Prenatal Factors Contribute to the Emergence of Kwashiorkor or Marasmus in Severe Undernutrition: Evidence for the Predictive Adaptation Model

Background Severe acute malnutrition in childhood manifests as oedematous (kwashiorkor, marasmic kwashiorkor) and non-oedematous (marasmus) syndromes with very different prognoses. Kwashiorkor differs from marasmus in the patterns of protein, amino acid and lipid metabolism when patients are acutely ill as well as after rehabilitation to ideal weight for height. Metabolic patterns among marasmic patients define them as metabolically thrifty, while kwashiorkor patients function as metabolically profligate. Such differences might underlie syndromic presentation and prognosis. However, no fundamental explanation exists for these differences in metabolism, nor clinical pictures, given similar exposures to undernutrition. We hypothesized that different developmental trajectories underlie these clinical-metabolic phenotypes: if so this would be strong evidence in support of predictive adaptation model of developmental plasticity. Methodology/Principal Findings We reviewed the records of all children admitted with severe acute malnutrition to the Tropical Metabolism Research Unit Ward of the University Hospital of the West Indies, Kingston, Jamaica during 1962–1992. We used Wellcome criteria to establish the diagnoses of kwashiorkor (n = 391), marasmus (n = 383), and marasmic-kwashiorkor (n = 375). We recorded participants' birth weights, as determined from maternal recall at the time of admission. Those who developed kwashiorkor had 333 g (95% confidence interval 217 to 449, p<0.001) higher mean birthweight than those who developed marasmus. Conclusions/Significance These data are consistent with a model suggesting that plastic mechanisms operative in utero induce potential marasmics to develop with a metabolic physiology more able to adapt to postnatal undernutrition than those of higher birthweight. Given the different mortality risks of these different syndromes, this observation is supportive of the predictive adaptive response hypothesis and is the first empirical demonstration of the advantageous effects of such a response in humans. The study has implications for understanding pathways to obesity and its cardio-metabolic co-morbidities in poor countries and for famine intervention programs.