A large and very well organised study from Denmark has shown that there is a significantly increased risk of infantile hypertrophic pyloric stenosis (IHPS) in infants exposed to macrolide antibiotics.

Data were collected over 16 years on nearly one million (999,378) singleton liveborn babies. Overall, 30,091 women (3%) took macrolides during pregnancy and 21,557 (2.2%) after birth. A total of 6,591 infants (0.6%) were given macrolide treatment directly.

Prenatal exposure was categorised as before or after 27 completed weeks’ gestation, and postnatal exposure as days 0-13 and 14-120 after delivery.

IHPS, defined as requiring a pyloromyotomy procedure during the first 120 days after birth, occurred in 880 infants (0.9/1,000 births).

The rate ratios (RR) for IHPS, compared with unexposed infants were:

• RR 29.8 (95% CI:16.4-54.1) for infants exposed to macrolides between days 0 and 13

• RR 3.24 (95% CI:1.20-8.74) in those exposed between days 14 and 120

For macrolide-exposed infants this equated to absolute risk differences of 24.4 (95% CI:13.0-44.1) and 0.65 (95% CI:0.06-2.21) cases per 1,000 exposed infants respectively.

Maternal macrolide use during the third trimester of pregnancy had a very modest effect on IHPS, RR 1.77 (95% CI:0.95-3.31). Exposure prior to 28 weeks had no measurable effect, RR 1.02 (95% CI:0.65-1.59)

Although 95% of women in Denmark breast-feed after delivery, and 75% still do so at four months, no data were available on breast-feeding rates in the women in this study. Macrolide excretion into milk is low, far less than infant therapy doses.

A sub-analysis looking at amoxicillin and penicillin exposure found no association between infant or maternal exposure and increased IHPS risk. Data on individual macrolide antibiotics suggested that the IHPS risk identified was a class effect.

A possible mechanism for the association between macrolide use and IHPS may relate to gut motilin receptors, which are concentrated in the gastric antrum. Macrolides have a prokinetic effect, which may promote muscular spasm and hypertrophy.

In future I shall avoid prescribing macrolides in late pregnancy, the puerperium in breast-feeding mothers, and of course to neonates where alternatives exist.

Dr Chris Barclay

• Lund M, Pasternak B, Davidsen RB et al. Use of macrolides in mother and child and risk of infantile hypertrophic pyloric stenosis: nationwide cohort study. BMJ 2014;348:g1908doi:10.1136/bmj.g1908