"Interest in all forms of eating disorder in children, adolescents and adults continues to grow steadily across the world. The number of publications grows from day to day but, despite the increasing body of knowledge, the causes and origins of eating disorders are as abffling and obscure as ever. Death from anorexia nervosa (AN) was recorded by both Lasegue (1873a,b) and Gull (1874), who first described it. In 1895, both Stephens and Marshall published postmortem finding in the Lancet for a 16-year-old and an 11-year-old patient respectively. In the last five decades there have been several reports of cases of AN with fatal outcomes and subsequent ausopsies (Siebenmann, 1955; Martin, 1955, 1958; Hack, 1959; Mosli, 1967; Chikasue et al., 1988). Gradually work began to be published which covered long observation periods and produced yet more mortality figures. In 1988 in Britain, Patton presented a study of a group of 460 consecutive patients with eating disorders covering the years from 1971-1981. These were divided into two groups for AN and bulimia nervosa (BN) with resulting crude mortality rates of 3.3 percent for the former and 3.1 percent for tha latter. He also carried out a critical evaluation of the methodological problems and the results obtained from earlier studies. Hsu et al. (1979) reported that more than 2 percent had died during an average follow-up period of 5.9 years; Isafer et al. (1985) gave a crude mortality rate of 8.2 percent with an average follow-up period of 12.5 years; and Theander (1985) a crude morality rate of 18 percent over 33 years. Even allowing for the differences in these data, it is clear that AN has the highest mortality rate of all the psychiatric illnesses (Licht et al., 1993). However these differences in crude mortality rates are unsatisfactory from a methodological point of view for a number of reasons. These include the way cases were selected and the differing lengths of the observation periods."