Cardiovascular Effects Of Incretin-Based Antihyperglycemic Drugs Relative To Treatment Alternatives In Older Adults Public Deposited

Downloadable Content

Download PDF
Last Modified
  • March 20, 2019
  • Gokhale, Mugdha
    • Affiliation: Gillings School of Global Public Health, Department of Epidemiology
  • Randomized placebo-controlled trials have examined the cardiovascular effects of dipeptidyl peptidase-4 inhibitors (DPP-4i), but limited data exist on the comparative incidence relative to therapeutic alternatives, including sulfonylureas (SU) and thiazolidinediones (TZD). In this study we therefore examined the comparative incidence of cardiovascular events with DPP-4i compared with relevant comparators using a new-user design. In recent years the use of SU was constant but DPP-4i use increased with a corresponding decrease in TZD use. Using hospitalization for heart failure (HF) as a positive control outcome we explored the use of calendar time as an instrumental variable (IV) and compared this approach to an active comparator new-user study comparing DPP-4i versus TZD. Using 2007-2013 US Medicare claims data, we identified two new user cohort pairs – DPP-4i versus SU and DPP-4i versus TZD. Since TZDs are contraindicated in patients with HF, we further excluded patients with diagnoses of HF or related conditions for the DPP-4i versus TZD analyses. Using propensity score-weighted survival analysis methods accounting for competing risk by death, we estimated hazard ratios (HR), risk differences (RD) and 95% confidence intervals (CI) for myocardial infarction (MI), stroke, HF hospitalization, and a composite outcome (MI, stroke, or all-cause mortality). For the IV analyses, we examined the IV strength and estimated RD for HF using Kaplan-Meier curves. The magnitude of RD per 100 patients for MI, stroke and HF hospitalization was <1 at one year after initiation with DPP-4i versus SU or TZD. The IV analysis compared patients initiating treatment during October 2010 to December 2013 versus January 2008 to May 2010 resulting in IV strength 40%. The 1- and 2-year RD of HF using the IV approach (scaled by IV-strength) and propensity score weighting were between 0 and -1 per 100 patients. Our well-controlled population based study suggests no increased short-term CV risk with DPP-4i relative to comparators. Both IV and propensity score-weighted approaches indicate lesser risk of HF hospitalizations among DPP-4i vs TZD initiators. The use of calendar time as an IV in settings where real-world market dynamics lead to profound changes in treatments is worth consideration.
Date of publication
Resource type
Rights statement
  • In Copyright
  • Jonsson Funk, Michele
  • Stürmer, Til
  • Buse, John
  • Lund, Jennifer
  • Simpson, Ross, Jr.
  • Doctor of Philosophy
Degree granting institution
  • University of North Carolina at Chapel Hill Graduate School
Graduation year
  • 2016

This work has no parents.