Methods

From an ongoing clinical registry of patients who underwent clinically indicated MDCT at the Leiden University Medical Center (Leiden, The Netherlands), 85 patients with a bicuspid aortic valve were identified. Additional 713 patients with a tricuspid aortic valve were identified. A propensity score was used to match in a 1:3 fashion patients with a bicuspid aortic valve to patients with a tricuspid aortic valve (see details in the “Statistical Analysis” section). The resulting population comprised 70 patients with bicuspid aortic valve and 210 patients with tricuspid aortic valve. MDCT data of aortic valve calcium and the presence of coronary atherosclerosis were evaluated and compared between patients with a bicuspid versus a tricuspid aortic valve. The presence and severity of aortic regurgitation and aortic stenosis were assessed with transthoracic echocardiography according to current recommendations. Data were prospectively collected in the departmental electronic clinical files (EPD Vision, version 11.3.26.0; Leiden, The Netherlands) and retrospectively analyzed. The institutional review board approved this retrospective analysis of clinically acquired data and waived the need of patient written informed consent.

MDCT scans were acquired with a 64-slice MDCT scanner (n = 6; Aquilion 64; Toshiba Medical Systems, Otawara, Japan) or with a volumetric 320-slice MDCT scanner (n = 274; AquilionOne; Toshiba Medical Systems, Tochigi-ken, Japan). For the 64-slice scanner, a collimation of 64 × 0.5 mm, rotation time of 400 ms, and tube voltages and currents (adjusted to the body mass index) of 120 to 135 kV and 250 to 500 mA were used and for the 320-slice scanner a collimation of 320 × 0.5 mm, rotation time of 350 ms, and tube voltages and currents of 100 to 135 kV and 200 to 580 mA. Unless contraindicated, oral β blockers were administered to patients with heart rates >65 beats/min after careful evaluation of the hemodynamic conditions. The MDCT acquisition protocol started with a prospective calcium scan (collimation 4 × 3.0 mm, tube voltage and current of 120 kV and 200 mA), followed by the contrast-enhanced MDCT coronary angiography. Based on the patient’s body weight, 80 to 115 ml of ionic contrast medium (Ultravist 370; Bayer, Whippany, New Jersey) was administered intravenously if the 64-slice system was used (flow rate of 5.0 ml/s) and 60 to 105 ml of contrast material was injected in 3 phases if the 320-slice system was used: first 60 to 80 ml of contrast material (flow rate 5.0 to 6.0 ml/s), followed by a 1:1 mixture of contrast and saline and additional 25 ml of saline (flow rate 3.0 ml/s). Triggering of the scan was synchronized with arrival of the contrast material in the left ventricle using automated peak enhancement detection (threshold of +180 Hounsfield units). For the 64-slice system, the electrocardiogram was simultaneously recorded during image acquisition for retrospective gating and image reconstruction. With the 320-slice scanner, the electrocardiogram was simultaneously registered for prospective triggering of the data. Phase windows were set at 70% to 80% or 30% to 80% of the RR interval in patients with regular heart rates of ≤65 or >65 beats/min, whereas target scans at 75% or 45% of the RR interval were performed for irregular heart rates of ≤65 or >65 beats/min. When the entire cardiac cycle was scanned (i.e., patients who underwent transcatheter aortic valve implantation), dose modulation was applied with the maximal tube currents at 75%, 65% to 85%, or 30% to 80% of the RR interval in patients whose heart rates were <60, 60 to 65, or >65 beats/min, respectively. MDCT data were reconstructed initially at 75% of the RR interval with a slice thickness and reconstruction interval of 0.5 and 0.25 mm. In the presence of multiple phases, additional phases with the least motion artifacts were reconstructed. Subsequently, for off-line image analysis, the reconstructed data sets were transferred to an external workstation (Vitrea 2; Vital Images, Plymouth, Minnesota).

The noncontrast calcium scans were used to assess the Agatston coronary artery calcium score and the calcium score and calcium volume of the aortic valve (including a volume from the aortic annulus until the level of the coronary ostia). To accurately exclude contiguous calcium located in the coronary arteries and mitral valve annulus from the calcium analysis of the aortic valve, the aortic valve calcium was also evaluated on the higher spatial, contrast-enhanced images enabling an accurate delineation of the calcium by multiplanar reformation planes. On the contrast-enhanced images, the presence of significant CAD was defined as ≥50% stenosis. The location of the aortic valve calcium was also assessed on the contrast-enhanced MDCT images.

Continuous data are presented as mean ± SD or as median and interquartile range (or with the minimal and maximal values), as appropriate. Categorical data are displayed as frequencies and percentages. To control the effects of confounding factors, propensity score matching was performed using a multivariate binary logistic regression model with the type of aortic valve (bicuspid or tricuspid) as dependent variable. Cardiovascular risk factors (age, gender, hypertension, hypercholesterolemia, diabetes, and smoking), the presence of chest pain symptoms and the MDCT clinical indication were added as covariates. The Hosmer-Lemeshow goodness-of-fit test was used to check the accuracy of the model. Subsequently, propensity score 1:3 (bicuspid:tricuspid) matching was performed with replacement and a caliper of 0.22 that was twice the SD of the probability. Differences between patients with a bicuspid aortic valve and those with a tricuspid aortic valve were analyzed using the unpaired Student t test or the Mann-Whitney U test, as appropriate, for continuous data and with the chi-square test for categorical data. Statistical tests were 2 sided, and p values <0.05 were considered significant. All statistical analyses were performed with the SPSS software (version 20.0; SPSS Inc., Chicago, Illinois).