Differences at surgery between patients with bicuspid and tricuspid aortic valves

In 2012 all patients with BAV or TAV disease being followed up after AVS were identified from the electronic medical record (EMR) system in Gelre Hospital, Apeldoorn, a non-academic teaching hospital without a cardiothoracic surgery unit (Fig. 1). AVS was defined as replacement or valvuloplasty of the aortic valve or replacement of the aortic root. Transcatheter aortic valve implantations were excluded as the aortic valve is not visualised. Clinical data were retrieved from the EMR; details of surgery and native valve anatomy were documented from the operative report. Surgical assessment was used to determine the number of cusps. The reliability of this method was discussed in a previous publication referring to many studies that used the intra-operative description by the surgeon for determination of valve anatomy [13].

Cardiovascular risk factors were assessed at the time of surgery. Hypertension was defined as being on antihypertensive treatment and diabetes as being on antidiabetic medication. Hypercholesterolaemia was defined as a serum total cholesterol >7.0 mmol/l or being on cholesterol-lowering medication.

Proximal aortic surgery was defined as comprising all surgical procedures involving the aortic root and/or the ascending aorta [i. e. aortic root replacement, Bentall procedure, supracoronary ascendens replacement (SCAR) and ascending aorta reefing/patch enlargement]. A Bentall procedure was indicated if there was an indication for AVR and a pre-operative aortic root or ascending aorta diameter of ≥4.5 cm during AVR, or a thin-walled aorta at inspection at the discretion of the surgeon. SCAR was usually performed in cases of a dilated ascending aorta but a non-dilated aortic root. Replacement of the aortic arch was usually performed if the dilatation included the distal ascending aorta (≥4.5 cm or 4.0 cm and thin-walled).

Normally distributed continuous variables were tested with the Student t-test, not normally distributed variables with the Mann-Whitney test. Dichotomous variables were tested with the chi-squared statistic or Fisher’s exact statistic in the case of small numbers. Logistic multivariable regression analyses were used to assess the association between the BAV/TAV status and the aortic valve procedure and additional surgery. Linear regression analysis was used to assess the association with the prosthesis diameter. Adjustment was performed for potential confounders, i. e. variables with a univariate association with BAV/TAV status indicated by a p-value <0.10, with a maximum of one confounder per 10 cases. A significance level of <0.05 was considered significant, providing 95% confidence intervals. We used IBM Statistical Package for the Social Sciences (SPSS) 24.

The institutional review board judged that this study fell outside the scope of the Dutch law of medical-scientific research with humans (WMO), and therefore patient consent was not required.

In our study, in many patients the presence of CAD and aortic valvular disease (AVD) required surgical intervention simultaneously, driven by cardiovascular risk factors, but only in TAV patients, as BAV patients less often underwent concomitant CABG and had less prevalent cardiovascular risk factors. A recent systematic review reported that at the time of AVR, in all 16 studies the need for concomitant CABG was lower for BAV than for TAV. A meta-regression analysis showed that this was explained by younger age and lower prevalence of diabetes in BAV compared to TAV patients [18]. In contrast, our study demonstrated that native valve anatomy determined the need for CABG in addition to age and hypercholesterolaemia. Davies et al. [11] and Boudoulas et al. [12] observed that the need for concomitant CABG was lower in BAV than in TAV patients of every age group. As in our study, the difference in the need for CABG was associated not only with age but also with native valve anatomy.

On the other hand, the difference in cardiovascular risk factors at the time of AVR also points to possible differences in the development of AVD between BAV and TAV. A study by Huntley et al. found that after age-matching, patients with TAV compared with BAV stenosis had more cardiovascular risk factors, including hypertension, hyperlipidaemia and diabetes [19]. Similarly, in our study patients with TAV had more cardiovascular risk factors than BAV patients, with hypertension and diabetes mellitus being associated with age, but hypercholesterolaemia remained more prevalent in TAV after multivariate analysis.

These results point towards hypercholesterolaemia as a possible major driver in tricuspid AVD, which is not or less the case in bicuspid AVD. In contrast to eight (of the nine) observational studies, the five randomised controlled trials (RCTs) with statin therapy failed to show any benefit regarding the progression of AS. However, it might be questioned whether the negative results of these RCTs exclude the possibility that hypercholesterolaemia might be a major driver in tricuspid AVD [20]. The two largest studies among these RCTs, the SEAS and the ASTRONOMER, comprised 5.3 and 49% patients with BAV [10, 21]. We are not aware of sub-studies reporting on the effect of statins in TAV patients only. Furthermore, it must be noted that the SEAS and the ASTRONOMER studies excluded patients with a clinical indication for cholesterol lowering, such as CAD and diabetes which diminished the prevalence of cardiovascular risk factors in the study and hence also diminished the amount of ‘typical tricuspid AS’ patients. We agree with the conclusion of a recent systematic review that the exclusion of patients with the greatest risk of atherosclerosis and the relative brief follow-up of the RCTs likely reduced the possibility that statins would produce a therapeutic effect [20]. A study of tricuspid AS patients which takes this into account might be considered, as the recent review article also suggested [20].

Valve prosthesis diameters were on average larger in BAV than in TAV, which was also reported by Huntley et al. [19]. The mean difference in diameter between BAV and TAV in their study was 0.8 mm, while this was 1 mm in ours, and after adjustment 0.5 mm. A difference between the studies must be noted: Huntley et al. compared BAV patients with age-adjusted TAV patients, whereas we adjusted the diameter for confounders, including age and gender.

The larger prosthesis in BAV compared with TAV is likely related to annulopathy as a part of BAV aortopathy [22, 23]. Furthermore, due to the BAV aortopathy, more additional ascending aorta or aortic arch procedures were also needed. The proximal aorta was repaired in 31% of our BAV patients. Two recent studies reported similar percentages for aortic replacement: 27 and 26% [24, 25]. Since 2002, after a report of a continuing dissection risk during follow-up after isolated AVR in BAV [13], there has been an increasing awareness of possible complications of an unrepaired ascending aorta in BAV. Based on a study published in 2004 [26], the 2006/2007 guidelines recommended replacing the ascending aorta concomitantly when its diameter exceeds 45 mm [27, 28]. However, among the patients in this study who developed an ascending aorta aneurysm requiring surgery, the majority also simultaneously underwent AVR for structural valve deterioration. In these patients it was difficult to retrospectively determine the principle reason for their reoperation [26]. More recent studies reported low incidences of adverse aortic events following isolated AVR in patients with BAV [15, 17, 29]; of these patients, those with BAV insufficiency might be a small subgroup with an increased risk [30]. The question may be posed as to whether too many ascending aorta replacements are performed, as the guideline recommendations were based mainly on one publication.

This study is limited by its retrospective design, including the lack of pre-specified criteria for diagnosing BAV at surgery. In our study, 445 out of 894 (50%) patients were excluded for various reasons (died, moved away, operative report missing) (Fig. 1). Our study population is not an ideal sample to study differences in initial patient characteristics. In contrast, it has been established that long-term post-operative survival of BAV patients, also after isolated AVR, is not lower than that in TAV patients, provided that the pre-operative diameter of the proximal aorta, the main risk factor for survival, does not exceed 5.0–5.5 cm [15, 16]. Furthermore this is a representative sample of patients with a history of AVS undergoing continued follow-up.