SAN FRANCISCO — Hyperglycemia within 48 hours of stroke onset is common and is associated with a significantly increased risk of poor outcome, judging from findings from the Virtual International Stroke Trials database presented at the 32nd International Stroke Conference.
A total of 2,645 patients from the Virtual International Stroke Trials (VISTA) database had glucose measurement data and diabetes status available; the VISTA database includes information on individual patients enrolled in more than a dozen trials over the past decade.
In these patients, hyperglycemia within 48 hours following stroke was associated with 30% reduced risk of favorable outcome as defined by a modified Rankin Scale (mRS) of 0–1 at 90 days, Dr. Keith W. Muir reported at the meeting, sponsored by the American Stroke Association.
Prior studies also have demonstrated associations between poststroke hyperglycemia and poor outcomes. For example, a systematic review showed that the condition increased mortality threefold, with an associated increase in poor functional outcome. But most of the prior studies have been single-center studies, involved glucose measurements at up to 72 hours following stroke, and have used varying definitions of hyperglycemia, said Dr. Muir of the University of Glasgow, Scotland.
A number of questions remained unanswered, including the prevalence of poststroke hyperglycemia, risk factors for developing poststroke hyperglycemia, the best timing for testing, and its association with stroke severity, he said.
The subjects included in the current study, who were an average age of 60 years, were treated upon admission, at a median of 5.5 hours after symptom onset. Poststroke hyperglycemia was defined in all patients as glucose greater than 7.0 mmol/L.
Initial glucose measurement done at presentation showed that of the 2,645 patients, 1,126 were hyperglycemic by this definition, representing 36% of nondiabetics and 81% of diabetics, and 1,421 who initially were normoglycemic developed hyperglycemia between 24 and 48 hours later, representing 48% of nondiabetics and 88% of diabetics in the study, he noted.
Nearly half of the patients experienced an increase in blood glucose after the time of admission, and 20% of patients who were normoglycemic at admission developed hyperglycemia within 48 hours.
Since only 1,264 patients in the study had more than one blood glucose measurement, it is possible the prevalence is higher than these findings suggest, Dr. Muir noted.
Most patients (about 85% on admission, and about 80% at 48 hours) had glucose measurements in the modest range below 10 mmol/L, and only a small number (5% on admission and 7% at 48 hours) had extreme levels above 15 mmol/L. A slight association was noted between stroke severity—as measured by the National Institutes of Health Stroke Scale (NIHSS) at admission—and development of hyperglycemia, with an odds ratio (OR) of 1.03 per point on the scale. This finding was significant but “not a huge factor,” Dr. Muir said.
Later time windows of patient presentation following symptom onset were associated with a significantly lower risk of hyperglycemia (OR of 0.49 for 6–12 hours vs. 0–3 hours, and 0.50 for 12–24 hours vs. 0–3 hours).
But diabetes status was the greatest predictor of hyperglycemia, conferring a statistically significant sevenfold increase in risk, he said.
Factors associated with a mRS score of 0–1 at 90 days included age (OR 0.98), recombinant tissue plasminogen activator treatment (OR 1.44), admission NIHSS score (0.81), smoking (OR 1.47), and, most dramatically, hyperglycemia within 48 hours (OR 0.69); the absolute difference in terms of favorable outcomes between those with and without hyperglycemia was 13%, which is greater than the effect seen with rTPA treatment, Dr. Muir noted. Admission glucose, admission hyperglycemia, diabetes, hypertension, myocardial infarction, sex, and transient ischemic attack history were not significant predictors of a mRS score of 0–1 at 90 days in this study.
When looking at only the subgroup of 1,264 patients with more than one blood glucose measurement, an increase in blood glucose from baseline at admission also was a significant and independent predictor of reduced risk of good outcome (OR 0.74). This was “over and above” the risk seen with hyperglycemia in first 48 hours, he said.
The study was limited by a lack of documentation of time of blood glucose measurements, a limited number of repeat measurements, and lack of defined methodology.
Nonetheless, the findings demonstrate a high prevalence of hyperglycemia on admission and show that diabetes is a major predictor of poststroke hyperglycemia. Additionally, poststroke hyperglycemia and increases in blood glucose levels within 48 hours of admission significantly reduce the likelihood of good recovery.
A diabetic diagnosis is not a surrogate for blood glucose measurement, however, and blood glucose documentation should be mandatory in acute stroke trials, Dr. Muir concluded.
FULL TEXT