Off-label use of AAs among elderly associated with increased risk of diabetes: Study

A new study yields important new evidence that the use of atypical antipsychotics among the elderly for off-label purposes is associated with increased risk of developing diabetes.

“The message for physicians who may be prescribing atypical antipsychotics for the elderly is to be careful not to discount the risk of new-onset diabetes, even when prescribing these drugs in small doses.”

Atypical antipsychotics (AAs) are a group of newer-generation antipsychotic drugs that are indicated for the treatment of psychiatric conditions including schizophrenia, bipolar mania and acute mania; but are often prescribed for unapproved indications ("off-label") by physicians.

The study, conducted by Prescription Solutions, a leading pharmacy benefits management organization and a UnitedHealth Group (NYSE: UNH) company, was presented today at the 2010 Annual Scientific Meeting of the American Geriatrics Society (AGS) in Orlando, Fla. It was selected for presentation as part of the Presidential Poster Session open only to abstracts that received the highest scores through the AGS' peer review process.

The correlation between the use of AAs and new-onset metabolic syndrome (which often leads to diabetes) has been widely documented in younger and middle-aged schizophrenic and bipolar patients. Previous studies suggested that because the elderly are usually given AAs in smaller doses than younger patients, the same weight gain and resulting diabetes would not be expected.

However, in this retrospective study of 78,450 elderly patients without diagnoses of schizophrenia, bipolar disorder or diabetes, an association was demonstrated between study patients being on an atypical antipsychotic and starting a diabetic medication. This was demonstrated even though 97 percent of the study patients with AA fills had doses within recommended guidelines for the treatment of dementia. Elderly patients with at least one fill of an AA drug had 32-percent greater odds of starting a medication for diabetes within one year compared to similar patients without AA exposure.

"We are fortunate to have a wealth of retrospective data that helps us study the effects of medications on this growing population," said Joseph Addiego, M.D., senior vice president and chief medical officer, Prescription Solutions. "The message for physicians who may be prescribing atypical antipsychotics for the elderly is to be careful not to discount the risk of new-onset diabetes, even when prescribing these drugs in small doses."

Currently, AAs are widely used for off-label indications in the elderly for conditions such as dementia. Earlier studies may not have shown the same correlation between AAs and diabetes due to smaller study sample sizes or missing data. For example, the CATIE-AD trial demonstrated that female patients with Alzheimer's disease had significant weight gain with AA treatment, but the study was not powered to measure outcomes related to diabetes.

Study Methodology

• The study was designed as a case-control study using medical and pharmacy claims of Medicare (MAPD) and commercial managed care plan members in the western United States during an identification period of Jan. 1, 2004, through Dec. 31, 2008.
• The study group included 78,450 patients older than 65 years of age, without a diagnosis of schizophrenia or bipolar disorder, who were continuously enrolled and eligible for medical and pharmacy benefit during a one-year pre-index period. (The date range of the pre-index period depended on the index date identified, so they were different for the different patients.)
• The study identified 13,075 patients who first filled a diabetes medication during an index date within the identification period. More than 65,000 control subjects were matched at a ratio of 5:1 to case subjects based on age, gender, health plan type and index date year.
• The primary statistical analysis was a conditional logistic regression to determine the odds of starting antidiabetic medication for patients with AA exposure compared to those with no AA exposure, after adjusting for differences in clinical characteristics.

Study Limitations

• Patients may acquire a diagnosis of obesity independent of the use of an AA; therefore the use of this diagnosis as a proxy for weight gain associated with AA use may not be valid.
• The researchers were unable to account for differences in effects among the various atypical antipsychotics due to the small numbers of patients for each individual medication.
• Limitations of this database study include the lack of validation of diagnoses identified from medical claims, laboratory HbA1C data to verify diabetes, and information on ethnicity which may be a confounding factor in new-onset treatment-dependent diabetes.