Outcomes in Carotid Endarterectomy Performed by Vascular Surgeons or Neurosurgeons
To the Editor:
The article on New York State carotid endarterectomy (CE) outcomes by Hannan et al1 reports an admirably low adverse outcome rate across all cases studied but also shows, as their title suggests, a surprising difference in outcome favoring the patients of vascular surgeons versus neurosurgeons (adjusted odds ratio for adverse events in neurosurgeons relative to vascular surgeons, 3.17 [95% CI, 1.26 to 7.97]). The authors themselves point out several reasons why this result may be erroneous, such as a possibly higher rate of detecting postoperative strokes by neurosurgeons. Another important difference that might have influenced the result was that the proportion of asymptomatic patients was 67% for vascular surgeons and 53.3% for neurosurgeons; additionally, the adverse outcome rate for asymptomatic patients is lower in this and other series.2 Perhaps more importantly, the study sample was clearly not representative of all CE procedures done in New York State. The authors studied a voluntary registry, including only 7% of the surgeons in the state doing CE, among whom the adverse event rate was substantially lower than for the rest of the state.
Most importantly, the authors acknowledge that this difference between surgeons was not found in four other studies of CE outcomes. Given the increasing recognition among research methodologists that Bayesian reasoning is required to interpret statistically significant findings from new research in the context of preexisting research,3,4⇓ we would propose that the finding of outcome differences across surgical specialties needs to be interpreted with great caution.
The study by Hannan et al1 uses elegant analytic methods and has shown that very low adverse outcome rates can be achieved with CE. However, we believe that the finding of a significant difference in outcomes between the two surgical specialties studied is not externally valid.
We would like to thank the authors for their comments on our article. We agree that another potential caveat is that the neurosurgeons had a lower percentage of asymptomatic patients, but we would like to emphasize that the presence of preoperative symptomatic stenosis was controlled for in the process of calculating risk-adjusted mortality rates for the surgeon specialties.
We agree that the study sample was not representative of the state as a whole, and that the findings are contrary to what was found in other studies. These facts were pointed out in the caveats to the study. Nevertheless, we believe that our findings are extremely interesting and worthy of further investigation because, unlike others, we were able to identify a strong relationship between risk-adjusted adverse outcomes and the use of various processes of care. Furthermore, the use of these processes of care was very strongly related to surgeon specialty, and this relationship was not investigated in other settings. Consequently, we believe that it remains to be seen whether these findings are reproducible in other settings.
Ultimately, we hope these data may suggest ways to improve outcomes for all patients undergoing carotid endarterectomy regardless of the specialty of the surgeon because the elements of care that we have identified as being associated with better results can be used by any qualified surgeon performing this procedure.