Mortality rates in patients with anorexia nervosa and other eating disorders¶
One-paragraph summary¶
Systematic review and random-effects meta-analysis of 36 quantitative mortality studies published from 1966 through September 2010. AN cohorts contributed 166,642 person-years. The weighted mortality rate was 5.1 per 1,000 person-years and the standardized mortality ratio was 5.86; one in five deaths among people with AN was suicide (Arcelus 2011, PMID 21727255).
Key findings¶
- AN: 5.1 deaths per 1,000 person-years.
- AN standardized mortality ratio: 5.86.
- Suicide: 20% of deaths in AN cohorts.
- Estimates are higher than those reported for bulimia nervosa and EDNOS in the same analysis.
Limitations¶
- Cohorts and eras were heterogeneous.
- Cause-of-death ascertainment varied.
- The proportion of deaths due to suicide is not individual cumulative suicide risk.
- Treatment and diagnostic systems have changed since the search period.
Replication and updating (audit, 2026-09-02)¶
Two independent 2026 syntheses tested this benchmark on much larger evidence and reproduced it rather than revising it.
| Source | Scope | AN estimate |
|---|---|---|
| Arcelus 2011, PMID 21727255 | 36 studies; 166,642 AN person-years; literature to Sept 2010 | 5.1 deaths/1,000 py; SMR 5.86; 1 in 5 deaths by suicide |
| Semchishen 2026, PMID 41536100 | 83 studies; 307,710 ED patients vs 15.7 million controls; mean follow-up 11.96 yr | All-cause RR 5.52 (4.47–6.82); suicide RR 9.86 (5.63–17.27) |
| Lai 2026, PMID 41277145 | 30 studies, 33,176 AN patients; 22 pooled; searched to May 2025 | SMR 5.06 (3.47–7.38); suicide 21% and cardiac 19% of deaths |
The three analyses share primary studies, so their agreement is not fully independent replication. What is nonetheless striking is that fifteen years of revised diagnostic criteria, expanded specialist services and a large randomized treatment literature have not moved the estimate — the single strongest argument in this knowledge base that mortality has never been treated as a treatment endpoint. Lai and colleagues add one genuinely new element: cardiac causes account for a further 19% of deaths, placing arrhythmia risk on a similar footing to suicide risk.
Why it matters¶
The paper established a quantitative mortality benchmark that continues to organize AN risk discussions and highlights that suicide prevention is distinct from medical stabilization. Its numbers have survived two subsequent meta-analyses on far larger samples.
Cited by wiki pages¶
- Overview
- Mortality and long-term outcome
- Red flags and safety concerns