Intended for healthcare professionals


Is there a risk in avoiding risk for younger patients with aortic valve disease?

BMJ 2011; 342 doi: (Published 26 May 2011) Cite this as: BMJ 2011;342:d2466
  1. Tom Treasure, professor of cardiothoracic surgery1,
  2. Asif Hasan, consultant cardiac surgeon2,
  3. Magdi Yacoub, professor of cardiac surgery3
  1. 1Clinical Operational Research Unit, UCL (Department of Mathematics), London WC1H 0BT, UK
  2. 2Freeman Hospital, Newcastle-upon-Tyne, UK
  3. 3Heart Science Centre, Harefield Hospital, Harefield, UK
  1. Correspondence to: T Treasure tom.treasure{at}
  • Accepted 14 March 2011

Tom Treasure, Asif Hasan, and Magdi Yacoub argue that a culture of risk avoidance in cardiac surgery may mean patients are not getting the most appropriate treatment

Around 30 000 patients in the UK and Ireland had open heart surgery to replace the aortic valve in the five years up to March 2008, 40% of whom also had coronary bypass grafts. The annual number of replacements steadily increased, reaching 7000 in the last recorded year.1 Valve replacement is highly effective in averting the risk of sudden death and fatal deterioration in cardiac function, and the risk of perioperative death is now below 2.5% for all patients under 80 years old and 1% in patients under 40 years.2 Nevertheless, patients living with a replacement valve face a time related accrual of life threatening and disabling events.3 Tissue valves fail over time, and recipients of mechanical valves must take anticoagulants for life, which reduces but does not entirely prevent thrombosis and embolism and increases the risk of bleeding. We pose the question whether in ensuring that surgeons’ and institutional death rates are brought to the lowest possible levels, as they have been superbly well for older patients, insufficient consideration may have been given to younger patients.

Downsides of valve replacement

The shortcomings of heart valve substitutes have not been overcome despite years of effort.4 Mechanical valves, typically made of pyrolitic carbon and titanium steel, are extremely durable, but lifelong anticoagulation is mandatory. In a meta-analysis of 35 observational studies, including 23 000 patients and over 100 000 patient years the annual rate for combined thrombotic events (valve thrombosis and embolic events) for mechanical aortic valves was 1.4% at the usual intensity of anticoagulation.5 Higher intensity anticoagulation (international normalised ratio >3) reduces the rate to 1%, but the gain is balanced by higher bleeding rates.6 Over years there is a substantial cumulative risk of stroke, bleeding, and death. Thus for patients under 40, who might realistically hope for another 40 years of life, the 1% perioperative risk of death is an ever diminishing proportion of lifetime hazard. Survival of these patients, who typically have moderately severe congenital abnormalities of the aortic valve, progressively falls below that of an age matched population.7

Both thromboembolic and bleeding risks can be minimised by using xenograft tissue valves and avoiding anticoagulation, but their limitation is the inevitability of valve failure. The only option to prolong survival is then reoperation. Deterioration is slower in older patients: only 10% of those who have surgery over the age of 65 and who survive to 80 are likely to need further surgery during that time.8 But for patients under 40 there is about 50% chance of having another aortic valve replacement within 15 years.8 Xenograft tissue valves have short term attractions as they permit freedom to travel, to engage in sport, and to have babies, but further valve replacements must then be included in expectations for life.

Alternative treatment

Given that younger patients face the virtual inevitability of at least one further operation or the hazards of thrombosis or bleeding, or any combination of the three, maybe it is time to reconsider the options. A candidate operation is the pulmonary autograft, first described by Donald Ross in 1967.9 The patient’s pulmonary valve is excised along with a cylinder of pulmonary artery and is used to replace the diseased aortic valve. The pulmonary valve and artery are then replaced with a human cadaver pulmonary artery and valve (known conventionally as a homograft).

Although the Ross pulmonary autograft operation is used by surgeons whose work includes children growing up with congenital disease of the aortic valve, it is rarely used for adults in Britain. Among 653 adults aged 18-39 years having an elective aortic valve replacement in the three years 2007-9, only 13 had this operation. The median age of the 13 was 24 years, and there were no perioperative deaths.2 Should the autograft operation be used more frequently? Should fully informed patients know that this surgery is one of their options, and if so what should they be told about their prospects with this operation compared with other solutions?

For information on outcomes there is a randomised trial and a systemic review and meta-analysis.10 11 Perioperative mortality for the Ross operation in the trial was 1% (1/108).10 In the meta-analysis the pooled early mortality was higher at 3% (95% confidence interval 1.8% to 4.9%). In the randomised trial (in which the control patients had homograft aortic valve replacement) all operations were done by one surgeon (MY), and there was 99% actuarial freedom from re-operation at 13 years among the 108 patients randomised to the autograft. One patient had an operation, and that was at 9.5 years for dilatation of the autograft causing regurgitation.10 In a meta-analysis of 1749 adults who had a pulmonary autograft in 29 centres, the rate of deterioration of the valve in its new aortic position was 0.8%, varying from 0.15% to 1.9%.11 Deterioration is more often caused by dilatation of the thin walled pulmonary artery than by cusp failure, and technical modifications to support the valve have reduced this problem. Although failure of the allograft valve used to substitute for the transferred native pulmonary valve was a substantial problem early on, in October 2010 the informally derived consensus at the Ross summit in Atlanta, Georgia, was that the problem has largely been solved by routine use of oversized pulmonary homografts.

In the primary objective of providing a living and enduring aortic valve replacement, proponents of the pulmonary autograft believe it has stood the test of time. There is no refuting that this operation requires a patient who presented with disease of one heart valve to undergo double valve replacement,12 and the skill and practice required to attain consistently good results should not be underestimated. Nevertheless, when implanted by an experienced surgeon in a centre with expertise the living pulmonary valve has proved capable of functioning for prolonged periods in the high pressure arterial circulation.

Whenever there is doubt, there is always a cry for more and better evidence. It is unlikely that a randomised trial can be used to resolve the areas of uncertainty. Although we support randomised studies and have used them to inform valve replacement strategies,10 13 14 this is not a choice of one valve versus another based on a short term and therefore retrievable outcome. A direct comparison of early outcomes and surrogates would favour simpler surgery because of the longer period of cardiopulmonary bypass required and the greater technical pitfalls to be negotiated with the autograft. Perioperative death and annual rates of valve failure are around 1%, and many years would be required for meaningful data to accrue. Data are already available for each adverse event (tissue valve failure, thromboembolism, bleeding), but finally we have to offer the patient a choice of risks: a haemorrhagic stroke after mechanical prosthesis implantation or a reoperation for biological prosthesis degeneration.3 A better research strategy might be modelling15 and individualised patient information by simulation.16 17

Taking the long view on risk

Even if meticulous analysis of what we know already defines a type of patient for whom this operation offers the best life time strategy, there would still be obstacles to implementation. Risk avoidance among surgeons is already thought to block the route to cardiac surgery for some adult patients when an operation might be in their best interests. In the case of the Ross operation, intolerance of even a small increase in immediate risk could impede access to a better long term solution for these patients. Cardiac surgeons have achieved remarkable reductions in perioperative risk, but they are aware that in a highly monitored practice an obsession with what can easily be counted—that is, perioperative deaths—might now be stifling innovation and be an obstacle to implementing improvement. It might be time to take a longer view and a more comprehensive perspective that considers the life time risks and benefits of a patient hoping for 40 years free of stroke, anticoagulant related bleeding, and re-operation, rather than minimising short term risk, but that change will have to be carefully managed.

Hasan went to considerable lengths 10 years ago to promulgate the Ross operation by its safe introduction in a structured programme of training and support between centres.18 Despite successful operations under his mentorship, surgeons did not adopt the technique, and the profession still has considerable wariness. This is undoubtedly a challenging operation to learn compared with the highly reproducible insertion of an off the shelf prosthetic valve. It would take many surgeons outside their comfort zone in these times when apparent poor performance is not forgiven, but the declining use of this operation is not fully justified by the available data.

The first step would be a dispassionate review of the data. At present it too often falls to individual cardiologists and surgeons to take their own view on the evidence, and that is not something they are trained to do, or can do objectively. Independent analysts might help.19 At present it falls to individual surgeons to champion innovation,18 but in a harshly critical world who will break ranks and take a risk? Implementation would have to be directed, controlled, and, most importantly, supported. Mentoring, supervision, and debriefing would be elements of such a process, not traditional in the culture of surgery. Not every surgeon has to learn uncommon operations such as this; they need only be performed by relatively few expert hands. It would be advisable to designate centres and individuals within those centres to provide this service.

In due course the Ross operation might be consigned to history for good reasons. Minimally invasive valve surgery may make a succession of tissue valve operations feasible. Home monitoring of warfarin has resulted in considerably better control and reduction of events.20 New oral drugs may simplify anticoagulation and even reduce thromboembolic risk without fear of bleeding. The dream of a mechanical valve that is thrombosis resistant may become reality. But the obstacles to innovation in a culture of caution will persist. We need the means to overcome them as part of the science of improvement,21 if not for the pulmonary autograft, perhaps for other interventions where obstacles to implementation, such as short term risk, might obstruct the way to better long term solutions.


Cite this as: BMJ 2011;342:d2466


  • doi:10.1136/bmj.d2871
  • , doi:10.1136/bmj.d2865
  • We thank Ben Bridgewater and members of the Society for Cardiothoracic Surgery in Great Britain and Ireland for access to the database and for several colleagues who have read drafts of the manuscript and given advice.

  • Contributors: MY, AH, and TT discussed the need to bring the evidence concerning the Ross Operation to those outside cardiac surgery. TT wrote the first draft. All authors commented on the draft and read and approved the final manuscript.

  • Competing interests: All authors have completed the ICJME uniform disclosure form at (available on request from the corresponding author) and declare no support from any organisation for the submitted work; no financial relationships with any organisation that might have an interest in the submitted work in the previous three years; and no other relationships or activities that could appear to have influenced the submitted work.

  • Provenance and peer review: Not commissioned; externally peer reviewed.


View Abstract