Coronary angiography: Has the gold standard become tarnished?

In 1958 Mason Sones famously, and apparently accidentally, performed the first coronary angiogram at the Cleveland Clinic. From then on much of the clinical care of patients with ischaemic heart disease was based on research that relied heavily on the visual interpretation of the coronary angiogram. However it wasn't long before papers stared to appear which questioned the accuracy and reproducibility of these visual estimates. One paper from 1976 reported than nearly half the time a group of experienced cardiologists could not agree on the presence of significant coronary artery disease. Other studies followed alleging to to demonstrate the benefits of performing quantitative coronary angiography using computers to assist in the measurement of the degree of narrowing. These methods were more reproducible than eyeballing the angiogram but still there was disagreement and the methodology was time consuming.

Take for example this angiogram on shown below. Do you think the this LAD stenosis is flow limiting? Why not vote here and see what others think?

By early 1990's the literature on the accuracy of angiogram went quiet. Cardiologists had other things on their mind - namely coronary angioplasty and stenting. Angiography was the test which fuelled the fire of angioplasty and so the problems with assessment of flow limiting lesions and lesion significance drifted into the background. The occulo-stenostic reflex was strong. Cardiologists needed their angiograms too much to call into question the ability of the test to diagnose and classify the severity of the coronary lesions.  This was the era of eminence-based medicine when expert opinion trumped anything else. The poor reproducibility and difficulty in assessment of lesion was forgotten.

When you learn angiography you quickly realise that the interpretation is difficult. When you work for a number of bosses you start to see a difference in their practice. Some always see a moderate lesion as severe, or a severe lesion as critical. The phrase "the angiogram often underestimates the severity of disease" is often be heard in the catheter lab control room as the guide catheter is being opened ready for angioplasty. Whilst there is usually agreement about the mild (<30%) and severe (>80%) lesions it is the moderate ones which are most difficult and unfortunately most common.

As I teach my fellows the percent stenosis is the wrong way to think about lesions, rather we should say whether we believe a lesion to be flow limiting or not. Flow limitation is dependent on the stenosis but also on the reference vessel size, the lesion length, the size of the territory supplied by the vessel and the presence or absence of collaterals. The issue of interpretation is vital to the individual patient since it determines what treatment is recommended, You don't want a cardiologist to put in a stent or offer bypass surgery if your coronary artery lesion is not flow limiting.

Recent studies have revealed what we knew all along namely that when coronary artery disease is moderate it is not possible to accurately know by visual assessment whether the lesion(s) are flow limiting or not. We need better methods not based on anatomy but rather on physiology. I have previously written about the RIPCORD trial but recently a large French registry has published its results which support the idea that the angiogram is difficult to interpret and that use of a pressure wire to measure fractional flow reserve (FFR) alters the cardiologists decision making.

The R3F study looked at 1000 people having a diagnostic angiogram. The vessels were assessed and significant lesions documented. The patients symptoms and the results of any non-invasive investigations were considered and a recommendation made as to whether the patient should have medical therapy, angioplasty or bypass surgery. After this the cardiologists performed a pressure wire measurement (FFR) of any stenosis. The results were then used to determine whether the stenosis was flow limiting and with this information in hand the treatment recommendation adjusted. So for example if a patient had a 40% stenosis on the angiogram with medical therapy recommended initially but then the pressure wire was significant (e.g. FFR 0.74) the recommended treatment would be to offer an angioplasty.

Using the pressure wire data the overall number of people recommended for medical therapy, angioplasty or bypass did not change but the decision for an individual patient changed 43% of the time. Overall the decision changed in 33% of patients initially recommended to have medical therapy and 50% of patients recommended to have angioplasty or bypass surgery.

These results are very important for individual patients since the treatment recommendation means the difference between just taking tablets versus having a procedure or an operation. We don't know yet whether a pressure wire guided approach makes a difference to clinical endpoints such as survival, mortality, rates of heart attacks and a large trial is needed to answer this question. 

For the moment when a moderate stenosis is diagnosed the patient should be asking their cardiologist what is the FFR?