Blood Sugar Rose 14 Times Faster in First 90 Days of Antipsychotics, UCL Study of 96,849 Adults Finds

Adults with severe mental illness who start a second-generation antipsychotic see their blood sugar and weight accelerate in the first three months, according to a University College London study published Oct. 14 in The Lancet Psychiatry. After day 90, the pace drops back to where it was before treatment. The damage already done does not reverse.

The work was funded by the National Institute for Health and Care Research.

What the researchers measured

Led by Dr. Juan Carlos Bazo-Alvarez of UCL Epidemiology and Health, the team used anonymized NHS general practice records from England's Clinical Practice Research Datalink Aurum database. The cohort: 96,849 adults aged 18 to 59 with schizophrenia, schizoaffective disorder or bipolar disorder who started an oral second-generation antipsychotic between 2000 and 2023.

They tracked each patient from two years before the first prescription to two years after. The four drugs studied were olanzapine, quetiapine, risperidone and aripiprazole.

Before treatment, HbA1c, a measure of average blood sugar, rose by about 0.01 mmol/mol per week. In the first 90 days on medication, that jumped to 0.12 mmol/mol per week, roughly 14 times faster. The excess increase was 1.42 mmol/mol over what the pre-treatment trend would have produced.

Weight went from 0.04 lb per week before treatment to 0.46 lb per week in the first 90 days, about nine times faster. That is an excess gain of 5.22 lb. From day 91 on, weight gain slowed to 0.07 lb per week.

Drug by drug

Olanzapine showed the largest early rises. HbA1c rose about 50 times faster and weight about 10 times faster than before treatment, per the UCL release. At day 90 the study put olanzapine's excess weight gain at 7.26 lb and its excess HbA1c increase at 2.33 mmol/mol.

The other three, at day 90:

• Aripiprazole: 3.84 lb excess weight gain, 2.27 mmol/mol excess HbA1c.
• Quetiapine: 4.98 lb excess weight gain, only 0.15 mmol/mol excess HbA1c.
• Risperidone: 4.92 lb excess weight gain, 0.54 mmol/mol excess HbA1c.

That quetiapine result is the one that undercuts the standard routine. "We were also struck by the fact that weight and HbA1c did not always move together," Bazo-Alvarez told MedPage Today. "This reinforces the message that weight alone does not fully capture early glycemic risk."

Who was hit hardest

Women saw steeper early surges than men. Excess 90-day HbA1c was 1.68 mmol/mol in women versus 1.18 in men. Excess weight gain was 5.73 lb versus 5.18 lb. Women starting olanzapine fared worst: 9.17 lb of excess weight and 3.27 mmol/mol of excess HbA1c.

Starting blood sugar mattered too. According to UCL, 19% of patients with prediabetes at the start of treatment developed type 2 diabetes within two years, compared with 3% of those who began with normal blood sugar.

What the authors say to do

Bazo-Alvarez was direct about the clinical takeaway. "Healthcare professionals should check both weight and blood sugar before treatment starts and again during the first few months afterwards," he said in UCL's release.

He also stressed the drugs' value: "Antipsychotic medicines are valuable treatments that help many people with serious mental illnesses." Schizophrenia affects about 23 million people worldwide and bipolar disorder about 37 million. People with these conditions typically die 10 to 20 years earlier than average, driven mainly by cardiovascular and metabolic disease, per UCL.

The limits of the data

This is an observational study. It shows what happened to blood sugar and weight after patients started these drugs. It cannot by itself prove the drug caused every change in every patient.

The testing gap is large. In the cohort, 76.6% of patients had a weight recorded during follow-up. Only 24.9% had an HbA1c measurement. Doctors were checking the scale far more often than blood sugar, which is the very mismatch the authors want fixed. It also means the HbA1c findings rest on a quarter of patients, and the study does not say how those patients were chosen for testing.

A harder line from critics of antipsychotic use

Separate research has pushed a more skeptical view of these drugs. Mad in America, a publication critical of psychiatric medication, recently summarized a Psychiatry Research systematic review led by Ramya Padmavathy Radha Krishnan of the University of Sydney. It covered 13 observational studies of long-term, off-label, low-dose antipsychotic use in adults without a psychotic disorder diagnosis.

The review reported weight gain in 28.8% to 77.9% of users versus 17.8% to 71.9% of non-users. High blood sugar showed up in 4.8% to 23.2% of users versus 0 to 9.7% of non-users. One included study linked low-dose use to a 21% higher risk of heart and metabolic death.

The authors flagged their own limits. The studies used varying methods, which ruled out a meta-analysis. They were observational, so they cannot show that antipsychotics cause the effects. The review found no randomized controlled trials that met its criteria. It also covers a different group of patients than the UCL study, which looked at people with severe mental illness.

What is still open

The UCL authors recommend baseline and early follow-up checks of both weight and blood sugar. No source in the research describes a change to NHS or U.S. monitoring guidance in response. Whether more HbA1c testing in the first 90 days actually prevents diabetes, rather than just detecting it earlier, is not something this study tested.