In 1995, a national trial revealed that surgery can reduce the risk of stroke in symptom-free men whose carotid arteries are narrowed by fatty deposits. This finding has led many physicians to examine apparently healthy patients who have risk factors for carotid artery disease. If a partial blockage is found, the artery is cleaned out.
The problem, says Colin P. Derdeyn, M.D., assistant professor of radiology at Washington University School of Medicine in St. Louis, is that many of these men soon die from other causes such as heart disease. Moreover, the diagnostic and surgical procedures themselves carry a small risk of stroke. So should doctors go looking for carotid artery stenosis? Or should they act only if symptoms appear?
Derdeyn and William J. Powers, M.D., associate professor of neurology and radiology, have addressed this dilemma with a computer model. "I think this is the first study that has looked at this problem from a cost-benefit standpoint," Powers says, "because the necessary data were not available before 1995."
The researchers asked if ultrasound screening of asymptomatic men either once at age 60 or every year for the next 20 years would be beneficial. The men would have risk factors such as heart disease, evidence of peripheral vascular disease or a bruit - a murmur heard through a stethoscope placed over the carotid artery. "Our model shows that, for men at high risk for carotid artery disease, a one-time screening would prevent strokes in a relatively cost-effective manner," Powers says. "Annual screening would be prohibitively expensive, even though it also would prevent strokes. And screening low-risk men every year would produce more strokes than it would prevent."
The model also shows that screening would be beneficial only if diagnostic and surgical standards are high. "Physicians need to get three pieces of information before they refer patients," Derdeyn says. "They need to know the accuracy of the ultrasound lab, the complication rate of the angiography facility that would subsequently image the artery, and the stroke or death rate at the surgical facility."
Derdeyn and Powers report their findings in the November 1996 issue of Stroke. Funds from the National Institute of Neurological Disorders and Stroke, Siemens Medical Systems and the Charles A. Dana Foundation supported the research.
The ACAS Trial
The carotid artery runs through the neck, branching to the face and brain. In people with atherosclerosis, fatty deposits may build up at the branch point - the place where you feel your pulse. With time, the deposits may block the artery completely or a piece of plaque may break off and lodge in the brain. The result is a stroke, which kills brain cells by depriving them of oxygen and nutrients. The dying cells then wreak havoc on surrounding parts of the brain.
Stroke is the leading cause of long-term disability in industrialized countries, accounting for the lingering symptoms of 2 to 3 million people in the United States. It also kills half a million Americans each year, and carotid artery disease is thought to be the cause 20 to 30 percent of the time.
The Asymptomatic Carotid Atherosclerosis Study, the trial that revealed that surgery can reduce the risk of stroke in asymptomatic men, assigned 1662 patients to a medical treatment group - daily aspirin and lifestyle changes - or to a group that also underwent endarterectomy, a procedure that opens up the artery and cleans out atherosclerotic plaque. All participants had carotid arteries whose diameter was reduced by at least 60 percent. After 2.7 years, 11 percent of the medically treated patients had suffered strokes compared with only 5.1 percent of patients in the surgical treatment group. The results were reported in JAMA on May 10, 1995.
Ultrasound And Angiography
Derdeyn and Powers combined the data from ACAS and other studies to look at the benefits and costs of using ultrasound to screen 1,000 asymptomatic men. They assumed that those who tested positive for carotid stenosis would be sent for angiography and that those who really did have the disease would then have one of their carotid arteries surgically cleaned.
Ultrasound is the most practical method for screening because it is noninvasive and inexpensive. Using a machine that resembles those used by obstetricians, the radiologist beams high-frequency sound waves at the neck, obtaining crude images that indicate whether the artery is normal. The sound waves bounce off red blood cells, measuring their speed. If the artery is blocked, the cells move faster than normal, just as water moves more quickly through a garden hose that is squeezed partly closed.
Because ultrasound cannot yield accurate images, patients with suspected carotid stenosis are subsequently sent for an angiogram, in which a catheter placed in the thigh is threaded to the neck.
X-ray dye injected into the catheter enables a radiologist to obtain detailed images of the constricted vessel.
Angiography cannot be used for screening because it is too expensive. Moreover, 1.2 percent of angiography patients in the ACAS trial suffered strokes, presumably because the catheter dislodged plaque or encouraged the formation of clots.
The Model's Predictions
Derdeyn and Powers take this angiography complication rate into account in their computer model. They also include the prevalence of carotid stenosis, the sensitivity and specificity of the ultrasound method, the stroke and death rates linked to endarterectomy, the risk of stroke with medical treatment, subsequent death rates from other causes, quality of life estimates and estimated costs of the procedures.
The simulation calculates the number of years of life that screening would save. These are expressed as quality-of-life years (QALYs), where a stroke-free year counts as 1 QALY and any year after a stroke counts as 0.8 QALY. The model also generates overall costs - costs incurred through ultrasound ($109 in 1995), angiography ($2,000) and surgery ($9,000) minus costs saved over 20 years by preventing strokes, deaths and the need for long-term care. Dividing overall costs by total QALYs gives the cost per QALY.
The model revealed that 1,000 asymptomatic, high-risk men who were screened for carotid stenosis once at age 60 would enjoy a total of 30 more QALYs than 1,000 similar men who were not screened. The cost of adding each QALY would be about $35,000 - "relatively cost-effective compared with some other medical procedures," Powers says.
Annual screening would provide a total of 7 QALYs, the model revealed, but at a cost of $458,000 per QALY. "That would be impossible to justify," Powers says.
One-time screening of low-risk men would add 7 QALYs at a cost of $53,000 per QALY. But annual screening of this population would be very detrimental because it would take away 9 QALYs, the model revealed.
Quality Control
By varying the data fed into the model, Derdeyn was able to see which factors most affect the
cost-benefit ratio. "One of the messages I hope people will take away from this study is that you need to address three basic issues," he says.
The first is the accuracy of the ultrasound lab. "Although the results here at Barnes-Jewish Hospital are very impressive, there can be incredibly lousy ultrasound in some otherwise good places," Derdeyn says. "So physicians and insurance providers need to demand some measure of quality assurance from wherever they send patients. There should be data that compare the performance of ultrasound with angiography on the same patients."
The model also was sensitive to complication rates from angiography. "Very low risks are achievable in places that do a lot of angiograms and do them well - the rate here is 0.3 percent," Derdeyn says. "But you would want to think twice about sending patients to facilities with complication rates 10 times higher."
Surgical risk was a third critical factor. "Centers that participated in the ACAS trial had only a 1.5 percent risk of stroke or death within 30 days of endarterectomy," Derdeyn says. "But this level of expertise is not available everywhere. Complication rates in the community can vary from 0.5 to 20 percent."
Symptom-Free Men Only
The Washington University model does not apply to men with symptomatic carotid stenosis - those who have had full-blown strokes or mini-strokes called transient ischemic attacks. Such patients face a 27 percent risk of having another stroke within two years unless they undergo surgical treatment. The model also is not applicable to women. "The data for women are more complicated because women's risks are lower and their surgical morbidity may be higher," Powers says.