Coronary angiography remains the gold standard imaging modality for the precise anatomical characterisation of coronary artery stenoses. However, despite the introduction of quantitative analysis, coronary angiography cannot, however, be relied upon to delineate the haemodynamic impact of intermediate diameter stenoses accurately (i.e., 30-80%) and so should not be used as a stand-alone guide in such circumstances. Measurements of coronary pressure and flow can provide essential data that complement the topography of stenotic lesions elucidated by coronary angiography and, when the two are combined, allow for more robust decisions on revascularisation to be made (Figure 1A, Figure 1B and Figure 2).
Figure 1. A proximal left circumflex (LCx) artery lesion of debatable anatomical significance by coronary angiography. A) Left anterior oblique caudal view in which the LCx artery appears to have a 40-50% proximal stenosis (arrow). B) Left anterior oblique cranial view in which the same LCx stenosis now looks in the order of 70-80% (arrow).
Figure 2. A subsequent pressure wire study of the same proximal LCx artery lesion gives an FFR of 0.72 suggesting the stenosis is of haemodynamic significance.
The physiological assessment of atherosclerotic lesions by measuring coronary pressure is not a revolutionary concept. In 1974, Gould and colleagues published their seminal work on the relationship between the severity of coronary flow restriction, resting and maximal coronary blood flow and regional flow distribution in a canine model using sodium diatrizoate to induce maximal arteriolar vasodilation1. Pressure and flow are intimately related to each other through the interplay between epicardial and myocardial vascular resistances. These resistances, in turn, are influenced by changes in arterial pressure, coronary vasomotion, myocardial oxygen demand and contrast injections, and are therefore in a constant state of flux. By inducing maximal arteriolar vasodilation across the vascular bed, these resistances are close to minimal but also, more importantly, constant, thereby negating their effect. This allows for inferences on coronary stenosis severity to be made with a high degree of confidence.
Andreas Gruentzig, during the inception of percutaneous transluminal coronary angioplasty, also used catheters that had an intrinsic capacity to measure coronary pressures via a lumen through which readings were taken and then used by the operator to determine the success of the angioplasty procedure2. This practice, however, was found to add little in the way of value to morphologic assessment post angioplasty. Moreover, there was a tendency to overestimate the trans-stenotic pressure gradient, since the balloon catheter would cause further reduction of the functional stenotic lumen as it crossed the lesion, and so the practice rapidly lost favour. This led to the development of an ultrathin 0.015-inch fluid-filled angioplasty pressure-monitoring guidewire to measure the distal coronary pressure and thereby overcome this limitation3. In an article published that same year, 1993, Pijls and colleagues used a canine model to report for the first time the use of this novel pressure wire to assess coronary stenosis severity, flow reserve, and collateral flow under maximal coronary vasodilation4. Most pertinently, and for the first time, they modelled the coronary circulation in flow-pressure terms based on the sum of flow through the epicardial arteries in addition to the contribution made by collateral flow. They went on to express the coronary flow reserve of a stenotic artery as a “fraction” of the normal value of that same vessel in the absence of an obstruction to flow – and so the term fractional flow reserve (FFR) was coined.
FFR is a lesion-specific index defined as the ratio of maximal hyperaemic myocardial blood flow across a stenotic artery to the maximal myocardial blood flow across the same artery in the theoretical absence of the stenosis. By applying Ohm’s Law, at maximal hyperaemia where resistance is assumed to be minimal and constant, pressure can be used as a surrogate marker of flow. Since the pressure in a normal coronary artery is equivalent to aortic pressure (Pa), the FFR can simply be derived as a ratio of the mean distal coronary pressure (Pd) at a point past the stenosis to the Pa during maximal hyperaemia where FFR=Pd/Pa. This calculation, however, is based on an assumption that central venous pressure (CVP) is close to zero at maximal arterial vasodilation. Some investigators would argue that the right atrial pressure (Pra) should be actively measured and used as a more precise estimate of CVP such that FFR=(Pd–Pra)/(Pa–Pra)5. The theoretical FFR of a normal coronary artery is 1.0, regardless of patient characteristics or vessel distribution. Moreover, a reduction in FFR below 1.0 can be interpreted without having to compare it against a normal reference distribution.
Pijls and colleagues went on to determine the range of FFR values that would help to identify the physiological significance of a particular coronary stenosis, and whether or not this would correspond to the presence of inducible myocardial ischaemia6. They studied changes in FFR in a cohort of 60 consecutive patients with known single-vessel coronary artery disease (CAD), normal left ventricular function, and a positive exercise tolerance test (ETT) 24 hours prior to planned percutaneous coronary intervention (PCI). Pre- and post-PCI FFR measurements were taken. A second ETT was performed five to seven days after successful PCI. Of the 58 patients who had had a successful PCI, 56 went on to have a normal ETT post PTCA with no evidence of inducible ischaemia. The investigators found an FFR of ≤0.74 in all patients pre PCI and observed it to be >0.74 following successful angioplasty. FFR measurements were also taken in a total of 18 normal coronary arteries from five patients with no discernible cardiovascular risk. The FFR was found to be near 1.0 in all cases, indicating no significant decline in pressure along the length of normal epicardial arteries.
By 1996, Pijls and colleagues published their landmark paper in the New England Journal of Medicine, in which they compared the validity of FFR (with an inducible ischaemia cut-off now set at <0.75) accurately to detect haemodynamically significant lesions that appeared only moderate in severity (defined as approximately 50% by visual assessment alone) against established non-invasive measures of myocardial ischaemia, namely thallium scintigraphy, bicycle exercise testing and dobutamine stress echocardiography7. They recruited 45 consecutive patients in whom chest pain could not be reliably linked to reversible ischaemia and was thought to be caused by a moderate proximal lesion in a large epicardial artery. All three non-invasive stress tests were performed on each patient who then had intracoronary pressure measurements taken during repeat angiography. Those patients with an FFR <0.75 (n=21) went on to have revascularisation and subsequent stress testing within six weeks of the procedure. In the 13 patients undergoing successful PTCA, FFR was seen to increase to a level >0.75 post procedure. Positive stress tests also reverted to normal post revascularisation. In 21 of the 24 patients with an FFR >0.75, all non-invasive tests were negative. Moreover, those not undergoing revascularisation (i.e., FFR >0.75) were left on medical therapy and found to have improved in functional angina class, and had suffered no ischaemic events at four-week follow-up. Overall, the investigators calculated the sensitivity, specificity, positive and negative predictive values and accuracy of FFR to be 88, 100, 100, 88, and 93 percent, respectively.
Two landmark randomised trials followed: DEFER (Fractional flow reserve to determine the appropriateness of angioplasty in moderate coronary stenosis: a randomized trial)8 and FAME (Fractional flow reserve versus angiography for guiding percutaneous coronary intervention)10. These would put the validity and reliability of FFR as a lesion-specific tool to determine the need for revascularisation beyond reproach. In brief, DEFER randomised patients awaiting planned PCI (but with no documented evidence of reversible ischaemia) with an FFR >0.75 to deferral (n=91) or performance (n=90) of PCI8. Those patients with an FFR <0.75 underwent PCI as planned and were deemed the Reference group (n=144). At two-year follow-up event-free survival was similar between the Defer group and the Perform group (89% vs. 83%, p=0.27). Event-free survival in the Reference group was 78%, which was similar to the Perform group but significantly lower than the Defer group (p=0.03). Freedom from angina at two years was higher in deferred patients compared to performed patients (p=0.02), and higher still in the Reference group (p<0.001). The trial demonstrated two fundamentally important points: firstly, that approximately half of those patients awaiting planned PCI without prior evidence of inducible ischaemia were found not to have physiologically significant lesions by FFR. Consequently, revascularisation did not give rise to a morbidity advantage over medical therapy in this patient cohort. Secondly, revascularisation in those patients found to have lesions that reduce FFR to <0.75 generated significantly greater improvements in functional class and freedom from angina. At five-year follow-up, event-free survival remained similar in both the Defer and Perform groups (79% vs. 73%, p=0.52) but was significantly lower in the Reference group compared to the overall functionally non-significant stenosis cohort (61% vs. 76%, p=0.03), thus suggesting that, despite revascularisation, haemodynamically significant lesions are associated with greater mortality and morbidity9. PCI in the Reference group continued to produce significant gains in freedom from angina at five-year follow-up (72% Reference vs. 67% Defer vs. 51% Perform; p=0.028 Reference compared to both other groups). It is clear from the DEFER study that PCI of lesions with an FFR >0.75 did not lead to prognostic or symptomatic benefit.
Prior to the FAME study, the utility of FFR had only been validated in patients with single-vessel CAD. FAME went on to compare an FFR-guided complete functional revascularisation strategy in multivessel CAD against a complete anatomical revascularisation strategy guided by coronary angiography alone10. In the former, lesions were only stented if they produced an FFR <0.80, whereas in the latter all lesions were stented based on morphological appearance. At one-year follow-up an FFR-guided strategy led to a significant reduction in the primary endpoint of death, non-fatal myocardial infarction (MI) and repeat revascularisation when compared with an angiography-guided strategy (18.3% vs. 13.2%, p=0.02). The benefit of using FFR to guide revascularisation persisted to two years in terms of the combined primary endpoint (22.4% vs. 17.9%, p=0.08) and was significant when mortality and MI were combined (8.4% vs. 12.9%, p=0.02)11.
Subsequently, FAME II was designed to compare an FFR-guided PCI strategy with optimal medical therapy (OMT) versus OMT alone in patients with stable CAD and FFR ≤0.80. Patients with FFR >0.80 were enrolled in a registry. Having randomised 1,220 patients at 28 sites across the world, an interim analysis revealed a significant advantage in favour of the FFR-guided approach (a difference in the primary end point which was a composite of death, myocardial infarction, or urgent revascularisation of 4.3% in the PCI group and 12.7%)12, thus forcing the independent data safety monitoring board to recommend cessation of further enrolment to the study in 2012. This was driven by the need for urgent revascularization in the patients randomised to medical therapy. Interestingly, outcomes in patients randomised to PCI were not significantly different to the registry group that had PCI deferred in the context of an FFR>0.8. In the wake of the controversy caused by the COURAGE trial (in which PCI was not FFR-guided), results from FAME II suggest that PCI in stable CAD still has a place, particularly when conducted in a judicious, and ultimately physiological, manner.