Bottle feeding and the sudden infant death syndrome

BMJ 1995; 311 doi: https://doi.org/10.1136/bmj.311.6997.122c (Published 08 July 1995) Cite this as: BMJ 1995;311:122
  1. E A Mitchell,
  2. A W Stewart,
  3. R P K Ford
  1. Associate professor of paediatrics Department of Paediatrics, University of Auckland Medical School, Private Bag 92019, Auckland, New Zealand
  2. Biostatistician Department of Community Health, University of Auckland
  3. Community paediatrician Community Child and Family Services, Christchurch, New Zealand

    EDITOR,--R E Gilbert and colleagues report that bottle feeding is not a significant independent risk factor for the sudden infant death syndrome.1 We believe that they may have made a type 2 error (stating that there is no difference when in fact there is one). They report that the risk of the syndrome in breast fed infants was almost half that seen in bottle fed infants after adjustment for a small number of potential confounders. As the reduction in risk did not reach significance, however, they conclude that bottle feeding was not an independent risk factor.

    The New Zealand cot death study, a large nationwide case-control study (485 cases and 1800 controls), found after adjustment for a wide range of potential confounders, that infants exclusively breast fed had a significantly reduced risk of the sudden infant death syndrome compared with infants who were bottle fed.2 The reduced risk was of a similar magnitude to that reported in Gilbert and colleagues' study. Residual confounding due to social or cultural factors is unlikely to explain the results from the New Zealand study. We controlled for a wide range of potential confounders (selection effect--age of infant, region, season, time; sociodemographic background--marital status, occupation, age at which mother left school, age of mother; pregnancy variables--number of previous pregnancies, attendance at antenatal clinics and education classes; infant factors--sex, ethnic group, birth weight, gestation; and postnatal factors--admission to neonatal unit, maternal smoking, sleep position, and cosleeping). As can be seen, these include a wide range of socioeconomic variables and measures of health behaviour. Although breast feeding rates are much higher in New Zealand than in the United Kingdom,3 the risk factors for not breast feeding are similar in New Zealand4 to those reported from the United Kingdom.3

    It is important to adjust not only for maternal smoking but also for mothers sharing a bed with their infants. Bed sharing is associated with breast feeding4 but increases the risk of the sudden infant death syndrome.2 Adjusting for bed sharing will tend to increase the beneficial effect of breast feeding on the risk of the syndrome (New Zealand cot death study, not exclusively breast feeding at discharge from obstetric hospital; odds ratio after adjustment for confounders, including bed sharing=1.89 (95% confidence interval 1.35 to 2.64)2; odds ratio after adjustment for confounders but excluding bed sharing=1.76 (1.25 to 2.45)). Use of dummies may also be a confounder.

    We conclude that the results of Gilbert and colleagues' study are consistent with those of our study; the very large national maternal and infant health survey,5 which also adjusted for socioeconomic factors and smoking5; and other studies, which show that breast feeding halves the risk of the sudden infant death syndrome.

    The New Zealand cot death prevention programme recommends that mothers should breast feed if possible. As breast feeding rates are lower in the United Kingdom than New Zealand the promotion of breast feeding may be even more important for the United Kingdom. We believe that breast feeding should be included in the “back to sleep” campaign.


    View Abstract

    Log in

    Log in through your institution


    * For online subscription