Takeaway
- Patients with severe mental illness receiving antipsychotic medication should be screened for diabetes with a 2-step testing algorithm, rather than clinical variables.
Why this matters
- Blood glucose is tested in only about half the patients under psychiatric care.
Study design
- Researchers examined consecutive psychiatric inpatients receiving antipsychotics (N=798; mean age, 37.7 years; men, 61.1%), comparing biochemical and clinical predictors of diabetes.
- Funding: EFSD/Lilly grant.
Key results
- No single clinical factor accurately confirmed a diagnosis of diabetes.
- Only 3 clinical factors could rule out diabetes with good accuracy: BMI (76.6% accurate), waist/hip ratio (74.2% accurate), and height (68.7% accurate).
- The most accurate (98.5%) assessment required a 2-step algorithm consisting of HBA1c ≥5.7% followed in 32.7% of patients by both fasting glucose and oral glucose tolerance testing.
Limitations
- Prospective data were not available on prediabetes, repeat testing, or end-organ dysfunction.
References
References
