Discussion

Article Outline

• Copyright

Dr von Segesser (Lausanne, Switzerland). I congratulate Dr Bryan and Prof Angelini’s group for this nice study that gives answers to some questions that have been around for many years, namely, whether there is a difference between the 2 types of valves. This brings me to the first question.

In the 1990s there have been at least a half a dozen design iterations between different CarboMedics and St. Jude valve designs. Can you specify the type of valve that was used in your study?

Dr Bryan. All of these valves were standard St. Jude Medical and CarboMedics prostheses.

Dr von Segesser. There seemed to be a slight difference with regard to thromboembolism between CarboMedics and St. Jude valves, not when you look at the single valve replacements but at the double valve replacements. Can you comment on that?

Dr Bryan. None of these differences attained anything like statistical significance. The number of thromboembolic events in the double valve group was very low, really, but even so, it didn’t attain statistical significance, and the confidence limits widely overlapped.

Dr von Segesser. There appears to be, at least graphically, a difference with regard to bleeding between the St. Jude Medical and CarboMedics valves. After 10 years of follow-up you have about 77% bleeding-free survival for St. Jude Medical versus 83% for CarboMedics. If we look at a simple test like Fisher’s exact, it comes up to about a 0.7 1-sided P value. I do not claim that this is significant, but did you explore this any further?

Dr Bryan. This is an unusual article for me, because actually the other 2 authors are both statistical advisors. So it is the first study I have been involved with 2 statistical advisors and not 1, and perhaps that is a reflection of the statistical nature of these kind of analyses. I am told that there are no differences in relation to the bleeding events, and, again, I haven’t presented P values on the slides because my advisors tell me that when the confidence limits overlap widely, it is not necessary.

Dr von Segesser. I agree that testwise there may not appear to be a difference, but graphically it seemed to be impressive. I wonder if you have an explanation why there was so much more bleeding in the St. Jude Medical group?

Dr Bryan. I think all I can say is although it might appear different, if it is not statistically significant, then we have to accept that it is not different.

Dr von Segesser. I would dare not to agree. Absence of proof is no proof of absence.

I have a final question. Did you have any objective measurement for valve performance between the 2 groups?

Dr Bryan. No, we did not. There were no echocardiographic data.

Dr K. Rasheed (Islamabad, Pakistan). Congratulations on this impressive article and the quality of the presentation.

Regarding such a low rate of thromboembolism both in patients with single valves and double valves, would you tell us what INR you were maintaining for single and double valve replacements? Thank you.

Dr Bryan. With reference to the thromboembolic rate, as we all know from observational studies, there is a wealth of information in relation to these 2 valves, and the thromboembolic rate that we have recorded fits perhaps toward the lower end of those recorded in the literature, but it certainly is not the lowest.

In terms of the anticoagulation, in our country, anticoagulation is essentially community monitored; we cover a wide geographic area. So the guidelines that are instituted really are general guidelines that are decided by the British Society of Hematology. At the start of this study period, the guidelines for mechanical prosthetic valves was that the INR should be maintained for all valve models in all valve positions in a range from 3 to 4.5. This was modified in 1997 to indicate that for modern prostheses this should be adjusted. One of the reasons why I presented some of the anticoagulation data, which we have a wealth, is that actually there is little evidence of penetration of the concepts that have been brought to our attention by people, such as Eric Butchart, that we should be anticoagulating in a prosthesis-specific manner, and certainly in our population there was no evidence of this.