Stroke Patients With Prediabetes Benefit From Pioglitazone

HONOLULU — The insulin-sensitizing agent pioglitazone (Actos) appeared effective for secondary prevention in stroke patients with prediabetes, according to a post hoc analysis of the IRIS randomized clinical trial reported here.

Over a median of 4.8 years, pioglitazone reduced a composite outcome of stroke and myocardial infarction by 40%, stroke alone by 33%, and new-onset diabetes by 80%, in a subgroup of stroke patients with prediabetes who adhered to the treatment protocol, reported J. David Spence, MD, of Western University in London, Ontario, and colleagues.

These associations were greater than the ones found in the full IRIS trial cohort and occurred despite a high proportion of men and smokers, those with high diastolic blood pressures, and low high-density lipoprotein (HDL) cholesterol levels in this subgroup, Spence and colleagues wrote in JAMA Neurology. Results were also presented here at the International Stroke Conference.

The IRIS trial, a study of 3,876 recent stroke or transient ischemic attack (TIA) patients with insulin resistance, reported in 2016 that pioglitazone reduced recurrent stroke or myocardial infarction by about one-fourth compared with placebo (pioglitazone 9% vs placebo 11.8%; HR 0.76, 95% CI 0.62-0.93), but raised the risk of weight gain, edema, and fractures.

Patients were included in IRIS based on their homeostatic model assessment of insulin resistance (HOMA-IR) score, and were randomized to 15 mg of pioglitazone, with dose titrated to a target of 45 mg daily, or matching placebo. For about 87% of IRIS participants, stroke was the index event.

In clinical practice, HOMA-IR scores are not commonly measured. To place IRIS results in a real-world setting, this post hoc analysis identified a subgroup of IRIS patients with prediabetes — defined by American Diabetes Association criteria of hemoglobin A1c of 5.7% to 6.4%, or fasting plasma glucose of 100 to 125 mg/dL — and analyzed effects in these patients with good adherence (taking 80% or more of the protocol dose), along with intention-to-treat results.

A total of 2,885 patients with prediabetes were included in the intention-to-treat analysis (pioglitazone n=1,456; placebo n=1,429). Mean age was 64 and about two-thirds were men. Overall, they had higher HOMA-IR scores and HbA1c levels than other IRIS participants.

Among prediabetes patients with good adherence (644 people taking pioglitazone and 810 taking placebo), hazard ratios were:

• Composite of stroke and myocardial infarction: 0.57, 95% CI 0.39-0.84
• Stroke alone: 0.64, 95% CI 0.42-0.99
• Acute coronary syndrome: 0.47, 95% CI 0.26-0.85
• Composite of stroke, myocardial infarction, and hospitalization for heart failure: 0.61, 95% CI 0.42-0.88
• Progression to diabetes: 0.18, 95% CI 0.10-0.33

In the intention-to-treat analysis, the pattern was similar but differences (also in terms of hazard ratios) were less pronounced:

• Composite of stroke and myocardial infarction: 0.70, 95% CI 0.56-0.88
• Stroke alone: 0.72, 95% CI 0.56-0.92
• Acute coronary syndrome: 0.72, 95% CI 0.52-1.00
• Composite of stroke, myocardial infarction, and hospitalization for heart failure: 0.78, 95% CI 0.63-0.96
• Progression to diabetes: 0.46, 95% CI 0.35-0.61

Compared with placebo, adherent patients had more bone fractures (3.6% vs 2.8%), weight gain of 10% or more (29.8% vs 12.0%), and edema (29.2% vs 21.6%). Adverse events in the intention-to-treat analysis followed similar patterns.

These results are globally in line with the findings of the whole IRIS trial, noted Leonardo Pantoni, MD, PhD, of the University of Milan in Italy, in an accompanying editorial.

But a point not investigated in the study — and potentially of great interest — is that drugs like pioglitazone may help prevent dementia, Pantoni observed. “Prediabetes is a condition associated with the presence of abnormalities on brain imaging that are strong predictors of cognitive decline, such as lacunar infarcts and white matter hyperintensities,” he noted. “The possibility of targeting this high-risk population might represent an appropriate approach to also prevent dementia linked with vascular brain changes.”

And because diabetes is a risk factor for Alzheimer’s disease, “the 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 added.

Pioglitazone struggles to be accepted as a secondary stroke prevention treatment, which may stem from fear of adverse effects including bladder cancer and heart failure, Pantoni pointed out. The FDA confirmed a possible link between bladder cancer and pioglitazone in 2016, though a 2017 meta-analysis found no evidence of increased risk of heart failure with the drug. In IRIS, patients with heart failure or bladder cancer were excluded.

Spence and colleagues noted several limitations to their analysis. The diabetes endpoint was based only on fasting glucose levels and patient reports. The researchers also did not perform hemoglobin A1c testing after baseline visits, and did not include oral glucose tolerance testing as part of the protocol.