associated with variations in molecular structure suggests that if the association between erythromycin and IHPS is causal it may not be a class effect.