The hypothesis that early life (pre-adult) factors may have a role in the aetiology of future illness, specifically cardiovascular disease, has attracted substantial research interest, particularly in the last two decades.1 The ecological (group-level) studies of Forsdahl2 were among the first empirical examinations of this association. His analyses of Norwegian counties revealed a positive association between higher infant death rates in 1896–1925 (regarded as a proxy of adverse environmental conditions) and increased rates of death from adult coronary arterial disease in 1964–1967 (ie, when the infants would have been nearing the age of disease onset for that period) (figure 1).

Replicating these findings in England and Wales,3 Barker and his team then extended this area of enquiry to individual-level indicators.4 To do so, they sourced unusually well-kept maternity hospital archives detailing the physical characteristics of mothers and their newborns, including birth weight. Via linkage with national mortality records, and subsequently clinical examination of surviving study members, birth weight was shown to be related to a range of adult cardiovascular outcomes: coronary heart disease, stroke and their risk factors, such as hypertension and type 2 diabetes.5

While this body of work was intriguing, there was doubt, some of which remains, as to the nature of these relationships. Critics noted that the findings were based on non-randomised (observational) data and were therefore subject to the perennial problem of confounding. Thus it was highly plausible that early life factors—maternal cigarette smoking or childhood social disadvantage were once strong candidates—could have led to both low birth weight and subsequent increased adulthood morbidity, so generating a spurious association. However, the findings of most studies suggest that …