Impaired aortic distensibility and elevated central blood pressure in Turner Syndrome: a cardiovascular magnetic resonance study

Females with karyotype-proven Turner Syndrome and age- and sex-matched controls were enrolled into this cross-sectional study from a prospective study of cardiovascular health in Turner Syndrome [11, 17, 18]. The participants had been recruited through the Danish National Society of Turner Syndrome Contact Group and an endocrine outpatient clinic. Exclusion criteria were malignancy, liver disease, contraindications to CMR, and pregnancy. Out of 67 eligible participants, fifty-seven completed this study, with exclusions due to CMR contraindications (n = 1), claustrophobia (n = 1), prior Bentall procedure (n = 1), suboptimal image quality with repeat CMR declined (n = 2), and aortic valve prosthesis (n = 6). Out of 39 eligible healthy age-matched controls, 36 completed the study with exclusions due to claustrophobia (n = 1) and suboptimal image quality with repeat CMR declined (n = 2). All examinations were performed during the same day.

In a first experiment pulse waves were recorded in the carotid and the femoral artery using a SphygmoCor (SPT-301B; Millar, Houston, Texas, USA) in combination with an applanation tonometer. Based on non-invasive recordings of the pulse waves, central blood pressure and carotid femoral pulse wave velocity can be determined. The investigation was performed between 7.30 A.M. and 9.30 A.M. following an overnight fast. Participants were instructed not to smoke or ingest caffeinated beverages at least 3 h before the examinations. Two of the investigators (JW and CT) performed all measurements with the patient in the supine position in a temperature-controlled room after > 5 min of rest.

At 11 A.M. CMR was performed on a 1.5 Tesla CMR (Achieva-dStream, Philips Healthcare, Best, The Netherlands) scanner using the standard anterior and posterior coils. Aortic distensibility was derived from 2D cine images acquired with a balanced steady-state-free-precession (bSSFP) sequence using retrospective electrocardiogram (ECG) gating. Repetition time was 3.6 ms and echo time was 1.8 ms. Depending on heart rate, 19 to 24 cardiac frames were acquired in a 14 s breath-hold. Slice thickness was 6 mm, field-of-view was 313 × 313 mm² and pixel size was 1.6 × 1.6 mm². Three imaging planes were acquired perpendicular to the aortic wall at: 1) the ascending aorta at the level of the main pulmonary artery bifurcation, 2) the transverse aortic arch, and 3) the descending aorta at the level of the main pulmonary artery bifurcation (Fig. 1a). The arterial blood pressure for distensibility calculation was measured before and after the scan with the patient lying on the scanner bed, using a CMR compatible sphygmomanometer (Aneroid, ERKA, Germany). Average blood pressure from the two measurements was used in subsequent computations. Aortic distensibility AoD was calculated as:where ΔA is the difference between the maximum and minimum cross-sectional aortic area over the cardiac cycle, ΔP is the difference between systolic and diastolic blood pressures, and A_dia is the diastolic cross-sectional aortic area [23]. The cross-sectional areas were obtained by semi-automatic segmentation of the vessel through all cardiac frames using the Siswin® software (Steffen Ringgaard, Aarhus, Denmark). The vessel lumen was manually selected by a single point within the vessel in one frame, and the vessel edge was determined as the maximum image intensity gradient found radially from this point. Edge point outliers were removed, and an ellipse was fitted to the edge points. The automatic vessel selections were visually inspected and manually corrected where necessary (Fig. 1b). The variation of aortic area over the cardiac cycle is shown in Fig. 1c. All participants were examined by the same staff and in the same scanner. The scans were analysed in two bulks. Scans were analysed in duplicates following anonymization using digest package in R (3.1.0 Foundation for Statistical Computing, Vienna, Austria). One scan was analysed 10 times to determine intra-observer variability of distensibility (coefficient of variation: 10% at the ascending aorta, 8% at the aortic arch, and 11% at the descending aorta).

As previously described and presented elsewhere [17, 24], transthoracic echocardiography was performed by a single observer on a GE Vivid 7 (GE Healthcare, Horten, Norway), with a 2.5 MHz transducer using second harmonic imaging. Aortic valve morphology and function were noted.

Aortic arch anomalies were determined from a 3D bSSFP diastolic-triggered and respiratory-gated non-contrast enhanced sequence [17, 24]. Aortic arch anomalies were diagnosed as 1) aortic coarctation when there was a shelf-like narrowing of the aortic lumen in the region of the aortic isthmus (Fig. 1d), and 2) elongated transverse aortic arch when the arch appeared elongated with a kink of the inferior curvature at the aortic isthmus. Echocardiography and CMR angiography were performed to define anomalies of the aortic valve and thoracic aorta [24], which included bicuspid aortic valve (BAV) and elongated transverse aortic arch (ETA).

In addition, we have examined these Turner Syndrome women over a period of 10 years and have determined the rate of dilation of the aorta from t = 0 to t = 10 years at nine positions: (i) aortic sinuses (measuring cusp-to-opposing-cusp diameter at the point of the maximum aortic diameter in the aortic sinus); (ii) the sinotubular junction; (iii) mid-ascending aorta at the level of the inferior margin of right pulmonary artery; (iv) distal ascending aorta immediately proximal to brachiocephalic artery; (v) proximal aortic arch between the brachiocephalic and left carotid artery arteries; (vi) distal aortic arch immediately proximal to left subclavian artery; (vii) aortic isthmus immediately distal to the left subclavian artery; (viii) proximal descending aorta between the left pulmonary artery and the top of left atrium; and (ix) distal descending aorta at the most caudal border of the left atrium [18]. These data have been added to the results.

Following experiment 2 and 3, 24 h ambulatory blood pressures were obtained with oscillometric measurements every 20 min (Spacelabs 90,217, Spacelabs Healthcare, Issaquah, Washington, USA). The cuff was placed on the left upper arm in all participants, and the cuff size was adjusted to the arm circumference. Fasting blood samples were drawn. Height, weight and medical history were recorded.

Statistical analysis was performed using Stata/IC 13.1 (StataCorp LP, College Station, Texas, USA). Normality was assessed by Q-Q plots of absolute or log-transformed values, and box-plots were scrutinized for outliers. Student’s independent t-test (given as mean ± SD or for transformed values, as geometric mean with confidence interval) or Mann–Whitney U-test (given as median with range) were used as appropriate. Repeated measurements were assessed using analysis of variance (ANOVA) testing the interaction between group (Turner Syndrome or control) and aortic position, reporting the Box’s conservative P-value. Assumptions were checked by Q-Q plot of the residuals, assessment of the covariance matrix, and assessment of sphericity. Comparisons of nominal variables were performed using the Fisher’s exact test. Bivariate correlations were assessed using Pearson’s coefficient of correlation. Explanatory models were constructed for aortic distensibility using multiple linear regression analyses. Independent variables were chosen from the bivariate correlation analyses of continuous variables. Independent variables were omitted from the models when p > 0.10. The contribution of each variable to the final model is stated as standardized coefficients (β). Assumptions behind the regression models were checked by Q-Q plots of the residuals, plotting residuals versus fitted, residuals versus each of the independent variables, box-plots, and leverage plots. A p-value < 0.05 was considered statistically significant.

Central diastolic blood pressure and night-time diastolic blood pressure were significantly raised in Turner Syndrome (Table 1), and 24-h ambulatory diastolic blood pressure correlated with central diastolic blood pressure (r = 0.6, p = 0.0001. Figure 2) and heart rate adjusted Aix. The 24-h systolic blood pressure measurements revealed comparable day and night-time values in Turner Syndrome and controls. Heart rate adjusted Aix (Table 2) was higher in Turner Syndrome compared to controls (p ≤ 0.03) even when excluding women with CoA, whilst PWV was comparable (p = 0.6).

The overall distensibility throughout the aorta differed between Turner Syndrome and controls (interaction between group and aortic position; p = 0.042). There was no site-specific difference in the ascending aorta and in the aortic arch, while aortic distensibility was reduced in the descending aorta in Turner Syndrome (p = 0.02) (Table 3, Fig. 3). Aortic distensibility was significantly lower in women with CoA (n = 8), both in the ascending aorta and descending aorta when compared to those without CoA (Table 3 and Additional file 1: Figure S1). Excluding women with CoA from the comparison of Turner Syndrome and controls left only a trend towards a lower distensibility at the descending aorta (p = 0.1; Table 3, Additional file 1: Figure S2). Turner Syndrome subgroups divided according to karyotype (p = 0.5), aortic valve morphology (p = 0.4) and the presence of an ETA (p = 0.5) were comparable for aortic distensibility, while Turner Syndrome with type 2 diabetes had significantly lower aortic distensibility at all sites (all p < 0.05), also when only studying patients with CoA in the ascending aorta and arch (all p < 0.05).

Within the Turner Syndrome group, central systolic blood pressure (ascending: r = − 0.45, p = 0.002; descending: r = − 0.34, p = 0.03), CoA (r = − 0.30, P = 0.02; r = − 0.41, p = 0.0013), and Aix (r = − 0.35, P = 0.02; r = − 0.31, p = 0.04) correlated with distensibility in the ascending and descending aorta, respectively, whereas BMI, BSA and central diastolic blood pressure did not (all p > 0.6). Only central systolic blood pressure correlated significantly with distensibility in the aortic arch. Age was strongly correlated with aortic distensibility at all levels in both Turner Syndrome and controls (all p < 0.001). When including these significant variables in a multiple linear regression model of the descending aorta, only CoA and age remained significantly explanatory variables of distensibility (β = − 0.42, p = 0.001; Table 4), while only age remained a significant explanatory variable of distensibility at the level of the ascending aorta and the arch (β = − 0.52 and β = − 0.47, both p < 0.001).

In separate analyses, we assessed the association between distensibility and the rate of change in aortic diameter during 10 years follow-up at anatomically-linked predefined thoracic sites. Ascending aortic distensibility did not correlate with rate of change in aortic diameter at any of the nearby assessed sites. Arch distensibility correlated with rate of change in aortic diameter at the aortic isthmus immediately distal to the left subclavian artery (r = − 0.36, p = 0.01). Finally, descending aortic distensibility correlated with rate of change in aortic diameter at proximal aortic arch between the brachiocephalic and left carotid arteries (− 0.36, p = 0.03) (Table 5). Omitting all cases with CoA did not materially change these results (results not shown).

Mean aortic area was similar among Turner Syndrome and controls at the ascending and descending sites, while significantly smaller at the arch (Table 3). However, when adjusted for BSA, the mean areas was similar at all sites. Note, however, the much larger variation in area of the aorta at all studied sites among females with Turner Syndrome, with some having a clearly smaller area at all sites, while others had a much larger area at all sites (Table 3).