Evaluating the short and long term outcomes of stentless valves in the literature is difficult for several reasons: many manufactured valves are no longer in clinical use; long-term data is scarce; there exists little randomized prospective data comparing stentless versus stented valves; studies generally mix various AVR indications together (patients with root pathology differ from those with a small annulus undergoing stentless valves for avoidance of patient-prosthesis mismatch); and most importantly, the majority of studies on stentless valves include both subcoronary and full root implants together in the "stentless" arm. This results in a heterogeneous population undergoing surgery for differing indications, by surgeons with bias towards one technique over another, and using techniques that can result in vastly different hemodynamic and potentially long-term results.
It has been well proven that stentless valves offer superior hemodynamic profiles compared with their stented counterparts. Kunadian et al performed a meta-analysis of randomized, prospective trials of stentless versus stented AVR and included 10 studies and 919 patients from European centers.21 They demonstrated an approximately 6-mm Hg lower peak gradient for the stentless group, but no survival advantage at 1 year. Importantly, the vast majority of the stentless valves were implanted by subcoronary technique. As a full root replacement, stentless valve hemodynamics compare favorably to homografts, with nonsignificant differences in mean gradients reported between the two prostheses.22 Freedom from aortic insufficiency and degenerative changes also favors the Freestyle root over the homograft.23 Compared with subcoronary implantation, stentless full root replacement achieves lower gradients, larger effective orifice area, and greater freedom from at least moderate aortic insufficiency after 10 years of follow up.9 Studies on the subcoronary implant technique have consistently shown a decrease in gradient and increase in EOA over time. Early and late gradients are both low for total root replacement and generally do not change over time. This discrepancy is most likely related to absorption of hematoma in the potential space between the porcine and native aortic walls, which occurs when performing a subcoronary implant.
Reverse Remodeling and Survival
Regardless of valve substitute, AVR in patients with severe, symptomatic AS consistently results in improvements in both structural and clinical outcome compared with medical treatment. However, long-term survival after AVR remains lower than age-matched populations, for known and unknown reasons.24-26 Known factors related to higher death rates include bleeding with anticoagulation required for mechanical valves, thrombotic complications, endocarditis, and structural valve degeneration. Additionally, in patients with advanced myocardial damage and fibrosis, AVR would not be expected to allow for complete reverse remodeling, and lower long-term survival in this condition is intuitive. Indeed, low preoperative ejection fraction is perhaps the strongest predictor of long-term outcome.27,28 In patients with less advanced structural disease, however, incomplete reverse remodeling of the pressure overload state is an unknown prognostic variable. Even seven years after AVR with stented or mechanical valves, increased muscle fiber diameter and percent interstitial fibrosis compared with normal is observed,29 possibly because of persistent prosthetic LV obstruction and residual gradient.
To overcome the problem of residual gradient after AVR, stentless valves have been proposed as ideal for allowing reverse remodeling to occur. Using LV mass as the predominant surrogate marker, multiple small randomized and nonrandomized studies have attempted to assess for differences in reverse remodeling after AVR with different valve substitutes. Although demonstrating conflicting results, regression of LV mass does appear to be associated with achieving valvular competence and a low transvalvular gradient, regardless of prosthetic choice, and with treatment of postoperative hypertension.30 The ASSERT trial31 is one of the few randomized studies comparing the effects of stented versus stentless AVR on LV mass. In this study performed in Europe, 190 patients with aortic annulus size ≤25 mm were randomized to receive a Medtronic stented Mosaic porcine prosthesis or Medtronic stentless Freestyle prosthesis, implanted using a subcoronary technique. After one year, both groups demonstrated similar reductions in LV mass despite significant differences in effective orifice areas and flow velocities in favor of the stentless valves. Operative and postoperative clinical outcomes were also similar. A significant weaknesses of the study, however, was the use of the subcoronary technique for implantation of the stentless valve, which leads to higher gradients, more aortic insufficiency, and higher rates of late valve failure compared with the total root technique.9,32 Additionally, one year follow-up may be an inadequate time to detect for differences in LV hypertrophy, because changes in LV geometry occur over many years after AVR.33,34 This concept is supported by a small, randomized study of stented versus stentless valves that found no difference in LV mass regression at 6 months but a significant difference in favor of the stentless valve at 32 months.35 The most recent large randomized study compared Edwards Lifesciences stented and nonstented subcoronary prostheses in 161 patients and again demonstrated similar reductions in LV mass at 1 year. However, in patients with LV ejection fraction less than 60%, subsequent improvement in ejection fraction was greater in the stentless group.36 Because of these conflicting data, some have questioned the overall clinical importance of LV mass regression after AVR. Despite the known adverse prognostic implication of LV hypertrophy in the population as a whole, and importance of regression with treatment of hypertension, it has not been conclusively determined that regression after AVR correlates with survival.37
Gender issues are important to the topic of remodeling in aortic stenosis. For example, left ventricular adaptation to severe aortic stenosis is gender specific, with less compensatory increase in LV mass and wall tension in females compared with males.38,39 Perhaps a better indicator of hypertrophy is the LV mass/volume ratio, a parameter that is increased in women with aortic stenosis compared with men.39,40 Consequently, women with aortic stenosis tend to have worse diastolic function and exercise capacity compared with men.41 Patient-prosthesis mismatch (PPM) is also more common in women and refers to implantation of a valve too small for a patient's size. Although PPM is generally defined as an in vivo effective orifice area indexed to body surface area (EOAI) of ≤ 0.75 to 0.85 cm2/m2,42 EOAI may be preferentially evaluated on a continuum, similar to native aortic valve stenosis. Depending on definition, PPM is present in 20 to 60% of patients after AVR.43-45 EOAI incorporates both size and functional considerations because postoperative gradients (ie, persistent LV pressure-overload state) are related to the area available for blood to leave the LV and to flow, or cardiac output in this instance, which correlates well with body size. Although controversial,46-48 the preponderance of data, all retrospective, suggests that the presence of PPM is associated with reduced mid- and long-term survival.45,49-54 Techniques to prevent PPM include annular enlargement procedures and implantation of stentless valves.
History teaches a good lesson in the story of the Toronto SPV stentless valve. Pioneered by David, this was the first prosthesis widely available for use. The Toronto SPV was designed to be implanted only by subcoronary technique. Numerous reports identified the hemodynamic advantages over stented valves, and remarkably, David reported improved survival over stented valves in midterm follow-up.55 At 9 years, durability and hemodynamics were found to be excellent, with 90% freedom from structural failure.56 At 12 years, however, freedom from structural valve degeneration was only 69%, and only 52% for patients under 65 years of age at time of implant.57 Freedom from moderate or severe aortic insufficiency was only 48% for the group as a whole. The mechanism for this finding was postulated to be dilatation of the sinotubular junction, and such a mechanism could theoretically affect late results of any subcoronary stentless valve. In a group of mostly subcoronary Freestyle implants, however, Bach et al reported 92% freedom from structural degeneration at 12 years with equivalent rates in patients less than 60 years of age at the time of implant.58 One possible explanation, although not proven, for the discrepancy between SPV and Freestyle longevity is that most Freestyles have been implanted with the modified subcoronary technique compared with complete subcoronary for the Toronto SPV. The preservation of two commissural posts may create more sinotubular junction stability over time. In general, though, the lessons to be learned from the SPV data are that all valves are not created equal, even when implanted in similar manners, and that long-term and ongoing follow up is required to assess for durability because all bioprosthetic valves will eventually fail.
Contemporary series of patients implanted with stentless valves do not support a higher operative risk, even with total root replacement. It is clear that the procedure requires slightly longer aortic cross-clamp and CPB times,21 but this has not translated into higher mortality. Florath et al examined factors related to operative mortality in 1400 patients undergoing AVR, and neither stentless nor stented valve use was found to be predictive.12