Coronary artery calcium burden, carotid atherosclerotic plaque burden, and myocardial blood flow in patients with end-stage renal disease: A non-invasive imaging study combining PET/CT and 3D ultrasound

Background Imaging-based measures of atherosclerosis such as coronary artery calcium score (CACS) and coronary flow reserve (CFR) as well as carotid atherosclerotic plaque burden (cPB) are predictors of cardiovascular events in the general population. The objective of this study was to correlate CACS, cPB, myocardial blood flow (MBF), and CFR in patients with end-stage renal disease (ESRD). Methods and results 39 patients (mean age 53 ± 12 years) with ESRD prior to kidney transplantation were enrolled. MBF and CFR were quantified at baseline and under hyperemia by 13N-NH3-PET/CT. CACS was calculated from low-dose CT scans acquired for PET attenuation correction. cPB was assessed by 3D ultrasound. Uni- and multivariate regression analyses between these and clinical parameters were performed. Median follow-up time for clinical events was 4.4 years. Kaplan–Meier survival estimates with log-rank test were performed with regards to cardiovascular (CV) events and death of any cause. CACS and cPB were associated in ESRD patients (r = 0.48; p ≤ 0.01). While cPB correlated with age (r = 0.43; p < 0.01), CACS did not. MBFstress was negatively associated with age (r = 0.44; p < 0.01) and time on dialysis (r = 0.42; p < 0.01). There were negative correlations between MBFstress and CACS (r = − 0.62; p < 0.001) and between MBFstress and cPB (r = − 0.43; p < 0.01). Age and CACS were the strongest predictors for MBFstress. CFR was impaired (< 2.0) in eight patients who also presented with higher cPB and higher CACS compared to those with a CFR > 2.0 (p = 0.06 and p = 0.4). In contrast to MBFstress, there was neither a significant correlation between CFR and CACS (r = − 0.2; p = 0.91) nor between CFR and cPB (r = − 0.1; p = 0.55). CV event-free survival was associated with reduced CFR and MBFstress (p = 0.001 and p < 0.001) but not with cPB or CACS. Conclusions CACS, cPB, and MBFstress are associated in patients with ESRD. Atherosclerosis is earlier detected by MBFstress than by CFR. CV event-free survival is associated with impaired CFR and MBFstress. Electronic supplementary material The online version of this article (10.1007/s12350-020-02080-w) contains supplementary material, which is available to authorized users.


INTRODUCTION
Chronic kidney disease (CKD) is a major risk factor for the development of atherosclerosis and coronary artery disease (CAD) and patients with CKD are at increased risk for myocardial infarction, heart failure, and cardiac death. 1,2 The pathophysiological interaction between the kidneys and the heart is termed the cardiorenal syndrome (CRS), with especially high risk for cardiovascular events in end-stage renal disease (ESRD) patients. 3 Besides CAD of the large coronary arteries which often remains asymptomatic, microvascular disease, i.e., coronary microvascular dysfunction, is a frequent finding in this high-risk group. 4 However, questions remain of how to assess cardiovascular risk in individual ESRD patients and how to identify patients who may benefit from further testing or a coronary intervention prior to transplantation. 5,6 In suspected or known CAD the role of non-invasive myocardial perfusion imaging (MPI) in preselecting patients for coronary angiography is well established. Single photon emission computed tomography (SPECT) is a widely used imaging method with good diagnostic accuracy for the detection of relevant epicardial coronary artery disease in CKD patients. 7 However, microvascular alterations may be missed by SPECT. Thus, quantitative myocardial perfusion positron emission computed tomography (PET) is the method of choice for quantitative MPI, as it additionally quantifies absolute myocardial blood flow (MBF) and coronary flow reserve (CFR) -increasing diagnostic accuracy and adding prognostic information. 4,8,9 Recently, global CFR was proven to be an independent factor for risk stratification for all-cause and cardiovascular mortality in ESRD. 4 Another imaging-derived parameter of atherosclerosis is the coronary artery calcium score (CACS), which also showed independent prognostic value in suspected CAD. [11][12][13] Atherosclerosis -as a systemic disease -also causes carotid artery stenosis and peripheral artery disease. 10 Apart from the calculation of intima to media thickness, 10,14 carotid ultrasound enables the quantitation of carotid artery atherosclerotic plaque burden (cPB). 15 cPB was the strongest predictor of coronary artery calcium score (CACS) in asymptomatic, otherwise healthy individuals. 16 In the general population, cPB and CACS can serve as imaging biomarkers of future cardiovascular events in asymptomatic adults. 16 However, the relation between myocardial perfusion, CACS, and cPB has not yet been studied in ESRD patients. Therefore, the aim of this study was to correlate the results from myocardial perfusion PET with CACS and cPB in a cohort of ESRD patients prior to kidney transplantation and to evaluate if CFR or absolute MBF are better indicators for coronary artery pathology.

Study Population
39 patients (14 female; 36%), mean age 53 ± 12 years (range 26-69) with ESRD on dialysis [12 peritoneal dialysis (PD); 27 hemodialysis (HD)], were prospectively enrolled in the study prior to kidney transplantation. Exclusion criteria were renal failure in a pre-dialysis state, prior kidney transplantation, or known CAD in patients individual history. All patients were asymptomatic at the time of imaging. Detailed patient characteristics are provided in Table 1.

PET Acquisition
A 128-slice PET/CT scanner (Biograph mCT, Siemens Healthcare) was used for PET imaging. Patients fasted for [ 4 hour and refrained from caffeine or theophylline for 24 hour before imaging. A low-dose CT scan of the heart (100 kV, 60 mA with CARE Dose, 0.5 second per rotation, See related editorial, pp. 2671-2675 Pitch 1.2; 3 mm reconstructed slice thickness) for attenuation correction of PET data was acquired during shallow breathing. For the rest study 13 N-NH 3 (352 ± 38 MBq) was given in an intravenous bolus. List-mode 3-dimensional PET images were acquired for 14 minutes. Infusion of adenosine was started through a separate intravenous line (0.14 lg/kg, 6 minutes) and a second dose of 13 N-NH 3 (354 ± 39 MBq) was infused at the half time of adenosine infusion, followed by a 14 minutes.
list-mode acquisition. Quality control of PET and CT for attenuation correction was performed. List-mode data were resampled to attenuation-corrected static, electrocardiographically gated, and dynamic images (17 frames; 12 9 10 seconds; 3 9 20 seconds and 2 9 30 seconds). Emission data were corrected for randoms, dead time, scatter, and attenuation. All images were reconstructed iteratively by a 3D attenuation-weighted ordered-subsets expectation

Analysis of Regional Perfusion
Static perfusion images were reangulated to short, horizontal, and vertical long axis slices for visual analysis. Each of the 17 myocardial segments was scored on a scale of 0-4 (normal, mildly, moderately, severely decreased, or absent perfusion) by two independent and experienced readers blinded to other data. Summed stress score (SSS), summed rest score (SRS), and summed difference score (SDS) were calculated. Abnormal perfusion was defined as a SSS C 4. Significant ischemia was defined as a SDS C 3. Coronary angiography was performed in these selected cases.

Quantification of MBF, CFR, and Left Ventricular Function
Absolute MBF (mlÁgÁmin) was calculated from the dynamic stress and rest scans corrected for motion. Automated factor analysis was used to generate arterial input function and myocardial tissue time-activity curves. Hyperemic MBF (MBF stress ) was corrected for myocardial 13 N-NH 3 activity (from the rest study). MBF stress and resting MBF (MBF rest ) were calculated by fitting the time-activity curves to a 2compartment kinetic model as described elsewhere. 17 MBF rest was normalized for cardiac work by dividing by the ratepressure product and multiplying the result by 10.000. 18 CFR was determined as the ratio of MBF stress to corrected MBF rest . Stress and resting left ventricular ejection fraction (LVEF stress and LVEF rest ) were calculated from the ECG-triggered PET data using the Corridor 4DM software (Invia, Ann Arbor, USA).

Coronary Artery Calcium Scoring
CACS assessment was performed using the low-dose, non-contrast multidetector CT scans for attenuation correction without prospective electrocardiography-gated acquisition. CACS was calculated by the Agatston method accounting for the area of coronary artery calcification in CT ([ 130 Hounsfield units, preset value by the calcium scoring software, Siemens Healthcare, Germany) and expressed as Agatston score 19 into the following categories: B 100; 101-400, 401-1000 and [ 1000. 20

3D Carotid Ultrasound
The Philips iU22 ultrasound system (Philips Healthcare, Bothell, Washington) with a dedicated 3D transducer was used. All scans were performed by physicians with advanced experience in head and neck ultrasound imaging. The protocol included 3D imaging of the left and right common carotid artery and its bifurcation as far as assessable. All 3D-ultrasound recordings were analyzed using dedicated software for semiautomatic plaque quantification (Philips QLAB-VPQ). Carotid plaque burden (cPB) was expressed in mm 3 . 16 cPB patients were stratified in the following tertiles: B 154 mm 3 , 155-344 mm 3 , and C 345 mm 3 (13 patients each).

Follow-Up
Clinical follow-up was performed every 6 months. This included clinical patient interview, laboratory data, and recording of clinical events. End-points were major cardiovascular (CV) events and death from any cause.

Statistical Analysis
Analysis of discrete variables was performed using the Chi squared test. Continuous variables were compared either by a two-sided Student's t test for unpaired samples or by one-way analysis of variance (ANOVA) with a Bonferroni/Tamhane correction. Univariate and multivariate logistic regression analyses were performed to test for associations between continuous variables. Pearson's correlation or Spearman-Rho tests were used for the comparison of parametric normally or not normally distributed variables. Variables significantly correlated in the univariate analysis were included in a multivariate stepwise regression model. Two-sided p values \ 0.05 were considered significant. Kaplan-Meier survival curves with logrank test were performed for the analysis of patients survival. Analyses were performed with SPSS software package (IBM SPSS Statistics, Version 22.0. Armonk, NY).
All procedures performed were in accordance with the ethical standards of the institutional and research committee and with the principles of the 1964 Declaration of Helsinki and its later amendments. The study was approved by the institutional review board of the Medical Association of Westfalen-Lippe and the Faculty of Medicine, University of Münster. Informed consent was given by all patients.

RESULTS
Detailed characteristics of the patient cohort are provided in Table 1.  Figure 1A) or with the time on dialysis (r = 0.28; p = 0.08; Figure 1B). CACS did not differ between the PD and HD subgroups (Table 1).
Significant correlations are summarized in Table 3. In a stepwise multivariate regression analysis hemoglobin proved as strongest predictor for MBF rest . Age and CACS were the strongest predictors for MBF stress whereas only age was predictive for cPB.

Ischemia
Reversible stress perfusion defects were observed in four patients (10%). In these cases ischemia ranged from 4 to 9%. Global CFR was C 2.0 in all cases and MBF stress was impaired (\ 2.0 mlÁgÁmin) in only one case. None of the patients had significant coronary artery stenoses [ 50% in subsequent coronary angiography.  Median follow-up time was 4.4 years (mean 3.6 ± 1.1 years). In the follow-up, there were four major cardiovascular events (MACE): 2 myocardial infarctions, 1 stroke, and 1 pulmonary embolism. There were six deaths, all by non-cardiovascular causes. CV event-free survival was associated by reduced CFR and MBF stress (p = 0.001 and p \ 0.001) but not with cPB or CACS ( Figure. 5). On the other hand overall survival could not be predicted by any of these variables (all p [ 0.25).

DISCUSSION
We investigated the association between quantitative MBF, CACS and cPB in patients with ESRD assigned for kidney transplantation evaluation. Our main  The BioImage Study investigated the relationship of several (imaging) biomarkers and cardiovascular outcome in a large population of asymptomatic adults without known cardiovascular or renal disease. 16 Here, a distinct correlation between CACS and cPB was described. In another study cPB was shown to be predictive of mortality and to occur more frequently in patients on HD than on PD, 21 the latter also observed in our study. In line with the BioImage Study, we found a significant correlation between CACS and cPB also in patients with ESRD. In a multivariate regression analysis the strongest predictors for CACS and cPB were MBF stress and age. Thus, the results of our study are in line with the current understanding of atherosclerosis as a systemic and diffuse disease. Interestingly, cPB was only moderately associated with and not predictive of MBF stress in the multivariate analysis. Instead, CACS turned out to be predictive for MBF stress . This is not surprising since CACS reflects the extent of coronary artery atherosclerosis which directly influences myocardial perfusion and the degree of atherosclerosis in the carotid arteries does not necessarily preclude coronary atherosclerosis with a similar extent. Our findings indicate that none of the imaging parameters is able to predict the others with sufficient accuracy in patients with ESRD. Importantly, CACS is not limited to intimal calcification but also superposed by smooth muscle cell mediated, metabolic media calcification in patients with ESRD. 22 This also could explain the only moderate correlation between CACS and cPB, since cPB is rather a marker of 'classical' atherosclerosis resulting from subintimal plaque development. Due to this different pathophysiology, the clinical value of CACS in CKD patients still remains controversial. 13 Interestingly, CACS did not correlate with age. This finding may be explained by the different pathophysiology of media calcification but also (at least partly) by the distinctly younger age of patients studied here as compared with a classical CAD population.
Another factor associated with MBF stress was the time on dialysis. This finding supports the assumption that the duration of renal replacement therapy has a strong impact on myocardial perfusion, mainly due to rarefication of small intramyocardial vessels. [23][24][25] We found impaired CFR in 20% of our patients, which is less than reported in other studies 33 and most probably due to the relatively young study cohort. Prior investigations have suggested that CKD is associated with abnormal coronary vasodilator function, 26 which results from multiple mechanisms including decreased capillary density, 27 vascular remodeling, 28 and endothelial dysfunction. 29,30 Early stage microvascular disease may be the reason for impaired MBF stress in our patient cohort. This can be assumed since global CFR was still preserved and significant epicardial CAD could be excluded by angiography in selected cases with mild to moderate regional ischemia. Thus, we interpret the data in that way that regional ischemia is most probably related to regional microvascular dysfunction, since technical image artifacts could be excluded. This can be missed by the assessment of global CFR.
Although preliminary results based on a rather small patient cohort should always be interpreted with care, our results indicate that the combination of low cPB and low CACS may have a high negative predictive value with regards to the exclusion of CAD. On the other hand the positive predictive value of CACS and cPB for CFR and MBF was low. Thus, it would be worth to evaluate in a larger study if the assessment of cPB and CACS may replace MPI in the evaluation prior to kidney transplantation listing in cases of normal findings. But on the other hand in this context importantly, our results indicate that cardiovascular event-free survival may be not predicted by cPB or CACS but only by CFR and MBF stress . These findings are in line with previous study results in distinctly larger ESRD patient cohorts already showing the prognostic value of CFR in this particular patient group, although in these studies a lower cutoff value for CFR than in our study was used. 4,31 Notably striking was the observation that CFR was independently associated with the current Hb value and was neither correlated with CACS or cPB nor with age or the time on dialysis. Increased MBF rest has been already observed in mild CKD and was correlated with the degree of GFR. 32 Anemia induces a compensatory increase in MBF rest via increased nitric oxide release 33 which naturally also affects CFR. Although in former investigations impaired CFR was highly prevalent in patients with CKD (4, 35; example Figure 6), our results suggest that MBF stress may be a better parameter for the early assessment of microvascular dysfunction. Variability of MBF rest in the individual person is widely known and is probably related to differences in myocardial work load. 35 Although corrected for cardiac work increased MBF rest due to other factors as metabolic oxygen demand or an enhanced activation of the sympathetic nervous and renin-angiotensin-aldosterone system 34 which is chronically activated in patients with ESRD cannot be ruled out. Therefore, when interpreting CFR, changes in MBF rest due to alterations in metabolic demand need to be considered. [36][37][38] Altogether, the results of our study raise the question if absolute MBF stress may be an earlier and more robust marker for CAD in patients with ESRD than CFR. This hypothesis is additionally supported by the finding that only MBF stress was significantly correlated with LVEF (rest) but not CFR. In former studies absolute MBF stress already shown to be more accurate than CFR in detecting CAD in non-CKD patients. 35,37 Interestingly, MBF rest was higher in patients without or minimal CACS and low cPB, a finding which is most probably caused by the higher frequency of females in these groups. Most studies report of higher MBF rest in females, although the mechanisms of this phenomenon remain uncertain. 38

Study Limitations
The small size of the patient cohort is certainly a limitation of our study. Furthermore, the study population was young and time on dialysis was relatively short. These factors may explain the relatively low number of pathological findings. Another limitation is the use of a non-ECG-triggered, low-dose CT for CACS although CACS on that basis is feasible. 39,40 Thus, a dedicated state-of-the-art CACS-CT was not performed additionally in order to avoid futile additional radiation exposure.

NEW KNOWLEDGE GAINED
Our results indicate that MBF stress is a marker of subclinical atherosclerosis in patients with ESRD, whereas CFR may miss early stages of vascular dysfunction in this patient cohort. Nevertheless, impaired CFR and MBF stress are predictors for CV event-free survival whereas cPB and CACS may be not.

CONCLUSIONS
Although CACS and cPB as parameters of coronary and systemic atherosclerosis are associated in patients with ESRD, only absolute MBF stress but not CFR correlates with CACS and cPB. Furthermore, MBF stress impairs with increasing time on renal replacement therapy. MBF rest and CFR are regularly influenced by actual hemoglobin values. Thus, MBF stress may serve as a robust parameter for the detection of microvascular disease in patients with ESRD than CFR.
Overall, our results indicate that quantitative MPI cannot be replaced by the assessment of CACS or cPB only. Nevertheless, there may be a good negative predictive value. CV event-free survival is associated with impaired CFR and MBF stress only.