Article Text

Download PDFPDF

A tragedy of modern cardiology: using ejection fraction to gauge left ventricular function in mitral regurgitation
  1. Blase A Carabello
  1. East Carolina Heart Institute, Brody School of Medicine, East Carolina University, Greenville, North Carolina, USA
  1. Correspondence to Dr Blase A Carabello, East Carolina Heart Institute, Brody School of Medicine, East Carolina University, Greenville, NC 27834, USA; carabellob16{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


The normal mitral valve provides unidirectional flow from left atrium (LA) to left ventricle (LV) in a smooth and efficient fashion. Mitral incompetence allows backward flow into the LA, requiring the LV to pump additional volume to compensate for that lost to regurgitation (MR). In its compensated state, even severe chronic MR is remarkably well tolerated at least for short periods of time. Patients are often totally asymptomatic, and many can achieve normal results during cardiopulmonary stress testing, objective proof that symptom absence is based upon true physiological compensation for the existing pathological state.1 How does the LV accomplish this feat? Logically, it must be able to generate adequate forward stroke volume at physiological filling pressure, lest the patient would develop dyspnoea on exertion.

Stroke volume is dependent on innate LV volume (elephants have bigger stroke volumes than mice), contractility, preload and afterload. Three of the four of these properties, volume, preload and contractility, are fairly straightforward. LV remodelling in chronic MR is unique, leading to an enlarged thin-walled chamber.2 Both in man and in the experimental animal, this remodelling leads to increased LV compliance (an unusual occurrence in cardiology) and enhanced LV filling, the mechanism whereby the large diastolic volume of MR is accommodated at near-normal filling pressure.3 Obviously, this same remodelling increases total LV volume, so that normal ejection creates increased total stroke volume in turn maintaining forward stroke volume in the normal range. Volume expansion of the LV also increases preload, using the Frank-Starling mechanism to further augment stroke volume. Contractility is initially normal or may even be augmented by increased sympathetic tone. It is the fourth property, afterload, that is, or at least should be, controversial.

Afterload myth one: the low impedance pathway

It has been reasoned that the mitral regurgitant orifice provides a low impedance pathway for ejection into the LA, in turn reducing afterload, enhancing LV ejection fraction. I am sure that I, in a less educated life, have made such a statement. Gaasch et al4 address this issue head on in the current edition of the journal. They found that the LV–LA pathway for ejection has higher impedance than the LV–aorta pathway, except when MR is in the very severe range which does occur in some patients. The finding is quite logical. Regurgitant flow is dependent upon the MR orifice size and the pressure gradient driving backward flow. Obviously, larger holes present less obstruction to flow than smaller ones, so that we might expect that LV–LA impedance is relatively high until the effective regurgitant orifice area exceeds about 0.5 cm.2 Where does this leave afterload? Gaasch and colleagues note that afterload in their study was normal, which is in fact the norm5 not reduced as urban legend often has it. Afterload is perhaps best quantified as wall stress, which by Laplace's law: stress (σ)=P× r/2 h, where P=LV pressure, r=radius and h=thickness. Wall stress may be reduced in acute MR (chordal rupture, for instance) when pressure is usually low and eccentric hypertrophy has not yet increased radius. However, in chronic MR, the radius term does increase, resetting afterload to normal.5 ,6 Little recognised is the fact that in MR, afterload may be actually increased as LV dysfunction ensues and LV radius becomes yet larger.6

Afterload myth two: ejection fraction falls after mitral surgery because the LV–LA pathway was closed, increasing afterload

Ejection fraction (EF) in MR is a witch's brew of misunderstanding. Ejection fraction is dependent upon preload, afterload and contractility, all of which can be altered in the same patient with MR, making it almost impossible to attribute changes in EF to any one property or a combination of properties. It seems clear that the ‘normal’ EF in MR is greater than the normal EF for normal subjects, at least judging the response of patient outcomes. For EF to be maintained in the normal range postoperatively, it should be >60% or even 64% preoperatively,7–9 and even EF in the 50%–60% range impacts survival negatively.7 ,8 Contractility can be normal and preload is increased, but what about afterload? From the above, it seems logical that afterload is usually normal in MR and occasionally increased. Following correction of MR, EF usually falls slightly, but may fall drastically if there was antecedent LV dysfunction or if the mitral apparatus is severed during surgery.7 Thus, while closing, the LV–LA pathway must logically make it harder for the LV to eject blood; according to the data from Gaasch et al, it usually does not remove a low impedance pathway for ejection, and any tendency for this to increase afterload is often offset by a reduction in the Laplace radius, as the LV reverse remodels after surgery.

What really happens to EF following surgery, and why?

As above, EF is dependent upon preload, afterload and contractility. Assume for the moment that preoperatively, end-diastolic volume (EDV) is 300 cc, LVEF is 0.60 and regurgitant fraction is 0.5. The LV is ejecting 180 cc, of which 90 cc is the forward stroke volume producing a normal cardiac output, and regurgitant flow is also 90 cc. After valve repair, if nothing else changed other than closing the regurgitant orifice, all 180 cc would be ejected forward producing a huge and un-needed cardiac output, and there is no evidence that such a state exists. If the circulation simply resets itself to a forward stroke volume of 90 cc, and the now absent volume overload of 90 cc reduced EDV by that amount, EDV would be 210 cc, end-systolic volume (ESV) would be 120 cc and LVEF 0.43. If afterload really did increase following surgery and ESV rose to 140 cc, EDV would become 230 cc and EF would be 0.39. While immediately after surgery these changes may occur, by patient discharge and at early postoperative follow-up, EF returns to normal, because ESV falls, not increases, because wall stress (afterload) actually falls as the radius term in the Laplace equation falls, provided the mitral apparatus is intact and LV contractility was preserved preoperatively (figure 1).5 ,10 ,11 From the data of Gaasch et al, it might be expected that systolic stress could increase immediately post repair in patients with very severe MR where the LV–LA pathway is truly of low impedance.

Figure 1

Demonstration of left ventricular response to mitral repair over time as constructed from human data obtained for Refs 5, 10, 11. EDV, end-diastolic volume; EF; ejection fraction; ESV, end-systolic volume; LV, left ventricular.

So what??

The goal in valve disease is to time surgery before LV dysfunction, and its consequences impact outcome. Currently, we are using a 50-year-old load-dependent tool, ejection fraction, to gauge function in the lesion, with the most confusing and least predictable changes in load. Understanding the mechanics of MR has the greatest impact on patients with the most severe MR, those with regurgitant fractions of 55% and above. It is in them that there is a low-impedance LV–LA pathway preoperatively, who are most likely impacted by an increase in afterload postoperatively, and especially in whom we need better tools than EF to measure LV function. We must develop biological tools that can peer into the workings of the myocardium to understand and predict when LV contractility is beginning to fall and when LV dysfunction will affect prognosis, tools that can and must replace our ancient, rusty, dull tool, ejection fraction.


View Abstract


  • Competing interests None.

  • Provenance and peer review Commissioned; internally peer reviewed.

Linked Articles