Predictors of paravalvular aortic regurgitation after surgery for Behcet’s disease-related severe aortic regurgitation

Behcet’s disease (BD) is a chronic systemic inflammatory disorder of unknown etiology, characterized by recurrent orogenital ulcers and other organ involvement such as skin and eyes. Cardiac involvement in BD is uncommon but is associated with life-threatening complications [1‐6]. Among various manifestations of cardiac BD, aortic regurgitation (AR) is the most serious complication, because early diagnosis and surgical treatment are challenging [3, 4]. The fragility of aortic structures and tissue inflammation in BD make surgical treatment of AR more difficult, frequently resulting in paravalvular leakage (PVL), pseudoaneurysm formation, and even death [3‐5, 7‐9]. Prognosticators have been suggested, however, the number of patients included in earlier studies was too small to draw robust conclusions. This study aimed to determine the predictors of PVL after corrective aortic valve (AV) surgery for BD-related AR at two large tertiary university hospitals.

The major findings of this study are summarized as follows: 1) The prevalence of severe BD-related AR is approximately 3.1% in all patients with severe AR of diverse etiologies requiring AV surgery. 2) The rate of complications such as PVL requiring repeat OHS, atrioventricular conduction disturbance, and death was not low. 3) Since the clinical introduction of pathognomonic echocardiographic findings suggesting severe BD-related AR, the incidence rate of PVL, the number of required repeated OHSs, and cardiac mortality tended to decrease. 4) BD-related PVL developed more frequently in patients not meeting the conventional ISG criteria. 5) Finally, concomitant ARR and postoperative IST were two independent predictors for less PVL development, whereas preoperative IST was not.

AR is a rare complication of BD [2‐4, 8, 12, 17], and its pathophysiology has yet to be completely understood [17]. Although annular dilation and sinus of Valsalva pseudoaneurysm formation secondary to aortitis had been suggested [18], direct valvular inflammation was also proposed as a critical mechanism causing AR [2, 12]. Valvular inflammation could lead to aneurysmal changes in the AV cusp, pseudoaneurysm of aortic root, and mass-like lesions, all of which are currently considered pathognomonic echocardiographic findings of AR in BD (Additional files 2, 3 and 4: Video S1 S2 and S3) [3‐5, 12].

Pathophysiology of recurrent PVL after corrective AV surgery is not completely understood in this population, although defective tissue-healing process due to inflammation was suggested as a main pathophysiology [19]. The frequency of early PVL after the first surgery in patients with BD-related AR was significantly higher (18 of 35 patients, 51.4%), compared to the rate of 2-year PVL (4 of 404 patients, 1.0%) reported in a previous study, where patients diagnosed with infective endocarditis who underwent AVR were included [20]. Thus, it is essential to find the appropriate management strategy (i.e., early diagnosis and appropriate perioperative treatment) to improve postoperative outcome. However, data has been insufficient, possibly owing to the rarity of the disease and difficulties in the diagnosis of BD-related AR prior to full-blown presentation of the ISG criteria.

Because BD has no pathognomonic single diagnostic test, ISG criteria have been used since 1990 [14]. Notwithstanding, there were multiple reports showing that the ISG criteria were not useful for the early diagnosis of BD-related AR [3, 4, 8, 11‐13]. Consistent with earlier reports, we also observed that only 6 patients (17.1%) fulfilled the ISG criteria prior to the first AV surgery. Given poor sensitivity of the ISG criteria in diagnosing BD-related AR [3, 4, 7, 8, 11‐13], diagnosis and treatment may be delayed, leading to poor prognosis. The absence of vascular manifestations in the ISG criteria may be one reason why the ISG criteria are considered less sensitive for the diagnosis of BD-related AR. Besides, the ISG criteria were originally suggested to elicit common consent for the diagnosis of BD [14]; thus, it is not suitable for the early detection of rare manifestations of BD on an individual basis. The ICBD introduced vascular and neurologic manifestations as additional criteria for the diagnosis of BD [16], and the application of this criteria in our study population could allow the diagnosis of BD in up to 26 patients (74.3%). Nevertheless, nine patients in our series were still not put on the diagnostic list of BD. Thus, we need to keep in mind that severe AR can be an initial manifestation of BD even in patients who do not meet the current diagnostic criteria. As the initial transthoracic echocardiography in all patients invariably showed pathognomonic findings reported in the literature [3‐5, 8, 10], we believe that transthoracic echocardiography can play a critical role in the early diagnosis of BD-related AR, even in patients who do not meet the ISG or ICBD criteria [3, 4, 8], and should be used clinically as a single specific diagnostic test for the early diagnosis of BD-related AR. We also observed that patients who did not definitely meet the ISG criteria had a poorer 2-year PVL-free survival rate, whereas the definite BD group carried better prognosis. This may be related to the early establishment of IST, which was confirmed by the multivariate analysis in the current study. In fact, we observed that postoperative IST, but not preoperative IST, emerged as an independent factor preventing PVL development after multivariate analysis. Besides, as preoperative IST was used in almost every case in which postoperative IST was used, the direct comparison of efficacy between preoperative and postoperative IST was not fair. To avoid this statistical fallacy, we performed the sensitivity analysis (Additional file 1: Figure S3) and confirmed that postoperative IST is an independent predictor along with concomitant ARR. This observation can give comfort to cardiologists or rheumatologists who need to decide whether to proceed with AV surgery, because most patients with BD-related severe AR can present with acute heart failure and require urgent surgery without effective control of systemic inflammation. Therefore, it may not be acceptable to delay corrective AV surgery in patients with BD-related severe AR to control for aortitis or valvulitis. We should also remember the importance of concomitant ARR as an effective surgical option to decrease PVL in this population. As the pathergy test implies, exaggerated inflammatory response to the injury to the unhealthy aorta can be elicited in patients with active BD [19, 21]. Thus, suturing artificial prosthesis such as mechanical or bioprosthetic valve to the inflamed aortic root could lead to unsuccessful healing, resulting in early PVL due to dehiscence. Concomitant ARR at the time of AVR using premade valved conduits or homografts can be the best option to avoid prosthetic valve dehiscence.

CAVB is a well-known complication of BD-related AR. We found that its incidence was 48.6%, whereas the incidence of CAVB requiring pacemaker implantation after general AR surgery was reported to be 3.2% [22]. Because the atrioventricular node is located just below non-coronary cusp of AV [23], the inflammation of aortic root, esp. non-coronary cusp of AV, and its expansion to the neighboring atrioventricular node can be a potential explanation regarding high frequency of CAVB. This explanation is advocated by the observation that the non-coronary cusp of AV was involved in 9 of 11 patients (81.8%) in our series. In addition, development of interventricular septal dissection (Additional file 3: Video S2 and Additional file 4: Video S3) mechanically blocks the atrioventricular node conduction, resulting in CAVB.

In the past, less attention was paid to BD-related AR, mainly because of its rarity and absence of pathognomonic imaging findings. Since the introduction of characteristic echocardiographic findings [3‐5, 8, 10], early diagnosis and early intervention have been possible, which is expected to have a favorable impact on outcomes. This is an important issue but is difficult to prove, because severe BD-related AR is a rare disease, and thus, a prospective study is extremely difficult to perform. Instead, we attempted to investigate the temporal trends of the incidence of PVL, number of required OHSs, and cardiac mortality by dividing the study population into three groups according to the time point when the first surgery was conducted. The decreasing trend of adverse events are shown in Fig. 3 suggesting that early detection and appropriate medical and surgical treatment should improve clinical outcomes.

Several limitations should be acknowledged. Firstly, only 35 patients with BD-related severe AR were found and analyzed, making the study susceptible to low statistical power and overfitting issue. However, BD-related AR is rare in prevalence, and to our knowledge, the present study included the largest population of patients with BD-related severe AR requiring AV surgery. Although we included 15 patients who were included in the previous study [4], the analysis of our study was completely different from the previous one. Secondly, nine patients in this study failed to fulfill the current ISG or ICBD criteria for a definite diagnosis of BD; hence, there might be an argument regarding the diagnostic accuracy in these patients. However, surgical and pathological findings for all patients showed characteristic histopathology suggesting AV involvement in BD, and thus, all relevant experts finally agreed on the diagnosis of BD. Thirdly, the institutions at which the study performed were tertiary referral centers, therefore the prevalence of the BD-related severe AR requiring AVR could be overestimated. Finally, we initially searched for patients with severe AR who underwent AV surgery; therefore, seriously ill patients who could not proceed with surgical intervention might not have been included in this study, leaving the possibility of underestimation of the prevalence of BD-related severe AR.

In conclusion, BD-related severe AR requiring AV surgery is a rare but clinically important disease entity. Postoperative IST and concomitant ARR, but not preoperative IST, are independent predictors of less PVL development. A special attention is required to early diagnose BD-related AR, especially in patients who did not satisfy the ISG criteria. The active use of pathognomonic echocardiographic findings could facilitate early diagnosis.