During the 1960–1970s we learnt how artificial feeding could rescue a few severely undernourished patients. Next, research characterised the integrative biochemistry of illness and sought ways by which the catabolic responses could be influenced nutritionally and hormonally. More recently the large proportion of hospital patients suffering from undernutrition, as judged by both weight for height and by micronutrient status, has become widely appreciated and there has been an interest in more routine supplementation of hospital patients. As the practice of supplementation increases there has been a growing need for randomised controlled trials to assess its clinical effectiveness. But applying rigorous clinical endpoints such as mortality, morbidity, or duration of stay in randomised controlled trials of a secondary treatment such as nutritional support is difficult because the endpoints are easily confounded by primary treatments and other factors.

The physiological effects of undernutrition in the otherwise normal human have been well known for about 50 years.1 As far as sip supplements are concerned there is now excellent evidence that they do not simply replace the intake of normal food but effectively increase nutrient intake and improve nutritional …