In summary, the association between malnutrition and infections, including respiratory infections, seems clear from consistent experience in developing nations. Young children are at the greatest risk, both of severe malnutrition and complicating infections. The cell-mediated immune system is the most affected by protein-calorie malnutrition, but antibody responses are also affected and complement levels are low. Infections with organisms handled by cell-mediated immunity would be the most predictable, but the immunoglobulin responses that are important for opsonization of invading microorganisms may also be impaired. The experience in developing nations has been extrapolated to patients in US hospitals, because hospitalized patients often have one or more abnormal nutritional parameters. However, severe malnutrition of the sort found in children in developing nations is uncommon in hospitalized patients, and the effects of malnutrition on host defenses in adults are likely to be less severe than in children. Whether the degrees of malnutrition that have been described in hospitalized patients produce clinically significant effects on antibacterial defenses in the lungs of adults remains uncertain. Despite the intuitive importance of nutritional support, and the repeated observation that nutritional parameters improve with nutritional support, a number of controlled trials have failed to show a clear improvement in patient outcome with aggressive nutritional therapy, including parenteral hyperalimentation. The results of these studies, together with the risks involved in parenteral alimentation have led some to suggest that "the emperor has no clothes," and that aggressive nutritional support is not worthwhile for most patients. The major problem in interpreting the data is the lack of clear clinical endpoints, and this may obscure potentially important responses to nutritional therapy. Nutritional status is only one of many interacting variables that may affect clinical outcome, particularly in patients in critical care units. Survival usually depends on many factors, particularly the status of major organ systems independent of nutrition, so that survival as an endpoint for nutritional studies is likely to be too insensitive. Prospective studies of the incidence and significance of infections, particularly pneumonia, in malnourished patients and the effects of nutritional therapy are lacking. At present, the prudent approach is to treat infections aggressively in malnourished patients, with antibiotics and drainage if necessary, and to provide nutritional supplementation in all patients via the gut as long as possible.(ABSTRACT TRUNCATED AT 400 WORDS)