An eating-disorder cohort linked diazepam with BMI change in a very small subgroup

A naturalistic inpatient study reported an exploratory diazepam association with BMI increase, with only four diazepam-exposed patients and overlapping treatments.

Two outcomes were tracked

Panarello and colleagues followed 127 inpatients with restrictive anorexia nervosa, binge–purge anorexia nervosa, or bulimia nervosa during a ten-week treatment program. They examined change in body mass index and an Average Morbidity Index intended to summarize weekly psychopathological symptom burden.

Medicines were prescribed as part of routine care alongside nutritional rehabilitation. Many patients received more than one drug class. Medication groups were therefore overlapping exposure categories, not independently randomized treatment arms.

The diazepam result needs its denominator

The authors reported a positive diazepam coefficient of 2.06 for BMI change, with a p value of 0.02, and no clear corresponding benefit on the morbidity index. Their individual-drug table listed only four diazepam-exposed patients. That small subgroup is essential to interpreting the result.

A model coefficient is not the same as the subgroup’s raw average change, and a positive BMI association is not proof of an effect on core eating-disorder symptoms. The paper measured nutritional and symptom outcomes separately. Its findings do not justify combining those endpoints into a single claim of treatment success.

Treatment choice can reflect clinical need

Patients were not assigned to diazepam at random. Anxiety, sleep disturbance, symptom severity, and concurrent medicines can influence both prescribing and recovery. The study’s naturalistic design cannot fully separate those influences from a drug effect.

Individual-compound findings were described as exploratory, with particular caution needed for medicines used in only a few patients. Overlapping exposure also means that the same person can contribute to more than one drug-class summary. Comparing those summaries as if they were separate trial arms would misrepresent the design.

The appropriate interpretation

The sample mainly represented young women with anorexia nervosa in specialized inpatient care. Generalization to men, older patients, outpatient care, or other eating-disorder populations is limited. The ten-week window also does not establish long-term benefit or harm.

For evidence review, the study raises a question about how anxiolytic exposure relates to nutritional trajectories during intensive care. It does not establish diazepam as a weight-restoration treatment, and it offers no individual impairment or dosing rule.

Source paper

Original analysis of the supplied journal article, identified as PMC12883831.