TORONTO—Preliminary evidence has provided some support for the use of a new stroke risk evaluation tool that incorporates the existing Framingham Stroke Risk Model and expands it to include other known risk factors and to take into account the effect of known preventive treatments. A small pilot validation study, which the researchers say is more a "proof of principle" study, was presented at the 2000 North American Stroke Meeting.

The new Stroke Risk Disk, available on CD-ROM through the National Stroke Association, builds on a currently available tool that focuses on the standard Framingham risk factors for stroke, said lead study investigator Don B. Smith, MD, of the Colorado Neurological Institute in Denver. "It had worked quite well, but it became increasingly clear that times had changed, and they needed to update it," he said. "It was at the American Academy of Neurology meeting last year that several of us got together and talked about a new risk disk," Dr. Smith commented.

Being able to quantitate a given patient's stroke risk has some practical advantages, Dr. Smith explained. These advantages include stratifying patients in stroke prevention trials, focusing available resources on patients at highest risk, and providing a cogent shorthand for communicating risk (among health care professionals). Also, a graphic representation of what a patient's stroke risk currently is and what it might be with risk factor modification could help in patient education.

EXPANDED MEASURES—FRAMINGHAM AND BEYOND

While the Framingham Stroke Risk Model is valid and widely known, it does not take into account important known risk factors such as prior stroke or transient ischemic attack, carotid stenosis, and cholesterol. Nor does it allow consideration of the impact of proven and widely used strategies to reduce stroke risk, including the use of warfarin for patients with atrial fibrillation and endarterectomy for patients with symptomatic carotid stenosis.

The new risk assessment tool represents an attempt to expand the comprehensiveness and clinical utility of the Framingham model, Dr. Smith explained. The Framingham model uses nine factors in the risk profile—sex, age, systolic blood pressure, antihypertensive treatment, diabetes, smoking, cardiovascular disease, atrial fibrillation, and left ventricular hypertrophy. Using these factors, a physician can calculate the patient's absolute stroke risk over a given period of time.

To include the new factors, the researchers decided to proceed by converting the absolute risk to a relative risk, that is, relative to the lowest possible risk in the Framingham profile for a person of the given age and sex. "You couldn't change your age or your sex but [theoretically] you could eliminate all the other risk factors, and that would give you your minimum risk," Dr. Smith explained.

The researchers then went to the literature to get relative risk values for a variety of other risk factors. "Then we simply multiplied it by that Framingham relative risk and came up with this expanded relative risk," Dr. Smith said.

To evaluate this expanded measure of stroke risk, they compared its predictive value in a small case-control study, with 25 cases randomly selected from a private practice database and 25 controls matched for age, sex, and race. Using risk factors culled from the files, they used the Risk Disk to calculate three measures: the absolute Framingham stroke risk, the relative Framingham stroke risk, and their expanded stroke risk.

HIGHER SENSITIVITY AND POSITIVE PREDICTIVE VALUE

In the context of this study at least, the expanded risk calculation had a higher sensitivity and positive predictive value than did the Framingham absolute risk model. Prior stroke and carotid stenosis appeared to be the most significant predictors within the expanded risk profile.

"I'm reluctant to call this a validation study," Dr. Smith said. "It's more a proof of concept—that it's plausible something like this might work." He added that it's very simplistic to assume that relative risks can be multiplied by relative risks; doing so ignores complex interactions among risk factors. In this setting, though, "it seemed to do a pretty good job." Nevertheless, establishing the validity and usefulness of this expanded risk profile will require evidence from larger, more robust datasets and studies with more rigorous design, he acknowledged.

The Risk Disk also includes educational materials and graphics that the physician can use to illustrate the process of stroke, including a patient quiz. Individual patient profiles can be saved and used to show patients changes in their risk over time. In addition, the physician can track these changes and search the saved database for research or analysis.

The Stroke Risk Disk is available through the National Stroke Association. More information can be obtained through their Web site (www.stroke.org) or by calling (303) 649-9299, ext 905.

NR