With regard to carotid stenting of asymptomatic patients, how is it that so many people can analyse the same data and draw such differing conclusions? In the United States, Medicare convened a panel of experts from professional societies to discuss whether carotid stenting is an appropriate treatment for asymptomatic carotid stenosis. Opinions ranged from definitely and unequivocally “yes” to equally emphatic “no”.
While carotid endarterectomy has been proved effective in preventing stroke in asymptomatic patients, the trials are ancient in terms of best medical therapy. Best medical therapy, with optimal (rather than routine) control of lipids, blood pressure, diabetes, and smoking, continues to improve but it is unlikely to reduce the stroke risk of an asymptomatic stenosis to zero. Therefore there should be a benefit to revascularisation, if performed with minimal morbidity and reasonable expense.
Comparing carotid stenting vs. carotid endarterectomy, CREST (Carotid revascularization endarterectomy vs. stenting trial) reported that the 30-day stroke/death rate for asymptomatic patients was 2.5% for carotid stenting vs. 1.4% for carotid endarterectomy. If post procedure myocardial infarction was included, carotid stenting and carotid endarterectomy had comparable outcomes, but it is unclear if myocardial infarction is a relevant outcome. It is a marker for long-term mortality, but it is possible that an asymptomatic “chemical” myocardial infarction does not necessarily cause increased mortality but instead simply indicates significant underlying coronary disease that itself causes late mortality. For symptomatic patients, the ICSS (International carotid stenting study) similarly showed roughly double the risk of stroke/death from carotid stenting as compared to carotid endarterectomy, as have analyses from national databases and from the Society for Vascular Surgery (SVS) carotid revascularisation registry. Furthermore, carotid stenting is significantly more expensive than carotid endarterectomy.
These data suggest that carotid stenting should not be performed if carotid endarterectomy is an option, and perhaps not performed at all if carotid stenting cannot meet the improved outcomes from best medical therapy. However, CREST also showed that stroke/death was far less common in the second half of the study, suggesting that truly contemporary carotid stenting outcomes easily match carotid endarterectomy and beat best medical therapy. In addition, cranial nerve injuries occur in about 5% of carotid endarterectomy patients. From a patient’s perspective it does not matter if a neurologic deficit was caused by a central or peripheral nerve injury. Most cranial nerve injuries are minor and resolve with time, but so do most minor strokes. If cranial nerve injuries were included in the 30-day stroke/death statistics carotid stenting and carotid endarterectomy would be comparable, although carotid stenting still has a higher mortality rate.
The results of CREST may not be generalisable. CREST recruited highly experienced operators with proven good outcomes and rejected half of the physicians who applied. Despite this selection, it was only after six years and half the patients had been enrolled that morbidity dropped. In ICSS, the physician could be performing his first case, albeit with a proctor, and could operate independently after 10 cases. ICSS is likely more representative of general practice and demonstrated carotid stenting outcomes for symptomatic patients that did not meet acceptable benchmarks. The learning curve reported in CREST was confirmed in ICSS and other studies. ICSS reported that centres treating <50 patients had 30-day stroke/death/myocardial infarction rate of 10% vs 5.9% for centres treating >50 patients. CAPTURE 2 (Carotid Acculink/Accunet post approval trial to uncover rare events) reported a facility threshold of 72 cases to consistently achieve a stroke/death rate rate <3%. An analysis of Medicare data from nearly 25,000 procedures reported the mean annual operator volume was only three carotid stenting cases/year. Low volume operators (<six carotid stenting cases/year) had nearly double the mortality (2.5% vs. 1.4%) of high volume operators (>24 carotid stenting /year).
The concern about lack of generalisability of good outcomes is being addressed in the US through independent accreditation of carotid stent facilities which requires meeting the American Heart Association thresholds of 30-day stroke/death rate of 3%/6% for asymptomatic/symptomatic patients. Outcomes-based accreditation is available through the ICACSF (www.icacsf.org) or ACE (bit.ly/wzj6US) but at the time of writing is not required by Medicare.
Should asymptomatic patients be treated with carotid stenting? It depends on how good the carotid stenting and endarterectomy operators are outside of trials, but more importantly, it depends on the outcomes from truly best medical therapy. These outcomes determine whether or not the procedure is suitable for reimbursement. The final answer to this question will probably end up depending on the proposed CREST 2 trial comparing best medical therapy to carotid stenting and carotid endarterectomy.
David Sacks is past president of the Society of Interventional Radiology (SIR) and president of the Intersocietal Commission for the Accreditation of Carotid Stent Facilities (ICACSF)