Talk to the Veterans Crisis Line now
U.S. flag
An official website of the United States government

VA Health Systems Research

Go to the VA ORD website
Go to the QUERI website
Publication Briefs

Compared to Thiazolidinediones, Sulfonylureas More Likely to Cause Death and Hospitalization for Veterans with Diabetes

With more than 12 classes of glucose-lowering drugs currently approved, choosing the correct medications is becoming increasingly complex for patients with diabetes and their providers. There are few systematic comparisons of long-term outcomes related to specific medications, especially beyond first-line treatments. This study compared long-term outcomes of the two most commonly used second-line oral hypoglycemic medications in the VA healthcare system – sulfonylureas (SUs) and thiazolidinediones (TZDs). Current diabetes treatment guidelines outside VA promote metformin (MET) as the preferred first drug, but no longer recommend a particular second-line agent; in contrast, VA guidelines continue to recommend SUs as the second-line agent. Using VA and Medicare data from 2000 to 2010, investigators identified Veterans with diabetes who were dually enrolled in VA and Medicare who received a prescription for metformin and then initiated either SUs (n=73,726) or TZDs (n=7,210). The initiation date of SU or TZD was the start of the study period (index date) for each patient, and the previous 12 months was the baseline period. Variables at baseline included patient demographics, HbA1c levels, serum creatinine, urine microalbumin, body mass index, and comorbidities. Outcomes included all-cause mortality, hospital admission (VA or Medicare) for any of 13 ACSCs (ambulatory care sensitive conditions), and acute myocardial infarction (AMI) or stroke.


  • Veterans with diabetes who started on sulfonylureas compared to thiazolidinediones as a second-line agent after metformin were significantly more likely to die or have an ACSC hospitalization.
  • Patients in this study were elderly (mean age 69), primarily white (88%), and had high rates of cardiovascular comorbidities (e.g., chronic pulmonary disease, hypertension), and obesity (41%).


  • It is possible that some uncontrolled provider factors were responsible for the results.
  • Although TZD use was associated with fewer adverse outcomes than SU use, this does not inform the relative safety profile of SUs and TZDs in comparison to other newer agents.

This study was partly funded by HSR&D (IIR 10-136). Drs. Prentice and Pizer are part of VA's Health Care Financing & Economics. Drs. Prentice, Conlin, and Pizer are part of the VA Boston Healthcare System. Dr. Gellad was supported by an HSR&D Career Development Award and is part of HSR&D's Center for Health Equity Research and Promotion in Pittsburgh, PA. Dr. Edelman is part of HSR&D's Center for Health Services Research in Primary Care, Durham, NC.

PubMed Logo Prentice J, Conlin P, Gellad W, Edelman D, Lee T, Pizer S. Capitalizing on Prescribing Pattern Variation to Compare Medications for Type 2 Diabetes. Value in Health. December 2014;17(8):854-62.

Related Briefs

» next 62 Diabetes Briefs...

» next 171 Medication Briefs...

What are HSR Publication Briefs?

HSR requires notification by HSR-funded investigators about all articles accepted for publication. These journal articles are reviewed by HSR and publication briefs or summaries are written for a select number of articles that are then forwarded to VHA Central Office leadership to keep them informed about important findings or information. Articles to be summarized are selected by HSR based on timeliness of the findings, interest of leadership, or potential impact on the organization. Publication briefs are written for only a small number of HSR published articles. Visit the HSR citations database for a complete listing of HSR articles and presentations.

Questions about the HSR website? Email the Web Team

Any health information on this website is strictly for informational purposes and is not intended as medical advice. It should not be used to diagnose or treat any condition.