Author + information
- John A. Elefteriades, MD* ()
- ↵*Reprint requests and correspondence:
Dr. John A. Elefteriades, Section of Cardiac Surgery, Boardman 2, Yale University School of Medicine, 333 Cedar Street, New Haven, Connecticut 06510
The paper by El-Hamamsy et al. (1) in this issue of the Journaloriginates from the group of the distinguished Sir Magdhi Yacoub, who has unsurpassed experience in the treatment of the diseased aortic root. This study appropriately identifies a question whose answer is unknown: specifically, whether an animal aortic root (Medtronic Freestyle graft, porcine in origin [Medtronic Inc., Minneapolis, Minnesota]) or a homograft is a better option when a biological root replacement is required.
This study by El-Hamamsy et al. (1) is valuable for a variety of reasons: not only is the main study question appropriate, but also the study is randomized (within a single center) and of long duration (median follow-up 7.6 years). Also, the Freestyle and homograft groups are remarkably similar in their baseline characteristics, setting an appropriate stage for outcome comparison.
The study was commercially funded (Medtronic makes the Freestyle graft), although none of the investigators identified personal financial conflicts of interest.
The study showed excellent overall results, with a 1% surgical mortality among patients undergoing only root replacement (without concomitant procedures) and 4.8% overall; the mortality is increased by the need for concomitant procedures (mainly coronary artery bypass graft surgery).
The main findings of the study include the following: 1) there was no difference in hospital mortality between Freestyle and homograft groups; 2) there was no difference in late valve-related complications; 3) there was no difference in late functional class; 4) there was no difference in late survival (which, in both groups, approached that of an age- and sex-matched general population); and 5) there was a marked difference in valve performance, with the homografts showing a much higher incidence of aortic valve dysfunction (86% vs. 37%, p < 0.001) and need for reoperation, which occurred only in the homograft group (all for valve dysfunction).
The longer follow-up is an important strength of this study. All too often in reports of biological valves, follow-up is inadequate to reveal valve deficiencies and deterioration, which are usually late phenomena. Indeed, a prior, short-term report of the same patients from the Yacoub group revealed no significant differences, even in valve function. In the present paper by El-Hamamsy et al. (1), thanks to the longer-term follow-up, valve deficiencies do come to light. These deficiencies are all confined to the homograft group, with the Freestyle valves showing consistently excellent performance in the longer follow-up.
Shortcomings of the study include the following. 1) The “n” was not great (166 total between the 2 groups). 2) The follow-up still falls within the range best described as “midterm,” with the longest patients only 11 years out from surgery, and as we know vividly from studies of biological valves of all types, 20-year follow-up is really necessary to know a valve's true colors. 3) Follow-up was not entirely complete (87% and 88% in the 2 groups). In this electronic era, we hope that computerized national registries of citizens can provide close to 100% follow-up, at least as regards a patient's being alive or dead. One always fears that lost patients may be dead patients or dissatisfied patients who have turned elsewhere for their subsequent care. 4) There are some potential sources of bias that may favor the Freestyle group. First, there is an exclusion criterion for “a known systemic illness affecting long-term survival”; this is a broad criterion, and subconscious investigator bias can occasionally operate via such exclusion criteria. Second, late echocardiographic follow-up was a bit more complete in the homograft group (83% vs. 76%), raising the possibility that there was better detection of valve problems in the homograft patients. 5) Not all homografts are the same. In fact, 2 different kinds were used in this study, 1 “fresh” and antibiotic sterilized and the other cryopreserved. A variety of techniques and manufacturers provide disparate types of human aortic roots, which fall under the heading of “homograft.” The long-term behavior of these different subtypes may vary. 6) There are 2 curious negative findings affecting the Freestyle group. Eight patients in the Freestyle group required renal replacement therapy compared to 1 in the homograft group; the reason for this difference between the 2 groups is unclear. Similarly, 8 patients in the Freestyle group (as opposed to none in the homograft group) had heart block. 7) For 7 patients in the Freestyle group, the cause of death was unknown, compared with only 2 unknown causes of death in the homograft group; this finding raises the possibility of an undetected cardiac problem in those Freestyle patients. When all is said and done, all these concerns represent relatively minor shortcomings for a study conducted in a “real-world” setting, and the findings of this study deserve notice and accommodation in our practices and policies.
Specifically, this study provides relatively strong evidence in favor of our using animal aortic roots over homografts. The late valve deterioration in the homograft roots pinpointed in this study deserves notice and respect from the cardiologic and cardiac surgical communities.
The curious susceptibility of homograft aortic roots to calcify has not been fully explained, but appears to be an immune reaction to cells remaining in the homograft (2–4). This propensity to structural deterioration in the homograft aortic wall is likely related to the tendency for the valve itself to deteriorate.
It is important to look at the general context of this paper, in addition to the specifics. One should consider the general choice of valve or root replacement prosthetics. El-Hamamsy et al. (1) point out, and illustrate in their last figure, the mechanical inferiority of stented biological valves, in terms of effective functional valve area, and they emphasize the long-term downstream benefits of maximizing prosthetic aortic valve functional cross-sectional area. However, most surgeons would perform a valve replacement rather than a full root replacement for isolated aortic stenosis, which was the primary problem in most of the patients in the present paper.
Another option for root replacement is, of course, a mechanical composite graft. Our own data (5), among those of others, indicate that a mechanical valved conduit is an almost “bullet proof” choice for the contingent with aortic valve disease and root aneurysm, yielding almost undetectable bleeding and embolism rates yet conferring the excellent hemodynamics and durability of a mechanical valve. It is worth noting that, in the present study, despite having a biological valve, nearly one-quarter of both the Freestyle and homograft patients were receiving warfarin therapy.
The present study cannot clarify or resolve the strong debates and prejudices regarding choice of stented versus nonstented or biological versus mechanical valves. This important study by El-Hamamsy et al. (1) does, however, show us that when we do wish to implant an unstented biological valve, we should choose a Freestyle graft over a homograft. Such a choice will optimize the likelihood of durable valve function in the medium term.
As the authors point out, endocarditis does provide the 1 exception that favors the use of a homograft, because of the “unique benefit due to the mitral flange,” which can be of immense technical utility in filling voids and infected spaces in an aortic root decimated by infection.
Dr. Elefteriades is a consultant for Baxter and Medtroinc; has received research support from Celera and NSF, and has principal in CoolSpine. Medtronic manufactures the graft reported in the paper to which this editorial applies.
↵* Editorials published in the Journal of the American College of Cardiologyreflect the views of the authors and do not necessarily represent the views of JACCor the American College of Cardiology.
- American College of Cardiology Foundation
- El-Hamamsy I.,
- Clark L.,
- Stevens L.M.,
- et al.
- ↵Farkas E, Tranquilli M, Anderson MS, Hatzaras I, Elefteriades J. Are thromboembolic and bleeding complications a drawback for composite aortic root replacement? Paper presented at: New England Surgical Society 90th Annual Meeting; September 11–13, 2009; Newport, RI.