Treatment effects of coronary bypass surgery—the math and the myths of the ejection fraction

It is common belief that improving blood supply to ischaemic myocardium (i.e. revascularization) is associated with an improvement in contractile function, presumably leading to improvement in survival [1]. It may therefore be surprising that coronary artery bypass grafting (CABG) may improve survival in ischaemic heart failure compared to medical therapy despite a lack of functional improvement and irrespective of the outcomes of myocardial viability or ischaemia testing [2–4]. Nakae et al. [5] addressed this belief in this issue of the EJCTS.

The authors assessed ejection fraction (EF) recovery in 490 patients having undergone CABG for significant coronary artery disease (CAD) and reduced EF (≤40%). They followed these patients after surgery for about 5 years and repeatedly measured EF by echocardiography. They determined EF changes and found almost 40% of patients who did not show EF recovery. They further showed that patients without EF recovery (defined as EF remaining below 40% during follow-up) had significantly worse survival than those who did show an improvement in EF. Thus, the results appear to support the notion that improvement in EF improves survival. However, assessing Nakae et al.’s patient population reveals that a substantial fraction of patients not only received CABG but concomitant mitral valve (up to 51%) and/or surgical ventricular reconstruction (SVR, up to 28%), which may directly affect the EF, independent of changes in contractility.

The EF is the quotient of stroke volume and end-diastolic volume. Cardiac surgery may affect both, the dividend and the divisor. Changes in EF may therefore arise from treatment effects on stroke volume and/or end-diastolic volume. Thus, using the EF as marker for contractile function requires that changes in these two parameters are limited to changes in contractility of the myocardium. Since both mitral valve surgery and SVR directly alter at least one component of the EF equation, assessment of a treatment effect based on EF changes is difficult.

EJECTION FRACTION-EFFECTS OF MITRAL VALVE SURGERY AND VENTRICULAR RECONSTRUCTION SURGERY

Severe mitral regurgitation usually requires a regurgitant fraction of at least 50% [6]. In other words, half of the ejected stroke volume does not contribute to systemic cardiac output. Consider an asymptomatic patient at rest with an end-diastolic volume of 200 ml and a total stroke volume of 100 ml, of which 50 ml is regurgitant volume over the mitral valve. The calculated EF is 50%. Mathematically, this EF will drop to 25% by simply eliminating regurgitation (e.g. by mitral valve repair) assuming that forward stroke volume and cardiac output do not change.

In contrast, SVR usually reduces end-diastolic volume. Consider a patient with a dilated left ventricle, no mitral regurgitation, but anterior dyskinesia and an end-diastolic volume of 240 ml. If stroke volume is 80 ml, EF will be 33%. If SVR were to reduce end-diastolic volume to 160 ml (without affecting contractility and stroke volume) the calculated EF would increase to 50%.

It is clear that these mathematical considerations are not immediately evident in daily practice, because differences in loading conditions during surgery, inotropic stimulation after cardioplegic arrest, etc. obscure their detection. However, they are important, specifically if EF is used a surrogate for contractile function.

EJECTION FRACTION AS MARKER OF CONTRACTILE FUNCTION

If the EF is to reflect contractile function of the myocardium, it requires the absence of mechanical changes in stroke volume and end-diastolic volume (as described above). An improvement in EF may then reflect an increase in stroke volume by improved contractility or a decrease in end-diastolic volume by reverse remodelling. Both treatment effects can be witnessed after CABG and satisfy patients and physicians for its perceived treatment success. Conversely, the absence of changes in EF may then be considered a lack of a treatment effect and explanations are discussed justifying the absence of or differences in EF improvement (different methods of EF assessment, different follow-ups or patient populations [5]). However, this consideration may not find support in the literature:

First, in Nakae et al.’s study, the differences in survival between patients with and without EF recovery are associated with so many other differences in patient risk profiles (e.g. renal function, mitral and redo-surgery) that it is impossible to causally link the difference in survival to the difference in EF. Second, the STICH trial demonstrated improved survival with CABG without changes in EF and independent of the presence of viability or ischaemia [2–4]. Third, the elimination of documented ischaemia (both inducible and hibernating) by revascularization may or may not improve EF [3, 7, 8]. Finally, we suggested that a CABG survival impact compared to medical therapy and percutaneous coronary intervention (PCI) is related to infarct-prevention and not the treatment of ischaemia [9]. Thus, a key mechanism for the most desired treatment effect of CABG in chronic CAD (i.e. the improvement of survival) may not depend on the elimination of ischaemia but rather on the prevention of new acute ischaemic events [10].

Seeing these mechanistic considerations in context with the above described mathematical influencers of EF illustrates a myth in using the EF as marker for operative success. While postoperative improvement in EF after cardiac surgery is likely to reflect a treatment success, a lack of EF improvement does not reflect the opposite. In many situations, EF improvement would not even be expected and this expectation should therefore not be part of decision-making for CABG. Nakae et al. provide an interesting investigation on a heterogeneous patient population, thereby stimulating thoughts on math and myths of the EF in cardiac surgery.

REFERENCES