Antioxidants for children with kwashiorkor

Protein energy malnutrition is the most deadly form of malnutrition. It is the primary or associated cause of around half of the nearly 11 million annual deaths among children under five, 30 000 each day.¹ The reasons for this tragedy are quite clearly poverty, underdevelopment, and inequality, yet knowing this does not translate into finding correspondingly obvious or immediate solutions.

Clinically, protein energy malnutrition presents broadly as one of two extremes: the severe loss of body weight of marasmus or the oedematous malnutrition of kwashiorkor. Conceptualisation of protein energy malnutrition in this way has some disadvantages in its simplicity, but it does have practical importance and reflects differences in epidemiology, pathophysiology, treatment, and presumably aetiology. Of the two, kwashiorkor is consistently the more lethal, with a high case fatality rate in many areas, and therefore merits special consideration.

Ever since Cicely Williams described the entity known as kwashiorkor in 1935 while working in west Africa, there has been much and sometimes vigorous debate on its pathogenesis.²,³ Why is it that among children destined to become malnourished some develop kwashiorkor while others develop marasmus? What are the determinants? Remarkably little research has been conducted over the past several decades on this question, despite the enormous practical implications of the answer. Painfully slow progress over the same period in reducing the global prevalence and childhood mortality of malnutrition reflect this lack of evidence.

Several insults have been proposed as principal causes of kwashiorkor including dietary protein deficiency, aflatoxins in food, and depressed cellular protein synthesis reinforced by infection. Nearly 20 years ago Golden et al proposed the free radical theory as a unifying hypothesis, which has become a main focus of conjecture.⁴,⁵ This theory proposes that kwashiorkor arises from excessive noxious insults, resulting in the generation of sufficient reactive oxidative free radicals to exceed the host's antioxidant capacity. An accumulating body of evidence, albeit indirect, lends substantial support for this concept.⁶,⁷ Various studies have shown that children with kwashiorkor, when compared to marasmic or normal controls, have greater concentrations of biomarkers of oxidative stress and damage as well as lower blood concentrations of antioxidants. Furthermore, clinical resolution of kwashiorkor coincides with the return to normal of these markers.

It is in this context that the study of Ciliberto et al in the current issue of the BMJ is most welcome.⁸ This study examined the impact of an antioxidant cocktail as a possible preventive treatment for kwashiorkor in children in a highly endemic area of Malawi and, notably, found no protective effect. Although these results may put a damper on the antioxidant hypothesis, it would be premature to discard the theory altogether. The specific antioxidants used in the Ciliberto study are known to have a high relative antioxidant capacity, but the amounts taken by the children may have been insufficient to overcome high oxidative stress and prevent kwashiorkor. Since neither antioxidant capacity nor oxidative stress was measured, an element of doubt about the adequacy of the antioxidant mix or dose is reasonable. Moreover, the study did not assess the children's HIV status, which may have contributed to or affected their responses to oxidative stress. More than 20% of Malawian children in hospital with kwashiorkor have HIV infection; indeed, kwashiorkor has been described as the final common pathway for children with AIDS.⁹

The aetiology of kwashiorkor is clearly multifactorial and includes in varying proportions food insecurity, inadequate weaning and other feeding practices, infection, and, the current study notwithstanding, possibly oxidative stress. The relative contribution of the elements that create the “perfect storm” of kwashiorkor may not be universal among different groups of children. That said, the results of this new study provide evidence that antioxidants might not play a central decisive role as originally proposed.⁵ They also serve to emphasise that 20 years is far too long to test the leading hypothesis for a condition that affects so many children with such devastating consequences.