Valve replacement: Present status of homograft valves

Abstract

We have surveyed the medical literature for reports of cardiac valve replacement. While the objectives of these reports are often diverse, they usually refer to the functional adequacy of the valve in question. Frequently the objective is to compare the function in two or more valve types. A variety of statistics can be derived from these patient data. Patient survival is the most consistently reported figure and is undoubtedly the most important and most accurate factor to consider. While actuarial curves are a well-described method of reporting these survival figures, there are still many centers that report flat percentage figures that do not consider the crucial period of postoperative follow-up. In addition to these actuarial survival curves, some attempt should be made to indicate which deaths are related to cardiac valve replacement and which are clearly unrelated. While there always must be some bias in this selection, it is probably most accurate to include in the valve-related category all sudden deaths, unknown and questionable deaths, and all patients lost to follow-up. This will clearly define those patients who have succumbed to causes completely unrelated to the valve. These valve-related figures obviously should also be reported on an actuarial basis. Along with an analysis of survival, the incidence of complications must be considered in evaluating the clinical results. While these data are often far less accurate than the survival data, they can reflect the general comparative trends in different valve types. For example, while definite thromboembolic episodes are usually reported, suspicious, unknown, or sudden episodes are rarely included. The incidence of hemorrhage from Coumadin is almost never indicated. It is clear from the assessment of these series that the complication rate must also be expressed on an actuarial basis as a percentage of those patients who are alive and well in a given follow-up interval. The combined incidence of deaths and complications will provide the most accurate impression of how well a particular valve can function. Figure 9 presents our total homograft experience as an example of actuarially reported survival, valve-related deaths, and valve-related complications. In statistical reporting of human cardiac valve data, the postoperative interval of follow-up is critical. With the types of valves presently being used, intervals shorter than 18–36 mo do not provide us with a significant follow-up period. Not until follow-up data reaches the 5–10-yr mark does it become meaningful to the assessment of long-term valve function. In reviewing the literature, no group with a 5–10-yr follow-up reports greater than 50% of the original patients to be alive and well without having suffered any significant complication since surgery. This applies to all types of valve replacement. This low incidence of perfect results is counterbalanced by the fact that deaths attributed directly to the valve are consistently less than 10%. This does not mean that in further follow-up many of these problems and complications are not going to result in a more significant mortality rate. It is also repeatedly stated that the recent advances and alterations in valve design will significantly influence the long-range outlook of valve replacement. it must be carefully indicated that this is speculation and is not supported by factual data. Attempting to draw finite conclusions from data comparing valve types is difficult. Two or three impressions stand out. It would appear, for example, that the fresh aortic homograft has a significantly better long-term record than other types of tissue valves. Both the frozen-irradiated homograft and the gluteraldehyde heterograft are reporting excellent early results. Of the prosthetics, the Smelloff-Cutter valve has probably the best long-term results, and, as opposed to most, it has not undergone recent design change. Cloth-covered valves have had a definite impact on the incidence of thromboembolism and can be used without the ongoing risk of anticoagulants. We are impressed by our observation that the earlier Starr-Edwards valve invariably had attached fibrin and thrombus deposits, even several yr after implantation. Cloth-covered valves are free of these persistent thrombi and significantly decrease the ongoing patient risk from embolism. Unfortunately, the effects of this design change may not be entirely beneficial. In all five of the cloth-covered valves we have inspected at reoperation, there has been definite fabric deterioration and in one instance the cloth dehisced completely from the cage. It is not possible for us, today, to state which valve type is best. The results of valve replacement with short-term follow-up has certainly improved markedly over the 10–12 yr of use. Statistics, however, do not yet indicate any significant improvement in late follow-up data.