IMPORTANCE

Eating disorders lead to increased mortality and reduced quality of life. While the acute presentations of eating disorders frequently involve electrolyte abnormalities, it remains unknown whether electrolyte abnormalities may precede the future diagnosis of an eating disorder.

OBJECTIVE

To determine whether outpatient electrolyte abnormalities are associated with the future diagnosis of an eating disorder.

DESIGN, SETTING, AND PARTICIPANTS: This population-level case-control study used provincial administrative health data for residents of Ontario, Canada aged 13 years or older from 2008 to 2020. Individuals without an eating disorder (controls) were matched 4:1 to individuals diagnosed with an incident eating disorder (cases) based on age and sex. Both groups had outpatient electrolyte measurements between 3 years and 30 days prior to index. Index was defined as the date of an eating disorder diagnosis in any inpatient or outpatient clinical setting for cases. Controls were assigned a pseudo-index date according to the distribution of index dates in the case population. Individuals with any prior eating disorder diagnosis were excluded. The data analyzed was from January 1, 2008, through June 30, 2020.

EXPOSURES

Any electrolyte abnormality, defined as abnormal test results for a composite of hypokalemia, hyperkalemia, hyponatremia, hypernatremia, hypomagnesemia, hypophosphatemia, metabolic acidosis, or metabolic alkalosis.

OUTCOMES AND MEASURES

Eating disorder diagnosis including anorexia nervosa, bulimia nervosa, and eating disorder not otherwise specified.

RESULTS

A total 6970 eligible Ontario residents with an eating disorder (mean [SD] age, 28 (19) years; 6075 [87.2%] female, 895 [12.8%] male) were matched with 27 878 age- and sex-matched residents without an eating disorder diagnosis (mean [SD] age, 28 [19] years; 24 300 [87.2%] female, 3578 [12.8%] male). Overall, 18.4% of individuals with an eating disorder had a preceding electrolyte abnormality vs 7.5% of individuals without an eating disorder (adjusted odds ratio [aOR], 2.12; [95% CI, 1.86-2.41]). The median (IQR) time from the earliest electrolyte abnormality to eating disorder diagnosis was 386 (157-716) days. Specific electrolyte abnormalities associated with a higher risk of an eating disorder were: hypokalemia (aOR, 1.98; 95% CI, 1.70-2.32), hyperkalemia (aOR, 1.97; 95% CI, 1.48-2.62), hyponatremia (aOR, 5.26; 95% CI, 3.32-8.31), hypernatremia (aOR, 3.09; 95% CI, 1.01-9.51), hypophosphatemia (aOR, 2.83; 95% CI, 1.82-4.40), and metabolic alkalosis (aOR, 2.60; 95% CI, 1.63-4.15).

CONCLUSIONS AND RELEVANCE

In this case-control study, individuals with an eating disorder were associated with a preceding outpatient electrolyte abnormality compared with matched controls. Otherwise unexplained electrolyte abnormalities may serve to identify individuals who may benefit from screening for an underlying eating disorder.