Mr. Kaiser, Mr. Pairolero, Members and Guests: In the name of my co-authors and myself, we thank you for the extreme privilege of being able to give this Maxwell Chamberlain Memorial Lecture.
Nine thousand six hundred consecutive primary coronary bypass patients were followed up with a completeness of follow-up of 99.9%. At 10 years after surgery, 86% of these patients were free from any post-CABG infarcts. One-third of these patients had received anginal warnings before their infarcts. The hazard function for first infarct after coronary bypass surgery is a three-phase hazard function. The early phase is not visible on this plot because it is so short; there is a constant phase and there is a late phase rising for the whole extent of the follow-up.
Please note that at two years after surgery, the late phase takes dominance over the constant phase. If we explode the timeframe for the first 30 days, it now becomes visible that the early hazard function stays active until around 5 days after surgery, when it reaches the plateau of the constant phase. A series of incremental risk factors were identified in their relationship with the event. They include: patient variables, procedure variables and institutional variables. On this table, you do not find the actual variables but grouping of variables by, e.g. cardiac comorbidity and vascular comorbidity. And you can identify that the patient-variables influence different phases with different strengths.
Let us explore the group of procedure variables. Under procedure variables, we identified general technical variables and arterial grafting variables. Patch grafting, coronary endarterectomy and incomplete revascularization were strongly related to the early and the constant phase. Non-arterial grafting and, in particular, non-arterial grafting in the presence of single-vessel disease, influenced strongly the early and the constant phase. Non-arterial grafting, in particular to the LAD, influenced the late phase.
We will first quantify the benefit of using at least one single arterial graftpreferably to the LAD for a median patient. On the horizontal axis you see the variation of the spectrum of age. The zero line of the vertical axis means no difference; the top is the benefit for arterial grafting. Please start by noting that the lines are horizontal, thereby stressing that, although age is statistically significant, it is clinically not so relevant. Please note also that the separation of the solid lines increases over time, thereby pointing out that the longer the extent of the follow-up, the larger the benefit. For example, at 15 years, this benefit for this median patient, whatever the age, is around a 10% reduction of infarct rate. The p value of that difference is statistically highly significant over the whole spectrum of age. You as well as we, were very disappointed that we could not identify a benefit for more extensive arterial grafting.
We would like to explore, together with you, some of the information we have identified. Let us start with the univariable non-time-related analysis of the number of events by the number of in-situ mammary artery anastomoses. If there is one arterial graft that is perfect or should be perfect, then it is the in-situ mammary artery aanstomosis. A gradual decrease is seen of the prevalence of infarcts. But, of course, this is univariate and non-time-related analysis. The use and method of use, as in most institutions, has been inconsistent over time, so we have to take care of that time relationship. Please note, the number of patients at risk at 5 years (alive and without having suffered the event). For example, at 5 years, there are still 583 patients with 2 in-situ anastomosis. Five years, that is already 3 years into the late phase, when the late-phase influences should already have become visible.
Another univariate projection is the time related one. We have corrected for some of the time relatedness and what you see is an actuarial presentation of the freedom from infarct. Please note the lower line is when there are no in-situ mammary artery anastomoses. But also note that the freedom curves for single-and two-mammary artery anastomoses are really one on top of the other. And the one for three mammary artery anastomoses stays on top of that same line for the whole extent of the follow-up; this is without correction for any patient variability.
What is the total gain for a median patient? Well, we can calculate a patient-specific prediction for a median patient. A median patient does not have that many risk factors, the use of one arterial graft improves his freedom from angina at 10 years from 86 to 91%, that is a 5% gain at 10 years. And if we are reasonable, we cannot expect that adding a second and a third arterial aanstomosis will each then add another 5% because then the numbers would get very difficult to understand.
Is infarct a clinically relevant problem for the patient? At the AATS, we presented the post-CABG return of angina. The return of angina is not strongly related to survival. The infarct, on the contrary, had a dramatic consequence on the life expectancy of the patient. This is survival after infarct, time 0 is now not the date of surgery but time 0 is the date of the infarct. We immediately see a drop of around 20% in the first month after the infarct and then there is a stable survival rate. We can stratify this by the presence or not of an arterial graft. Indeed, before correction for any patient variability, this arterial graft improved survival. Other variables as age, left ventricular dysfunction and the presence of diabetes influenced this survival and with singular strengths.
In conclusion, the first post-CABG infarct is a rare event. For the whole group it is 14% at 10 years and one-third of the patients received anginal warnings before this infarct. The one-month's mortality after the first post-CABG infarct is 21%. The occurrence of this infarct has to be avoided by all therapeutic means. Arterial grafting, preferably to the LAD seems mandatory in most patients. More extensive arterial grafting is unlikely ever to reduce the hazard due to the rarity of the event, the limited life expectancy of the current CABG patient with an average age of ± 72 years, the influence of other variables add the not-yet identifiable benefit .
I thank the STS for the privilege of this Maxwell Chamberlain Memorial Lecture.