Serum Calcification Propensity and Coronary Artery Calcification Among Patients With CKD: The CRIC (Chronic Renal Insufficiency Cohort) Study

Abstract

Rationale & Objective:

Coronary artery calcification (CAC) is prevalent among patients with CKD and increases risks of cardiovascular disease (CVD) events and mortality. We hypothesized that a novel serum measure of calcification propensity is associated with CAC among patients with CKD stages 2–4.

Study Design:

Prospective cohort study.

Setting & Participants:

Participants from the Chronic Renal Insufficiency Cohort (CRIC) Study with baseline (n=1274) and follow-up (n=780) CAC measurements.

Predictors:

Calcification propensity, quantified as the transformation time (T₅₀) from primary to secondary calciprotein particles, with lower T₅₀ corresponding to higher calcification propensity. Covariates included age, sex, race/ethnicity, clinical site, estimated glomerular filtration rate, proteinuria, diabetes, systolic blood pressure, number of antihypertensive medications, current smoking, history of CVD, total cholesterol, and use of statin medications.

Outcomes:

CAC prevalence, severity, incidence, and progression.

Analytical Approach:

Multivariable-adjusted generalized linear models.

Results:

At baseline, 824 (65%) participants had prevalent CAC. After multivariable adjustment, T₅₀ was not associated with prevalence of CAC but was significantly associated with greater CAC severity among participants with prevalent CAC: one standard deviation (SD) lower T₅₀ was associated with 21% (95% confidence interval [CI], 6% to 38%) greater CAC severity. Among 780 participants followed up an average of 3 years later, 65 (20%) without baseline CAC developed incident CAC while 89 (19%) with baseline CAC had progression, defined as an annual increase of ≥100 Agatston units. After multivariable adjustment, T₅₀ was not associated with incident CAC but was significantly associated with CAC progression: one SD lower T₅₀ was associated with 28% (95% CI, 7% to 53%) higher risk of CAC progression.

Limitations:

Potential selection bias in follow-up analyses; inability to distinguish intimal from medial calcification.

Conclusions:

Among patients with CKD stages 2–4, higher serum calcification propensity is associated with more severe CAC and progression of CAC.

Introduction

Cardiovascular disease (CVD) is the leading cause of death among patients with chronic kidney disease (CKD) and is a major public health challenge.^1,2 Vascular calcification is common in CKD and is one mechanism by which CVD risk is increased in patients with CKD.³ In addition to developing medial calcification, which is associated with increased arterial stiffness⁴ and heart failure,⁵ patients with CKD also develop intimal calcification, indicative of atherosclerosis. Both types of calcification can contribute to the coronary artery calcium (CAC) score. The presence^6,7 and progression^8,9 of CAC are strongly associated with CVD in the general population and previous studies have shown that reduced kidney function is associated with more severe calcification¹⁰ and more rapid CAC progression.^3,11 Among patients with CKD stages 2–4, the CAC score independently predicts risks of CVD and all-cause mortality.¹²

The gold standard for quantifying vascular calcification is computed tomography (CT), but radiation exposure limits the utility of longitudinal CT measurements of calcification in clinical practice. A novel in vitro assay quantifies the propensity for calcification in serum by evaluating the transformation time (T₅₀) from primary to secondary calciprotein particles when challenged with additional calcium and phosphate.¹³ Primary calciprotein particles are amorphous accumulations of calcium and phosphate. Their transformation to secondary calciprotein particles, composed of crystalline calcium phosphate, may provide information about the status of the humoral calcification-regulating system.¹⁴ Under normal homeostatic conditions, calcification promoters (e.g. calcium and phosphate) and inhibitors (e.g. albumin, fetuin-A, magnesium, and pyrophosphate) are balanced such that vascular calcification does not occur. The T₅₀ test represents a composite, functional measure of this promoter-inhibitor balance. Higher calcification propensity (denoted by lower T₅₀) may reflect decreased inhibitory capacity to remove excess mineral from the circulation and has previously been associated with cardiovascular and all-cause mortality among individuals with advanced CKD, kidney transplant recipients, and patients with kidney failure undergoing hemodialysis.^15–18 However, associations with CVD, including CAC, in patients with mild-to-moderate CKD are unknown.

The Chronic Renal Insufficiency Cohort (CRIC) Study provides a unique opportunity to examine the associations of T₅₀ with the presence and progression of CAC among a diverse sample of patients with CKD stages 2–4. We tested the hypothesis that low levels of T₅₀ would be associated with prevalent and incident CAC among patients with CKD stages 2–4.

Methods

Study Design and Participants

The CRIC Study is a prospective cohort study of a racially and ethnically diverse group of men and women aged 21 to 74 years with mild-to-moderate CKD (estimated glomerular filtration rate [eGFR] entry criteria 20 to 70 mL/min/1.73 m²). A total of 3,939 participants were enrolled from 7 centers in the US between May 2003 and August 2008.¹⁹ Patients with cirrhosis, HIV infection, polycystic kidney disease, or renal cell carcinoma; those receiving dialysis or an organ transplant; or those taking immunosuppressive medications were excluded. Participants with a history of coronary artery revascularization did not undergo CT examination. The study was approved by the institutional review boards from each clinical center, and all participants provided written informed consent.

Computed Tomography Measurements

Of the entire cohort, 1,142 participants were randomly selected, stratified by age, sex, race/ethnicity, diabetes status, and eGFR, for electron-beam or multidetector CT. In addition, all eligible participants from 3 centers were scanned as part of an ancillary study, yielding 1,964 total participants scanned within the first 3 years of the original baseline examination (Figure 1). Of these participants, 1,274 had T₅₀ measured at the same study visit as their first CT scan (i.e. “baseline” for the present study) as part of an ancillary study. A repeated CT measurement was obtained among 1,123 participants an average of 3.2 +/− 0.6 (standard deviation [SD]) years later, 780 of whom had T₅₀ data.

Figure 1.

Flowchart describing the study sample selection for cross-sectional and longitudinal analyses.

Abbreviations: CRIC, Chronic Renal Insufficiency Cohort Study; CT, computed tomography

Trained and certified technologists scanned participants twice using phantoms of known physical calcium concentrations. A cardiologist read all scans at a central reading center (Los Angeles Biomedical Research Institute at Harbor-UCLA Medical Center) to quantify calcification according to the Agatston score.²⁰ The total CAC score was calculated as the sum of scores from the left main, left anterior descending, left circumflex, and right coronary arteries. Final scores are the mean of 2 scans.²¹

Exposure Assessment

We quantified calcification propensity as the transformation time (T₅₀) from primary to secondary calciprotein particles in vitro, with lower T₅₀ corresponding to higher calcification propensity.¹³ Serum samples, stored at −80°C and shipped with sufficient dry ice, were used for the test, which was performed using a Nephelostar nephelometer at the Calciscon Laboratory in Switzerland.¹³ The mean intraassay coefficient of variation is 2.2% and the mean interassay coefficient of variation is 3.4%. The reference range is 270 to 460 minutes, as determined in 253 healthy Swiss adults. T₅₀ values were reported in other populations, including 184 patients with CKD stages 3–4 (mean 329 +/− 95 minutes)¹⁵ and 2,785 patients undergoing hemodialysis (median 212 [10^th to 90^th percentiles, 109 to 328] minutes).¹⁸

Covariate Assessment

We obtained covariate data from the same study visit as the first CT scan, or the most recent previous annual visit if missing (<2% missing for all covariates except 24-hour urinary protein [8%]). Self-reported sociodemographic characteristics, medical history, and current medications were obtained via questionnaire. Body weight, height, and blood pressure (BP) were measured using standard protocols.¹⁹ Diabetes was defined as fasting glucose ≥126 mg/dL, nonfasting glucose ≥200 mg/dL, and/or the use of antidiabetic medications. History of CVD was defined as self-reported prior coronary artery disease, heart failure, stroke, or peripheral vascular disease.

Glucose, cholesterol, bicarbonate, phosphate, calcium, magnesium, serum albumin, and total parathyroid hormone (PTH) were measured using standard laboratory methods. 24-hour urinary protein was measured using the turbidometric method with benzethonium chloride. Fibroblast growth factor 23 (FGF-23) was measured by a second-generation carboxy-terminal assay (Immutopics). High-sensitivity C-reactive protein (hsCRP) and interleukin 6 (IL-6) were measured at the original baseline examination using the particle enhanced immunonephelometry method. Fetuin-A concentration was measured at the original baseline examination using the quantitative sandwich enzyme immunoassay technique. We calculated eGFR using the equation derived in the CRIC cohort.²²

Statistical Analysis

We summarized baseline characteristics of study participants as the mean +/− SD or median (interquartile range) for continuous variables and percentages for categorical variables, by quartile of T₅₀. We evaluated the cross-sectional association of T₅₀ with CAC using a two-part model.²¹ First, we modeled the prevalence of CAC >0 among all participants using Poisson regression with robust variance estimation. Second, among those with CAC >0, we modeled severity of CAC using linear regression and natural log-transformed CAC score. We exponentiated regression coefficients and expressed them as the percent difference in CAC per SD-decrease in T₅₀, or between quartiles of T₅₀ compared with the highest quartile (reference). Additionally, we modeled the prevalence of moderate (≥200 units) and severe (≥400 units) CAC.

We evaluated the longitudinal association of T₅₀ with CAC stratifying by presence of baseline CAC.²³ Among those with no baseline CAC, we defined incidence as CAC >0 at follow-up. Among those with baseline CAC, we defined progression as an annual increase in CAC ≥100 units, which is significantly associated with higher risk of coronary heart disease.⁹ Additionally, we assessed progression defined as an annual increase in CAC ≥200 units. We evaluated incidence and progression using Poisson regression with robust variance estimation, employing an offset to account for the time between CT scans.

We included covariates in sequential regression models based on prior clinical knowledge. In addition to unadjusted analyses, two multivariable-adjusted models were used: 1) adjusted for age, sex, race/ethnicity, and clinical site; and 2) adjusted for variables in model 1 plus eGFR, proteinuria, diabetes, systolic BP, number of antihypertensive medications, current smoking, history of CVD, total cholesterol, and use of statin medications. We included the baseline CAC score in models analyzing participants with baseline CAC. In additional analyses, we evaluated the impact of adjusting for variables potentially affecting T₅₀ (calcium, phosphate, bicarbonate, magnesium, serum albumin, fetuin-A, FGF-23, PTH, and use of medications including warfarin, active vitamin D, phosphate binders, and calciferols) and inflammatory variables (IL-6 and hsCRP) on associations of T₅₀ with CAC. Magnesium, IL-6, hsCRP, and fetuin-A were measured at the original baseline examination. Because onset of ESRD may increase the risk of calcification,²⁴ we conducted a sensitivity analysis excluding those with ESRD at baseline (ie, at the time of the scan; cross-sectional analyses) and during follow-up (longitudinal analyses).

We tested effect modification by including T₅₀-by-subgroup interaction terms (defined by age, sex, race/ethnicity, and diabetes) in the regression models. All analyses were conducted using SAS version 9.4 (SAS Institute, Inc) and R version 3.4.2 (The R Foundation). All tests were 2-sided and statistical significance was defined as P<0.05.

Results

Among 1,274 participants with CT and T₅₀ data, the mean age was 57.5 +/− 11.7 years, 46.9% were female, 44.3% had diabetes, 27.2% had a history of CVD, and the mean eGFR was 44.5 +/− 17.8 ml/min/1.73 m². Participants included in the current analyses were, on average, healthier compared to those not included in the current analysis (Table S1). The median T₅₀ was 321 (IQR, 270 to 366) minutes. Those with low T₅₀ were more likely to be non-Hispanic black race/ethnicity (P<0.001), have a history of CVD (P=0.004) and diabetes (P<0.001), and be taking antihypertensive (P<0.001), statin (P=0.01), and active vitamin D medications (P<0.001) (Table 1). On average, those with low T₅₀ had higher systolic BP (P=0.006), 24-hour urine protein (P<0.001), phosphate (P<0.001), FGF-23 (P<0.001), PTH (P<0.001), IL-6 (P=0.001), and hsCRP (P=0.02), and lower eGFR (P<0.001), bicarbonate (P<0.001), calcium (P=0.006), magnesium (P=0.005), serum albumin (P<0.001), and fetuin-A (P<0.001). Of the 1,274 participants, 824 (64.7%) had baseline CAC. Figure 2 shows the distribution of baseline CAC by quartiles of T₅₀. Mild CAC severity (<100 Agatston units) was similar among quartiles of T₅₀ while lower quartiles of T₅₀ were more likely to have severe CAC (>400 Agatston units).

Table 1.

Baseline Characteristics of 1274 CRIC Participants by Quartiles of T₅₀

Variables
	T50 Q4	T50 Q3	T50 Q2	T50 Q1
No. of patients	316	320	322	316
T50 range, min	367–600	322–366	270–321	72–269
Age, years	57 ± 12	58 ± 12	57 ± 12	58 ± 11
Female sex	146 (46)	145 (45)	158 (49)	148 (47)
Race/ethnicity
Non-Hispanic White	183 (58)	150 (47)	150 (47)	127 (40)
Non-Hispanic Black	97 (31)	124 (39)	134 (42)	161 (51)
Hispanic	9 (3)	22 (7)	10 (3)	16 (5)
Other	27 (9)	24 (8)	28 (9)	12 (4)
Body mass index, kg/m2	30 ± 7	31 ± 7	31 ± 7	31 ± 7
Current cigarette smoking	32 (10)	24 (8)	36 (11)	45 (14)
History of cardiovascular disease	77 (24)	78 (24)	80 (25)	111 (35)
Diabetes mellitus	120 (38)	134 (42)	139 (43)	171 (54)
Systolic blood pressure, mmHg	122 ± 19	125 ± 20	126 ± 21	128 ± 21
No. of antihypertensive medications	2 [1, 3]	3 [2, 4]	2 [1, 3]	3 [2, 4]
Total cholesterol, mg/dL	186 ± 40	181 ± 39	186 ± 42	182 ± 45
Statin medication use	158 (50)	194 (61)	173 (54)	192 (61)
Warfarin medication use	16 (5)	16 (5)	18 (6)	12 (4)
eGFR, ml/min/1.73 m2	50 ± 16	45 ± 16	45 ± 18	38 ± 18
Urinary protein, g/24h	0.1 [0.1, 0.4]	0.2 [0.1, 0.8]	0.2 [0.1, 1.0]	0.2 [0.1, 1.5]
Bicarbonate, mmol/L	25 ± 3	24 ± 3	24 ± 3	23 ± 4
Active vitamin D medication use	6 (2)	18 (6)	24 (8)	33 (10)
Phosphate binder medication use	20 (6)	13 (4)	28 (9)	26 (8)
Calciferol medication use	43 (14)	40 (13)	48 (15)	40 (13)
Calcium, mg/dL	9.4 ± 0.4	9.3 ± 0.5	9.3 ± 0.5	9.3 ± 0.6
Phosphate, mg/dL	3.6 ± 0.9	3.8 ± 1.1	3.8 ± 0.7	4.2 ± 1.0
Magnesium, mg/dL	2.0 ± 0.3	2.0 ± 0.3	1.9 ± 0.2	1.9 ± 0.3
Serum albumin, g/dL	4.2 ± 0.4	4.1 ± 0.4	4.1 ± 0.4	4.0 ± 0.5
Fetuin-A, mg/L	582±116	529 ±104	517 ± 97	492 ±111
FGF-23, RU/mL	115 [77, 175]	135 [84, 246]	139 [97, 276]	170 [101, 358]
Parathyroid hormone, pg/mL	55 [36, 85]	64 [43, 93]	56 [36, 90]	65 [42, 122]
Interleukin-6, pg/mL	1.4 [0.8, 2.5]	1.7 [1.1, 2.6]	1.7 [1.1, 2.8]	1.8 [1.1, 2.9]
hsCRP, mg/L	1.6 [0.8, 4.3]	2.4 [0.9, 5.7]	2.3 [1.1, 5.3]	2.4 [0.9, 5.5]

Values are presented as mean ± standard deviation, median [interquartile range], or number (%). eGFR, ____; FGF-23, _____; hsCRP, ____; Q, quartile.

Figure 2.

Bar chart describing the proportions of participants in CAC categories, by quartile of T₅₀.

Abbreviations: CAC, coronary artery calcium

Table 2 shows the cross-sectional associations of T₅₀ with prevalence and severity of CAC. T₅₀ was not associated with prevalence of CAC >0 after multivariable adjustment (P=0.6 for linear trend across quartiles). However, lower T₅₀ was associated with greater CAC severity among participants with baseline CAC. After multivariable adjustment, one SD lower T₅₀ was associated with 21% (95% CI, 6% to 38%) greater CAC severity, and we observed graded associations across quartiles of T₅₀. Additionally, lower T₅₀ was significantly associated with greater prevalence of moderate and severe CAC (Table S2).

Table 2.

Association of T₅₀ with Prevalence and Severity of Coronary Artery Calcification at Baseline

	Continuous: Per 1-SD* ↓ T50	Categorical				P Value for Linear Trend
		T50 Q4 (≥367 min)	T50 Q3 (322–366 min)	T50 Q2 (271–321 min)	T50 Q1 (≤270 min)
All Participants (N=1274): Prevalence of CAC >0, Prevalence Ratio (95% CI)
n/N**		194 / 316	210/320	211/322	209/316
Unadjusted	1.01 (0.97–1.05)	1.00 (reference)	1.07 (0.95–1.20)	1.07 (0.95–1.20)	1.08 (0.96–1.21)	0.2
Model 1a	1.03 (0.99–1.06)	1.00 (reference)	1.05 (0.95–1.17)	1.10 (0.99–1.22)	1.11 (1.00–1.23)	0.04
Model 2b	0.99 (0.96–1.03)	1.00 (reference)	1.03 (0.93–1.14)	1.09 (0.98–1.21)	1.01 (0.91–1.13)	0.6
Participants with Baseline CAC >0 (n=824): CAC Severity, % Difference (95% CI)
Unadjusted	29% (12% to 49%)	1.00 (reference)	21% (−20% to 82%)	23% (−18% to 84%)	88% (25% to 183%)	0.004
Model 1a	38% (21% to 58%)	1.00 (reference)	21% (−17% to 77%)	39% (−5% to 104%)	118% (48% to 221%)	<0.001
Model 2b	21% (6% to 38%)	1.00 (reference)	19% (−17% to 71%)	36% (−5% to 96%)	58% (8% to 129%)	0.01

Abbreviations: BP, blood pressure; CAC, coronary artery calcium; CVD, cardiovascular disease; eGFR, estimated glomerular filtration rate; CI, ____________; Q, quartile; SD, standard deviation

^aModel 1: adjusted for age, sex, race/ethnicity, and clinical site

^bModel 2: adjusted for model 1 + eGFR, proteinuria, diabetes, systolic BP, number of antihypertensive medications, current smoking, history of CVD, total cholesterol, and use of statin medications

^*1 SD is 77 min

^**Events/total number.

Table 3 shows the longitudinal associations of T₅₀ with incidence and progression of CAC among 780 participants with a follow-up CT scan an average of 3.2 +/− 0.6 years later. Among 320 participants without baseline CAC, 65 developed incident CAC over follow-up; T₅₀ was not associated with incident CAC. Among 460 participants with baseline CAC, 89 had an annual increase of ≥100 Agatston units and 37 had an annual increase of ≥200 Agatston units. T₅₀ was significantly associated with both definitions of progression. After multivariable adjustment, one SD lower T₅₀ was associated with 28% (95% CI, 7% to 53%) higher risk of progressing ≥100 Agatston units per year and we observed graded associations across quartiles of T₅₀. Corresponding associations for CAC progression defined as increase ≥200 Agatston units per year were similar, but stronger and larger in magnitude.

Table 3.

Association of T₅₀ with Incidence and Progression of Coronary Artery Calcification

	Continuous: Per 1-SD* ↓ T50	Categorical				P Value for Linear Trend
		T50 Q4 (≥367 min)	T50 Q3 (322–366 min)	T50 Q2 (271–321 min)	T50 Q1 (≤270 min)
Participants with Baseline CAC=0 (n=320): Incident CAC, RR (95% CI)
n/N**		13 / 81	16 / 81	16 / 77	20 / 81
Unadjusted	1.15 (0.931.41)	1.00 (reference)	1.31 (0.682.54)	1.31 (0.682.52)	1.54 (0.832.88)	0.2
Model 1a	1.07 (0.841.36)	1.00 (reference)	1.34 (0.712.53)	1.29 (0.702.36)	1.33 (0.702.53)	0.4
Model 2b	0.95 (0.761.18)	1.00 (reference)	1.13 (0.592.17)	1.18 (0.632.24)	0.93 (0.491.76)	0.9
Participants with Baseline CAC >0 (n=460): Increase >100 Agatston U/y, RR (95% CI)
n/N**		12 / 113	20 / 114	23 / 118	34 / 115
Unadjusted	1.48 (1.251.74)	1.00 (reference)	1.72 (0.883.37)	1.86 (0.973.57)	2.86 (1.555.25)	<0.001
Model 1a	1.46 (1.241.73)	1.00 (reference)	1.73 (0.923.26)	1.94 (1.113.39)	2.66 (1.534.62)	<0.001
Model 2b	1.28 (1.071.53)	1.00 (reference)	1.39 (0.722.69)	1.81 (1.053.12)	1.86 (1.053.32)	0.02
Participants with Baseline CAC >0 (n=460): Increase >200 Agatston U/y, RR (95% CI)
n/N**		4 / 113	4 / 114	12 / 118	17 / 115
Unadjusted	1.98 (1.482.65)	1.00 (reference)	1.03 (0.274.03)	2.91 (0.978.74)	4.29 (1.4912.35)	<0.001
Model 1a	1.92 (1.392.66)	1.00 (reference)	1.03 (0.313.35)	3.15 (1.347.38)	3.08 (1.277.47)	0.003
Model 2b	1.81 (1.362.41)	1.00 (reference)	1.19 (0.334.32)	3.30 (1.537.13)	2.95 (1.256.97)	0.005

Abbreviations: BP, blood pressure; CAC, coronary artery calcium; CVD, cardiovascular disease; eGFR, estimated glomerular filtration rate CI, ____________; Q, quartile; RR, relative risk; SD, standard deviation

^aModel 1: adjusted for age, sex, race/ethnicity, clinical site, and baseline CAC score (among those with CAC >0 only)

^bModel 2: adjusted for model 1 + eGFR, proteinuria, diabetes, systolic BP, number of antihypertensive medications, current smoking, history of CVD, total cholesterol, and use of statin medications

^*1 SD is 77 min

^**Events/total number.

We did not detect any significant interactions between T₅₀ and age, sex, race/ethnicity, nor diabetes. Results were similar to the primary analyses following adjustment for variables potentially affecting T₅₀ (calcium, phosphate, bicarbonate, magnesium, serum albumin, fetuin-A, FGF-23, PTH, and use of medications including warfarin, active vitamin D, phosphate binders, and calciferols) and inflammatory variables (IL-6 and hsCRP) (Table S3). A sensitivity analysis excluding participants with ESRD did not substantively impact the results (Table S4).

Discussion

High serum calcification propensity, denoted by low T₅₀, was independently associated with greater CAC severity and progression among patients with CKD stages 2–4 and established CAC at baseline. Conversely, T₅₀ was not associated with CAC prevalence nor with the incidence of CAC among patients without CAC. These findings implicate T₅₀ as a measure of the severity of calcification and the risk for progression. Given the strength of associations and consistency of results in several additional analyses, T₅₀ and its components may be involved in the calcification pathway, identify patients with CKD who are at high risk for CVD, and offer insights into mechanisms of calcification and targets for therapy to reduce the burden of CVD among patients with CKD.

There are several possible mechanisms that could explain our findings. First, information gained from the T₅₀ test could relate to medial vascular calcification and its determinants. Disordered mineral metabolism is common in CKD,²⁵ marked by abnormal levels of the calcification promoters phosphate and calcium, which are themselves associated with presence²⁶ and progression²³ of CAC. However, vascular calcification is a tightly-regulated and dynamic process that begins with the deposition of amorphous calcium phosphates on an organic template. While the mechanisms of formation of the initial nidus are complex, the subsequent ripening and transformation of the nidus to more crystalline forms are likely governed by calcification promoters and inhibitors.^27,28 Given that the T₅₀ test represents a composite measure that captures summary information about calcification promoters and inhibitors in the serum,¹³ we surmise that T₅₀ may not reflect the initiation of vascular calcification but may provide information on the terminal step of transformation once the nidus is already formed. Thus, the T₅₀ test may identify individuals prone to propagation, but not necessarily initiation, of vascular calcification. It is likely that once calcification is established, it progresses rapidly, especially among patients with high calcification propensity (i.e. low T₅₀). Our findings support this hypothesis, showing that while low T₅₀ was not associated with prevalence of CAC >0 nor incidence, it was strongly associated with severity and progression, with larger magnitudes of association observed for larger increases in CAC. Components of the calcification process, particularly fetuin-A, a potent inhibitor of calcification, have previously shown a similar association with CAC severity among patients with established calcification.²⁹

The T₅₀ test could also reflect factors that promote atherosclerotic intimal calcification, which is characterized by endothelial damage, lipid deposition, macrophage accumulation, and inflammation.³⁰ Prior in vitro studies suggest that secondary calciprotein particles, the maturation of which is captured, in part, by T₅₀, stimulate inflammation and apoptosis of macrophages, which may promote ectopic calcification.^31,32 Furthermore, secondary calciprotein particles may trigger atherosclerosis via endothelial damage and increase local inflammation through production of pro-atherosclerotic cytokines, including IL-6.³³ In our analysis, additional adjustment for the inflammatory variables hsCRP and IL-6 did not significantly change the associations we observed, despite being associated with both T₅₀ and CAC. While these measurements were not collected concurrently with T₅₀, the T₅₀ test may capture information inherent in inflammatory variables. While we could not distinguish intimal vs. medial calcification, the relationships of T₅₀ with both types of calcification warrant further mechanistic and translational research approaches.

Finally, we acknowledge the possibility that T₅₀ and its constituents may not be directly on the causal pathway to calcification. Given that T₅₀ was not associated with incidence of CAC, it is possible that T₅₀ captures a disease state that is already in progress and that is mediated via different causal mechanisms. T₅₀ may also reflect another component that itself upsets promoter-inhibitor homeostasis. Our findings point to a threshold effect characterized by severe calcification in those with low T₅₀ and established CAC. However, in the current analysis, it is impossible to distinguish whether such a threshold effect directly involves T₅₀ and its components or other mechanisms. Nevertheless, our findings point to the utility of T₅₀ in providing information about calcification severity and progression. Further analyses are necessary in broader patient populations and larger sample sizes to determine if T₅₀ can predict future calcification and whether its components are potential therapeutic targets. Additionally, evaluation of complementary assays of promoter-inhibitor balance and calciprotein particles, such as the hydrodynamic radius (R_h),³⁴ may offer additional insights.

The present study has several strengths. First, it is the first to estimate the associations of the T₅₀ test, a novel measure of calcification propensity, with CAC among patients with mild-to-moderate CKD. One previous analysis among 73 patients with diabetes and CKD identified an association between fetuin-A-mineral complex and CAC,³⁵ but we show for the first time associations of a composite calcification propensity measure with CAC in a larger, longitudinal analysis. Second, the CRIC Study employs standardization of methods and measurements across clinical sites which minimizes bias. Third, our results were robust to adjustment for several covariates, including variables related to T₅₀ (i.e. mineral metabolism variables) and inflammatory variables, and various sensitivity analyses. However, the present study has potential limitations. First, the T₅₀ test is conducted in vitro with supersaturation of calcium and phosphate, which results in synthetic calciprotein particles. Since physiochemical properties and predicted pathogenic effects of synthetic and endogenous calciprotein particles appear to be comparable,³⁶ it is reasonable to investigate T₅₀ as a functional read-out of the calciprotein transformation process. Second, we cannot rule out the possibility of selection bias, especially for progression analyses which required participants to survive long enough for a follow-up scan. While participants included in the present analysis were, on average, healthier compared with the full cohort, the population under study is still of clinical and public health relevance. Third, the present study represents a relatively small sample size with short follow-up, but are the first data of its kind and, importantly, included longitudinal analyses. Fourth, while we considered many covariates in our analyses, we were unable to evaluate some additional variables important to mineral metabolism and calcification in CKD, including vitamin K and pH. Fifth, we were unable to distinguish between intimal and medial calcification owing to limitations in CT technology. While each may represent different developmental pathways, both are associated with higher risk of CVD and mortality in patients with CKD.³⁷ Finally, one CT measurement over follow-up does not allow us to pinpoint the exact time point of calcification incidence or progression.

The findings presented here have important clinical and research implications. Calcification is a dynamic process involving complex pathophysiology. Our results suggest that once subclinical disease is established, inherent characteristics of a patient’s serum, captured by the T₅₀ test, may be useful in determining both the extent of calcification and risk of significant progression. However, given that we did not observe associations with CAC prevalence nor incidence, it is unclear in the current study whether a low T₅₀ value precedes calcification, or vice versa. One possibility is that the ongoing calcification process consumes inhibitors, like fetuin-A, resulting in a lower T₅₀ value.³⁸ Alternatively, it is possible that consequences of vascular disease, including inflammation, suppress the synthesis of inhibitors, also prompting a lower T₅₀ value.³⁹ Regardless, a previous analysis in the CRIC Study found that severe CAC was significantly associated with CVD.¹² Thus, in patients with low T₅₀, increased vigilance may be warranted to mitigate the potential for adverse cardiovascular health outcomes. Future research is warranted in those with high calcification propensity to determine if promoter-inhibitor homeostasis can be improved using novel drug interventions.

In conclusion, higher serum calcification propensity, denoted by lower T₅₀, was significantly associated with severity and progression of CAC among patients with CKD. However, T₅₀ was not associated with incidence of CAC. These findings provide valuable insights into the development of calcification and atherosclerosis in patients with CKD and highlight potential pathways for risk stratification and therapeutic intervention. Future research should evaluate these associations in other CKD populations and the general population and clinical trials may be warranted to establish causality.

Supplementary Material

1

Table S1. Comparison of CRIC participants included and not included in analyses of T50 and CAC.

Table S2. Associations of T50 with prevalence of moderate and severe CAC at baseline.

Table S3. Impact of adjustment for additional variables on associations of T50 with CAC.

Table S4. Impact of excluding participants with ESRD on associations of T50 with CAC.