Hospital case-fatality rates for severe malnutrition in the developing world remain high, particularly in Africa where they have not changed much over recent decades. In an effort to improve case management, WHO has developed treatment guidelines. The aim of this review is to critically appraise the evidence for the guidelines and review important recent advances in the management of severe malnutrition. We conclude that not only is the evidence base deficient, but also the external generalisability of even good-quality studies is seriously compromised by the great variability in clinical practice between regions and types of health facilities in the developing world, which is much greater than between developed countries. The diagnosis of severe wasting is complicated by the dramatic change in reference standards (from CDC/WHO 1978 to CDC 2000 in EpiNut) and also by difficulties in accurate measurement of length. Although following treatment guidelines has resulted in improved outcomes, there is evidence against the statement that case-fatality rates (particularly in African hospitals) can be reduced below 5% and that higher rates are proof of poor practice, because there is wide variation in severity of illness factors. The practice of prolonged hospital treatment of severe malnutrition until wasting and/or oedema has resolved is being replaced by shorter hospital stays combined with outpatient or community follow-up because of advances in dietary management outside of hospital.