Many epidemiological studies have now shown a strongly increased risk of developing type 2 diabetes and the metabolic syndrome in adults who as neonates showed signs of poor early (fetal and early postnatal) growth. The thrifty phenotype hypothesis was proposed to provide a conceptual and experimentally testable basis of these relationships. We have used protein restriction of rat dams, as a means to test this hypothesis. In vivo and in vitro studies of the growth-restricted offspring of such pregnancies have provided findings showing remarkable parallels with the human conditions. Permanent changes in the expression of regulatory proteins in liver, muscle and adipose tissue provide at least part of the explanation of the changes observed and offer potential markers for testing in the human context. These studies have also raised the question as to whether 'catch up' growth following early growth retardation may add to the risks posed by this early handicap. Male rats growth-retarded during fetal life and cross-fostered shortly after birth to normal lactating dams reach normal body and organ weights by weaning but have a reduced longevity. This finding raises the possibility that catch up growth, whilst potentially beneficial in the short term, may be detrimental to long-term survival. Human epidemiological studies may point in the same direction. Work by others on other models of early growth restriction have produced similar, although more limited, data. These findings raise the interesting possibility that the response to fetal stress, be it nutritional or other, may evoke a somewhat restricted and uniform pattern of adaptive response.