HbA1c and Weight Rise Fastest Early After Antipsychotic Initiation

Key Takeaways
- Among adults with severe mental illness starting oral second-generation antipsychotics in England, HbA1c and weight rose fastest during the first 90 days.
- Olanzapine had the largest reported early increases in both HbA1c and weight.
- Quetiapine was associated with early weight gain but minimal HbA1c change.
- Estimated two-year type 2 diabetes incidence was higher with baseline prediabetes than with normal baseline HbA1c.
Investigators conducted an individual-level interrupted time-series analysis of HbA1c and weight after antipsychotic initiation using Clinical Practice Research Datalink Aurum electronic health records from England dated January 1, 2000, to June 7, 2023. The analysis included 96,849 eligible adults aged 18–59 with severe mental illness and a first recorded oral antipsychotic initiation. Eligibility required another oral second-generation antipsychotic prescription within 90 days. Mixed models with linear splines estimated within-person weekly HbA1c and weight trajectories before initiation, during days 0–90, and from day 91 to two years. Analyses were stratified by drug, sex, and baseline glycemic status; time to incident type 2 diabetes was another outcome. People with lived experience contributed to the research question, approach, interpretation, and dissemination.
During days 0–90, the estimated HbA1c slope was +0.12 mmol/mol/week (95% confidence interval [CI] 0.09–0.15). The weight slope was +0.21 kg/week (95% CI 0.20–0.22). Both were slower before treatment. Later HbA1c change returned to its pretreatment pace, while weight gain slowed relative to the early period. Olanzapine had the largest weight increase (0.28 kg/week) and, with aripiprazole, the largest HbA1c increase (0.18 mmol/mol/week). Rates were generally higher among women. With quetiapine, early weight gain was +0.19 kg/week despite minimal HbA1c change.
Estimated two-year cumulative type 2 diabetes incidence was 19% among participants with baseline prediabetes versus 3% among those whose baseline HbA1c was normal. These estimates describe risk over follow-up, not the observed proportions of eligible participants recorded as developing diabetes.
Only 24.9% of eligible adults had at least one recorded HbA1c measurement; not everyone had serial glycemic measurements. The trajectories and drug-stratified patterns are observational associations, not causal comparisons. The authors proposed assessing both HbA1c and weight at baseline and 12 weeks, followed by risk-stratified monitoring and intervention. This was their interpretation of the findings, not a monitoring strategy tested for clinical benefit.
The authors concluded that HbA1c and weight increased fastest early after antipsychotic initiation but did not always move together.
Clinician Questions
Do HbA1c changes after second-generation antipsychotic initiation apply to adults who already have type 2 diabetes?
The English analysis excluded people with existing type 2 diabetes, as well as those with dementia or antipsychotic prescribing before age 18. Its post-initiation HbA1c trajectories therefore do not directly describe adults who already had type 2 diabetes.
How was new type 2 diabetes identified after oral antipsychotic initiation?
In these English electronic health records, the additional outcome was time to a recorded type 2 diabetes diagnosis or a prescription for antidiabetic medication.
Recommended Reading
- For more on cardiovascular risk in prediabetes: Addressing Cardiovascular Risk in Prediabetes: Why Earlier Action Matters