Should Pioglitazone Be Used More for Secondary Stroke Prevention in Prediabetes?

Post hoc analysis of the IRIS trial suggests that it should.

Should Pioglitazone Be Used More for Secondary Stroke Prevention in Prediabetes?

HONOLULU, HI—Patients with a prior stroke or TIA who have prediabetes detected using widely available clinical assessments seem to derive a benefit from pioglitazone, post hoc analysis of the IRIS trial suggests.

Among all prediabetic patients, pioglitazone was associated with lower risks of a variety of vascular outcomes—including recurrent stroke and a composite of stroke, MI, or hospitalization for heart failure—as well as new-onset diabetes, J. David Spence, MD (Robarts Research Institute, London, Canada), reported last week at the International Stroke Conference. The magnitude of the apparent benefits was even greater in patients with good adherence to therapy.

Although treatment was associated with greater risks of weight gain, edema, and bone fracture, Spence indicated that those could be managed and would be outweighed by the expected reductions in adverse clinical outcomes.

“A hell of a lot more people should be using pioglitazone,” he told TCTMD.

Pioglitazone is a thiazolidinedione approved for use in the United States as an adjunct to diet and exercise to improve glycemic control in adults with type 2 diabetes (but not for secondary stroke prevention in prediabetes).

The main results of the placebo-controlled IRIS trial, showing that pioglitazone reduced risks of stroke/MI, stroke, ACS, and new-onset diabetes in nondiabetic patients with a recent stroke or TIA and insulin resistance, were published in the New England Journal of Medicine in April 2016. “I’ve been fighting ever since then with the Ontario drug-benefit program to pay for pioglitazone for secondary stroke prevention in my patients with prediabetes and insulin resistance,” Spence said.

A hell of a lot more people should be using pioglitazone. J. David Spence

He has not been successful in his quest, and Spence explained that the reluctance on the part of the government stems from the way insulin resistance was assessed in the trial—with HOMA-IR, which is not commonly used in clinical practice.

This new post hoc analysis focused on patients with prediabetes is meant to make the trial findings more clinically relevant because this condition is commonly assessed in practice using glycated hemoglobin (HbA1c) or fasting glucose. “And because the Ontario drug benefit won’t have to pay for pioglitazone in people who don’t take it, we focused on an on-treatment analysis defined by adherence of 80% instead of focusing on an ITT [intention-to-treat] analysis,” Spence said during a poster presentation.

This analysis, published simultaneously online in JAMA Neurology, examined data on 2,885 patients with prediabetes according to American Diabetes Association criteria—an HbA1c of 5.7% to 6.4% or a fasting glucose of 100 to 125 mg/dL. Roughly half of those patients had good adherence to treatment.

Pioglitazone reduced fasting blood glucose, blood pressure, and triglycerides and raised HDL cholesterol. In addition, it reduced risks of variety of adverse clinical outcomes in both the ITT and on-treatment (good adherence) analyses.

Risks of Adverse Outcomes With Pioglitazone Versus Placebo



95% CI






     Stroke/MI/HF Hospitalization

     New-Onset Diabetes



















Good Adherence




     Stroke/MI/HF Hospitalization

     New-Onset Diabetes



















Abbreviations: HF, heart failure; ITT, intention to treat; NNT, number needed to treat.

Among the subgroup of patients with good adherence to pioglitazone, treatment was associated with increased rates of weight gain of at least 10% (29.8% vs 12.0%; P < 0.0001) and edema (29.2% vs 21.6%; P = 0.0009). Bone fracture was numerically—but not significantly—higher (3.6% vs 2.8%; P = 0.43). Similar differences were seen in the entire prediabetic group, except that the difference in bone fracture was significant (4.9% vs 3.2%; P = 0.02).

Still, Spence argued that pioglitazone should be used more widely. He pointed to the fact that the number needed to harm associated with fractures in patients with good adherence was 125, whereas the numbers needed to treat to prevent new-onset diabetes or stroke/MI were just 12 and 24, respectively.

The main adverse event affecting tolerability of the drug is fluid retention, which can be managed either by reducing the dose of pioglitazone or by using amiloride, Spence said.

Commenting for TCTMD, Seemant Chaturvedi, MD (University of Maryland School of Medicine, Baltimore), called the analysis intriguing. “In the past, we haven’t had any definite therapies for patients with prediabetes—especially with regard to stroke reduction—and I think neurologists probably haven’t given it a lot of thought as to whether we should be adding pharmacologic therapy specifically based on a history of prediabetes or of elevated glucose [in patients] not really meeting the criteria for diabetes,” he said.

Chaturvedi questioned whether clinicians would buy into the idea of using pioglitazone in this population, noting that the drug is perceived to have tolerability issues related to edema, weight gain, fractures, and—potentially—heart failure.

“Based on that, my sense is that probably a lot of neurologists are not actively using pioglitazone right now, but I think maybe this paper would force us to reexamine our practices and begin to think about whether we should consider it, especially for higher-risk patients,” Chaturvedi said. He added that if clinicians were to use pioglitazone in their patients, they would have to monitor the situation to make sure the patients were tolerating the medication and sticking with treatment.

I think clinicians should take a fresh look at it and not go in with too many preconceived notions that the medication is hard to tolerate. Seemant Chaturvedi

Considering these data, Chaturvedi said, “I think clinicians should take a fresh look at it and not go in with too many preconceived notions that the medication is hard to tolerate.” He added that he would consider using pioglitazone in more patients moving forward.

In an accompanying editorial, Leonardo Pantoni, MD, PhD (University of Milan, Italy), talks about the links between stroke and dementia, between prediabetes and abnormalities on brain imaging that have been tied to cognitive decline, and between diabetes and Alzheimer disease independent of stroke to highlight the potential for the treatment approach examined in this post hoc analysis to prevent dementia.

[T]he availability of insulin resistance drugs might represent an opportunity to test for preventing other forms of dementia that are not immediately linked with cerebrovascular events,” he says, noting that trials in this area are ongoing.

Todd Neale is the Associate News Editor for TCTMD and a Senior Medical Journalist. He got his start in journalism at …

Read Full Bio
  • IRIS was supported by a grant from the US National Institute of Neurological Disorders and Stroke. Takeda Pharmaceuticals International donated pioglitazone and matching placebo for the trial.
  • Spence and Pantoni report no relevant conflicts of interest.