European Journal of Vascular & Endovascular Surgery
Volume 39, Issue 5 , Pages 523-526, May 2010

Riding on the CREST of a Wave!

  • A.R. Naylor

      Affiliations

    • Corresponding Author InformationTel.: +44 116 2523252; fax: +44 116 2523179.

Department of Vascular Surgery, Clinical Sciences Building, Leicester Royal Infirmary, Leicester LE2 7LX, UK

Received 11 March 2010; accepted 12 March 2010. published online 19 April 2010.

Article Outline

 

“It doesn't surprise me that we got better results with stenting in the US where we have had much broader experience. North American surgeons do the job better.”

Gary Roubin, Los Angeles Times February 26th 20101

The much heralded CREST trial (Carotid Revascularization Endarterectomy vs Stenting Trial) reported outcome data to the American Stroke Association International Stroke Conference on Friday 26th February 2010.2 This study (funded by the National Institute of Neurological Diseases & Stroke with support from Abbott Laboratories) randomised 2502 patients (asymptomatic n = 1176, symptomatic n = 1326) to undergo either carotid endarterectomy (CEA) or carotid artery stenting (CAS). The trial started with the intention of only randomising symptomatic patients, but because recruitment was slow, the protocol was amended to include asymptomatic patients as well.3 CREST was not, however, powered to be a ‘stand alone’ randomised trial in asymptomatic patients. As with earlier randomised trials, interventionists and surgeons had to submit a track record before their centre was accepted; CEA was performed using the technique preferred by individual surgeons, while CAS practitioners had to use an obligatory stent (Rx Acculink, Abbott Laboratories) and embolic protection device (Rx Accunet System, Abbott Laboratories). Primary endpoints were; (i) the 30-day risk of death/stroke and/or myocardial infarction (MI) and (ii) the rate of ipsilateral stroke out to 4 years.

The main conclusions from CREST1, 2, 3, 4, 5, 6, 7, 8 were that; (i) the primary endpoint (30-day risk of death/stroke/MI) was similar 5.2% after CAS vs 4.5% after CEA, (Hazard Ratio: 1.18 (95%CI 0.82–1.68) p = 0.38), (ii) peri-operative stroke was significantly higher following CAS 4.1% after CAS vs 2.3% after CEA (Hazard Ratio: 1.79 (95% CI 1.14–2.82) p = 0.01), (iii) peri-operative MI was significantly higher following CEA 1.1% after CAS vs 2.3% after CEA (Hazard Ratio: 2.0 (95%CI 1.06–3.8) p = 0.03), (iv) more procedural strokes after CAS tended to be classed as major and ipsilateral, (v) one year after the procedure, patients who suffered a peri-operative stroke reported a poorer quality of life compared to those suffering a peri-operative MI, (vi) cranial nerve injury was significantly higher following CEA (4.8% vs 0.3%, p < 0.0001), (vii) patients aged <70 years had better results following CAS, while patients aged >70 years gained better outcomes after CEA and (viii) the four year rate of ipsilateral stroke (excluding peri-operative events) was very similar (2.0% after CAS vs 2.3% after CEA, p = 0.85).1, 2, 3, 4, 5, 6, 7, 8

Not surprisingly, headlines in national newspapers concluded that “CAS and CEA were equally as safe and effective in terms of stroke prevention”. 1, 9 The latter statement is, however, factually incorrect because EVA-3S, ICSS and now CREST have all shown that CAS was associated with a significantly higher risk of procedural stroke (almost a two fold excess risk). Notwithstanding this, the Millenium Research Group (quoted as the global authority on medical technology market intelligence5) predicted that CREST would “support growth in carotid stenting by more than 10% over the next five years”. If past experience is anything to go by, that could be a significant underestimation?

Given the occasionally rancorous past history regarding the performance and interpretation of randomised trials comparing CEA with CAS, CREST will inevitably attract a lot of professional, industry and media scrutiny, not least because their data were released on exactly the same day that ICSS (still the largest randomised trial in symptomatic patients) published completely different findings.10, 11 In direct contrast to CREST, ICSS (1713 recently symptomatic patients randomised to CEA or CAS) observed significantly higher rates of procedural death/stroke after CAS. On an ‘intention to treat’ analysis, the 30-day risk of stroke/death was 8.5% after CAS vs 4.7% after CEA (Hazard Ratio: 1.86 (95%CI 1.26–2.74), p = 0.00110). Using a ‘per-protocol analysis’, the 30-day rate of death/stroke was 7.4% after CAS vs 3.4% after CEA (Hazard Ratio: 2.16 (95%CI 1.4–3.310)). ICSS also reported the results of a substudy involving 231 patients and observed a statistically significant, threefold excess risk of new ischaemic lesions on diffusion weighted imaging on post-treatment MRI scans in patients randomised to CAS.11 Interestingly; embolic protection devices did not protect CAS patients from developing new ischaemic lesions.

When ICSS published its data, it became the fourth consecutive large-scale, multicentre randomised trial that failed to show statistical superiority for CAS (over CEA) in the prevention of stroke in recently symptomatic patients.10, 12, 13, 14 In the New York Times, however, the ICSS results were described as being “dismal”.9 To other observers, ICSS and the preceding trials reinforced intuitively held beliefs that CAS was inevitably going to incur higher procedural risks in recently symptomatic patients, while for others, the simple reason that CAS had not proved superiority (or even equivalence) to CEA was because European and Australasian CAS practitioners were somehow technically inferior to their US counterparts.15 The latter theme, undoubtedly provocative, was continued by Gary Roubin who observed that “North American surgeons do the job better” when asked by the Los Angeles Times to explain why the ICSS and CREST results were so discordant.1 However, given that the most experienced CAS practitioners incurred the highest procedural risks in the French EVA-3S study,13 this does seem to be an unduly inflammatory explanation for differences in published risk.

So has CREST settled the debate? Can we now (as suggested by the deputy director of NINDS2) offer individualized management decisions because “CREST provided doctors and patients with much needed risk/benefit information to help choose the best carotid procedure based on an individual's health history”? To this observer, the data do not (yet) support widespread and uncritical changes in practice (especially in asymptomatic patients) until we are provided with missing, vital information. Some readers will, of course, dismiss this opinion as being that of a surgeon with a biased agenda. However, while this observer believes that CAS will evolve an increasingly important role in the management of patients with carotid disease, there are important reasons (so far ignored) why CREST reported ‘the lowest published risks in any randomised trial to date”,2, 6 there are many unanswered questions that need to be addressed before clinicians can safely offer ‘individualized management decisions’ and it is not inconceivable that uncritical interpretation of the CREST data could fatally undermine ongoing randomised trials in asymptomatic patients.

In comments attributed to Thomas Brott (CREST Principle Investigator), the New York Times9 reported that “prior to the CREST trial, we did not really have the best evidence, but that these results indicate that we have two very safe and effective methods to prevent stroke”. In the same article, the New York Times offered no response to comments by Martin Brown (ICSS Principle Investigator) that three other large-scale, international trials had suggested that “surgery was safer”, while even less interest was directed towards Brown's assertion that the trial cohorts were not statistically comparable.9 Accordingly, the reader should not just consider what CREST (and the other trials) have now told you. More important is to ask what they have not told you. This missing information will, ultimately, form a crucial element in helping to “individualize management decisions”.

The obvious reason why CREST reported ‘the lowest published risks in any randomised trial to date”2, 6 was because it combined 30-day procedural risks for patients with recent onset symptoms together with patients who were neurologically asymptomatic. Procedural risks will always be lower in asymptomatic patients and (given that almost 50% of the CREST cohort was asymptomatic) it is disingenuous to uncritically claim that CREST had much lower procedural risks than SPACE, EVA-3S or ICSS. This is because the latter studies only randomised recently symptomatic patients, with no dilution of risk through the inclusion of asymptomatic patients. Accordingly, and before widespread changes in clinical practice can be justified, it is imperative that CREST provides details about the 30-day risk of death/stroke and death/stroke/MI after stratification for symptom status. If emerging ‘grape vine’ rumours prove to be correct (which suggest that the 30-day death/stroke rates were once again significantly higher in symptomatic patients undergoing CAS), it is not inconceivable that the procedural risks in symptomatic patients in CREST may be more similar to those observed in ICSS, SPACE & EVA-3S than some American surgeons and interventionists might wish to accept. It is also important to note that most of the peri-operative strokes following CAS in ICSS were classed as ‘non-disabling’ and some observers consider that this might be a relatively small price to pay for a less invasive intervention. In CREST, the opposite seems to be the case (ie more major, ipsilateral strokes after CAS). This could explain why one year quality of life measures were poorer in patients suffering a procedural stroke than in patients suffering an MI.

The second important issue relating to CREST is how these data will influence ongoing randomised trials. For example, three (statistically powered) randomised trials are currently evaluating the roles of CEA and CAS in asymptomatic patients (ACT-1, SPACE II and ACST II). However, uncritical interpretation of the CREST data could fatally alter the equipoise of some trial participants (ie leading to reduced trial recruitment because CAS is preferentially offered to asymptomatic patients) at exactly the same time that Neurologists and Stroke Physicians are suggesting that improvements in best medical therapy have significantly reduced the annual risk of stroke16 (ie moving towards a strategy of restricting access to CEA and CAS for the majority of asymptomatic patients). Until (if) TACIT secures funding (sadly, less likely now), SPACE II remains the only randomised trial in asymptomatic patients with a third limb for ‘best medical therapy’. To this observer, it would be disastrous if data from an underpowered cohort of asymptomatic patients in CREST was used as mitigation for not completing this study.

The third important issue is that CREST has provided no information about one of the most important current issues regarding stroke prevention; ie the need to offer CEA or CAS within the hyperacute period after a patient suffers their index stroke or TIA. Most health systems around the world are now beginning to recognise that the highest risk period for suffering a stroke is the first few days/weeks after symptom onset, rather than just considering the first six months as the high risk period. In the United Kingdom, NICE now recommends that CEA be performed within 14 days of symptom onset,17 while the UK Department of Health are aiming for a 48 h treatment threshold.18 There is no evidence to support a policy which encourages delays to treatment (so as to reduce the procedural risk) as this confers no benefit to the patient in terms of late stroke prevention.19, 20, 21 Evidence derived from a reworking of the pooled data from ECST, NASCET and the VA Trials suggest that the surgeon who performs CEA within 14 days with a 10% procedural risk might well attract uncritical, professional censure, but he/she will prevent more strokes in the long term than the surgeon who defers any intervention for 4 weeks and then operates with a 0% procedural risk.21

Accordingly, while CREST has been the first randomised trial to report more favourable outcomes following CAS, it would be inappropriate to completely ignore the results from the preceding randomised studies in symptomatic patients, inappropriate to conclude that there was no need to continue with randomised trials in asymptomatic patients and inappropriate to conclude that provided symptomatic patients were offered CEA or CAS within 6 months of suffering their index symptom, surgeons/interventionists were somehow offering their patients optimal practice. It is an incontestable (though sometimes unpalatable) fact that many more strokes will be prevented in the community by focussing attention upon treating symptomatic patients soon after they suffer their index TIA or minor stroke, rather than trying to treat vast numbers of asymptomatic patients. Notwithstanding the logistical problems of identifying patients with asymptomatic carotid disease, even if all were subjected to CEA or CAS, at least 95% of all strokes destined to occur in the community will still happen.22 Moreover, because 94% of CEA/CAS procedures in asymptomatic patients are ultimately unnecessary, the United States will continue to spend $2.1 billion each year on (ultimately) unnecessary interventions.22 It is, therefore, essential that we identify clinical, imaging and plasma biomarkers that will identify a high risk cohort of patients for suffering a stroke. It then remains to be seen whether CAS or CEA becomes the preferred intervention. Uncritical increases in the number of interventions in asymptomatic patients (highly likely following CREST) will do little to reduce the incidence of stroke in the community, but it will contribute to an ever increasing financial burden to health systems around the world.

In conclusion, the rapid treatment of symptomatic patients should remain the first priority of every health system. We, therefore, need to pool data from CREST, ICSS, EVA-3S and SPACE in order to perform individual patient meta-analyses to determine which patient subgroups might benefit most (or least) from CAS in the first few days after symptom onset and those who might benefit most (or least) from undergoing CEA in the same time frame. To suggest that we now have enough evidence to individualize management strategies based upon data from one trial that pooled 30-day outcomes from symptomatic and asymptomatic patients (whilst excluding important data from three other trials incorporating 3312 patients) is just not sustainable. To paraphrase Knut Rockne (b1888–d1931), “When you are riding on the crest of a wave, you are more likely to miss something”.

Back to Article Outline

Conflict of Interest 

None.

Back to Article Outline

References 

  1. Maugh TH. Stroke study puts two procedures on equal footing. http://articles.latimes.com/2010/feb/26/science/la-sci-stroke27-2010feb27[accessed 05.03.10]
  2. Landmark NIH . Trial comparing two stroke prevention procedures shows surgery and stenting equally safe and effective. http://www.nih.gov/news/health/feb2010/ninds-26.htm[accessed 05.03.10]
  3. Heartwire . CREST: stenting and endarterectomy show similar net safety and efficacy for carotid stenosis. http://www.theheart.org/article/1050701.do[accessed 03.03.10]
  4. Bhatt DL. Cardiosource for the American College of Cardiology. http://www.cardiosource.com/clinicaltrials/trial.asp?trialID=1907[accessed 05.03.10]
  5. Medical News Today . CREST clinical trial results find that carotid artery stenting is as safe and effective as conventional surgery. http://www.medicalnewstoday.com/printerfriendlynews.php?newsid=180975[accessed 05.03.10]
  6. CREST trial results announced at . 2010 International Stroke Conference. Endarterectomy and stenting are similar in safety and efficacy but show differences in stroke, heart attack and death rates at certain ages. http://vasculardiseasemanagement.com/content/crest-trial-shows-comparable-results-cea-and-cas-to-prevent-stroke[accessed 05.03.10]
  7. Medscape medical News . CREST shows similar net safety and efficacy with stenting vs endarterectomy for carotid stenosis. http://medscape.com/viewarticle/717669?src=mp&spon=35&uac=85459SG[accessed 10.03.10]
  8. Vascular News . CREST puts carotid stenting back in the game. http://www.cxvascular.com/vn-latest-news/crest-puts-carotid-stenting-back-in-the-game[accessed 10.03.10]
  9. Rabin RC. Study finds stents effective in preventing strokes. http://www.nytimes.com/2010/02/27/health/27stroke.html?ref=health
  10. International Carotid Stenting Study Investigators. Carotid artery stenting compared with endarterectomy in patients with symptomatic carotid stenosis (International Carotid Stenting Study): an interim analysis of a randomised controlled trial. Lancet 2010; 375:985–97.
  11. Bonati LH, Jongen LM, Haller S, Flach HZ, Dobson J; Nederkoorn PJ et al. New ischaemic brain lesionsnon MRI after stenting or endarterectomy for symptomatic carotid stenosis: a substudy of the International Carotid Stenting Study (ICSS). Lancet Neurol 2010;9:353–62.
  12. CAVATAS Investigators . Endovascular versus surgical treatment in patients with carotid stenosis in the Carotid and Vertebral Artery Transluminal Angioplasty Study (CAVITAS): a randomised trial. Lancet. 2001;357:1729–1737
  13. Mas J-L, Chatellier G, Beyssen B, Branchereau A, Moulin T, Becquemin J-P, et al. Endarterectomy versus stenting in patients with severe symptomatic stenosis. N Engl J Med. 2006;355:1660–1671
  14. SPACE Collaborators . Stent Protected Angioplasty versus Carotid Endarterectomy in symptomatic patients: 30 days results from the SPACE Trial. Lancet. 2006;368:1239–1247
  15. Brandler E. Doctor here says European carotid stent studies flawed. www.allbusiness.com/management-companies-enterprises/4058985-1.html[accessed 8.07.10]
  16. Abbott AL. Medical (non-surgical) intervention alone is now the best for prevention of stroke associated with asymptomatic severe carotid stenosis: results of a systematic review and analysis. Stroke. 2009;40:e573–e583
  17. www.nice.org.uk/CG068[accessed 02.07.09]
  18. www.dh.gov.uk/en/Publicationsandstatistics/Publications/PublicationsPolicyandGuidance/DH_081062[accessed 02.07.09]
  19. Rothwell PM, Eliasziw M, Gutnikov SA, Warlow CP, Barnett HJM Carotid Endarterectomy Trialists Collaboration. Endarterectomy for symptomatic carotid stenosis in relation to clinical subgroups and timing of surgery. Lancet. 2004;363:915–924
  20. Giles MF, Rothwell PM. Risk of stroke after transient ischaemic attack: a systematic review and meta-analysis. Lancet. 2007;6:1063–1072
  21. Naylor AR. Delay may reduce the procedural risk, but at what price to the patient?. Eur J Vasc Endovasc Surg. 2008;35:383–391
  22. Naylor AR, Gaines PA, Rothwell PM. Who benefits most from treating asymptomatic carotid disease: patients or professionals?. Eur J Vasc Endovasc Surg. 2009;37:625–632

PII: S1078-5884(10)00197-8

doi:10.1016/j.ejvs.2010.03.015

European Journal of Vascular & Endovascular Surgery
Volume 39, Issue 5 , Pages 523-526, May 2010

Article Tools