Conference Coverage

COMMENTARY—Inappropriate Control Group Compromises Study Results


There is a known risk of stroke whenever surgery or intravascular treatments are performed on brain-supplying blood vessels. That risk involves complications immediately or soon after the intervention, as well as a subsequent risk of stroke that relates to the original vascular disease and implies that the procedure was not effective in preventing a new stroke. The surgical and interventional procedures invariably involve one artery and do not prevent strokes caused by abnormalities in other vessels, the heart, or the aorta.

Drs. Olena Seminog and Michael Goldacre of Oxford University studied the frequency of subsequent ischemic stroke in patients with ischemic stroke who had surgery or interventional treatment of brain-supplying arteries. These authors attempted to quantify the risk of stroke in these patients by comparison with the occurrence of stroke in a reference cohort of patients who had minor medical and surgical conditions stratified by age, sex, admission year, area of residence, and socioeconomic status.

Unfortunately, they chose the wrong group to use as comparators. The comparator group’s surgical procedures were not related to blood vessels that supply the brain or vascular conditions that pose a stroke risk. This group had little immediate risk of stroke, and the patients’ conditions were likely to be irrelevant to long-term stroke risk. The control group should have been patients with vascular lesions similar to those treated surgically or interventionally. Also, an analysis of strokes that developed in the same vascular territory would be more meaningful, because the surgical and interventional procedures only treat one vessel.

—Louis R. Caplan, MD
Professor of Neurology
Harvard Medical School
Beth Israel Deaconess Medical Center

Next Article: