Most infants in the UK start solids before the recommended age of 6 months. We assessed the independent effects of solids and breast feeding on the risk of hospitalisation for infection in term, singleton infants in the Millennium Cohort Study (n = 15 980). For both diarrhoea and lower respiratory tract infection (LRTI), the monthly risk of hospitalisation was significantly lower in those receiving breast milk compared with those receiving formula. The monthly risk of hospitalisation was not significantly higher in those who had received solids compared with those not on solids (for diarrhoea, adjusted odds ratio 1.39, 95% CI 0.75 to 2.59; for LRTI, adjusted odds ratio 1.14, 95% CI 0.76 to 1.70), and the risk did not vary significantly according to the age of starting solids.
Statistics from Altmetric.com
The World Health Organization recommends that all infants are exclusively breast fed for 6 months. In the UK 2005 Infant Feeding Survey, <1% of babies were exclusively breast fed for 6 months and 51% had started solids by 4 months.1 A possible adverse effect of introducing solids too early is the vulnerability of the gut to infection. The high permeability of the immature gut may permit large foreign proteins to penetrate and provoke immune sensitisation.2 There is also a risk of infection through contaminated foods or feeding equipment. However, evidence on these possible adverse effects from developed countries is scarce.3 We hypothesised that (i) there will be a greater risk of infection associated with consuming solids or with introducing solids too early, and (ii) that any adverse effect of solids might be different in infants on breast milk compared with those on formula. Here we assess the independent effects of solids and breast feeding on the risk of hospitalisation for infection in the UK Millennium Cohort Study (MCS).
The study methods have been reported elsewhere.4 In brief, the MCS is a nationally representative UK longitudinal study of 18 819 infants born in the UK between 2000 and 2002 who were alive and living in the UK at age 9 months. Ethical approval for the MCS was granted from the Multi-centre Research Ethics Committee. Information was collected on socio-economic and health factors from a parental interview when infants were aged 9 months. Hospitalised morbidity was assessed by the reported age and diagnosis at the time of any hospital admissions since birth. Diarrhoea was defined as “gastroenteritis” (n = 201) and lower respiratory tract infection (LRTI) as “chest infection or pneumonia” (n = 552).
The analysis focused on hospitalisation in the first 8 months after birth in term, singleton infants who did not have major problems at birth (n = 15 980 infants). Breastfeeding duration and the introduction of other milk and solids were estimated from interview responses about the age of the infant when last given breast milk, and when first given formula, other types of milk and solids. Infants were categorised separately according to whether they were on solids (yes, no) and the type of milk they were receiving (formula only, breast milk and formula, breast milk only).
The data for each infant were divided into 1-month age bands in order to incorporate time-changing variables as follows. (i) For the analysis of having received solids that month, all data from months 1–8 were included. The monthly risk of hospitalisation was estimated according to whether the infant had received solids in that month (a time-dependent variable), the type of milk they had received in that month (a time-dependent variable) and potential confounders (including age). (ii) For the analysis of age at the introduction of solids, the outcome was the risk of hospitalisation at any time after the introduction of solids. Hence, data were included from the month in which the infant started solids until they were 8 months old. The monthly risk of hospitalisation was estimated according to the month at which they started solids (a fixed variable), the type of milk they received in that month (a time-dependent variable) and potential confounders (including age).
For both of these analyses, logistic regression was used to estimate odds ratios adjusted for potential confounders, and to test whether there was an interaction between milk and solids. For categorical variables, the largest group was chosen as the baseline. Adjustment was made for age and confounders that were significantly (p<0.05) associated with the outcomes (tables 1 and 2). All confidence intervals presented are adjusted for the clustered survey design. All proportions and odds ratios are weighted to allow for the stratified survey design.
The characteristics of the infants have been reported elsewhere.4 Among the 15 980 infants included, 1.1% were hospitalised in the first 8 months for diarrhoea and 3.2% for LRTI. Overall, 25% were still breast fed at 6 months and the mean (median) age of introducing solids was 3.8 (4) months. Infants who were ever breast fed tended to start solids about 1 week later (ie, a mean of 0.20 months later, 95% CI 0.16 to 0.25) than those who were never breast fed (p<0.001).
Effect of having received solids that month
For both diarrhoea and LRTI, the monthly risk of hospitalisation was significantly lower in those receiving breast milk only compared with those receiving formula only, and this was independent of whether or not the infants had received solids (table 1). The monthly risk of hospitalisation for diarrhoea was significantly higher in those who had received solids compared with those not on solids (unadjusted odds ratio 1.57, 95% CI 1.05 to 2.33), but this was effect was smaller and not statistically significant after adjustment for milk group (adjusted odds ratio 1.34, 95% CI 0.90 to 2.00) and after further adjustment for confounders (table 1). The monthly risk of hospitalisation for LRTI did not differ between those on solids and those not on solids (table 1). The odds ratio for solids was the same in all three milk groups (ie, there was no interaction). In order to test whether the effect of solids might be stronger in younger infants, we repeated the analysis restricting the data to infants aged ⩽4 months and obtained similar results.
Effect of age at the introduction of solids
For both diarrhoea and LRTI, the monthly risk of hospitalisation following the introduction of solids was significantly lower in those receiving breast milk only compared with those receiving formula only, and this was independent of the age the infants had started solids (table 2). The age of starting solids was not significantly associated with diarrhoea or LRTI.
Infants aged ⩽8 months who are receiving breast milk as their only source of milk are at a lower risk of hospitalisation for diarrhoea and LRTI compared with those on formula only. This reduction in risk does not appear to be altered by the introduction of solids, and there is no increased risk of hospitalisation associated with solids per se, or with introducing solids at a relatively early age.
Our large, longitudinal study identified fixed and time-dependent risk factors for hospitalisation for infection. The fact that formula milk and other variables were strong and significant risk factors suggests that we had the power to detect an effect of solids, if one existed. Our findings for diarrhoea support those from a Scottish study which found no effect of the early introduction of solids.5 The same study observed an increased risk of respiratory illness in those who started solids before 8 weeks compared with those who started solids after 12 weeks. This was the only study of infant infection in a developed country which was identified from a systematic review of the effect of age of the introduction on solids.3
The optimal age for the introduction of solids should be derived using evidence on all the potential harms and benefits, and these may vary according to the setting (for example, developed and developing countries), the type of milk (breast only, breast and formula, formula only), and the needs of the individual infant. In our large UK study, the strongest risk factor for hospitalisation for infection was formula milk; the introduction of solids was not associated with an increased risk in either breastfed or formula fed infants. Our study only examined hospitalisation following infection in infancy; there may or may not be effects of early or delayed weaning later in childhood. Our findings should be considered together with the effect of solids on other outcomes.
See Editorial,, p 79
Funding: The National Perinatal Epidemiology Unit is funded by the Department of Health in England.
Competing interests: None.
Ethics approval: Ethical approval for the MCS was granted from the Multi-centre Research Ethics Committee.
Disclaimer: The views expressed in this paper are those of the authors and do not necessarily reflect the views of the Department of Health.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.