Death, survival and recovery in anorexia nervosa: a thirty-five year study

INTRODUCTION An excess of premature deaths at long-term followup, in people who have presented to secondary/tertiary referral centres over the last sixty years or so with diagnoses of anorexia nervosa, is well documented. At 20þyears of follow-up, it often stands at around 10–20% of those first presenting with a mean age of around 20 years (Kay 1953; Ratnasuriya, Eisler, Szmuckler, & Russell, 1991; Seidensticker & Tzagournis 1968; Theander, 1985). In two other such series (Crisp, Callender, Halek, & Hsu, 1992), clinically comparable to the others at presentation, death at 20-year follow-up had arisen in only 4% of the anorectic population, no more than expected within the general population over the same time span. Nielsen et al. (1998) suggested that this might be due to the treatment package delivered to the majority of that population. Subsequent deaths of patients presenting in this way and within this time scale are predominantly reported as being due to malnutrition or frank suicide. Nearly seventy years ago Crichton-Miller (1938) construed anorexia nervosa itself as ‘a compromise with suicide’ and, to some other clinicians, the disorder appears to be ‘a form of psychic suicide’ (Brill, 1939) and primarily self-destructive (e.g. Strauss 1956). Simply searching national registers for deaths recorded as due to or contributed to by anorexia nervosa, is a seriously faulty procedure. In the United Kingdom, nearly 30 years ago, two of us pressed Government to increase service provision for people with anorexia nervosa. Government responded by claiming that the national register of deaths revealed between only 21 and 26 such deaths per year over the previous five years. Government concluded that it was not a seriously crippling or fatal condition. But it is. Anorexia nervosa, egosyntonic and often marked by secrecy, like alcoholism (Bell & Cremona, 1989; Maxwell & Knapman, 1985), was almost certainly frequently overlooked as a background cause to death during the period in question (Crisp, 1978). In the case of anorexia nervosa such deaths are often due to malnutrition or suicide. Meanwhile, mislabelling of malnutrition in terminal illness in the elderly as anorexia nervosa, occasionally introduces false positives