Naylor writes that the “much publicised rationale is flawed by the simple fact that the poorer survival rates observed in CREST were not attributable to a greater proportion of endarterectomy patients dying following their procedural myocardial infarction. In fact, a relatively higher proportion of carotid artery stenting patients suffering a perioperative myocardial infarction died during follow-up. The clinical reality is that up to 10% of patients will suffer a stroke within seven days of their index transient ischaemic attack and the benefits of intervening in the hyperacute period after onset of symptoms (ie. offering greater stroke prevention) will far outweigh any potential consequences of perioperative myocardial infarction and reduced life expectancy.”
Naylor summarised the review of the guidelines, by stating that concerns about perioperative myocardial infarction should certainly inform, but not drive the debate as to whether endarterectomy or stenting is preferable in ‘average risk’ patients. “Unfortunately the subject has now assumed a level of influence that is not justified on the basis of evidence. The choice of intervention will inevitably depend upon a number of factors, most notably the need to intervene early. However, the CREST MI study has not shown that the risk of perioperative myocardial infarction should deflect attention away from the most important priority; the prevention of stroke. This is one situation where the heart should not rule the head!”
Naylor spoke to Interventional News about his editorial.
In a nutshell why is myocardial infarction not as important as stroke, in your view?
The sole purpose for intervening in any patient with carotid artery disease is to prevent stroke. Perioperative myocardial infarction may have some relevance regarding late survival and should certainly inform the debate, but it cannot assume equivalence or superiority over the primary goal of stroke prevention.
Some physicians feel that myocardial infarction is a significant predictor of long-term survival, however your editorial does not seem to take this view. Why?
A typical interpretation of the data from CREST (as was evident in Sumaira Macdonald’s piece in the November 2011 issue of Vascular News) is that “carotid endarterectomy is associated with a higher myocardial infarction rate (that impacts significantly on four year mortality) than carotid stenting”. The clear implication from this kind of statement is that the excess mortality observed in patients suffering a procedural myocardial infarction was attributable to an excess of late deaths in the endarterectomy group. Actually, the converse was true. A greater proportion of carotid artery stenting patients who suffered a perioperative myocardial infarction in CREST (27%) died during follow-up as compared to those suffering a myocardial infarction after endarterectomy (17%). Only 13 patients in the entire CREST cohort of 2,500 patients died prematurely during follow-up after having suffered a perioperative myocardial infarction. This represents a tiny fraction of those randomised within the study (<1%) and it would be inappropriate to allow this type of statistic to dictate management strategies. There are other more important issues to be considered.
Why do you think the American Heart Association (AHA) guidelines were liberalised on the basis of this “flawed” logic?
The AHA guidelines inexplicably ignored the findings of the ICSS trial (which was published one year before the CREST data were released) and based its decision to liberalise carotid artery stenting indications almost solely on the CREST data. With few exceptions, however, every large randomised trial (including CREST) has found that endarterectomy is associated with significantly lower procedural death/stroke rates than carotid artery stenting in symptomatic patients. The perioperative myocardial infarction issue has clouded the debate (wrongly), for the reasons alluded to above. There is no doubt in my mind that carotid artery stenting has a role in the management of patients with carotid disease, but the way the data have been interpreted and portrayed in the AHA guidelines is misleading.
Why do you think some of the important trials in the past did not include myocardial infarction as an endpoint?
Historical. Virtually every preceding trial used 30-day death/stroke as its primary endpoint. Personally speaking, I have no problem including perioperative myocardial infarction within any endpoint provided that the data are interpreted appropriately. In the case of CREST, I believe that the way the data were interpreted (regarding the impact of perioperative myocardial infarction) led to the misconception that most (all) of the premature deaths were occurring in endarterectomy patients who suffered a perioperative myocardial infarction. That was not the case.
In your view, what message should be promoted now?
There is a complementary role for carotid endarterectomy and carotid artery stenting in managing patients with carotid disease. Speaking personally, it is no secret that I believe that too many asymptomatic patients undergo unnecessary interventions. It is essential that we identify a high-risk (for stroke) cohort in whom to target endarterectomy or stenting. However, the absolute priority must be the rapid treatment of symptomatic patients. Intervening as soon as possible after onset of symptoms will prevent far more strokes in the long-term and will completely negate the relatively minor impact of perioperative myocardial infarction on late survival. At present, the available evidence suggests that endarterectomy can probably be performed more safely within the first 14 days after symptom onset, but that may change as technology for carotid artery stenting improves (eg flow reversal protection systems). The key message, however, should be that encouraging delays to treatment in order to get lower procedural risks is not beneficial to patients in the long term because the highest risk period for stroke is the first few days after onset of symptoms. It is a salutary fact that the available evidence suggests that a surgeon performing carotid endarterectomy within two weeks with a 10% death/stroke rate will probably prevent more strokes in the long term than the surgeon who delays intervention for four weeks and then operates with a zero per cent procedural risk. The same rationale will apply for stenting.