Clinical and psychological phenotypes of type 1 diabetes and disordered eating derived from a case vignette series:...

Type 1 diabetes and disordered eating (T1DE) is a common, complex comorbidity of type 1 diabetes with high morbidity and mortality risk. T1DE currently lacks a clear definition, diagnostic differentiation into phenotypes and clinical severity grading, which results in exclusion of people with T1DE from accessing appropriate treatments to improve outcomes.

We developed anonymised T1DE case vignettes from the baseline assessments of the Safe management of people with Type 1 diabetes and EAting Disorders studY (STEADY) and the London-T1DE pilot service at King's College Hospital for people with diabetes and severe eating disorder, integrating detailed biomedical and psychiatric data. In an iterative process, phenotypes of T1DE were developed, the T1DE vignettes were sorted into the different phenotypes and a clinical severity grading framework was developed.

The 70 participants (94.3% women) were 34.5 ± 10.6 years old with a diabetes duration of 18.4 ± 12.5 years and HbA_1c of 84.8 ± 32.9 mmol/mol (10.2 ± 3.0%). All but one fulfilled ICD-10/Diagnostic and Statistical Manual of Mental Disorders, 5th Edition criteria for a formal eating disorder diagnosis, 34 met criteria for other specified feeding or eating disorder, 27 for bulimia nervosa, six for anorexia nervosa and two for binge eating disorder. Over 50% had a mood disorder, 44.3% an anxiety disorder and 20.0% a trauma-related disorder. From the anonymised T1DE vignettes generated, three main T1DE phenotypes with subtypes emerged: insulin-omission (diabulimia) T1DE (28 participants, 40.0%), characterised by near total insulin omission in intention to lose weight, with (binge-)purge, anorectic and cyclical subtypes; restrict T1DE (17, 24.3%), characterised by restriction of food/carbohydrates (with [binge-]purge, anorectic and hypo subtypes); and binge T1DE (25, 35.7%), characterised by subjective or objective binge eating with or without omission of bolus insulin (with binge-purge, binge-only and hypo subtypes). Severity grading categorised 19 cases (27.1%) as 'low', 23 (32.9%) as 'moderate' and 28 (40.0%) as 'severe' clinical risk.

The three T1DE phenotypes will serve as a basis for more precise clinical diagnosis and severity grading and will be used to develop future diagnostic screening tools, improved psychological interventions and integrated treatment care pathways to match the individual clinical presentation and severity in a personalised medicine approach.