Reduced glomerular filtration rate is associated with ascending aortic dilatation in South African chronic kidney disease patients

Background: Ascending aortic dilatation (AAD) is an adverse prognostic cardiovascular marker in the general population. There are few data reporting its presence or clinical significance in chronic kidney disease (CKD) patients. The aim of this study was to evaluate ascending aorta dimensions and their correlates in a population of South African CKD patients. Methods: A total of 124 CKD patients and 40 healthy controls were enrolled. Cardiac dimensions, systolic and diastolic function indices, and aortic root diameters were assessed by transthoracic echocardiography. The ascending aorta was measured at four levels (aortic annulus, sinuses of Valsalva, sino-tubular junction, and ascending aorta) and was normalised for body surface area. The prevalence of AAD was assessed in CKD patients compared with the control group. results: In CKD patients, the ascending aorta dimension was significantly larger than in controls at all four sites of the aorta that were measured. The prevalence of AAD was 6.5% at the annulus, 12.9% at the sinuses, 15.3% at the sino-tubular junction, and 8.9% at the ascending aorta. Overall, 29 patients (23%) had AAD. On multivariate analyses, eGFR was independently associated with AAD (odds ratio 0.980; 95% confidence interval 0.965–0.996; P = 0.014). conclusion: AAD is a common cardiovascular phenotype in South African CKD patients. Low eGFR was independently associated with AAD, suggesting a direct link between CKD and the development of AAD in South African CKD patients.

Even though it has been widely acknowledged that CKD is a predisposing condition for enhanced cardiovascular risk, reports of the link between estimated glomerular filtration rate (eGFR) and ascending aorta dimension have not been consistent [6][7][8].
Differences resulting from a lack of a uniform method of measurement of ascending aorta diameter as well as in the definition of AAD [9][10][11][12][13] and heterogeneity of study populations may have contributed to the dissimilar findings in previous studies. Further, there are no existing data on the prevalence of AAD and its association with reduction in GFR among South African CKD patients. The study reported here evaluated ascending aorta dimensions at the aortic annulus, sinuses of Valsalva, sino-tubular junction, and ascending aorta, and determined the prevalence and associated factors for AAD in a population of South African patients with CKD.

Study population, anthropometric and clinical data
We enrolled 124 CKD patients (comprising 41 peritoneal dialysis, 41 haemodialysis and 42 stage 3 CKD patients). Forty age-and gender-matched apparently healthy subjects, who were staff and students of the Charlotte Maxeke Johannesburg Academic Hospital (CMJAH), were recruited to serve as controls. Inclusion criteria were age 18-60 years, receipt of haemodialysis (HD) and peritoneal dialysis (PD) for at least three months at the dialysis unit or attending the renal clinic for at least three months with stage 3 CKD (defined as an eGFR of 30-59 mL/min/ 1.73 m 2 , present for at least three months). The eGFR was calculated from the serum creatinine using the CKD-Epidemiology Collaboration equation (CKD-EPI) formula, without the African American ethnicity factor. Patients were excluded if they had a history of moderate to severe valvular heart disease, congenital or structural heart disease (such as bicuspid aortic valve and cardiomyopathy), atrial fibrillation, diabetes mellitus, previous open-heart surgery or aortic root replacement surgery, or a disorder known to be associated with aortic diseases, including Marfan's syndrome [14].
Information regarding demographic data and tobacco use were documented. Race was self-reported by study participants. Height and weight were measured using the Seca 220 telescopic measuring rod (Seca GmbH, Germany) and body mass index (BMI) was calculated as the ratio of weight (in kg) to height (in metres) squared. Blood pressure was measured non-invasively using a GE Dinamap ® DPC121X-IA automated blood pressure monitor (GE Medical Systems Information Technologies, Inc, Milwaukee, WI, USA). Blood pressure (BP) for HD patients was recorded non-invasively in the arm without the A-V fistula in the sitting position. BP was estimated by averaging all pre-dialysis and post-dialysis BP recordings taken during the month before the study (three measurements per week for a total of 12 measurements). Among PD and CKD patients, BP was recorded at the time of the clinic visit. The BP average of four clinic visits was taken as the patient's actual BP. In the controls, BP was taken in the sitting position after resting for five minutes. The average of three readings taken five minutes apart was then recorded.

blood sample collection and laboratory analysis
Venous blood was collected in heparinised tubes and tubes containing ethylenediaminetetraacetic acid (EDTA) after a 12-hour overnight fast. Samples for measurements of haemoglobin concentration, creatinine, total cholesterol, high-density lipoprotein (HDL), low-density lipoprotein (LDL), and triglycerides were analysed by the National Health Laboratory Service as part of the routine standard of care using ADVIAR auto-analysers (Siemens Healthcare Diagnostics Inc, USA).

Echocardiography
As previously described [15,16], echocardiography was performed according to the American Society of Echocardiography recommendations using the Philips iE33 machine equipped with a S5-1 1-5 MHz transducer (Philips Corporation, USA); M-mode, two-dimensional and colour Doppler measurements were recorded. Left ventricular end-diastolic diameter (LVEDD), end-diastolic thickness of the septum and posterior wall were measured in the parasternal long axis view using M-mode, while the linear method and the parasternal long axis approach were used to calculate left ventricular (LV) mass. Left ventricular enddiastolic volume (LVEDV) and left ventricular ejection fraction (LVEF) were measured using the modified Simpson's method from images obtained in the apical 2-and 4chamber views in end-systole and end-diastole. LV mass was calculated from internal diastolic diameter, posterior wall dimension, and septal dimension. Body surface area (BSA) was calculated by the DuBois formula [17] and was used to index the measurements of LA, LVEDV and LV mass. Mitral inflow velocities during early (E wave) and late (A wave) diastole were recorded at the level of the mitral valve leaflet tips using pulsed-wave Doppler ultrasound. Peak early myocardial diastolic relaxation velocity was measured using tissue Doppler imaging at the base of lateral and septal annulus during early (e' velocity) diastole. Four consecutive cardiac cycles were recorded for offline analysis and were presented as the mean of the three consecutive heartbeats. The E/e' ratio was then calculated, and the value was considered to reflect the filling pressures to the LV.
Ascending aorta measurements were obtained in the parasternal long axis view using an inner edge to inner edge technique during systole. Ascending aorta measurements were acquired at the sinus of Valsalva, annulus, sino-tubular junction, and ascending aorta. Ascending aorta dimensions were indexed using body surface area. In adults, it should be noted that ascending aorta dimensions may be influenced by age, sex and height [18]. According to guidelines, normal values for ascending aorta dimensions in adults are: annulus ≤1.4 cm/m 2 (male and female), sinus of Valsalva ≤1.9 cm (male), ≤2.0 cm (female), sinotubular junction ≤1.7 cm (male and female), proximal ascending aorta ≤1.7 cm (male), ≤1.9 cm (female) [19,20]. Upper limits of ascending aorta dimensions were considered to be the mean ± 2 standard deviation (SD) of the values found in normal controls at each level. All measurements were performed by the same echocardiographer, who was blind to the clinical details and laboratory data of the participants.

Data analysis
All data analyses were performed using the statistical package SPSS version 16 (SPSS, Inc., Chicago, IL). Variables were presented as mean ± SD for continuous data, and percentages and frequencies for categorical data. Results were analysed using the t-test and Mann-Whitney U test where appropriate, with the one-way ANOVA tests for comparison across the groups. Correlation between variables was assessed by the Spearman or Pearson correlation coefficients accordingly. A multivariate logistic regression model was created to identify associations of AAD. Factors that are known to be associated with AAD were included in the model: age, gender, mean arterial blood pressure (MABP), and eGFR. A 2-tailed P value less than 0.05 was considered statistically significant. All data were completed, with none missing, for all participants.

EtHIcAl cONSIDERAtIONS
The study was approved by the Human Research Ethics Committee of the University of the Witwatersrand (Protocol M130127).

Prevalence of ascending aorta dilatation and echocardiographic findings
In CKD patients, the ascending aorta diameter was significantly greater than in controls at the annulus, sinuses of Valsalva, sino-tubular junction, and at the level of the ascending aorta (Table 1 and Figure 1). The overall prevalence of AAD in those with CKD, including those on dialysis, was 6.5% at the annulus, 12.9% at the sinuses, 15.3% at the sino-tubular junction and 8.9% at the ascending aorta. Overall, 23% of CKD patients and 15% of healthy controls had AAD (P = 0.260).
Cardiac chamber enlargement was greater among CKD patients compared to the controls (LV mass index: 118.4 ± 40.0 vs 84.4 ± 16.0 g/m 2 , P < 0.001; left atrial dimension: 3.9 ± 0.7 cm vs 3.2 ± 0.4 cm, P < 0.001). LVEF was significantly lower in CKD patients than in the controls (58.5 ± 9.0 % vs 65.0 ± 8.1 %, P < 0.001). Patients with CKD had LV diastolic dysfunction, with a significantly lower E/A ratio and higher LV filling pressure (E/e'), than the control group; P < 0.001. However, there was no significant difference in the deceleration time between CKD patients and those with normal kidney function. comparison between patients with ascending aortic dilatation and those with normal ascending aorta dimension Table 2 compares the parameters between patients with AAD and those with normal ascending aorta dimension. Subjects with AAD were anaemic compared with patients with normal ascending aorta dimension (haemoglobin: 10.9 ± 2.7 vs 12.2 ± 2.2 g/dL, P = 0.009) and had severe kidney failure (eGFR: 12.0 ± 4.0 vs 23.3 ± 19.8 mL/min/1.73 m 2 , P = 0.005). Table 3 shows the correlations between ascending aorta dimensions and demographic, clinical and echocardiographic characteristics of the study population. Mean arterial pressure revealed a weak positive correlation with ascending aorta dimensions whereas eGFR and haemoglobin levels were only slightly negatively correlated.

DIScUSSION
This study shows that AAD is common in South African CKD patients, among whom we found an inverse relation-reduced gFr and ascending aortic dilation ship between eGFR and ascending aorta diameter corrected for body surface area. Furthermore, after adjustment for multiple confounders, a low eGFR was found to be significantly associated with AAD in our South African CKD patients. This suggests another mechanism for the link between CKD and the increased burden of cardiovascular disease.

Prevalence of ascending aortic dilatation in South African cKD patients
We found a 23% prevalence of AAD in our CKD population. To date, there have been no studies on the ascending aorta size and its associated risk factors in indigenous African CKD populations. Mulé et al. [21], in their study of hypertensive CKD patients, reported a prevalence of 39% among patients with stage 5 CKD and demonstrated that aortic root size increased with the severity of kidney failure. According to these authors, aortic root dilatation was found in 31% and 18% of the patients when aortic root diameter was indexed for height and normalised to body surface area, respectively.

Relationship between aortic root dilatation and eGFR
We found an association between AAD and low GFR in our CKD patients. Moreover, these patients had a higher prevalence of AAD and a larger ascending aorta diameter than the healthy controls, suggesting that advanced kidney failure may represent a novel mediator of thoracic aortopathy in adult CKD patients. While it remains to be seen whether CKD patients with associated AAD are categorically at risk of thoracic aortic aneurysms, previous studies have reported an association between chronic renal insufficiency and risk of developing abdominal aortic aneurysms [23,24]. Considering this evidence, it can be argued that the harmful effect of impaired kidney function may also be extended to the structure of the aorta.
Corroborating the findings of our study, Mulé et al. [21] reported an inverse relationship between GFR and indexed values of aortic root size after adjusting for multiple confounders, thereby highlighting the link between renal insufficiency and increased cardiovascular risk mediated by the effect of AAD. However, Li et al. [8], in their study of 5,538 Chinese hypertensive patients, did not find any reduced gFr and ascending aortic dilation  and of our study (which are procedurally similar) may suggest a specific direct link between persistently low eGFR and elevated cardiovascular risk represented by AAD. The differences observed between us and by Li et al. [8] underscore the need to highlight possible explanations for the seemingly different conclusions between the two investigations.
For example, unlike in the study by Li et al. [8], where ascending aorta diameter was measured at only the annulus level using two-dimensional echocardiography, we employed a more robust method which was based on ascending aorta measurements at four different levels by M-mode tracings, including sinuses of Valsalva that have been commonly used to determine the clinical significance and prognostic value of ascending aorta size in previous studies [11][12][13]21]. Second, it has been well documented that determinants of the ascending aorta dimension may reduced gFr and ascending aortic dilation Abbreviations: AAD, ascending aortic dilatation; MABP, mean arterial blood pressure; CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; HDL, high-density lipoprotein; LDL, low-density lipoprotein; LVMI, left ventricular mass index. * P values were calculated by comparing patients with aortic root dilatation (n = 29) to those with normal aortic dimension (n = 95) using chi-squared and t-tests where appropriate.
differ if the latter is recorded at different ascending aorta levels [9,10]. Finally, Li et al. [8] estimated the GFR using the original version of the CKD-EPI equation that not only included a two-level variable for race (African American and non-African American) but has also been shown to be less accurate in Asian people [25]. However, in our study, we estimated the GFR using the CKD-EPI formula, without the African American correction factor, as it has been shown previously that the CKD-EPI equation is more accurate in African populations when the African American ethnicity factors are omitted [26]. It is therefore plausible that the aforementioned differences may help provide explanations for the seemingly different conclusions of the present study and those of Li et al. [8].

Pathophysiology of ascending aortic dilatation in cKD
Although a close relationship has been established between AAD and renal impairment [2,21], the mechanism for such association is not known. However, a logical pathophysiological explanation can be offered on the basis of the various contributions from multiple risk factors for aortopathy such as volume overload, chronic anaemia, anorexia, hypertension, and the presence of arteriovenous fistulae, which may coexist in patients with CKD [2]. Furthermore, ESKD patients on maintenance haemodialysis are exposed to significant haemodynamic instability as a result of high pulsating pressure and low resistance to flow from alterations in the intravascular volume and myocardial stunning following frequent bouts of myocardial ischaemia [27,28]. It is therefore plausible that these alterations may induce degenerative remodelling of the myocardium and vasculature, including the aorta.
Moreover, AAD represents a maladaptive vascular repair which has been linked to CKD. Elevated serum levels of parathyroid hormone in CKD patients could reflect a more advanced chronic kidney disease mineral bone disorder (CKD-MBD), which is closely associated with cardiovascular remodelling in CKD [2,29]. Besides, parathyroid hormone also has direct vascular effects. Previous experimental work has shown the presence of aortic dilatation in uraemic rat models with vascular calcification and ele-reduced gFr and ascending aortic dilation  vated parathyroid hormone levels, and demonstrated that aortic dilatation was ameliorated by modest PTH suppression [30]. Thus, the experimental evidence suggests a role for parathyroid hormone and CKD-MBD in the development of AAD in CKD patients.

Relationship between ascending aorta size and hypertensive target organ damage
Even though our data did not indicate any association between AAD and blood pressure on multivariable analysis, we found that AAD correlated with hypertension and markers of hypertension-related cardiac damage such as LV mass index. Taken together, these findings suggest that perhaps both AAD and hypertensive heart disease are manifestations of kidney impairment in CKD patients.

StRENGtHS AND lIMItAtIONS
Our study has notable strengths. In contrast to the previous works, which determined aortic dimension only at the level of annulus and Valsalva's sinuses, we evaluated ascending aorta dimensions at the aortic annulus, sinuses of Valsalva, sino-tubular junction, and ascending aorta as equal reflections of abnormal pathology. We also measured aortic root size using M-mode tracings because available evidence about the clinical significance and prognostic value of aortic root size is mostly based on aortic M-mode measurements. Further, we estimated the GFR by using the CKD-EPI equation, without the African American ethnicity factor, thus providing a more accurate estimate of GFR and significantly improving its performance, particularly in Black individuals. However, there were some limitations which need to be considered in interpreting the results. This study was observational and cross-sectional in design; any causeand-effect relationship between reduced eGFR levels and AAD remains to be established. Moreover, this was a single-centre study that included a relatively small population of CKD patients; thus, the results may not be generalised to all CKD patients. The ascending aorta measurements were performed using the inner edge to inner edge technique during systole. The 2017 guidelines for standardisation of adult transthoracic echocardiography recommend measuring the aortic annulus in mid-systole, and the other ascending aorta measurements at enddiastole. Also, the guideline recommends measuring from the inner edge to inner edge only for the aortic annulus and the leading edge to leading edge measurement techniques for the other ascending aorta measurements [19]. However, considering the recruitment period of our study (April-December 2013), it was not possible to have the ascending aorta measured according to the new guidelines because, at that time, how to measure the aorta in 2D had not been defined. Moreover, because the study population was predominantly young Blacks, with a relatively low burden of atherosclerotic CVD and non-calcified arterial wall layers, the inner edge to inner edge measurements were easily obtained. On the other hand, left atrial volume is a more accurate marker of left atrial size than left atrial diameter [15], and it has also been shown to be a better predictor of cardiovascular events [31]; however, left atrial diameter is readily recorded and has been used extensively in clinical practice. Larger longitudinal follow-up studies are required to investigate potential effects of kidney function on the occurrence and prognosis of aortopathy in this population. We did not perform CT scanning, cardiac magnetic resonance imaging or three-dimensional echocardiography, which could have helped to validate our results further. Nevertheless, available evidence suggests that the clinical significance and prognostic value of ascending aorta size is commonly based on aortic M-mode measurements using echocardiography [11][12][13]. At present, CT scanning is used more in measuring the aortic annulus before transcatheter aortic valve implantation (TAVI) and transcatheter aortic valve replacement (TAVR) procedures [20]. The classification of CKD may have been different if cystatin C had been used to calculate eGFR instead of creatinine only. In spite of these drawbacks, we believe that our findings justify a need for additional future prospective studies to screen noninvasively and to carry out detailed evaluation of ascending aorta structure with the sole aim of early recognition in CKD patients with attendant high cardiovascular risk. Moreover, the effect of parathyroid hormone and CKD-MBD was not analysed because of the study design. A long-term observational study is needed to determine levels of parathyroid hormone and other markers of CKD-MBD and their associations with AAD in patients with CKD.

cONclUSION
This study reports the first data on ascending aorta dimensions and the prevalence of AAD is in South African CKD patients. AAD appears relatively common in this population and eGFR was inversely related to ascending aorta diameter. Persistently low GFR was independently associated with AAD, thus suggesting a direct link between kidney failure and the development of AAD in South African patients with CKD.