**Background:** Eating disorders (EDs) are considered among the most lethal psychiatric disorders, but reported mortality rates vary widely across studies. Standardized mortality ratios (SMRs) for anorexia nervosa (AN) have ranged from 2.89 to 15.9 depending on study design, setting, and follow-up duration. The authors note that most prior studies included selected clinical populations rather than geographically representative cohorts, and that treatment organization may significantly influence outcomes. This study aimed to compare mortality in a representative cohort of ED patients from the Florence area with the general population of Tuscany.
**Methods:** Patients were enrolled between October 1994 and December 2018 from two main public ED centers in Tuscany: the ED Service of the Local Health Unit 'Tuscany Centre' and the Eating Disorders Unit of the Florence University Hospital 'Careggi'. Inclusion criteria were age >18 years and a DSM-5 diagnosis of AN, BN, or BED. Exclusion criteria included intellectual disability, illiteracy, bipolar or psychotic disorder, residence outside Tuscany, and lack of consent. Life status was determined through linkage with the Regional Mortality Registry of Tuscany through December 31, 2018. SMRs were calculated using indirect standardization, stratified by gender, age group (5-year intervals), and calendar time (1-year intervals). Person-years were computed from first clinical encounter until death or end of follow-up.
**Key Results:** A total of 1,277 patients were included: 368 with AN (28.8%), 312 with BN (24.4%), and 597 with BED (46.8%). Females comprised 92.4% of the sample. Median age at evaluation was 31.1 years (IQR 21.7–44.9), varying by diagnosis (AN: 22.5 years, BN: 26.8 years, BED: 43.0 years). Median follow-up was 7.4 years (IQR 2.9–12.0, range <0.1 to 24.2 years). Twenty-two patients (1.72%) died: 3 AN (0.82%), 3 BN (0.96%), and 16 BED (2.68%). The overall SMR was 1.19 (95% CI 0.79–1.81; 22 observed vs. 18.43 expected deaths). By diagnosis, SMR was 2.49 for AN (95% CI 0.80–7.74), 2.07 for BN (95% CI 0.67–6.42), and 1.01 for BED (95% CI 0.62–1.66). Among BN patients, mortality was significantly increased after age 60 (SMR 4.22, 95% CI 1.06–16.88) and after 10 years from clinical evaluation (SMR 11.24, 95% CI 3.62–34.84). Causes of death included neoplasms (n=11, 50%), diseases of the circulatory system (n=6, 27.3%), endocrine/nutritional/metabolic diseases (n=2), injury/poisoning (n=2), and respiratory disease (n=1). No suicides were recorded. SMR for circulatory system deaths was 1.70 (95% CI 0.76–3.78) and for neoplasms was 1.14 (95% CI 0.63–2.06).
**Clinical Implications:** The overall mortality in this cohort was lower than most previously published studies, particularly for AN where SMRs of 5–15 have been commonly reported. The authors attribute this to the integrated, multidisciplinary treatment network (EDTN) that provides continuity of care across outpatient, day hospital, inpatient, and residential settings, with regular monitoring even after treatment discontinuation. The absence of suicide deaths is notable and contrasts with meta-analytic findings of 7-fold (BN) to 31-fold (AN) increased suicide risk. The significantly elevated late mortality in BN patients (after 10 years) suggests that chronicity is a crucial factor in premature death for this subgroup. Limitations include the exclusion of patients relying on private care, potential selection bias from 158 patients who declined participation, and limited statistical power for detecting late increases in SMR. The study supports the hypothesis that specialized, coordinated, community-based treatment networks may reduce mortality in eating disorders.