Eating disorder diagnosis marks sharply higher risks for organ injury, self-harm and death
A diagnosis of eating disorder in primary care was followed by substantially higher rates of coded physical illness, psychiatric harm and death across linked hospital and mortality records. The cohort study drew on English primary care electronic health records from 1 January 1998 to 30 November 2018 and included 24,709 patients aged 10 to 44 years with an eating disorder diagnosis, each matched by age, sex and practice to as many as 20 comparators without an eating disorder, for a total comparison group of 493,001. Disorders included anorexia nervosa, bulimia nervosa, binge eating disorder and other eating disorders grouped together.
The pattern was not confined to one clinical domain. Short-term excess risk emerged in the kidneys, liver, mental health and mortality, then remained elevated years later. In the first year after diagnosis, renal failure occurred about six times more often than in matched comparators, with a hazard ratio of 6.0 and an excess of 15 events per 10,000 individuals. Liver disease followed a similar pattern, with a hazard ratio of 6.7 and an excess of 6 events per 10,000. Those early differences are clinically hard to dismiss: they point to more than background comorbidity and suggest that the period immediately after recognition of an eating disorder is a time of concentrated medical vulnerability.
Early psychiatric and renal complications concentrate soon after diagnosis
The first year after diagnosis carried especially heavy psychiatric burden. Recorded depression was more than seven times as common as in the matched cohort, with a hazard ratio of 7.3 and an excess of 596 per 10,000 individuals. Self-harm was even more pronounced, with a hazard ratio of 9.4 and an excess of 309 per 10,000. These are not subtle separations between groups; they describe a rapid divergence in risk that would matter in day-to-day primary care, emergency assessment and community mental health follow-up.
The renal findings deserve parallel attention because they were not fleeting. Although the earliest year showed the steepest relative increase, the hazard ratio for renal failure remained elevated at 2.6 after five years, and the cumulative excess at 10 years reached 110 per 10,000 individuals. Liver disease also persisted, with a hazard ratio of 3.7 after five years and a cumulative excess of 26 per 10,000 at 10 years. The likely clinical explanation is familiar even if the study design cannot prove mechanism: malnutrition, electrolyte disturbance, dehydration, purging behaviors and metabolic stress can all strain renal and hepatic function. The data show the downstream clinical imprint of those stresses in real-world records.
Mortality differences persisted long after the first presentation
Mortality separated early and stayed separated. All-cause mortality was more than fourfold higher in the first year after diagnosis, with a hazard ratio of 4.6 and an excess of 10 deaths per 10,000 individuals. Unnatural deaths were five times more common, with a hazard ratio of 5.1 and an excess of 30 per 100,000. Suicide risk was especially striking: the first-year hazard ratio reached 13.7, with 24 excess deaths per 100,000 individuals.
Those excesses did not disappear with time. After five years, all-cause mortality remained elevated with a hazard ratio of 2.2 and an excess of 43 deaths per 10,000 individuals. Unnatural deaths stayed high as well, with a hazard ratio of 3.2 and an excess of 184 per 100,000. At 10 years, the study still found 95 excess all-cause deaths per 10,000 individuals and 341 excess unnatural deaths per 100,000. Suicide remained increased at 10 years, with a hazard ratio of 2.7 and 169 excess deaths per 100,000. One sentence captures the clinical meaning plainly: an eating disorder diagnosis is not an isolated episode in care, but the start of a prolonged period of excess mortality risk.
Primary care records capture a high-risk population that needs integrated follow-up
The value of this study lies partly in its setting. By using routine primary care records linked to hospital admissions and national mortality data, it traced outcomes across the settings where patients actually move, rather than limiting analysis to a single clinic or inpatient service. That approach is especially relevant for eating disorders, where patients may present first to primary care, be followed intermittently in the community, and then appear in secondary care only when medical or psychiatric deterioration becomes unavoidable.
The authors compared individuals with an eating disorder diagnosis against matched peers without such a diagnosis, which strengthens the case that the observed differences reflect more than age, sex or practice-level variation. Even so, the analysis depends on coded diagnoses in routine records. Some people with eating disorders will never receive a code, and some severity details, symptom duration and treatment exposure are not available in the data source. The study therefore likely captures a clinically meaningful subset, but not the full burden in the community.
What the numbers do support is a broader model of care. Physical surveillance and mental health support need to be linked from the moment the diagnosis is recorded. Renal and liver monitoring, self-harm assessment, suicide risk review and coordinated psychiatric care are not separate tasks in this population; they are part of the same clinical response. The excess risks were largest soon after diagnosis, but they remained visible years later, which argues for follow-up that is sustained rather than episodic.
What the excess risks imply for bedside care
The outcomes in this cohort point toward a narrow, practical message. Patients with eating disorders face elevated risks of organ failure, depression, self-harm and death, and those risks extend well beyond the initial diagnosis window. The study does not identify which interventions reduce those hazards, but it does establish the scale of the problem with enough clarity to shape practice.
Primary care clinicians are often the first to see the weight loss, gastrointestinal complaints, electrolyte disturbance, anxiety, syncope or self-harm that sit around an eating disorder diagnosis. Once the diagnosis is made, the record should trigger active medical review rather than passive observation. Hospital teams, mental health services and primary care need to function as a single pathway for this group, because the risk profile spans all three.
The remaining question is not whether the excess risk exists. It is how quickly integrated monitoring, psychiatric treatment and medical surveillance can be delivered after diagnosis, and how consistently those supports can be maintained over the following years.