Have EVAR outcomes changed practice?


The greatest of the Great Debates at Charing Cross this year was arguably the one centered on the EVAR 1 trial. The proposition was ‘EVAR 1 and DREAM trial results have swung the choice towards endovascular repair and away from open repair in patients anatomically suitable for both’, and there were big hitters on both sides.

Peter Harris (Liverpool), speaking first for the motion, started by complementing his opponent Jan Blankensteijn (Nijmegen, the Netherlands) on “rushing to print the results of a small trial” and in so doing “promulgating this bizarre conclusion”, namely that EVAR offers no life-expectancy or health-related quality of life advantage, is more expensive, and leads to a greater number of complications and re-interventions – the conclusion drawn by The Lancet and later picked up by the British Medical Journal. These types of conclusions are “manna to healthcare economists who are more interested in pounds, shillings and pence than patient welfare”, such as those at the Belgian Federal Centre for Healthcare Expertise who recommended that “Endovascular aneurysm repair (EVAR) should not be made available in standard healthcare”, he stated.

Harris then went on, in his own words, to “confuse the argument with facts and figures”. There is a three-times greater risk of death with surgery compared to endovascular repair, and a twice as great a risk of aneurysm-related mortality. “Don’t tell me that’s not important.” The argument against endovascular repair, based on the fact that it offers no advantage in all-cause mortality, is equivalent to telling your patient that it doesn’t matter which procedure they have “because you’re going to die anyway”.

Aneurysm-related mortality is the important factor; “nobody suggested that endovascular repair conferred immortality”, he commented. Furthermore, talk of EVAR’s cost is premature; there are no valid data for post-trial practice. One by one he took apart The Lancet’s conclusions, branding them “completely inappropriate”. He suggested that a better conclusion would be: ‘EVAR offers reduced risk of death associated with abdominal aortic aneurysm with more rapid recovery and shorter hospital stay’. “Therefore, EVAR wins,” he concluded.

Venturing into ‘enemy territory’, Blankensteijn admitted that the opponents to the motion were the underdogs in this debate. He immediately went on the offensive, highlighting how even the specialist medical media were affected by ‘spin doctoring’, claiming they “cherry-picked” results. “The fact of the matter is that the outcomes of DREAM and EVAR 1 are the same.” However, referring to the motion, Blankensteijn added “the debate is not about the data but about the interpretation of the results”. Crucially, about what impact they have had on surgeons’ practices.

To this end, Blankensteijn told how he had conducted a survey of Dutch surgeons. Each participant was presented with the situation of having two patients with AAAs of 6.5cm who are anatomically and clinically perfectly suitable for both open and endovascular repair. Patient A is 65 years old with no co-morbidity, patient B is 77 years old with mild co-morbidity. The survey asked “Which treatment for either patient did you prefer before the publications of the DREAM and EVAR 1 trials, and which do you prefer after?” Results from the 192 relevant respondents showed that while around a third of surgeons changed their AAA policy to some extent, “there was no swing in choice to endovascular repair. If anything, the trials have swung the choice of vascular surgeons away from endovascular repair for relatively young patients without co-morbidity”, he finished.

Closer analysis Krassi Ivancev (Malmo, Sweden) spoke for the motion. He took the audience through a closer examination of the facts of EVAR. “There are more complications with endovascular repair, but most of them are type 2 endoleaks, which are not real problems,” Ivancev said. He also noted that that the number of reinterventions for endovascular repair did not match the number of complications.

Ivancev proposed that the skills of the endovascular surgeons involved played a huge part in the outcomes of the trial. For example, the average number of endovascular repairs per hospital at the start of the trial worked out at four per year. “Working like that is like living in a fog. But yet, in spite of the difficulties, the trials still showed good results. And that’s compared to surgical techniques that have been around for years.”

He also highlighted trends in surgeons’ practice: both from Sweden and European Vascular and Endovascular Monitor (EVEM) data. These showed that rates of endovascular repair are increasing while those for surgery are falling. In conclusion, “the only thing that matters is patient choice”, and a patient will say “I prefer a less invasive treatment from which I would have a lower risk of dying and faster recovery”, he said.

Summing up the arguments against the motion was Richard Gibbs (London, UK). He said that while both trials had level 1 evidence there was equivocation in their interpretation. “It’s difficult for us clinicians to view them objectively,” Gibbs observed. “The prism of our own prejudice is based on our own practice.” However, he admitted that as an endovascular specialist, the fact that he had adopted a position against the motion made him unusual. “It is doubly painful when I look at the data as it doesn’t prove the superiority of endovascular repair.”

Gibbs concentrated on the “headline news of mortality rate”. At four years there is no difference between endovascular and open repair. “The patients are still dying of the same things, it’s just that endovascular repair postponed it,” he opined. Almost a third of deaths in the trials were from cardiovascular causes. He said that very few patients in the trials were on statins or even aspirin. “It’s perfectly reasonable to say that this figure should have been 100%.” Therefore, this argument about aneurysm-related mortality is a red herring, he claimed. Furthermore, debating this issues is “like fiddling while Rome burns” as we are failing to take care of important risk factors.

Greenhalgh opened the discussion with a comment about how endovascular repair fits in with the “natural progression” of all treatments – from big, complicated procedures down through minimally invasive interventions all the way to prevention.

Roy Greenberg (Cleveland, OH) took to the microphone (refusing Greenhalgh’s call to take a “machismo approach to endovascular repair”), and asked if EVAR 1 was done too early, given that all the devices used in the trial have either changed dramatically since then or fallen out of use altogether. Greenhalgh replied that if you wait until the time is right, it will inevitably be too late to affect opinion. Blankensteijn agreed, adding that perhaps another trial was called for, one with better risk management.

There was nothing left for it then but to vote on the motion, and a gasp went around as the result of the close vote was revealed. However, 55% of the audience went against the motion, believing that in patients suitable for both procedures there is no swing towards endovascular repair over surgery.


For the second EVAR trial, the question under debate was “In patients unfit for open repair, the focus should be on improving fitness rather than performing EVAR early”. Opening the presentations for the motion was Janet Powell (London, UK). She commented that her slides were principally in black and white as, “this is such a black and white, cut and dried situation”, namely whether early intervention with endovascular repair was suitable for patients with large aortic aneurysms and high-risk or unfit for surgery.

Eurostar and RETA data, using an ASA category of IV, have 30-day mortality rates of just under 12% and 15% respectively. “EVAR 2 did not assess fitness using ASA criteria. Instead there were guidelines for cardiac, pulmonary and renal function, but all the patients equate to ASA III or IV. And the 30-day mortality was 9% – no-one can say the British surgeon was not doing the job competently,” she stated. However, this 9% for unfit patients “contrasts starkly” to the 1.7% 30-day mortality that EVAR 1 achieved in fit patients.

However, Powell highlighted the comparable mortality rates for the mid-term results from the EVAR 2 trial between EVAR patients and those who had no intervention. “Overall, after four years, the mortality in the no intervention group was 62% versus 66% in the EVAR group. These patients do not have a long life expectancy.” The figures remain similar whether analyzed by intention-to-treat or per-protocol. “Early EVAR is not offering any benefit,” she stressed. “These patients have a short life expectancy, therefore quality of life is important. There was no difference in quality of life between the groups.”

So, the key is to examine the data to see if there are any clues as to how to improve results of EVAR. Powell highlighted the fact that some patients in the ‘no intervention’ group crossed over to have open repair. “The reasons why these patients underwent intervention are relatively poorly described. I think there is a strong suspicion from reading the case notes that fitness had been improved.” She called for a change of thinking to allow early EVAR if the risk of aneurysm rupture exceeds 10% in the following 3-4 months. In other words, if the aneurysm is larger than 7cm in diameter. “For AAA less than 7cm – the majority in the EVAR 2 trial – the risk of rupture is likely to be less than that of an early EVAR.”

To improve fitness, Powell recommended that patients be referred to a cardiologist for treatment of ischemia, arrhythmias and heart failure. Statin therapy should also be considered, as should a peri-operative beta-blockade, and exercise is good as well. “These patients are elderly and frail, we can’t put all these things together at once. We need to do it in phases and stages. It will take at least three months to get these patients ready.”

Dr Rodney White opened the case opposing the motion. “Dr Powell is too respectable for me to do the usual debacle,” he joked. White stressed that the question raised by EVAR 2 is really ‘what do we tell these patients?’. “They are not convinced that any of these data are any good. What they’re worried about is their risk of rupture – how we are going to be able to prevent that and get them through an operation.” This relates to the “primary goal of aneurysm repair: to prevent rupture,” he observed.

Using data from the Lifeline Registry – pooled results from US clinical trials on commercially available devices – White showed that operative mortality for both EVAR and surgical controls were both less than 2%. “How many of these people would fit a high-risk or unfit for surgery category? About 30-40% of patients meet the EVAR 2 criteria.” So where does the 9% come from? “There’s a total mismatch in the data; either it is the auditing, the follow-up, or somehow the definitions do not make sense,” White opined.

It is clear from looking at the Kaplan-Meier curves for all-cause mortality between endograft patients and surgical controls that they look the same. “These patients all die at a similar rate,” he said. “Clinically, these are all high-risk folks. By the time they get to be 76 with an aneurysm and co-morbidities their curves match like that then. They are very comparable.”

White described EVAR as “the only reasonable, randomized trial that addressed high-risk patients; everything else has been a registry”. The 9% mortality rate was actually affected by a large number who died before intervention. Removing those brings the figures closer into line with those from Lifeline. “It may be that randomization itself is an independent risk factor,” he supposed. “We have to treat these patients with some sort of urgency.”

In summary, doctors need to look at co-morbid conditions, although they do not have an impact other than “minor medical tuning up”. EVAR should be done early, he claimed, “because 50% of them in EVAR 2 who had to wait died before an intervention. That in itself is a clear indication that high-risk patients should not wait”.

Ken Ouriel (Cleveland Clinic, OH), started his talk by dealing with some criticisms of the EVAR 2 trial. “Criticism #1: The rate of aneurysm rupture was lower than expected in the ‘no treatment’ group.” He answered this by saying that in this group the mortality was balanced between death from rupture and peri-operative deaths. “So if you will, the endovascular mortality rate just never caught up.”

The second criticism is that surgeons were unskilled. However, it was basically the same surgeons in EVAR 2 as EVAR 1, where a very respectable 1.7% mortality rate was achieved. The third criticism Ouriel highlighted is one that White mentioned: “There was an inordinately long period of time between randomization and repair in the EVAR treatment group.” The median time was 57 days, during which time eight died. “However, the non-aneurysm related deaths would have occurred anyway, maybe even being hastened by the procedure.” Removing ruptures prior to treatment, there would have been absolutely identical deaths in the two groups at 67 each, he claimed.

In conclusion, Ouriel estimated that “overall, some proportion of the 48 deaths would have been preventable with improved fitness – maybe not a lot, but some”.

Summing up for the opposition was Kim Hodgson (Southern Illinois University). “To argue the case for early repair and against watchful waiting, I must impugn the results of EVAR 2.” This means showing that the trial is flawed, that its conduct and interpretation are flawed, and that there are other data available that refute its findings. “Why should I want to do this, other than to just make an enemy of Roger [Greenhalgh]?” he asked. “Because the results of trials like EVAR 2 are being used to formulate public health policy.”

Hodgson started by highlighting the different interpretations of ‘high-risk’, emphasizing that although EVAR 2 authors praised the ASA system for its ‘simplicity’, they had gone on to say it had ‘proved too difficult to use’ for the trial. “They instead chose to substitute it with something that, in my opinion, is even more open to interpretation.” In other words, leaving the decision on the whole to the surgeon.

“EVAR 2 results themselves speak to the subjectivity of these kinds of classifications,” he added. Other trials have suffered from the same problems, said Hodgson. “High-risk in many centers is sometimes not as high-risk as they would have us believe.” This also begs the question of whether it is a high-risk patient or high-risk surgical technique, for example general versus local anesthetic. Or, indeed, high-risk surgeons. “EVAR 2 participation only required 20 previous EVAR cases – many would consider that still well within the learning curve for this relatively complex procedure.” He also speculated that many of the patients recruited to the trial should not have been EVAR candidates in the first place, suggesting that the trial investigators may have been “pushing this technology”.

Hodgson then explained that in his experience, “fitness rarely improves with age. It takes decades to become unfit”. He also asked “what in fact is the endpoint?” Patients ‘becoming fitter’ according to Powell might in fact have had aneurysms enlarge to the point where the patients themselves demanded treatment. Few patients were on statins, aspirin or beta-blockers.

He concluded by saying that EVAR 2 provides “level 1 evidence, but with many demerits”. Instead, the Lifeline Registry provides a much better picture (see front page). “To close, it’s not uncommon for patients to say ‘If I was your mother what would you recommend?’ And I always um and ah and give them an answer. But then I ask them ‘Don’t you want to know if I like my mother?’ Basically, patients need to know that we’re on their side. Leave these sorts of decisions to the politicians.”