Breastmilk feeding and lipoprotein profile in adolescents born preterm: follow-up of a prospective randomised study

Summary

Background

Breastfeeding is associated with reduced cholesterol concentration later in life, but previous studies have not used random assignment of infant diet with prospective follow-up. We tested the hypothesis that breastmilk feeding benefits the lipoprotein profile in adolescents born preterm, in whom randomisation to different diets at birth is feasible.

Methods

926 infants born preterm were randomly assigned in two parallel trials to receive (trial 1) donated banked breastmilk or preterm formula, or (trial 2) standard term formula or preterm formula, as sole diet or as supplements to mother's milk in both trials. We followed up 216 participants at age 13–16 years and measured ratio of low-density to high-density lipoprotein cholesterol (LDL to HDL), ratio of apolipoprotein B to apolipoprotein A-1 (apoB to apoA-1), and concentration of C-reactive protein (CRP; a measure of the inflammatory process associated with atherosclerosis).

Results

Adolescents who had been randomised to banked breastmilk had a lower CRP concentration (p=0·006) and LDL to HDL ratio (mean difference 0·34 [14% lower], 95% Cl −0·67 to −0·01; p=0·04) than those given preterm formula. A greater proportion of human milk intake in infancy was associated with lower ratios of LDL to HDL (p=0·03) and apoB to apoA-1 (p=0·004)—independent of gestation and potential confounding factors—and with lower CRP concentration (p=0·03). CRP concentration correlated with the two lipoprotein ratios (p<0·0001 and p=0·003, respectively).

Interpretation

Our data provide experimental evidence for the long-term benefits of breastmilk feeding on the risk of atherosclerosis.

Introduction

The short-term benefits of breastfeeding are well known. Few studies, however, have addressed its effects on health beyond the first year of life. Breastfeeding has been associated with a protective effect against atherosclerotic cardiovascular disease (CVD)¹ and its risk factors,2, 4 and especially the lipoprotein profile later in life,5, 6, 7, 8 but the evidence is inconclusive.9, 10, 11, 12 These previous studies, like much of the data that support the long-term benefits of human milk, were affected by problems of retrospective design, recall bias, and, most importantly, potential confounding by educational, socio-economic, and lifestyle factors associated with the mother's decision to breastfeed. Clearly, experimental studies are needed, but randomisation to breastfeeding or formula-feeding in human babies is problematic.

Nevertheless, 20 years ago such an experimental approach was possible in infants born preterm. At the time, banked breastmilk (from unrelated donors) was commonly available and randomisation of infants to donor breastmilk or formula was feasible and ethical. Our randomised trial of banked breastmilk versus formula, initiated in 1982, therefore provided a unique opportunity to test the hypothesis that breastmilk consumption permanently affects, or programmes, later risk factors for CVD with a formal experimental approach.13, 14, 15 Our planned trial follow-up allowed us to test the generic hypothesis that consumption of human milk has favourable effects on later risk factors for CVD and the a priori specific hypothesis that breastmilk feeding benefits the lipoprotein profile later in life. We have shown, in the same cohort, that breastmilk consumption reduces later blood pressure.⁴ We now report our findings on the lipoprotein profile, a major and established determinant of atherosclerosis¹⁶ and CVD.¹⁷