Comment on Fractional Flow Reserve-guided PCI as compared with coronary bypass surgery (FAME-3) trial

Background:

Most studies have demonstrated a benefit of CABG over PCI for the treatment of multivessel coronary artery disease. Nevertheless, FFR-guided PCI yielded better results than medical therapy in patients with stable coronary artery disease. The difference was driven by a lower rate of urgent repeat revascularization, which suggests that optimal clinical results for PCI can be achieved by functional assessment. Therefore, studies comparing FFR-guided PCI with modern coronary surgery were lacking.

Study design:

The FAME-3 study was designed as a multicenter, prospective-randomized, non-inferiority trial including 1500 patients in 48 centers worldwide. Of the 1500 patients enrolled, 757 underwent FFR-guided PCI and 743 were assigned to CABG. The primary endpoint was the occurrence of Major Adverse Cardiac or Cerebrovascular Events (including death from any cause, myocardial infarction, stroke, and repeat revascularization) at 1 year. Secondary endpoints included the composite of death, myocardial infarction, or stroke as well as safety endpoints such as stent thrombosis, bleeding, significant arrhythmias and acute renal failure. The non-inferiority criterion was set to an HR upper boundary of less than 1.45 with a 95% confidence interval. Due to additional data published from the BEST, NOBLE and EXCEL trial, the upper boundary for non-inferiority was set to HR>1.65 during the recruitment period.

Major results:

The mean SYNTAX score was 26.0 and 25.8, for PCI and CABG patients respectively. In the PCI group a mean number of 3.7 stents were used with a median total length of 80 mm. In the CABG group a mean of 3.4 peripheral anastomoses were performed, multiple arterial grafts were used in 24.5% of the patients. Regarding the primary endpoint, the incidence of MACCE at 1 year was 10.% in the FFR-guided PCI and 6.9% in the CABG group providing a HR of 1.5, 95% CI 1.1-2.2 (p=0.35). As a result, FFR-guided PCI did not meet the non-inferiority criterion set by the investigators.

There was a significant difference in MACCE after stratifying according to the SYNTAX score. The primary endpoint rates were similar in patients with less complex disease. In patients with intermediate or high SYNTAX score, the benefit of CABG over FFR-guided PCI in terms of MACCE was evident.

The differences between the groups favoured CABG particularly in patients >65years of age.

There were no significant differences between the groups with regard to the secondary endpoints:
Death: HR=1.7, 95%CI=0.7-4.3
Myocardial infarction: HR=1.5, 95%CI=0.9-2.5
Stroke: HR=0.9 95%CI=0.3-2.4
Death, Myocardial Infarction or Stroke: HR=1.4, 95%CI=0.9-2.1
Repeat revascularization: HR=1.5, 95%CI=0.9-2.3
Regarding the safety endpoints, the incidences of bleeding, acute kidney injury, atrial fibrillation and rehospitalization were higher in the CABG group. Definite stent thrombosis in the PCI group and symptomatic graft occlusion in the CABG group were documented in 0.8 and 1.3% of the cases, respectively.

Interpretation:

Several studies delved into the value of functional assessment of coronary stenosis as a method to improve the indication and outcome of revascularization in patients with coronary artery disease.

The FAME 1 study demonstrated that measurement of FFR in elective PCI candidates reduces the rate of MACE (composite of death, myocardial infarction and repeat revascularization) by reducing the number of unnecessary stents by almost 30%. The main positive effect was observed on the need for repeat revascularization and on the composite death and myocardial infarction.

The FAME 2 trial addressed the question whether FFR improves outcome of PCI over optimal medical therapy in patients with chronic coronary syndromes. The study was prematurely stopped after enrolling 888 randomized patients, due to the significant benefit of FFR-guided PCI in terms of the composite of death, myocardial infarction or urgent revascularization. The major driver leading to the statistical advantage of functional measurement of stenosis was an 87% reduction on urgent revascularization at 1 and 5 years. The favourable outcome of patients without signs of functional stenoses treated with medical treatment highlighted the importance of FFR measurements as discriminator of prognostically relevant stenoses.

The effect of FFR evaluation in the setting of acute myocardial infarction was investigated in the FLOWER-MI trial. The trial compared a conventional angiographic vs FFR-based strategy for the revascularization of non-culprit lesions. In this setting the FFR-strategy was not found to be beneficial over angiography-guidance in terms of death, myocardial infarction or urgent revascularization at 1 year. This finding is not consistent with the findings of the original FAME trial but emphasizes the need for completeness of revascularization in a setting of acute myocardial ischemia. Both repeat myocardial infarction rates and repeat urgent revascularization tended to be higher in the FFR-guided group. This fact supports an older observation which demonstrated the role of coronary lesions which have been stratified as non-significant in previous angiographies. Those initially non-significant stenoses can turn to culprit lesions and in the course of the disease cause acute myocardial infarction due to plaque rupture or thrombosis.

There are several take-home messages stemming from the FAME 3 trial:

1. CABG remains the treatment of choice for complex coronary artery disease. The procedure has evolved not only with regard to the higher use of arterial grafts which set as target an improved coronary perfusion, translated to lower myocardial infarction and mortality rates in the long-run. We now see that even in the short-term (1 year) outcomes of CABG the benefit is getting obvious: perioperative complications such as bleeding (3.8%), acute kidney injury (0.9%) and rehospitalizations (10.2%) are significantly reduced over time. Perioperative myocardial infarction defined in a clinically relevant and fair way is extremely low. The same is true for the rates of symptomatic graft occlusion (1.3%) and repeat revascularization (3.9%). The results may be attributed to the standardization of the use of arterial grafts, the expansion of quality measures such as transit flow time measurements, and optimization of risk stratification (e.g. the use of preoperative CT scan).
2. The differences between FFR-guided-PCI and CABG in terms of clinically relevant endpoints at 1 year were unexpected. Several prospective randomized trials such as the SYNTAX, BEST and FREEDOM trial have already highlighted the benefits of surgical revascularization over PCI in different clinical settings. All trials mentioned above, followed a similar pattern: The curves between CABG and PCI started separating between 2-3 years after treatment and the differences in clinical events became more evident at 3 and 5 years. With regard to the endpoints, it all starts with differences in repeat revascularization and turns to differences in myocardial infarction and mortality. In case of the FAME 3 trial, the events seem to start within 3-6 months after treatment and major clinical endpoints -though not powered for the analysis- demonstrate a clear trend in favor of CABG (Myocardial, infarction: CI=1.46, 95%CI=0.92-2.33, Cardiac death: CI=1.75, 95%CI=0.99-3.08). Taken together, the PCI results are better than in the SYNTAX trial, but the results of CABG have shown a remarkable improvement which makes the difference at 1 year.
3. The main concept of FFR-guided PCI to increase the efficacy of treated lesions by targeting those which are responsible for the events and by avoiding complications from the treatment of unnecessary lesions, does not seem to work. The same is true for FFR-guided surgical revascularization, as it has already been shown by the GRAFFITI and the IMPAG trial. Undoubtedly, the use of FFR-guidance leads to a reduction of the number of stents or bypass grafts used and significantly simplify the procedure. However, this is associated with lower rates of complete revascularization at least regarding the anatomical point of view. T
4. The differences between anatomical and functional completeness of revascularization still remain unclear. Both conditions are snapshots of a specific measurement timepoint and are prone to short-run changes. For historical reasons, assessing the completeness of anatomical revascularization is easier and more standardized based on long experience of quantitative coronary angiography. We have known for years that stable plaques without causing significant stenosis may still undergo abrupt rupture, thus leading to myocardial infarction and vessel thrombosis. Confirming functional relevance of an angiographic stenosis is a different procedure. As we know only 14% of the patients with angiographic 3 vessel-disease had a functional impairment of all three territories. The FAME 3 trial demonstrates that the arguments to address angiographic significant non-functional stenoses are predominant.
5. Why is CABG more beneficial than PCI regardless of the functional degree of the stenosis? First, because apart from the anastomotic area (usually 5-6 mm of length) the target vessel remains intact; as a result, the target vessel is exposed to a limited risk regardless from the functional degree of the stenosis. The opposite is true for PCI targets, the patency of which is closely related to the total length and the diameter of stents used. Second, because the success of coronary grafting is not affected by the progression of coronary artery disease, as long as the distal run-off is not compromised. On the contrary, PCI is both prone to the progression of disease distally to the inserted stents, let alone in-stent-restenosis.
6. This hypothesis is further supported by the equipoise between FFR-guided-PCI and CABG in less complex coronary artery disease (low SYNTAX score). Those patients suffer from more focal stenoses rather than severe diffuse disease. The latter requires implantation of long stents or neglection of lesion which are supposed to be innocent at the time of treatment but still causing events in the near future.

Taken together, although the data on completeness of revascularization in FAME 3 are underway, one can hypothesize that the anatomical completeness of revascularization is the key to midterm and probably also long-term success. It has been demonstrated in several studies that this principle is better served under the rules of surgical rather than percutaneous revascularization in patients with chronic coronary syndromes.