Recently, an article published in Circulation: Cardiovascular Interventions by Kuehnl et al suggested the lack of a clear relationship between operator experience and outcomes for the vast majority of surgical and endovascular carotid interventions in Germany over a 6-year period.1 This type of analysis would likely be almost impossible in a country the size of the United States, with a vastly greater number of centers performing these procedures as compared to a relatively small country like Germany, with dissimilar hospital systems and referral patterns, and differing capacities for data collection and adjudication. The fact that this on the surface appears to be a very comprehensive and all-inclusive analysis does not, however, imply that the reported outcomes are entirely accurate or wholly applicable to the daily practice of carotid intervention. As with any retrospective observational analysis, there are numerous pitfalls and opportunities for bias in the data that cannot be avoided.
Coronary stents when first released were often touted as the “great equalizer” among coronary interventionists. Subsequently, however, literally dozens of well-conducted studies have repeatedly disproved this rather naïve notion. Although this might hold true for the simplest type A lesions, we know convincingly that high-volume operators have the highest success rates and the fewest number of complications. This is particularly true for high-risk patients and complex-lesion subsets. Low-volume operators are routinely associated with poorer outcomes, including lower success rates, higher complication rates, longer hospital lengths of stay, and higher costs per procedure. This is true for most procedures requiring advanced technical skill sets and has also been reported repeatedly and convincingly in the surgical literature. Why then, would it seem to appear that these very consistent observations supported by solid data would not hold true for carotid percutaneous intervention, where the slightest mistake could result in major neurologic consequences? Damage or infarction of even a small amount of neurologic tissue is likely to have far greater daily consequences – including functional capacity and cognitive ability – than a small bump in a troponin level.
Most importantly, this is not a randomized controlled and blinded study. As such, selection bias is without question present in these data. We do not know the basis on which the surgical or endovascular option was chosen: Was it based on carotid anatomy? Was it based on arch anatomy? Was it based on referring physician preference? Was it based on operator preference? Was it based on patient preference? Was it based on regional or hospital preferences and patterns? What devices were used – open or closed stent design? Was distal protection used? Was the operator a low-volume carotid operator, but a high-volume endovascular operator? The outcomes of such a physician would likely be superior to a low-volume carotid/low-volume overall endovascular operator. Was the “low-volume” operator a novice or a seasoned veteran with years of endovascular interventions and experience under their belt in spite of their low annual volume?
Self-reported outcomes, as tabulated in this publication, are also notoriously plagued by bias and underreporting as well. This is particularly true with carotid interventions, where subtle motor, sensory, and cognitive changes may be entirely missed without independent and experienced neurologic assessment and adjudication. In 1996, Rothwell et al2 reviewed 51 studies reporting on the incidence of stroke and/or death following carotid intervention. Interestingly, stroke rates were averaged at 2.3% with a single surgeon author, 5.5% with multiple surgeon authors, 6.4% with a neurologist author included, and 7.7% with independent neurologic assessment and adjudication. In the publication by Kuehnl et al, we know nothing about who or how the neurologic assessment was made. We do not know if the same parameters and variables were used and reported by each of the hospitals and physicians from whom these data were obtained, nor do we know the accuracy of these data. How rigorously is quality assurance maintained by each of the hospitals and physicians included in this report? Furthermore, we know only the outcomes during the periprocedural period and not in follow-up. We know nothing of the lengths of stay or other procedural complications, such as hematomas, retroperitoneal bleeds, transfusions, acute renal insufficiency, etc, all of which are likely to be higher with an inexperienced operator, and all of which have acute and long-term consequences. Finally, we know nothing of the patient co-morbidities and demographics, including their cardiovascular risk factors or prior cardiovascular history and events. We do not know how rigorously these patients were managed medically pre or post procedure. The devil is in the details and we don’t have very many.
Of the few strengths of this study, the most important is that it at least raises the question for discussion of the importance (or lack thereof) that carotid interventionists maintain a high volume of interventions in order to remain competent and to safely maintain acceptable outcomes. Unfortunately, this current study still does not definitively answer this question. Also unfortunately, in the United States, this has been a nagging question, as under the current FDA regulations and CMS reimbursement policies, percutaneous carotid interventional volumes for asymptomatic patients have dropped off dramatically in recent years, turning many highly skilled interventional physicians into low-volume carotid operators. Although most interventionists would firmly agree that the currently available data support and prove clearly that carotid stenting is at least as safe and effective as carotid endarterectomy (in appropriately chosen cases), if not more so, politics, both medical and governmental, have continued to trump data and logic. Hopefully, the CREST 2 study, which is currently enrolling, will finally provide the data and ammunition to convince even the staunchest critics of carotid artery stenting that this is a safe, valuable, and necessary service that must be made routinely available for our patients.
Address for correspondence: Richard C. Kovach, MD, Deborah Heart & Lung Center, 200 Trenton Road, Browns Mills, NJ 08015. Email: firstname.lastname@example.org
- Kuehnl A, Tsantilas P, Knappich C, et al. Significant association of annual hospital volume with the risk of inhospital stroke or death Following carotid endarterectomy but likely not after carotid stenting. Circ Cardiovasc Interv. 2016;9(11).
- Rothwell PM, Slattery J, Warlow CP. A systematic review of the risks of stroke and death due to endarterectomy for symptomatic carotid stenosis. Stroke. 1996;27:260-265.