Introduction
In patients with end-stage chronic kidney disease (CKD) requiring dialysis, cardiovascular mortality is 5 times higher than in the general population [
1]. Furthermore, studies have demonstrated that even early stages of CKD, which are more prevalent than end-stage CKD requiring dialysis [
2], are associated with elevated risk of cardiovascular disease and mortality [
3,
4]. Therefore, accurate detection of cardiovascular disease in patients with early stages of CKD is important for accurate risk stratification.
The coronary calcium score (CCS) measured by computed tomography can predict the presence of significant coronary artery disease (CAD) in the general population [
5‐
10]. While end-stage CKD is associated with elevated CCS as compared with the general population, several reports have demonstrated conflicting results in the correlation between CCS and the presence of significant CAD [
10‐
13]. Moreover, the role of CCS to detect CAD in patients with moderate CKD, who are at risk for future cardiovascular events, is unknown. Therefore the aim of the current study was to evaluate the predictive role of CCS for diagnosing CAD by computed tomography coronary angiography (CTA) in patients with moderate CKD compared with patients without significant CKD.
Methods
Study population
The study population consisted of 704 patients who underwent CCS and CTA assessment for suspected CAD. Patients were enrolled at the Leiden University Medical Center. Exclusion criteria included cardiac arrhythmias, severe renal insufficiency (defined as an eGFR <30 mL/min/m2), known hypersensitivity to iodine contrast media and pregnancy.
Classification of moderate chronic kidney disease
Serum creatinine levels were used to assess the eGFR calculated with the Modified Diet in Renal Disease equation [
14]. In order to prevent contrast-induced renal dysfunction affecting the analysis, only serum creatinine levels obtained prior to the CTA examination (up to 180 days prior to CTA) were used. Of note, patients with suspected acute renal failure (defined by an increase in serum creatinine of ≥0.5 mg/dl in <2 weeks or an increase of >20 % over baseline if baseline serum creatinine was ≥2.5 mg/dl) were excluded.
Patients were stratified into those with moderate CKD and those without significant CKD. The definition of moderate CKD was based on the recommendation from the National Kidney Foundation [
15] using a value of eGFR between 30 and 59 mL/min/1.73m
2. Patients with eGFR ≥60 mL/min/1.73m
2 were considered to have no significant CKD. The mean duration between renal function assessment and CTA was 37 ± 14 days.
Computed tomography coronary angiography protocol
Examinations of CCS and CTA were performed using a 64-row (Aquillion64, Toshiba Medical Systems, Tokyo Japan) computed tomography scanner. Descriptions of scan parameters for CCS and CTA assessment have been published previously [
16,
17].
Data analysis
Post-processing of the CCS and CTA was performed on dedicated workstations (Vitrea2, Vital Images, Minneapolis, Minnesota, USA). The CCS was calculated using the Agatston method and patients were stratified as CCS 0, CCS 1 to 399 and CCS ≥ 400. Coronary anatomy was assessed in a standardised method by dividing the coronary arteries into 17 segments [
18]. All CTA were interpreted by two experienced cardiologists blinded to the results of the CCS and eGFR. Classification of CTA results was made between non-obstructive and obstructive CAD using a luminal narrowing ≥50 % as a threshold for obstructive CAD lesions. In addition, the number of segments (among the 17 segments) and vessels (among the 3 coronaries) involved in each category of CAD (presence of CAD, non-obstructive CAD and obstructive CAD) was measured.
Statistical analysis
Continuous variables are presented as mean ± standard deviation and compared using either Student’s t or Wilcoxon’s rank-sum test as appropriate. Categorical data are presented as frequencies and percentages and compared using the chi-square or Fisher’s exact test. Receiver-operator characteristic (ROC) curve analysis was performed to determine the value of CCS in diagnosing obstructive CAD. The optimal cut-off value was defined as the maximised value for the sum of sensitivity and specificity. In addition, the CCS values to diagnose obstructive CAD for a predefined 1) sensitivity of 80 % and 2) specificity of 80 % in patients with moderate CKD and without significant CKD were evaluated. All statistical analyses were performed using the statistical package SPSS for Windows (Version 15.0, SPSS, Chicago, USA). A p value <0.05 was considered to be statistically significant.
Discussion
The present report demonstrates that patients with moderate CKD had a higher prevalence, more diffuse and greater extent of coronary calcium and CAD than patients without significant CKD. Moreover, CCS predicted the presence of obstructive CAD in both patients with moderate CKD and patients without significant CKD. Importantly, the optimal cut-off value of CCS for predicting the presence of obstructive CAD was higher in patients with moderate CKD than patients without significant CKD.
Recent studies using computed tomography demonstrated that patients with early stages of CKD had a higher prevalence of CCS and CAD than patients without CKD [
19,
20]. This is in concordance with the present results, demonstrating a higher prevalence, more diffuse and greater extent of coronary calcium and CAD in patients with moderate CKD than patients without significant CKD. Accordingly, these results provided further supporting evidence for the relationship between elevated cardiovascular risk and moderate CKD [
3,
4]. Interestingly, the increased prevalence of CAD in patients with moderate CKD was mainly contributed by the presence of non-obstructive CAD. The observation may partially be explained by the thickening and calcification in the media layer, instead of the intima layer, of the coronaries in patients with moderate CKD, a phenomenon that is classically described as Mönckeberg’s calcification or medial calcinosis [
21,
22]. Therefore, in patients with moderate CKD, the elevated CCS located in the medial layer may be associated with non-obstructive coronary atherosclerosis, rather than significant luminal obstruction [
23].
Although the coronary calcification and degree of CAD differed between patients with and without CKD, a close relation between CCS and presence of obstructive CAD has been shown in the general population [
24,
25] as well as in patients with end-stage CKD [
10‐
12]. The present results further showed that high CCS was associated with higher prevalence, more diffuse and greater extent of CAD in both patients with moderate CKD and patients without significant CKD.
In addition to the significant relation with CAD, previous reports have demonstrated that the use of CCS can predict the presence of obstructive CAD in the general population [
6,
7]. Moreover, few studies have evaluated the predictive role of CCS for obstructive CAD in patients with end-stage CKD. In the report by Sharples evaluating 18 patients with CKD requiring dialysis, the presence of CCS was not correlated with obstructive CAD diagnosed by coronary angiogram [
13]. Conversely, Fujimoto et al. demonstrated a CCS value of 1000 provided a sensitivity of 68 % and specificity of 69 % to diagnose obstructive CAD in 76 patients with CKD requiring dialysis [
10]. In addition, Robinson and colleagues showed a CCS value of 400 provided a sensitivity of 86 % and specificity of 83 % in diagnosing obstructive CAD in 37 patients with CKD (17 of them requiring dialysis) [
12]. While these studies mainly evaluated patients with end-stage CKD, none of them explored the role of CCS to predict obstructive CAD specifically in patients with moderate CKD, which is more common in the general population [
2]. Nonetheless, patients with moderate CKD have a greater cardiovascular risk than the general population, highlighting the need for appropriate strategies to detect obstructive CAD [
3,
4]. The results of the present study have thus confirmed the predictive role of CCS in diagnosing obstructive CAD in patients with moderate CKD. However, the optimal cut-off value of CCS was 2.8 fold higher in patients with moderate CKD than patients without significant CKD. For the same sensitivity and specificity to diagnose obstructive CAD, the value of CCS was consistently higher in patients with moderate CKD than patients without significant CKD (Fig.
2). The reason for this observation has not been fully elucidated, but may possibly be explained by the different morphology and distribution of coronary calcification in patients with and without CKD. Patients with CKD appear to have more diffuse calcified plaque burden that tends to be located in the media wall rather than protruding into the lumen [
26]. Therefore, while the relation between elevated CCS and obstructive CAD existed in patients with moderate CKD, the degree of this relationship is different as compared with patients without moderate CKD. Accordingly, the optimal cut-off value of CCS to predict the presence of obstructive CAD should be different in patients with moderate CKD as compared with patients without significant CKD. However, the exact value of CCS for this purpose requires further evaluation based on larger study populations.
Limitation
The present study consisted of patients with moderate CKD and the predictive role of CCS to diagnose obstructive CAD in patients with severe CKD (eGFR < 30 mL/min/1.73m
2) cannot be extrapolated. Similar to previous evaluations exploring the CTA results of patients with and without CKD, the present study confirmed that patients with moderate CKD were older as compared with patients without significant CKD [
17]. However, due to the small population, the independent association between moderate CKD and high CCS/significant CAD could not be evaluated. Although CCS is proven to be a valuable tool to detect coronary calcification, it is unable to differentiate media calcification, which is common in patients with CKD, from intima calcification. Moreover, whether the use of computed tomography fractional flow reserve will provide additional diagnostic information in patients with CKD would require further studies [
27].
Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.