• echocardiography
• cardiac magnetic resonance (cmr) imaging
• valve disease surgery
• heart failure with reduced ejection fraction
• aortic regurgitation

Chronic aortic regurgitation (AR) is associated with increased preload (regurgitant volume overload) and afterload (increased end-diastolic volume and wall stress), leading to left ventricular (LV) dilatation and systolic impairment that precede the onset of symptoms. Progression is slow—surgery is required in <4% of asymptomatic patients with normal LV dimensions per annum,1 yet risk of mortality rises to 10%–20% per annum once symptoms surface.

Chronic AR may result from root enlargement (for example, from hypertension or Marfan syndrome) or gradual disruption of the normal valve leaflets secondary to congenital aortic valve abnormalities, rheumatic fever, collagen vascular disease, infective endocarditis or atherosclerotic degeneration. AR may be classified into three subtypes2: type 1 where annuloaortic ectasia leads to failure of coaptation and central AR, type 2 where cusp prolapse or fenestration leads to eccentric AR and type 3 where cusp retraction with poor tissue quality results in central or eccentric AR. Comprehensive echocardiography is the mainstay of evaluation and surveillance allowing assessment of aetiology, severity and haemodynamic effects on the LV. Severe AR is defined by the integration of numerous echocardiographic parameters: regurgitant volume ≥60 mL/beat, regurgitant fraction ≥50% and an orifice area ≥0.3 cm², jet width: LV outflow tract diameter ≥65%, pressure half time <200 ms and vena contracta width >0.6 cm. In patients where echocardiography is challenging or equivocal (or LV function is borderline), cardiac magnetic resonance (CMR) imaging is of value and can also provide additional detail concerning concomitant aortic pathology. Nevertheless, the prediction of symptom onset and LV decompensation remains difficult and careful surveillance is recommended in specialist valve clinics.3

The importance of AR has been overlooked in the furore of interest surrounding aortic stenosis since the advent of percutaneous valve intervention. Only one in six surgical aortic valve replacements (AVRs) are undertaken for AR (Society of Cardiothoracic Surgery Blue Book, 2009, http://www.bluebook.scts.org) and the development of specific transcatheter devices for AR remains an elusive engineering goal. Yet moderate–severe AR is common with an incidence of 1.6% in asymptomatic patients aged >65 years4. Unsurprisingly, medical therapy has limited impact on a fundamentally mechanical problem, not least since initiation is often deferred until late in the natural history of the disease. Beta-blockers and ACE inhibitors (or angiotensin receptor blockers) can be beneficial for patients with aortic root dilatation, and vasodilator therapy may be considered in patients with LV dilatation and preserved systolic function. While historical data suggest that vasodilator therapy can slow the natural history of AR and delay the need for AVR, evidence is of low quality and inconsistent.

For several decades, there has been consensus that surgery should be undertaken in patients with chronic severe AR associated with symptoms or asymptomatic LV dilatation or dysfunction. Surgical outcomes are favourable if prompt surgery is performed as soon as LV dysfunction is detected. Conservative management is appropriate in those who are asymptomatic with normal LV parameters (since progression is slow over 5-year follow-up), but close follow-up and serial observation are mandatory. LV dilatation (LV end-systolic dimension (LVESD) ≥50 mm or LV end-diastolic dimension (LVEDD) ≥70 mm) is strongly predictive of the development of symptoms or systolic dysfunction and recently updated European Society of Cardiology (ESC)/European Association for Cardio-Thoracic Surgery guidelines recommend surgery for severe AR in any patient with LV ejection fraction (EF) ≤50%, LVEDD >70 mm or LVESD >50 mm, or in any symptomatic patient irrespective of LVEF, so long as operative risk is not prohibitive.3

In their Heart paper, , Fiedler et al 5 report a single-centre, retrospective analysis of 40 patients with severe AR and LVEF <35%, of whom 18 (45%) underwent surgery and 22 (55%) were treated conservatively. Although selection bias could not be excluded, AVR was associated with prolonged survival (mean 6.3 vs 0.5 years) and reduced mortality (HR 0.143, P<0.05) compared with medical management, after adjustment for end-stage renal failure, atrial fibrillation, age >65 years and peripheral vascular disease. Remarkably, severe AR was relatively rare in this large hospital series—only 40 patients among 145 785 echocardiograms screened met the criteria of severe AR and severe LV impairment. Those with significant concomitant valve disease or need for additional cardiac surgery were excluded and there was no difference in the prevalence of coronary artery disease in the two study groups.

Although the study suggests a survival benefit associated with surgery in patients with severe AR and poor LV function, it remains difficult to draw any firm conclusions. The retrospective, observational design and small study numbers mean that the patient groups were not matched and there were resulting significant differences in age (conservative management vs surgery: 69 vs 51 years, P=0.002) and the prevalence of systemic hypertension (conservative management vs surgery 91% vs 44%, P=0.0014). The aetiology of LV dysfunction could have been multifactorial, related to a coexistent dilated cardiomyopathy or purely due to organic valve disease. Moreover, reasons underlying the decision to offer surgical AVR or conservative management were unclear—there was no difference in New York Heart Association class III/IV symptomatic status but surgical risk scores (The Society of Thoracic Surgery Risk Score or European System for Cardiac Operative Risk Evaluation) and frailty indices were unavailable. Information regarding heart failure therapy (including the use of cardiac resynchronisation devices), raw parameters of LV function and AR severity, follow-up imaging data and the causes of death may provide further insight. Notwithstanding these limitations, the study findings are provocative in suggesting that selected high-risk patients with AR and severe LV impairment experience significant survival benefit with surgery when compared with medical treatment alone. These outcomes suggest that some of the pathophysiological changes in LV function imposed by chronic severe AR may be better tolerated than previously thought and that reverse remodelling can be achieved even in the late stages of this condition.

Advances in cardiac imaging and the use of serum biomarkers may enhance current clinical and echocardiographic management algorithms and encourage earlier intervention before potentially irreversible LV dysfunction is established. Multimodality imaging has potential to delineate the underlying phenotype and determine which patients with severe AR and poor LV function are likely to respond more favourably to surgery. Quantification of AR using CMR can help stratify those asymptomatic patients in whom early surgery should be considered from those who should remain under careful follow-up in a specialist valve clinic: patients with a regurgitant fraction ≤33% have a significantly higher 2-year survival without surgery than those with a regurgitant fraction >33% (97% vs 39%, P<0.0001).6 Levels of brain natriuretic peptide ≥130 pg/mL may also predict clinical events in asymptomatic patients with severe AR and normal LV function.

LVEF is a strong and independent predictor of overall long-term survival in severe AR7 and patients with LV dysfunction that is prolonged or associated with poor exercise tolerance often display no significant LV recovery after AVR.8 Long-term improvement in EF following AVR correlates closely with reduction of LV dimensions in the early and late postoperative period (r=0.63 and r=0.69, respectively) and LV function can continue to improve for several years (although late recovery is less likely if there is no initial improvement within 6 months).8 LV remodelling and eccentric hypertrophy maintain normal wall stress but chronic volume loading induces fibronectin production by cardiac fibroblasts while collagen I synthesis remains virtually unchanged. Although late gadolinium enhancement on CMR has not been shown to correlate with the degree of AR, several CMR studies have demonstrated that diffuse myocardial fibrosis contributes to impaired LV recovery following AVR. Importantly, impaired global longitudinal strain prior to AVR confers a higher risk of death at 5 years and global longitudinal and circumferential strain may remain abnormal following surgery, even when overall ejection fraction returns to normal.