Recently, I examined a beautifully excised severely stenotic congenitally bicuspid aortic valve in a 64-year-old man who, at the same operation, had undergone excision of the ascending aorta, which, by preoperative imaging studies, had a maximal diameter of 4.3 cm. Histologically, the wall of the aorta was normal. Although the aortic valve replacement (AVR) was clearly appropriate, was excision of the ascending aorta justified on a prophylactic basis? I would argue that the latter should be reconsidered.

Before the 1990s, relatively few patients with stenotic or purely regurgitant congenitally malformed aortic valves also underwent concomitant resection of the dilated ascending aorta. During the 1990s, some prominent and persuasive cardiac surgeons began advocating prophylactic concomitant ascending aortic resection in these circumstances, mainly for 3 reasons. First, to prevent postoperative dilation of the ascending aorta in a previously normal sized aorta or to prevent additional dilation in an already dilated aorta; second, to prevent aortic dissection, which is recognized to occur more commonly in patients with congenitally unicuspid and bicuspid aortic valves than in patients with tricuspid aortic valves ; and, finally, to prevent rupture of the ascending aorta late postoperatively. The latter reason was because it is recognized that the media of the ascending aorta in patients with congenitally bicuspid aortic valves can be depleted of certain elements (elastic fibers, smooth muscle cells, collagen) and can be repleted with certain other elements (mucopolysaccharide material) compared to the aorta’s structure in adults with tricuspid aortic valves (except in those patients with the Marfan syndrome and the forme fruste variety of it, cardiovascular syphilis, and ankylosing spondylitis, and giant-cell aortitis ).

But how often do these potential aortic complications (additional dilation, dissection, rupture) occur in patients who have had AVR because of a dysfunctioning congenitally malformed aortic valve (either unicuspid or bicuspid) without concomitant resection of the dilated ascending aorta? The answer is infrequently. The best data (largest number of patients and longest follow-up) supporting that view comes from an analysis of 1,286 patients (1,180 [92%] with aortic stenosis [AS], 85 [7%] with aortic regurgitation [AR], and 20 [1%] aortic valve dysfunction type not stated) who had undergone AVR at the Mayo Clinic from 1960 to 1995 because of a dysfunctioning bicuspid aortic valve, without concomitant resection of the ascending aorta, irrespective of its size. The mean age at operation was 58 years (72% men). The follow-up period was 0 to 38 years (median 12). During that postoperative period (excluding the 36 who died early [2.8%] and the 26 lost to follow-up [2.6%]), documented aortic dissection occurred in 13 (1%) and additional aortic dilation (ascending aorta diameter reached ≥5.0 cm or >1.0 from that at the time of AVR) occurred in 127 (9.9%). Additionally, 11 (0.08%) of the 1,227 patients not dying early or lost to follow-up later underwent resection of the ascending aorta; in 9 of the 11, however, it was at the time of replacement of a malfunctioning mechanical prosthesis or a bioprosthesis in the aortic valve position.

Although 151 events affecting the ascending aorta occurred during the long follow-up period, 127 (84%) consisted of dilation of the aorta, something that can occur after any aortotomy, irrespective of the structure of the aortic valve, and dilation of the ascending aorta after an aortotomy in and of itself infrequently produces clinical consequences. Indeed, the 15-year all-cause survival was greater in the patients with “enlarged” aortas than in those with normal-size aortas at the time of AVR (56% vs 50%). Although the ascending aorta was later excised in 11 patients, that aortic resection probably would not have been done in 9 if excision of a dysfunctioning prosthesis or bioprosthesis had not been required. Additionally, aortotomy, performed with isolated AVR, causes proliferation of fibrous tissue around the aorta, making the aorta thicker than normal, and the resulting thicker wall almost surely is more difficult to rupture than an aortic wall of normal or less than normal thickness. Furthermore, any type of interruption of the wall of ascending aorta (be it for insertion of a saphenous vein conduit or arterial cannula or incision for AVR or for resection of an aortic tear in aortic dissection) increases the risk of late dissection or rupture, be it in patients with unicuspid, bicuspid or tricuspid aortic valves.

Similar long-term follow-up after isolated AVR in patients with dysfunctioning tricuspid aortic valves without concomitant resection of ascending aorta was recently presented by Gaudino et al. They described 93 consecutive patients (mean age 67 years; 74% men) undergoing isolated AVR for AS at their institution (Catholic University Rome) from January 1990 to December 2000. Preoperatively, the maximal diameter of the ascending aorta ranged from 5.0 to 5.9 cm (mean 5.6). The diagnosis of tricuspid aortic valve was confirmed in each patient at operation by visual inspection. The ascending aorta was neither replaced nor plicated. The 92 patients surviving the operation were followed up for a mean of 14.7 ± 4.8 years. During that period, 16 patients died (17%). In no case was death attributed to an aortic event. Indeed, no patient during the long follow-up experienced an acute ascending aortic rupture, dissection, or pseudoaneurysm formation. Also, during the follow-up, no patient required reoperation for thoracic aortic complications. Additionally, the diameter of the ascending aorta during the 10 years after AVR changed insignificantly (56 ± 2 mm preoperatively vs 57 ± 11 mm at the end of follow-up). The mean expansion rate of the ascending aorta was 0.3 ± 0.2 mm/year. Patients with connective tissue disorders such as the Marfan syndrome were excluded from their study. The report by Gaudino et al included the largest number of patients with dilated ascending aortas and stenotic 3-cuspid aortic valves undergoing AVR alone and followed up for ≥10 years. The frequency of late ascending aortic complications is lacking in patients who have had combined AVR for dysfunctioning unicuspid, bicuspid, or tricuspid aortic valves and resection of the dilated ascending aortas with similarly long (10 to 15 years) follow-up.

Adding ascending aortic resection to AVR usually significantly increases the operative mortality. Data from the Society of Thoracic Surgeons database indicate that ascending aortic resection increases the operative risk by a factor of 2.78 compared to AVR alone. I agree with McKeller et al from the Mayo Clinic who stated: “The widespread adoption of an aggressive attitude toward [aortic] root replacement, therefore, may create more problems than it solves.”

Another recent report also provides some evidence that prophylactic resection of the ascending aorta is probably performed too frequently (unnecessarily) in patients undergoing AVR for dysfunctioning congenitally malformed aortic valves. The evidence is provided by a histologic study of operatively excised dilated (>4.5-cm) aortas in patients undergoing AVR for dysfunctioning unicuspid or bicuspid aortic valves. My colleagues and I examined histologically the operatively excised ascending aortas in 58 patients undergoing AVR for AS involving a congenitally unicuspid or bicuspid aortic valve and in 38 patients undergoing AVR for pure AR involving a bicuspid aortic valve. All 96 patients by preoperative imaging studies had ascending aortas with diameters >4.5 cm. We found that dividing the patients into those with AS and those with pure AR, something that had not been done previously, was helpful in predicting whether the structure of the ascending aorta was likely to be normal, in which case its resection probably was not indicated, or abnormal, in which case its resection probably was indicated. In the 58 patients with AS, the structure of the aortic media was normal or nearly normal (consistent with aging) in all 14 patients with unicuspid aortic valves and in 39 (87%) of the 44 patients with bicuspid valves. In contrast, of the 38 patients with purely regurgitant congenitally bicuspid aortic valves (none had unicuspid valves), the aortic media was normal in 20 (53%) and abnormal in 18 (47%). These findings suggest that among patients with preoperative imaging studies showing the ascending aorta to have a diameter >4.5 cm, it is unlikely the ascending aorta requires resection if the congenitally malformed aortic valve is stenotic. However, aortic resection will be necessary in about 1/2 of the patients with purely regurgitant congenitally bicuspid valves. Unfortunately, among the patients with pure AR, no preoperative features were found to distinguish those with normal from those with abnormal aortic media histologically (judged exclusively by the degree, if any, of the loss of medial elastic fibers).

In summary, it is likely that too many dilated ascending aortas are being excised prophylactically in patients with aortic valve dysfunction involving a congenitally unicuspid or bicuspid aortic valve. The rates of late postoperative occurrence of aortic dissection and/or rupture are extremely low, and additional dilation of the ascending aorta after AVR infrequently produces deleterious clinical events. The histologic structure of the dilated ascending aorta is usually normal in patients with stenotic congenitally unicuspid or bicuspid aortic valves. In contrast, the aortic histologic structure is often (about 50%) abnormal in patients with dilated ascending aortas and purely regurgitant congenitally bicuspid aortic valves.