The prevention of child mortality is a commonly stated health goal in developing countries and the target of much international assistance in the health sector. Over the past decade the primary strategy for accelerating the reduction in child mortality has been the dissemination of simple, low-cost technologies, such as immunization, oral rehydration therapy and antibiotics, that target specific diseases (Huffmann and Steel 1994). This is done despite the knowledge that malnutrition and disease have a synergistic relationship (Scrimshaw et al. 1968) and that the optimal strategy may involve a combination of health and nutrition interventions. In the 1970s, for instance, it was estimated that malnutrition (notably protein-energy malnutrition--PEM) was the underlying or contributing cause of death for roughly half of all deaths to children aged 1-4 years in several Latin American countries (Puffer and Serrano 1973). Apart from this early study, however, there has been little effort to quantify the contribution of malnutrition to child mortality in other regions of the world in ways which are meaningful to policy. This paper reviews the results of 28 community-based, prospective studies, in 12 Asian and Sub-Saharan African countries, which examined the relationship between anthropometric indicators of malnutrition and child mortality. One purpose is to estimate the contribution of malnutrition to child mortality--distinguishing the effects of severe malnutrition from mild-to-moderate malnutrition--and to examine a number of related issues relevant to policy, programs and research in this area. The accumulated results are consistent in showing that the risk of mortality is inversely related to anthropometric indicators of nutritional status and that there is elevated risk even in the mild-to-moderate range of malnutrition. This latter result contradicts the findings from an earlier, landmark study which suggested that mild-to-moderate malnutrition was not associated with an increased risk of mortality (Chen et al. 1980). The present results indicate that somewhere between 20% and 75% of child deaths are statistically attributable to anthropometric deficits, with most estimates falling in the range 25-50%. When taking account of the relative proportions of severe versus mild-to-moderate malnutrition in the population, the results show further than 16-80% of all nutrition-related deaths are associated with mild-to-moderate malnutrition rather than severe malnutrition. In most studies 46-80% of all nutrition-related deaths are in the mild-to-moderate category.(ABSTRACT TRUNCATED AT 400 WORDS)