OUP user menu

Impact of serum calcium and phosphate on coronary atherosclerosis detected by cardiac computed tomography

Sanghoon Shin, Kwang-Joon Kim, Hyuk-Jae Chang, Iksung Cho, Young Jin Kim, Byoung-Wook Choi, Yumie Rhee, Sung-Kil Lim, Woo-In Yang, Chi-Young Shim, Jong-Won Ha, Yangsoo Jang, Namsik Chung
DOI: http://dx.doi.org/10.1093/eurheartj/ehs152 2873-2881 First published online: 19 June 2012

Abstract

Aims High calcium (Ca), phosphate (P), and Ca–P product (CPP) are associated with cardiovascular disease in patients with chronic kidney disease. Whether this relationship persists in individuals with normal kidney function is not yet elucidated. We explored the relationship of serum Ca, P, and CPP to coronary atherosclerosis assessed by cardiac computed tomography angiography (cCTA) in participants with normal kidney function.

Methods and results This study included 7553 participants (52 ± 10 years, male 57%) with near-normal kidney function (estimated glomerular filtration rate > 60 mL/min/1.73 m2) who underwent cCTA. The relationship of Ca, P, and CPP to coronary atherosclerosis [coronary artery Ca score (CACS) >100 and the presence of coronary artery disease (CAD)] was evaluated. Higher Ca, P, and CPP were significantly associated with CACS > 100 continuously [adjusted odds ratio (OR) per mg/dL: Ca 1.21, P = 0.026; P 1.29, P < 0.001; CPP 1.03, P < 0.001]. However, they correlate only weakly with the presence of CAD (OR: Ca 1.17, P = 0.001; P 1.05, P = 0.173; CPP 1.01, P = 0.034). This discrepancy was because calcified or mixed plaque and non-calcified plaque (NCP) were included in CAD. A significant relationship was demonstrated between calcified or mixed plaque and Ca, P, and CPP (OR: Ca 1.20, P = 0.001; P 1.13, P = 0.003; CPP 1.02, P = 0.001), but not NCP.

Conclusion Elevated serum levels of Ca, P, and CPP are significantly associated with the presence of calcified coronary atherosclerotic plaque. It is unclear if there is a causal relationship. This relationship is thought to contribute to vascular calcification, but is less closely associated with NCP.

  • Atherosclerosis
  • Computed tomography
  • Calcium
  • Phosphate

Introduction

Serum calcium (Ca) and inorganic phosphorous (P) levels are tightly regulated within a narrow range in healthy individuals, and maintenance of normal serum Ca and P levels is a pre-requisite for various physiological processes, including bone formation, vascular function, several metabolic pathways, and intracellular signalling. Therefore, abnormalities in Ca and P homeostasis have been directly or indirectly linked to various skeletal, endocrine, and cardiovascular disorders.13 In clinical settings, abnormalities in Ca and P metabolism are most commonly found in individuals with impaired renal function. Several previous studies have shown that higher serum Ca and P levels are associated with unaccountably high rates of cardiovascular disease (CVD) in patients with chronic kidney disease (CKD),46 and recent studies have suggested that Ca–P homeostasis plays a role in the pathogenesis of atherosclerosis, especially in the coronary arteries.7,8

Major consequences of elevated serum Ca and P in patients with CKD are vascular calcification and increased risk of cardiovascular morbidity and mortality.9 In addition some previous studies suggested that even with normal kidney function, higher serum Ca is associated with increased cardiovascular morbidity. Large-scale epidemiological studies have shown that in middle-aged men, serum Ca levels are an independent, prospective risk factor for CVD,10,11 and high-normal serum Ca levels are associated with increased cardiovascular mortality.12 Similarly, recent studies also suggest that higher serum P is associated with abnormal vascular phenotypes, such as increased carotid intima-media thickness,13 arterial stiffness,14 and cardiovascular mortality.15 Reports that Ca supplementation increases cardiovascular risk, especially the risk of myocardial infarction16 have raised interest in the association between Ca, P, and CVD.

Coronary artery Ca score (CACS) is a good marker of atherosclerosis which represents the degree of atheromatous plaque burden,17,18 and coronary computed tomography angiography (cCTA) has achieved high diagnostic accuracy for detecting coronary artery disease (CAD).1921 Therefore, we investigated the relationship of serum Ca and P with coronary atherosclerosis using non-invasive cCTA, a useful method in predicting cardiovascular disease22,23 in participants with near-normal kidney function. Furthermore, we have also observed how the relationship between the two factors changes under the presence of various risk factors after considering the increase in the incidence of cardiovascular disease due to the risk factors such as diabetes.24

Methods

Study design and population

This observational retrospective single-centre study consisted of 8648 consecutive participants who underwent cCTA evaluation using 64-slice multi-detector computed tomography (MDCT) from January 2004 to April 2009 at Severance Hospital. Patients were referred for a variety of indications, including asymptomatic subject for a general health evaluation, evaluation of symptoms, signs of cardiac disease such as abnormal resting or stress electrocardiographic test, or asymptomatic patients with peripheral arterial disease, cerebrovascular disease, or multiple CAD risk factors. Symptoms included typical angina, atypical angina, dyspnea, and excessive fatigue. Participants were excluded if any one of the following criteria was met: (i) age < 30 years (n= 69), (ii) modification of diet in renal disease (MDRD) glomerular filtration rate (GFR) under 60 mL/min/1.73 m2 (n = 958), or (iii) insufficient medical records (n = 68). As a result, 7553 participants were finally included.

Multi-detector computed tomography protocol

Patients with initial heart rate higher than 65 beats/min before MDCT examination received a single oral dose of 50 mg metoprolol tartrate (Betaloc, Yuhan, Seoul, Korea) 1–2 h before CT examination unless β-adrenergic-blocking agents were contraindicated (overt heart failure or atrioventricular conduction abnormalities). Patients were scanned using a 64-slice CT scanner (Sensation 64; Siemens Medical Systems, Forchheim, Germany). Before the helical scan for CCTA, a non-enhanced prospective electrocardiogram (ECG)-gated scan was performed to measure CACS with the following parameters: rotation time of 330 ms, slice collimation of 0.6 mm, slice width of 3.0 mm, tube voltage of 120 kV, tube current of 50 mA, and table feed/scan of 18 mm. Afterward, cCTA was performed using retrospective ECG-gating with the following scan parameters: rotation time of 330 ms, slice collimation of 64 × 0.6 mm, tube voltage of 100–120 kV, tube current of 600–800 mA depending on patient size, table feed/scan of 3.8 mm, and pitch factor of 0.2. ECG-based tube current modulation was applied to 65% of the R–R interval. A real-time bolus-tracking technique was used to trigger the initiation of the scan. The total estimated average radiation dose for the multislice CT protocol (CACS + cCTA) was 10.9 ± 1.9 mSv. The average radiation doses were 1.7 ± 0.3 mSv for the CACS protocol and 8.8 ± 1.6 mSv for the CCTA protocol.

Contrast enhancement was achieved with 60 mL iopamidol (370 mg iodine/mL, Iopamiro; Bracco, Milan, Italy) injected at 5 mL/s, followed by an injection of 30 mL diluted contrast (contrast agent : saline = 3:7), and then 30 mL saline at 5 mL/s using a power injector (Envision CT; Medrad, Indianola, Pa) via an antecubital vein. The estimated volume CT dose index (CTDIvol) in milligrays (mGy) determined from the scanner's output parameters (kVp and mA per slice) was recorded for each patient. The product of CTDIvol and scanning length (dose–length product, mGy × cm) was calculated, and effective dose (ED, in millisieverts [mSv]) was estimated using a normalization factor for the adult chest (0.017 mSv × mGy−1× cm−1).

Measurement of outcome variables

Coronary CT angiograms were evaluated by two experienced cardiac radiologists (Y.J.K. and B.W.C., with 6 and 9 years of experience in cardiac CT, respectively) who were unaware of the clinical history of the patients. We selected two outcome variables: CAD, defined as presence of any plaque and CACS, measured using the scoring system previously described by Agatston et al.25

Plaque were defined as structures >1 mm2 within and/or adjacent to the vessel lumen, which could be clearly distinguished from the vessel lumen and surrounding pericardial tissue. Plaque occupied by calcified tissue for more than 50% of the plaque area (density >130 Hounsfield Unit in native scans) were classified as calcified plaque. Plaque with <50% Ca were classified as mixed calcified plaque, and plaque without any Ca were classified as NCP.26 The subjects with plaque were divided based on the plaque characteristics into two groups: individual with exclusively NCP and individual with calcified or mixed plaque. In addition, participants were further categorized based on the CACS in the following manner: CACS ≤ 100 and CACS > 100 as main categories; and CACS = 0, 0 < CACS ≤ 10, 10 < CACS ≤ 100, and CACS > 100 as subcategories. There were few individuals with CACS > 400 (3%); as a result, they were combined with those with CACS > 100.

Measurement of clinical variables and confounding variables

Medical history of hypertension, diabetes mellitus (DM), and smoking status were systematically acquired. Height, body weight, and blood pressure were measured during visits. Serum Ca, P, total cholesterol, triglyceride (TG), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), fasting plasma glucose, glycated haemoglobin, blood urea nitrogen, and serum creatinine level were measured after a minimum of 12-h fasting period on the day of CT scan as the part of the clinical work-up. Creatinine was measured using rate-blanked and compensated Jaffé method with calibration for automated systems (Roche Diagnostics, Basel, Switzerland). The level of kidney function was ascertained by estimated GFR calculated using the formula developed and validated in the MDRD study 27 as follows:Embedded Image

Serum C-reactive protein (CRP) was measured by latex agglutination with an auto-analyzer (Toshiba-200 FR, Tokyo, Japan). Diabetes was defined as treatment with hypoglycaemic agents or insulin, or fasting glucose ≥126 mg/dL. Hypertension was defined as systolic blood pressure ≥140 mmHg and/or diastolic blood pressure ≥90 mmHg or treatment with antihypertensive agents. Smoking history was considered present if patients currently smoked or smoked until 1 month before the study. Dyslipidaemia was defined as total cholesterol ≥240 mg/dL, LDL ≥130 mg/dL, HDL ≤40 mg/dL, TG ≥150 mg/dL and/or treatment with lipid lowering agents. Laboratory values for lipid parameters were obtained within 1 month prior to multi-slice CT examination.

Old age (≥60 years), male gender, dyslipidaemia, hypertension, and diabetes were considered as classical risk factors of cardiovascular diseases, and patients with 0 or 1 risk factors were classified as low-risk group, while those with 2 or more risk factors were classified as high-risk group.

Statistical analysis

Continuous variables are expressed as medians and interquartile range, whereas categorical variables are presented as absolute values and percentages. Differences between continuous variables were analysed by independent t-test or Mann-Whitney U-test and those between categorical variables by χ2 test. Variables that showed significant relationships in previous studies and those with a P < 0.3 on univariate analysis were defined as confounding variables related to the dependent and independent variables. These included age, male gender, hypertension, DM, body mass index (BMI), MDRD GFR, and proteinuria. The covariate-adjusted odds ratios (OR) and their 95% confidence intervals (CI) for each dependent variable (CAD and CACS > 100) were calculated. All analyses were performed using the Statistical Package for the Social Sciences (SPSS, version 18; SPSS, Chicago, IL, USA).

Results

The clinical characteristics of participants are shown in Table 1. The median age was 52 years, and 57% of the participants were male. Hypertension, DM, and dyslipidaemia was present in 30, 11, and 53% of the patients, respectively, and 21% were smokers. The medians and interquartile ranges of serum Ca, P, and CPP were 9.00 [8.70, 9.40] mg/dL, 3.70 [3.30, 4.10] mg/dL, and 33.3 [29.5, 37.8] mg2/dL2, respectively. The presence of plaque was confirmed by cCTA in 32.3% of all patients: when sub-classified, 25.2% showed calcified plaque, and 7.1% showed NCP. Coronary artery Ca score of 0 was seen in 73.9% of the patients and 8% had CACS > 100 (Figure 1).

View this table:
Table 1

Baseline characteristics

CACS ≤ 100, (n = 6950)CACS > 100, (n = 603)P-valueNo CAD (n = 5111)CAD (n = 2442)P-value
Age, years50 (44, 58)63 (57, 69)<0.00148 (43, 56)58 (51, 65)<0.001
Male3864 (56)446 (74)<0.0012630 (52)1680 (69)<0.001
Asymptomatic4179 (60)187 (31)<0.0013380 (66)986 (40)<0.001
BMI, kg/m224.1 (22.2, 26.0)24.4 (22.7, 26.5)<0.00123.9 (22.0, 25.8)24.5 (22.8, 26.4)<0.001
Hypertension1951 (28)301 (51)<0.0011227 (24)1025 (43)<0.001
DM688 (10)165 (28)<0.001359 (7)494 (20)<0.001
Smoking1484 (22)104 (18)0.0231108 (22)480 (20)0.070
Dyslipidaemia3603 (52)349 (59)0.0022491 (49)1461 (61)<0.001
T-chol, mg/dL195 (172, 218)190 (166, 213)<0.001194 (172, 217)194 (170, 219)0.619
HDL, mg/dL52 (44, 61)49 (42, 57)<0.00153 (45, 62)49 (43, 58)<0.001
LDL, mg/dL111 (93, 133)113 (93, 135)0.593110 (92, 130)115 (96, 137)<0.001
TG, mg/dL113 (79, 164)126 (91, 184)<0.001109 (75, 158)128 (91, 182)<0.001
High-sensitivity CRP0.10 (0.01, 0.31)0.29 (0.06, 0.82)<0.0010.01 (0.01, 0.24)0.20 (0.01, 0.64)<0.001
Serum Ca, mg/dL9.00 (8.70, 9.40)9.10 (8.80, 9.50)<0.0019.00 (8.60, 9.30)9.10 (8.80, 9.50)<0.001
Serum P, mg/dL3.70 (3.30, 4.10)3.70 (3.30, 4.10)0.7323.70 (3.30, 4.10)3.70 (3.30, 4.10)0.057
CPP, mg2/dL233.3 (29.4, 37.8)33.8 (30.4, 37.8)0.03233.1 (29.4, 37.8)33.6 (29.9, 37.8)0.108
Serum chromium, mg/dL1.00 (0.90, 1.10)1.00 (0.90, 1.10)0.5971.00 (0.90, 1.10)1.00 (0.90, 1.10)0.018
Albumin creatinine ratio6.10 (4.10, 9.80)7.20 (4.98, 7.20)<0.0016.10 (6.00, 9.50)6.80 (4.40, 12.6)<0.001
GFR, mL/min/1.73 m276 (69, 84)76 (69, 86)0.28576 (69, 83)76 (69, 85)0.004
Proteinuria420 (6.6)62 (11.3)<0.001270 (5.8)212 (9.6)<0.001
  • Data are expressed as median [interquartile] or n (%). BMI, body mass index; CACS, coronary artery calcium score; CAD, coronary artery disease; Ca, calcium; CPP, calcium –phosphate product; DM, diabetes mellitus; GFR, glomerular filtration rate; HDL, high density lipoprotein; hsCRP, high-sensitivity C-reactive protein; LDL, low density lipoprotein; P, phosphate; TG, triglyceride.

Figure 1

The distribution of coronary artery disease and coronary artery calcium score.

As seen in Table 2, age, male gender, BMI, hypertension, diabetes, dyslipidaemia, and proteinuria demonstrated association with both CACS > 100 and the presence of CAD on univariate analysis. On the contrary, while serum Ca level was associated with both CACS > 100 and the presence of CAD (OR = 1.41, 95% CI: 1.23–1.61, P < 0.001; OR = 1.49, 95% CI: 1.38–1.61, P < 0.001, respectively), serum P level was negatively associated with the presence of CAD, and CPP was not significantly associated with either CACS > 100 or presence of CAD (Table 2). However, after adjusting for all confounding factors, including age, male gender, BMI, hypertension, diabetes, smoking status, dyslipidaemia and proteinuria, serum Ca, P, and CPP level were independent risk factors of CACS > 100, a marker of coronary atherosclerosis (OR: 1.22, 95% CI: 1.03–1.43, P = 0.019; OR: 1.28, 95% CI: 1.12–1.45, P < 0.001; and OR: 1.03, 95% CI: 1.02–1.05, P < 0.001, respectively), and the association was more strongly seen in higher quartiles. However, when compared with CACS > 100, the presence of CAD, another marker of coronary atherosclerosis, was more weakly associated with serum Ca, P, and CPP levels (OR: 1.19, 95% CI: 1.08–1.31, P < 0.001; OR: 1.05, 95% CI: 0.98–1.13, P = 0.242; and OR: 1.01, 95% CI: 1.00–1.02, P = 0.034, respectively; Table 3).

View this table:
Table 2

Univariate correlations of calcium, phosphate, and calcium–phosphate product

CACSCAD
OR (95% CI)P-valueOR (95% CI)P-value
Age, years1.13 (1.12–1.15)<0.0011.10 (1.09–1.11)<0.001
Male gender2.28 (1.89–2.75)<0.0012.08 (1.88–2.30)<0.001
BMI, kg/m21.05 (1.03–1.08)<0.0011.07 (1.05–1.09)<0.001
Hypertension2.64 (2.22–3.12)<0.0012.36 (2.13–2.62)<0.001
DM3.45 (2.83–4.19)<0.0013.36 (2.90–3.88)<0.001
Current smoking0.78 (0.62–0.97)0.0230.90 (0.79–1.01)0.070
Dyslipidaemia1.31 (1.11–1.56)0.0021.61 (1.46–1.78)<0.001
Creatinine, mg/dL1.07 (0.65–1.75)0.7911.28 (0.96–1.71)0.090
High-sensitivity CRP, IU1.06 (1.04–1.09)<0.0011.12 (1.08–1.16)<0.001
Proteinuria1.79 (1.35–2.37)<0.0011.73 (1.43–2.08)<0.001
Ca, mg/dL1.41 (1.23–1.61)<0.0011.49 (1.38–1.61)<0.001
P, mg/dL0.96 (0.87–1.07)0.4470.86 (0.83–0.94)<0.001
CPP, mg2/dL21.01 (1.00–1.02)0.2911.00 (0.99–1.01)0.690
  • CRP, c-reactive protein.

View this table:
Table 3

Association between serum calcium, phosphate, and calcium–phosphate product with coronary atherosclerosis

CACS>100CAD
OR (95% CI)P-valueOR (95% CI)P-value
Per 1 mg/dL of Ca1.22 (1.03–1.43)0.0191.19 (1.08–1.31)<0.001
By Ca quartile
 I (∼8.6 mg/dL)11
 II (8.7–8.9 mg/dL)1.19 (0.86–1.64)0.3081.21 (1.01–1.44)0.040
 III (9.0–9.3 mg/dL)1.68 (1.20–2.33)0.0021.33 (1.10–1.61)0.003
 IV (9.4 mg/dL)1.43 (1.04–1.96)0.0291.35 (1.13–1.61)0.001
Per 1 mg/dL of P1.28 (1.12–1.45)<0.0011.05 (0.97–1.13)0.242
By P quartile
 I (∼3.2 mg/dL)11
 II (3.3–3.6 mg/dL)1.12 (0.83–1.52)0.4551.03 (0.86–1.24)0.721
 III (3.7∼4.0 mg/dL)1.41 (1.07–1.85)0.0151.28 (1.08–1.51)0.004
 IV (4.1 mg/dL)1.76 (1.29–2.38)<0.0011.17 (0.98–1.40)0.079
Per 1 mg2/dL2 of CPP1.03 (1.02–1.05)<0.0011.01 (1.00–1.02)0.045
By CPP quartile
 I (∼29.4 mg2/dL2)11
 II (29.5–33.2 mg2/dL2)1.37 (1.02–1.85)0.0361.12 (0.94–1.33)0.205
 III (33.3–37.7 mg2/dL2)1.68 (1.25–2.27)0.0011.31 (1.10–1.56)0.002
 IV (37.5 mg2/dL2)1.92 (1.41–2.62)<0.0011.28 (1.07–1.53)0.006
  • Adjusted for age, male gender, BMI, hypertension, DM, smoking, dyslipidaemia and proteinuria. OR, odds ratio; CI, confidence interval.

To elucidate the reason for the weak associations between serum Ca, P, and CPP, and the presence of CAD compared with those of CACS > 100, the presence of CAD was further divided into NCP and calcified or mixed plaque, and each group's association to serum Ca, P, and CPP was evaluated. Interestingly, serum Ca, P, and CPP are independent risk factors for calcified or mixed plaque (OR: 1.22, 95% CI: 1.09–1.35, P < 0.001; OR: 1.12, 95% CI: 1.03–1.22, P = 0.006; OR: 1.02, 95% CI: 1.01–1.03, P < 0.001, respectively), but their association with NCP was close to zero (OR: 1.07, 95% CI 0.91–1.25, P = 0.423; OR: 0.91, 95% CI: 0.80–1.04, P = 0.155; and OR: 1.00, 95% CI: 0.98–1.01, P = 0.205, respectively; Figure 2).

Figure 2

The association between calcium, phosphate, and calcium–phosphate product with the presence of coronary artery disease divided with calcified or mixed plaque and non-calcified plaque. The reason behind a relatively weak association between serum calcium, phosphate, and calcium–phosphate product; and the presence of coronary artery disease is inclusion of the group of non-calcified plaque in coronary artery disease.

When the groups are sub-divided into various risk factors including age ≥60, gender, presence of symptom, hypertension, and diabetes, it can be seen that Ca, P, and CPP show associations with coronary atherosclerosis (see Supplementary material online, Table S1), although the significance of each factor may differ. Old age (≥60 years), male gender, dyslipidaemia, hypertension, and diabetes were considered as classical risk factors of cardiovascular diseases, and patients with 0 or 1 risk factors were classified as low-risk group, while those with 2 or more risk factors were classified as high-risk group. When subdivided into risk groups based on classical risk factors, the association between serum Ca, P, CPP, and coronary atherosclerosis was not seen in low-risk group, while it was seen in high-risk group. Upon further examination, Ca level was found to be significantly associated with CAD and calcified or mixed plaques in high-risk group—this association was not seen in NCPs of the same patient group (Table 4).

View this table:
Table 4

Association between calcium, phosphate, calcium–phosphate product, and coronary atherosclerosis based on classical risk factors of cardiovascular disease

CACS > 100CADCalcified or mixed plaqueNCP
OR (95% CI)P-valueOR (95% CI)P-valueOR (95% CI)P-valueOR (95% CI)P-value
Low risk
 Per 1 mg/dL of Ca1.25 (0.82–1.90)0.3091.21 (1.01–1.44)0.0351.26 (1.02–1.55)0.0321.05 (0.79–1.40)0.727
 Per 1 mg/dL of P1.27 (0.94–1.72)0.1240.97 (0.85–1.11)0.6461.03 (0.88–1.21)0.7020.90 (0.73–1.10)0.302
 Per 1 mg2/dL2 of CPP1.03 (0.10–1.07)0.0861.00 (0.99–1.02)0.9961.01 (0.99–1.03)0.3500.99 (0.96–1.01)0.987
High risk
 Per 1 mg/dL of Ca1.18 (0.99–1.41)0.0661.17 (1.04–1.31)0.0081.18 (1.05–1.33)0.0071.01 (0.84–1.22)0.899
 Per 1 mg/dL of P1.27 (1.10–1.46)0.0011.08 (0.99–1.19)0.0931.16 (1.05–1.28)0.0040.89 (0.76–1.05)0.156
 Per 1 mg2/dL2 of CPP1.03 (1.01–1.05)<0.0011.01 (1.00–1.02)0.0221.02 (1.01–1.03)0.0010.99 (0.97–1.01)0.176
  • OR, odds ratio; CI, confidence interval.

  • Classical risk factor: old age (≥ 60 years), male gender, hypertension, diabetes, dyslipidaemia.

  • Low risk, zero or one risk factor; high risk, more than two risk factors.

Discussion

Four clinical implications were demonstrated through this study. First, serum Ca, P, and CPP level are independent risk factors of coronary atherosclerosis, even in participants with near-normal kidney function. Second, this relationship is stronger in coronary calcification and is more pronounced in participants in the higher quartiles of serum Ca, P, and CPP. Third, the reason behind a relatively weak association between serum Ca, P, and CPP and the presence of CAD is inclusion of the group of NCP in CAD. The serum markers are independent risk factors for calcified or mixed plaque, similar to CACS. Lastly, while association between serum Ca, P, CPP, and coronary atherosclerosis was observed in high-risk group with 2 or more classical risk factors, no such association was found in low-risk group with 0 or 1 risk factor.

Serum calcium and phosphate induces coronary atherosclerosis by increasing coronary calcification in participants with near-normal kidney function

The relationship between serum Ca, P, and CPP and CVD was first studied in patients with renal failure. It is well-known that mineral regulation is impaired in patients with reduced kidney function, leading to Ca deposits that cause vascular calcification and increased risk of cardiovascular events. However, evidence of serum Ca and P levels affecting the incidence of cardiovascular events in patients with near-normal kidney function is lacking. Leifsson et al.12 monitored 21 131 patients under 50 years old for 10.8 years, and have concluded that the mortality rate is higher in patients with high serum Ca, even within the normal range, and that cardiovascular disease was the major cause of mortality. In addition, Dihngra et al.15 reported a 1.55-fold increased risk of CVD in patients in the highest serum P quartile among 3368 participants of the Framingham Offspring study over 16.1 years: Larsson et al.28 showed that circulating Ca, P, and CPP are related to total, cardiovascular, and non-cardiovascular mortality rates in 2176 cohort participants over 29.8 years. Such longitudinal studies have consistently pointed to the association between serum Ca, P, and CPP and cardiovascular events in participants with near-normal kidney function, but its pathophysiology is still not clearly understood.

Roberts et al.29 performed autopsies on 18 participants with chronic hypercalcaemia and found cardiac calcific deposits and progression of coronary atherosclerosis, demonstrating that hypercalcaemia in itself can be a risk for coronary atherosclerosis. In addition, increased cardiovascular mortality is seen in hyperthyroidism or chronic kidney disease, which is related to dysregulation of Ca metabolism, suggesting an association between Ca and cardiovascular mortality. Nonetheless, a causal mechanism between Ca and atherosclerosis has not been clearly found. Several articles assume that higher serum Ca affects the incidence of coronary atherosclerosis through its relationship with various metabolic diseases and other risk factors known to be related to the atherosclerotic process.2,10

Elevated serum P is also known to increase cardiovascular events and mortality in patients with CKD.30 In addition, it is associated with mortality and recurrent events in CAD patients,30 and is known to increase the incidence of CVD in communities of participants with near-normal kidney function.15 The mechanism behind serum P inducing atherosclerosis has yet to be elucidated, but several possible explanations were suggested. First, high phosphorus levels were shown to inhibit 1,25-dihydroxyvitamin D synthesis.31 Lower levels of 1,25-dihydroxyvitamin D are hypothesized to increase coronary calcification.32 Additional studies are required to evaluate whether lower 1,25-dihydroxyvitamin D levels contribute to greater calcification on coronary CT in individuals with higher serum P. Secondly, higher serum P levels may directly promote vascular injury. Investigators have reported that higher P levels increase the propensity of mineral deposition in vascular smooth muscle cells in vitro,33 which can be partly explained by increased osteopontin expression or decreased fibroblast growth factor 23.34,35 Serum P may also directly increase vascular calcification, especially when levels of CPP are high, as seen in patients with CKD36 in most observational studies. However, this study showed that serum P is associated with increased coronary calcification also in patients with near-normal kidney function.

Serum calcium and phosphate do not increase the frequency of non-calcified plaque

Serum Ca, P, and CPP were independent risk factors for coronary atherosclerosis in patients with near-normal kidney function. However, while these serum markers show a consistently significant relationship with CACS, a surrogate marker of coronary atherosclerosis on CT, they are relatively weakly associated with the presence of CAD. We hypothesized that such discrepancy was because of the mechanism through which serum Ca, P, and CPP affect coronary atherosclerosis—namely, vascular calcification. Non-calcified plaque, which is included in CAD, may be thought of as a factor weakening the association between the serum markers and CAD. Therefore, the presence of CAD was further subdivided into NCP and calcified or mixed plaque, and similar to CACS, calcified or mixed plaque showed strong associations with serum Ca, P, and CPP, while NCP did not show such associations.

These findings may add to hypotheses explaining the conflicting results on the effect of Ca supplements on the incidence of cardiovascular events. Bolland et al.16 concluded that Ca supplementation increases the incidence of cardiovascular events, especially myocardial infarction, whereas Lewis et al.37 argued that Ca supplements do not increase mortality in patients with atherosclerotic cardiovascular events. Our study may explain such contradictory findings: while higher serum Ca, P, and CPP increase vascular calcification and coronary atherosclerosis progression, causing a modest increase in mortality, they are independent of NCPs, which are more strongly associated with acute coronary syndrome and increased mortality.

The association between calcium, phosphate, calcium–phosphate product, and coronary atherosclerosis based on classical risk factors of cardiovascular disease

We found that association between serum Ca, P, CPP, and coronary atherosclerosis is observed in high-risk group, but not in low-risk group. This implies that an isolated increase in Ca or P level does not lead to increased risk of coronary atherosclerosis across all patients. In other words, high Ca level may act as a trigger of coronary atherosclerosis under the presence of several risk factors. Indeed, Ca content was significantly greater among non-culprit lesions of patients with diabetes, who are at higher risk of cardiovascular diseases.24,35 In diabetes, advanced glycation end products promote mineralization,38 and the incidence of coronary atherosclerosis may increase in the setting of high Ca level. Similarly, in the patients with high risk of systemic inflammation such as patients with dyslipidaemia or patients older than 55 years, high Ca level may possibly accelerate vascular mineralization.39 These results suggest that the presence of risk factors may affect the result of studies on association between Ca supplements and cardiovascular diseases, which has recently undergone controversies. Furthermore, the result also suggest that when determining the management of patients with osteoporosis, to whom Ca supplementation is essential, analysis of risk factors must be made in addition to simple serum Ca level. Prohibiting Ca supplemntation, based on the claim that it may increase Ca level, is thought to be inappropriate for osteoporotic patients with no cardiovascular risk factors.

Limitations

Several limitations to the study should be acknowledged. First, this is a retrospective, cross-sectional study performed at a single centre, and the patients had variable indications for MDCT, which makes it uncertain whether the results will be equally applicable to the general clinical practice. Second, parathormone levels and vitamin D levels, which could also be associated with adverse outcomes in renal failure, were not measured. However, the serum level of these hormones do not start to change until the estimated glomerular filtration rate falls below 60–70 mL/min/1.73 m2, and we excluded all patients with such degree of renal insufficiency.40 Third, we cannot exclude the possibility that our findings were influenced by dietary intake of P or Ca supplements, and that therefore, dietary habits or medications may confound the association between serum Ca and P and the results of coronary CT.

There are also several points unique to this study. First, this study examined patients of Asian ethnicity only. According to Nasir et al.,41 the frequency of CACS > 100 in Asian populations is low compared with Caucasians, African-Americans, and Hispanics. Compared with other similar studies performed in Western nations, this study is novel in that only participants of Asian ethnicity were studied. In addition, this study enrolled nearly 8000 patients exclusively with near-normal kidney function and obtained large-scale CT data: the finding that serum Ca, P, and CPP were independent risk factors of coronary atherosclerosis confirmed by coronary CT scan is meaningful. Furthermore, within coronary atherosclerosis, Ca, P, and CPP affect relatively stable calcified or mixed plaques only, and are unrelated to NCP. Such findings may contribute to additional hypotheses to explain the conflicting results of studies examining the effect of Ca supplements on cardiovascular events.

Whether a strict control of serum Ca or P level is necessary for patients with normal kidney function requires a careful approach. Previous large-scale cohort studies have also demonstrated associations between serum Ca and P and CVD12,15; however, unlike in CKD patients, there is no clear evidence suggesting benefits of serum mineral control in those with normal kidney function. Furthermore, there is no established target range for serum mineral levels, and evidence demonstrating hazards of such supplements is lacking compared with their clear benefits on osteoporosis. For further studies, a longitudinal study monitoring the change in the atherosclerosis of the patients with high serum Ca and P using cCTA seems plausible, as well as a follow-up cohort study observing the relationship serum Ca and P distribution and the differences between initial and follow-up cCTA. In addition, these studies must analyse the effect of Ca supplementation on cardiovascular disease after stratifying the risk factors of the disease.

Conclusions

Elevated serum levels of Ca, P, and CPP are significantly associated with the presence of calcified coronary atherosclerotic plaque. It is unclear if there is a causal relationship. This relationship is thought to contribute to vascular calcification, but is less closely associated with the formation of NCP.

Funding

Technology Innovation Program (Industrial Strategic Technology Development Program, 10038419, Intelligent Image Diagnosis and Therapy-support System) granted from Ministry of Knowledge Economy (MKE, Korea)

Conflict of interest: none declared.

Footnotes

  • These authors contributed equally to this work as first authors.

References

View Abstract