Surgery for Severe Aortic Regurgitation


The timing of surgery for aortic regurgitation (AR) is still controversial, especially with regard to the role of symptoms. Thus, Klodas and colleagues reviewed a 10-year experience with surgery for AR in 289 patients seen at the Mayo Clinic. All patients had confirmed grade 3 or 4/4 AR severity without stenosis or other valve disease requiring surgery. Patients who had aorta replacement or coronary bypass surgery were not excluded. The major symptoms assessed were dyspnea (NYHA grading) and angina (CCS grading). Combining both symptoms, 80 were class I (no symptoms), 81 were class II, 101 were class III, and 27 were class IV (rest symptoms). A multivariate statistical approach was used to adjust for other variables such as left ventricular function, surgical factors, and other clinical features. Comparing those with class I-II symptoms to the class III-IV patients, the latter were older, more often female, hypertensive, diabetic or had post-myocardial infarction, and were more likely to have concomitant coronary artery bypass. Operative mortality was higher in class III-IV (7.8%) vs. class I-II (1.2%; P = 0.005), and 10-year post-operative survival was worse (45% vs 78%; P < 0.001). Multivariate analysis showed that class III-IV symptoms were independent predictors of operative mortality (odds ratio 5.5; P < 0.04) and worse long-term survival (odds ratio 1.8; P < 0.01). Excluding the 40 patients with coronary artery bypass did not change the results. Consideration of left ventricular ejection fraction (LVEF) did not alter the prognosis of the two symptoms groups. Klodas et al conclude that in patients with grade 3 or 4 AR, class III or IV symptoms are predictive of higher operative mortality and reduced long-term survival. Thus, the presence of class II symptoms and moderately severe to severe AR should prompt consideration of surgery. (Klodas E, et al. J Am Coll Cardiol 1997;30:746-752.)


Older natural history studies of AR showed that the development of severe symptoms was associated with a poor prognosis. Hence, when an effective surgical approach was developed (valve replacement), such patients were referred for surgery. Unfortunately, some patients did not improve post-operatively because of persistent LV dysfunction. Thus, the idea that reduced LV function may predict poor surgical outcome was advanced and tested in the 1970s. Those studies showed that marked increases in LV size or wall stress and reduced LV function were predictive of poor surgical outcomes and resulted in the concept that patients who reached certain LV performance measures should undergo surgery irrespective of symptoms. The application of this concept required trying to convince asymptomatic patients that they needed heart surgery—no easy feat.

Studies in the 1980s raised two new issues. First, the results of the older surgical series could not be replicated. Symptomatic patients usually experienced an improvement in functional class, and post-operative LV dysfunction was unusual. These results were mainly attributed to improved myocardial preservation techniques. Second, new natural history studies showed that about 5% of patients with moderate-to-severe AR developed LV dysfunction before symptoms—raising the issue that frequent repeated measurements of LV function may not be warranted, and waiting for symptoms to develop is acceptable.

In this context, the Mayo Clinic study is of interest because it re-explores the role of symptoms in the timing of surgery. Since the operative and long-term mortality of class I and II patients was the same and markedly different from class III-IV patients, it seems reasonable to wait for class II symptoms and then consider surgery. However, no mention is made in this article about the role of medical therapy. If vasodilator drugs abrogate mild (class II) symptoms, can one afford to wait until they return on therapy or should LV function be religiously measured serially? On the other hand, the medical therapy studies have been small and largely involved asymptomatic patients. Thus, I would favor surgery for anyone with symptoms unless their risk for surgery was high for other reasons.

This study also evaluated the effect of other risk predictors that have been advocated in the past. Class IV patients did not explain the results since class III patients had a similar prognosis. Also, the presence of coronary artery disease did not explain the results in the more symptomatic (III-IV) patients, since eliminating them did not influence the observed outcomes. In addition, LVEF was not predictive of poor outcomes as had been observed in previous studies. Klodas et al explain that this is because of the high number of women in this study (almost 50%). According to their analysis in women, this relationship is not observed, but it was in the men. This is consistent with other studies that show that in women, the LV response to a volume or pressure load is different from that in men. However, there was also not much difference between the EFs in the two symptom groups (52% vs 48%), diminishing the discriminating power of LV function measurements. Thus, there were symptom class III-IV patients who had normal LVEF and I-II patients with reduced function. Presumably, patients with no symptoms (class I) and reduced LVEF or markedly dilated LVs should still be encouraged to undergo surgery, but this study did not specifically address this issue.