Patients With Eating Disorders Have an Elevated Rate of Death



Individuals who have eating disorders have an elevated mortality rate, especially those with anorexia nervosa (AN), according to a meta-analysis of previous studies, reported in the July issue of Archives of General Psychiatry, one of the JAMA/Archives journals.

According to background information in the article, the majority of studies of eating disorder–related death rates focus on AN. Some research of bulimia nervosa (BN) implies that death rates are low for this condition. Moreover, there is limited information about the likelihood of death with eating disorders not otherwise specified (EDNOS), disordered eating that does not meet the criteria for AN or BN, say the authors: “Despite EDNOS being a common presentation in eating disorders services, few published data exist regarding mortality rates in patients given this diagnosis.” In this study, the investigators sought to determine the death rates (and the factors thereof) for each of these types of eating disorders.

Jon Arcelus, L.M.S., M.Sc., M.R.C.Psych., Ph.D., from Leicester General Hospital in Leicester, England, and colleagues systematically searched, assessed and analyzed studies related to eating disorder death rates. They included English-language, peer reviewed articles published between January 1966 and September 2010 from a variety of scientific-literature databases and collections. The researchers examined 36 studies; almost all included AN, one-third included BN, and one-fifth included EDNOS. The studies included 17,272 different patients with eating disorders and reported a total of 755 deaths.

The total number of person-years represented by the studies was 166,642 for AN; 32,798 for BN; and 22,644 for EDNOS. For each 1,000 person-years, the AN group experienced 5.1 deaths (1.3 of which were from suicide), the BN group experienced 1.7 deaths, and the EDNOS group experienced 3.3 deaths. The standardized mortality ratio (the number of actual deaths compared with the number of expected deaths) was 5.86 for AN, 1.93 for BN and 1.92 for EDNOS. The age at which a patient presented for treatment was found to correlate with AN deaths, with patients in their late teens and 20s having a higher death rate than younger patients or those in their 30s.

The authors acknowledge that some of the deaths included in the studies may be due to factors other than an eating disorder. However, they found that death rates for eating disorders—especially AN—are higher than for some other psychiatric disorders, such as schizophrenia and depression. The elevated death rate for BN and EDNOS as well “highlights the seriousness of those conditions,” write the authors. They call for more research into what factors might predict death risk in those two conditions in particular.

(Arch Gen Psychiatry. 2011;68[7]:724-731.