Severe aortic stenosis (AS) is often characterized by myocardial interstitial fibrosis. Myocardial interstitial fibrosis, classically measured by magnetic resonance imaging, was also shown to be accurately measured by computed tomography (CT)-derived extracellular volume fraction (ECVF). Serum albumin (SA) level (g/dl) has been shown to correlate with ECVF among patients with heart failure and preserved ejection fraction. Our objective was to evaluate the association between SA and ECVF among patients with severe symptomatic AS. Patients with symptomatic severe AS who were evaluated as candidates for intervention between 2016 and 2018 were enrolled prospectively. All patients underwent precontrast and postcontrast CT for estimating myocardial ECVF. Valid ambulatory SA within 6 weeks of the cardiac CT were obtained and classified as (tertiles) <3.8, 3.8 to 4.19, and ≥4.2 g/dl. Patients with acute systemic illness at the time of the albumin test were excluded. The study included 68 patients, mean age 81 ± 6 years, 53% women. Patients with lower SA were more likely to have chronic renal failure, previous percutaneous coronary interventions, and a reduced functional class. The mean ECVF (%) in the study cohort was 41 ± 12%, significantly higher among the patients in the lower SA level groups (50 ± 12% vs 38 ± 7% vs 33 ± 9% in the <3.8 g/dl, 3.8 to 4.19 g/dl and ≥4.2 g/dl groups respectively, p for trend <0.001). A statistically significant inverse correlation was found between SA levels and ECVF ( r −0.7, p <0.001). Multivariable analysis showed significant independent association between low SA and ECVF. In conclusion, the SA level is inversely associated with CT-derived ECVF in patients with severe AS.
Albumin is the main protein in human plasma with many physiologic roles. Hypoalbuminemia is emerging as a powerful cardiovascular prognostic marker in a wide variety of cardiovascular diseases. Low serum albumin (SA) levels were reported to be associated with worse outcomes in patients with severe aortic stenosis (AS) who underwent transcatheter aortic valve implantation (TAVI). , In patients with severe AS, overt heart failure was characterized by increased myocardial interstitial fibrosis (MIF) that was shown to be a negative prognostic marker in these patients. MIF is the end point of various cardiovascular diseases and is often evaluated noninvasively by measurement of the myocardial extracellular volume (ECV) fraction (ECVF) using a cardiac magnetic resonance imaging (MRI). ECVF measured by MRI has been shown to correlate with histopathologic findings. , However, in most institutions, patients with severe AS who are candidates for intervention routinely undergo a preprocedural cardiac computed tomography (CT) scan. Recent studies have shown the feasibility and accuracy of CT-derived ECVF measurement in these patients. Furthermore, a recent study has found the lower SA level is associated with more MIF as measured by ECVF in patients with heart failure and preserved left ventricular ejection fraction. We aimed to evaluate the association between SA level and CT-derived ECVF in patients with severe AS.
Methods
The present study included patients with symptomatic severe AS who were evaluated as candidates for valvular intervention (surgical aortic valve replacement [SAVR] or TAVI) between 2016 and 2018. The enrollment for this study was performed prospectively as previously reported. In the present study, we included patients that had a valid ambulatory SA result within 6 weeks of the cardiac CT. Patients were excluded because of the presence of pacemakers, implantable cardioverter defibrillators, metallic foreign objects in the proximity of the heart, SAVR, active malignancy, chronic liver failure, major surgery in the last 6 months, active inflammatory/rheumatic disease, sepsis, or acute infection when SA levels were drawn. Patients were referred for cardiac CT (see details about protocol in the following paragraphs) by their part of the routine preprocedural evaluation process in our institute.
Severe AS was defined as per contemporary valvular heart disease guidelines (symptoms typical of AS, mean aortic gradient >40 mm Hg, valve area <1.0 cm 2 ). Furthermore, AS was stratified into 5 stages according to the extent of cardiac involvement as previously described by Généreux et al. Stage 0 = no cardiac damage; stage 1= left ventricular (LV) damage as defined by presence of LV hypertrophy (LV mass index >95 g/m 2 for women, >115 g/m 2 for men), severe LV diastolic dysfunction (E/e′ >14), or LV systolic dysfunction (LV ejection fraction [LVEF] <50%); stage 2 = left atrial or mitral valve damage or dysfunction as defined by the presence of an enlarged left atrium (>34 ml/m 2 ), presence of atrial fibrillation, or presence of moderate or severe mitral regurgitation; stage 3: pulmonary artery vasculature or tricuspid valve damage or dysfunction, defined by the presence of systolic pulmonary hypertension (systolic pulmonary arterial pressure >60 mm Hg) or presence of moderate or severe tricuspid regurgitation; stage 4 = right ventricular damage as defined by the presence of moderate or severe right ventricular dysfunction.
SA levels were measured using a photometric color test for clinical chemistry analyzers and were categorized into 3 categories (tertiles) as <3.8 g/dl, 3.8 to 4.19 g/dl, and ≥4.2 g/dl because the distribution of the SA values the groups were not exactly identical.
The institutional review board approved the study which was performed in accordance with the Declaration of Helsinki. All patients signed a written informed consent.
All patients underwent CT angiography as a part of a routine evaluation before intervention for severe AS using a 256-slice system (Brilliance iCT, Philips Healthcare, Cleveland, Ohio). Precontrast CT scans had a tube voltage of 120 kV, tube current 337 mA, and gantry rotation time of 330 milliseconds. Acquisition was performed during an inspiratory breath hold, while the electrocardiogram was recorded simultaneously to allow for prospective gating of the data. Besides the precontrast scan, an additional postcontrast scan with the same scan parameters was added 7 minutes after contrast infusion. Images were reconstructed using iterative model reconstruction (level 2) with a slice thickness of 2 mm. To avoid contrast agent exposure and potential risk of contrast nephropathy, we used intravenous injection of 50 to 60 ml of nonionic contrast agent (iopromide 370, Bayer Schering, Berlin, Germany) at a flow rate of 3 ml/s, followed by a 30-ml saline chase bolus (3 ml/s). The 3-dimensional dataset of the contrast-enhanced CT scan was reconstructed at 10% intervals over the cardiac cycle, generating a 4-dimentional CT dataset.
We measured myocardial and blood pool attenuation values in the precontrast and postcontrast CT scan. This measurement was performed by an experienced cardiovascular imaging specialist, blinded to the clinical data and other test results. Region of interest was drawn on myocardial septum with the greatest area and within the descending aorta. The regions of interest were first drawn on the CT scan obtained during contrast injection and then copied to the precontrast and postcontrast CT scans as previously described. Thereafter, the mean attenuation of the myocardium and blood were recorded and expressed in Hounsfield units. The recorded myocardial ECVF (%) was calculated using the following formula:
ECVF=(1−hematocrit)X(ΔHUtissueΔHUb)
ΔHU=HUpost−HUpre
Statistical analyses were performed with IBM SPSS Statistics 26 software. Baseline characteristics of the study cohort were presented as mean and SD for continuous variables and n (%) for categorical data. Comparisons of the baseline characteristics between the study groups were performed using chi-square or chi-square for linear trend and analysis of variance (ANOVA) or ANOVA for linear trend tests. The relation between SA levels (as the continuous variable) and ECVF was assessed using Pearson correlation. In addition, the association between SA categories and the outcome was assessed using a linear regression, in which the lower category was used as the reference group and polynomial second order (quadratic) regression. Two models were built: the first model (univariate) included only the variable of the SA group ; the second model included the variable of SA category and the patients’ baseline characteristics, which were found to be significant in the univariate model and in addition to age and LVEF. The results of the models were presented as the coefficients with standard errors and beta. In addition, the coefficient of determination (adjusted R-squared) for each model was presented. For each test, two-sided p <0.05 was considered significant.
Results
Overall, the study included 68 patients, mean age 81 ± 6 years, 53% females. Baseline characteristics according to the SA tertiles are presented in Table 1 . Patients in the lower SA level tertiles were more likely to have chronic renal failure, previous percutaneous coronary interventions, and a reduced New York Heart Association functional class. No other statistically significant differences were found between the study groups. Specifically, no statistically significant differences were found in the peak and mean aortic valve pressure gradients and/or aortic valve area between the groups, distinguished by the SA tertiles. Similarly, the CT-derived calcium score measure was not significantly different between the study groups. The mean ECVF in the study cohort was 40.96 ± 11.6%, significantly higher among the patients in the lower SA level groups (50 ± 11.6% vs 38.2 ± 7.3% vs 32.8 ± 8.6% in the <3.8 g/dl, 3.8 to 4.19 g/dl, and ≥4.2 g/dl groups, respectively, p for trend <0.001). Furthermore, as shown in Figure 1 , a statistically significant inverse correlation was found between SA levels and ECVF (linear regression: r −0.7, p <0.001) and polynominal order 2 [quadratic] regression ( r −0.715), p <0.001. The results of the univariate and multivariate models ( Table 2 ) have shown statistically significant inverse association between SA and ECVF with the lower categories of SA independently associated with higher ECVF. The additional parameters in the multivariate model were age, LVEF (with no statistical significance), and grade 2 or higher diastolic dysfunction. The goodness-of-fit measure of the multivariate model demonstrated that about 40% of the variance in ECVF are explained collectively by these independent variables.
