Kevin S. Attenhofer, MD
Diabetes is becoming increasingly prevalent worldwide, with over 30 million people diagnosed in the United States as of 2015. It is no secret that diabetes is an independent risk factor for stroke. In fact, mortality is higher and post-stroke outcomes are poorer in patients with stroke and uncontrolled glucose levels.
In some patients, a phenomenon of stress hyperglycemia is observed at the time of stroke. This is a relative increase in glucose during an acute critical illness. It is an ill-defined metric with no consistent definition in the literature. Previous studies have shown that stress hyperglycemia is a better predictive biomarker of critical illness than absolute hyperglycemia. The authors of this paper sought to determine an association between stress hyperglycemia and incidence of new stroke or TIA following index ischemic stroke.
The authors analyzed a subgroup of 3000 patients from the CHANCE (Clopidogrel in High-Risk Patients With Acute Nondisabling Cerebrovascular Events) trial. For their purposes, they defined stress hyperglycemia as a ratio of fasting plasma glucose (FPG) within the first 72 hours after index event (within 48 hours of randomization) to the glycated albumin (GA). They did not use Hemoglobin A1c because chronic kidney disease and anemia can affect the accuracy of this number. After calculating the FPG/GA ratio, patients were then categorized by quartiles for comparison. The primary outcome was new stroke during the 90-day follow-up period (same as the CHANCE trial).
Patients in quartile 1 had the lowest median FPG and highest median GA (4.9 mmol/L (88 mg/dl) and 18.9%, respectively). Patients in quartile 4 had the highest median FPG and lowest median GA (6.2 mmol/L (111 mg/dl) and 14.2%, respectively). Compared with patients in quartile 1, patients in quartile 4 had ~ 1.5-fold increase in risk of stroke recurrence by 3 months.
The authors conclude that stress hyperglycemia is associated with an increased risk of early stroke recurrence. They state that this association is independent of absolute hyperglycemia on admission; however, their own data reveals that patients in quartile 4 had statistically significantly higher fasting plasma glucose levels in the acute phase when compared with quartile 1.
This study is limited in many ways. It is a post-hoc subgroup analysis that is limited in the information available to the authors. There is a lack of data regarding stroke cause, duration of hyperglycemia, intervention prior to FPG measurement, and subsequent glycemic control during the study.
In the end, it seems that higher glucose levels at time of stroke portend poorer outcomes. The next question revolves around how to best obtain glycemic control. Currently, much of our glucose management literature comes from trials in general medical patients, with new and promising stroke-based trials of glucose control on the horizon.
