Examining Mortality Rates

Eating disorders, particularly anorexia nervosa, have one of the highest mortality rates of all mental illnesses; there is extensive evidence of this effect. In a landmark meta-analysis of worldwide eating disorder mortality rates (JMIR Res Protoc. 2017. 6:e146), the standard mortality ratio, or SMR, of anorexia nervosa patients was 5.1 deaths per 1000 person-years. Young people between the ages of 15 and 24 years with AN have 10 times the risk of dying as their peers (Curr Psychiatry Rep. 2012. 14:406). People with eating disorders also have an overall worse quality of life, and their treatment costs are 48% higher than in the general population (Curr Opin Psychiatry. 2020. 33:521).

But, a recent study by Italian researchers has produced very different mortality rates for patients with eating disorders. In their study of 1277 persons with eating disorders, Dr. Giovanni Castellini and his colleagues at the University of Florence, Italy, found that the mortality rate of people with eating disorders did not differ from that of the general population. The 1277 people with eating disorders in their study included 368 with AN, 312 with BN, and 597 with BED. Twenty-two patients, or 1.72%, died during a median follow-up over 7.4 years (Int J Eat Disord. 2022. 17:122). The standardized mortality ratio, or SMR, was 1.19. Only among BN patients did the mortality significantly increase 10 years after clinical evaluation. The life status of participants, according to the authors, was determined through linkage with the Regional Mortality Registry.

Age was the most significant factor influencing mortality, according to the authors. Deaths occurred at the ages of 35, 57, and 64 years among the 3 AN patients; at the ages of 49, 72, and 82 years among the 3 BN patients, and at a median age of 64 years among those with BED.

Some possible explanations for the lower mortality rate

The authors suggest that the low mortality rate in persons with eating disorders in their study, compared to other published studies, could well be attributed to the eating disorders treatment network of the Florence area (EDTN), which they describe as a coordinated regional multidisciplinary treatment service.

The study also had limitations, according to the researchers. First, it did not include patients who used private care services; thus, the SMRs were not applicable to this group. And the SMRs were calculated considering gender, age group, and calendar time. To improve accuracy, according to Dr. Castellini and colleagues, future studies should add additional clinical characteristics of the patients and the type and duration of treatment.

Several other factors might be considered as contributing to the low mortality rates. Some are technical or statistical: the sensitivity and specificity (essentially, the “accuracy”) of the mortality database was not described; it appears to be a regional rather than national database. The duration was relatively short for a mortality study, and the sample was relatively small. In addition, about half of the sample had been diagnosed with BED; as one might predict based on prior work, excess mortality was observed in those with AN or BN.

While those factors are probably critical to interpreting the results, they are nonetheless important in suggesting that higher mortality is common in those with eating disorders.