By Michael H. Crawford, MD, Editor

SYNOPSIS: A retrospective observational study of patients with moderate to severe aortic regurgitation followed for five years showed that symptoms or a left ventricular end-systolic dimension index > 20 mm/m2 were multivariate predictors of death.

SOURCES: Yang LT, Michelena HI, Scott CG, et al. Outcomes in chronic hemodynamically significant aortic regurgitation and limitations of current guidelines. J Am Coll Cardiol 2019;73:1741-1752.

O’Gara PT, Sun YP. Timing of valve interventions in patients with chronic aortic regurgitation: Are we waiting too long? J Am Coll Cardiol 2019;73:1753-1755.

Deciding when to intervene surgically in patients with chronic aortic regurgitation (AR) remains challenging. Investigators from the Mayo Clinic conducted a retrospective observational study of patients with moderate to severe AR. They excluded patients with aortic dissection, active endocarditis, valve stenosis, more than mild mitral regurgitation, previous valve surgery, or other cardiac diseases such as coronary artery disease. All patients underwent a complete transthoracic echocardiogram. The primary outcome was all-cause mortality. The study population included 748 patients who met the criteria, and the authors studied this cohort from 2006 to 2017. Mean age was 58 years, and 82% were men. Aortic valve replacement (AVR) was performed in 361 patients, of whom 93% met class I guideline indications. Only 10% had left ventricular ejection fraction (LVEF) < 50%. In 14% of this group, ascending aortic aneurysm was the main indication. The median time from echo to surgery was 34 days. In 65% of patients, the valve was repaired or replaced with a bioprosthesis. The most common concomitant procedure was root or ascending aorta repair (30%). Over a median follow-up of five years, 17% of patients died. Multivariate predictors of death were symptoms and LV end-systolic dimension index (LVESDi). Also, AV repair or replacement within six months of the baseline evaluation was associated with better survival (hazard ratio [HR], 0.36; 95% confidence interval [CI], 0.24-0.53; P < 0.0001). In addition, when dichotomized for class I surgical indications, patients undergoing surgery for class I indications (symptoms, dilated aorta, or LVEF < 50% demonstrated a worse survival rate (HR, 7.98; 95% CI, 1.71-142; P = 0.003). The authors concluded that class II indications for surgery are associated with better survival postoperatively and should be strongly considered. LVESDi was the only echo measure independently associated with survival. The ideal cutoff appears to be > 20 mm/m2.


When to intervene in chronic AR remains a challenge because it almost always means valve replacement surgery. It is difficult to sell surgery to asymptomatic patients. This retrospective, observational experience at one tertiary center confirms current practice in that 61% of the patients undergoing surgery exhibited symptoms. In only 2% of patients was LVEF < 50% the only criterion. In 4% of patients, aortic dilation was the sole criterion. Accounting for overlap, 284 patients met one or more of the three class I guideline indications for surgery. In the remaining 77 patients, 50 met class II indications: LVESD > 50 mm in seven; LVEDD > 65 mm in 39; and four met both LV dimensional indications. The remaining 27 patients met no class I or II indications. In surgical patients, only one died and three suffered a stroke during the first 30 days (total death or stroke, 1.1%). Counterintuitively, patients with only class I indications for surgery had a higher risk of post-AVR all-cause mortality than those with class II indications. Among class II indications, only LVESDi was a predictor of mortality, and a measurement of > 20 mm/m2 seemed to be the inflection point, which is lower than > 25 mm/m2 — the cutpoint in the European Society of Cardiology guidelines. U.S. guidelines still use LVESD > 50 mm, which was not a significant predictor in this study. This observation speaks to the importance of adjusting LV dimension measurements to the size of the individual.

Waiting for symptoms in patients with chronic moderate to severe AR may not be the best approach as symptoms seem to be a late manifestation. However, most patients with symptoms did not meet LV dimensions or functional criteria for surgery. Perhaps some symptomatic patients had other comorbidities that adversely affected survival, and their symptoms were attributed falsely to AR. If symptomatic patients do not meet dimensional criteria for surgery, a search for other explanations for their symptoms should be undertaken to avoid symptom persistence after aortic valve surgery. LVEDD alone is not particularly predictive of outcomes, and an LVESDi of 20-25 mm/m2 should occasion consideration of surgery.

There were limitations to this study besides its retrospective, observational design. Patients with coronary artery disease were excluded. Stress tests were not performed to confirm symptom status. There are no data on symptom duration, hypertension, atrial fibrillation, or medical therapy. Also, there are no data on parameters other than those in the current guidelines such as LV volumes, regurgitant volume, late gadolinium enhancement on cardiac MR, and brain natriuretic peptide levels. In addition, perhaps three-dimensional echo or cardiac MR-determined LV size and function, would be superior to echo LV dimensions. Since surgical mortality was so low in this study, a push for earlier surgery in such low-risk patients is justifiable. Waiting for clearly valve-related symptoms or LV dysfunction probably is too conservative.