ABSTRACT
Background
The serum T50 test quantifies calcification propensity, with lower values indicating a higher propensity for calcification. Lower T50 values have been associated with adverse cardiovascular outcomes, particularly in advanced chronic kidney disease. Whether serum T50 is associated with coronary artery calcification (CAC) in individuals with mildly to moderately reduced kidney function remains unclear. We examined this association in a population-based cross-sectional study.
Methods
We analysed 1945 participants from the population-based Lifelines ImaLife cohort with mildly to moderately reduced kidney function (estimated glomerular filtration rate, 37.5–87.6 ml/min/1.73 m²). Serum T50 was measured using a standardized nephelometric assay. CAC was assessed by electrocardiographically triggered non-contrast computed tomography and categorized into 0, 1–99, 100–399, and ≥400 Agatston units. Associations between continuous T50 (per 10-min increase) and CAC categories were examined using multinomial logistic regression.
Results
Of the 1945 participants, 56.5% had detectable CAC (CAC 1–99: 30.9%, 100–399: 14.4%, ≥400: 11.2%). In crude analyses, higher T50 values were associated with the presence of CAC [odds ratio (OR) 1.02, 95% confidence interval 1.01–1.03] and with CAC categories 1–99, 100–399, and ≥400 versus 0 (ORs 1.01, 1.02, and 1.03, respectively). These associations were no longer significant after adjustment for age, body mass index, and sex. Subgroup analyses did not identify risk groups in which T50 was significantly associated with the presence or severity of CAC.
Conclusion
In this cohort with mildly to moderately reduced kidney function, T50 was not independently associated with prevalent CAC after multi-variable adjustment.
Keywords: calcification propensity, chronic kidney disease, coronary artery calcification, T50 test, vascular calcification
Graphical Abstract
Graphical Abstract.
KEY LEARNING POINTS.
What was known:
Lower T50 is associated with cardiovascular outcomes in advanced chronic kidney disease (CKD), but its relationship with coronary artery calcification (CAC) in earlier stages of CKD remains unclear.
This study adds:
In a large population-based cohort with mildly to moderately reduced kidney function, T50 was not independently associated with the presence or severity of CAC after adjustment for cardiovascular risk factors.
Potential impact:
T50 may have limited value for CAC-based cardiovascular risk stratification in early CKD.
INTRODUCTION
Chronic kidney disease (CKD) affects ∼10%–15% of adults worldwide [1]. Patients with CKD have a markedly elevated risk of cardiovascular disease (CVD) and the presence of CKD is associated with significantly higher all-cause and cardiovascular (CV) mortality [2, 3]. Vascular calcification, the pathological deposition of calcium-phosphate in vessel walls, is an important predictor of CVD risk in CKD patients [2]. As kidney function declines, medial vascular calcification becomes more prevalent [4, 5].
Given the high CV burden in CKD, improved risk stratification tools are needed, particularly in earlier disease stages. If patients susceptible to accelerated vascular calcification could be recognized at the earlier stages of CKD, this may assist in selecting patients for early intervention. Early risk stratification may help inform clinical decision-making and healthcare planning [6]. One biomarker for CV risk assessment is the serum calcification propensity assay, known as the T50 test [7]. It measures the time it takes for secondary calciprotein particles (CPPs) to form in vitro in patient serum after the addition of supersaturating concentrations of calcium and phosphate, reflecting the resistance to calcification [7–9]. A longer T50 indicates greater resistance to calcium-phosphate crystallization, whereas a shorter T50 signifies a higher propensity for calcification. Coronary artery calcification (CAC) is a well-established marker of (subclinical) atherosclerosis and is strongly associated with future CV events [10, 11]. Unlike clinical end-points, CAC provides a graded measure of disease burden, enabling the study of earlier and more subtle stages of vascular pathology [12].
Recent studies have linked serum T50 with CV risk factors and outcomes [13–16]. Specifically, in moderate CKD (stages 3–4) patients, a lower T50 was independently associated with higher aortic pulse wave velocity and progressive arterial stiffening [17]. Additionally, in kidney transplant recipients, a lower T50 was associated with higher all-cause and CV mortality [18] and similarly, in haemodialysis patients, a lower T50 correlated with increased rates of myocardial infarction and all-cause death [19]. Notably, in a general population cohort, individuals with lower T50 values showed a higher risk of CV mortality [13].
However, it remains unclear whether serum T50 is associated with the cross-sectional presence and severity of CAC in individuals with mildly to moderately reduced kidney function. We therefore examined this association in a large population-based cohort. We hypothesized that lower T50 values are associated with the presence and severity of CAC.
MATERIALS AND METHODS
This population-based cross-sectional study is reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.
Study population
This cross-sectional analysis was conducted within the ImaLife subcohort of the Lifelines study, a large population-based cohort in the Northern Netherlands [20–23]. Lifelines is a multi-disciplinary prospective population-based cohort study examining in a unique three-generation design the health and health-related behaviours of 167 729 persons living in the north of the Netherlands. It employs a broad range of investigative procedures in assessing the biomedical, socio-demographic, behavioural, physical, and psychological factors which contribute to the health and disease of the general population, with a special focus on multi-morbidity and complex genetics. Ethics approval and informed consent were obtained. The Lifelines study, including all subcohorts such as ImaLife, was approved by the Medical Ethical Review Committee of the University Medical Center Groningen (METC UMCG 2007/152). All participants provided written informed consent prior to participation in the study [23]. Between 2017 and 2022, 11 678 participants underwent cardiac computed tomography (CT) for CAC evaluation. After excluding individuals with missing data on sex and prior cardiac interventions (n = 55), we selected participants with the lowest estimated glomerular filtration rate (eGFR) values among those who underwent CAC scanning to enrich the study population for individuals with mildly to moderately reduced kidney function. This resulted in a study population of 1945 participants with an eGFR range of ∼38–88 ml/min/1.73 m². A formal power calculation was not performed; the study sample reflects the maximum number of eligible participants available within the cohort. Participant selection is summarized in Fig. 1.
Figure 1:
Flowchart of participant selection for the present analysis. Participants were selected from the Lifelines ImaLife cohort based on availability of CAC imaging and kidney function.
Serum calcification propensity (T50)
Fasting blood samples were collected during Lifelines assessment visits and processed according to standardized protocols, and serum was stored at −80°C in the Lifelines biobank until shipment under frozen conditions in July 2024, followed by analysis in September 2024 without an intervening freeze–thaw cycle. The T50 assay was performed by Calciscon (Nidau, Switzerland) according to a standardized nephelometric protocol [7]. The serum T50 assay quantifies calcification propensity by incubating serum with excess calcium and phosphate to trigger CPP formation. In practice, 80 µl of serum is mixed with 20 µl saline in each well of a 96-well plate; pre-warmed phosphate (50 µl) and calcium (50 µl) stock solutions are then rapidly injected (yielding final Ca²⁺ concentration ≈10 mM and phosphate concentration ≈6 mM at pH 7.4). The plate was sealed and placed in a temperature‐controlled nephelometer at ∼36°C–37°C. The excess calcium and phosphate first form amorphous primary CPPs, which gradually convert into larger crystalline secondary CPPs; this phase change increases the sample turbidity. Light scattering is recorded continuously (up to 570 min) to follow the rise in turbidity as particles mature. The T50 value is defined as the time (in minutes) at which the nephelometric signal reaches 50% of its total increase. Samples were measured in duplicate, and the mean value was used for analysis.
CAC scoring
CAC was quantified using non-contrast electrocardiogram (ECG)-triggered cardiac CT imaging, following the standardized protocol of the ImaLife study [24–26]. Scans were acquired on a third-generation dual-source CT scanner (SOMATOM Force, Siemens Healthineers) in high-pitch spiral mode with prospective ECG triggering at minimal coronary motion (typically 60%–65% of the R–R interval) during an inspiratory breath-hold. Acquisition parameters included 120 kVp tube voltage, 64 mAs reference tube current (automatically adjusted for body size), 250 ms rotation time, and pitch of ∼3.2. Images were reconstructed using a 250 mm field of view, filtered back projection with a slice thickness of 3.0 mm, increment of 1.5 mm, and a medium-sharp quantitative kernel (Qr36). CAC was quantified using the Agatston scoring method [27]. Voxels with an attenuation value ≥130 Hounsfield Units were identified and summed using syngo.via VB30A CaScoring software (Siemens Healthineers). Each scan was scored by trained analysts according to standardized procedures. Participants were categorized into four CAC groups based on their Agatston score: 0 (reference), 1–99, 100–399, and ≥400 units.
Covariates
Covariates were selected from the Lifelines Cohort Study design, as previously described by Scholtens et al. [22]. Baseline variables were assessed via standardized questionnaires, clinical examination, and fasting blood sampling. Covariates included age, body mass index (BMI), sex, diabetes status (self-reported; missing values supplemented using HbA1c measurements, with HbA1c >53 mmol/mol indicating diabetes when available), systolic blood pressure (measured), lipid profile [triglycerides, low-density lipoprotein (LDL) cholesterol, and total cholesterol], education level, smoking status, and relevant medication use (coded according to ATC classifications C01–C05, C07–C10, A10, and thiazide-related subclasses). Medication use was coded as present if any were reported. Sex was defined as biological sex as recorded at baseline.
eGFR was calculated using the 2021 Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) creatinine equation based on serum creatinine, age, and sex [28]. Serum phosphate concentrations were measured spectrophotometrically.
Missing-data handling
Missing data were handled using multiple imputation with the mice package in R, with imputation methods selected according to variable type and distribution. BMI had 0.05% missing data and was mean-imputed because of the very small proportion of missing values and its approximately normal distribution. Diabetes status (1.03% missing) was imputed using logistic regression, whereas education level (2.42% missing) was imputed using polytomous regression. Predictive mean matching was used for other continuous variables with missing values. Blood pressure and total cholesterol were transformed before imputation using lambda and square-root transformations, respectively. Medication use was coded as a binary variable according to ATC classifications C01 to C05, C07 to C10, A10, and the thiazide-related subgroups C03AA, C07BB, and C09DX; participants were classified as medication users if at least one relevant ATC code was present.
Statistical analysis
The primary exposure variable was serum calcification propensity (T50), analysed as a continuous variable. The primary outcome was CAC category (0, 1–99, 100–399, ≥400), as described above. In secondary analyses, CAC presence was defined as CAC >0. CAC was analysed using categories because many participants had a CAC score of zero and the values were not normally distributed. Four nested models were constructed: Model 1: adjusted for age, BMI, and sex; Model 2: additionally adjusted for diabetes, systolic blood pressure, triglycerides, LDL cholesterol, total cholesterol, education, smoking, and medication use (as defined above); Model 3: additionally adjusted for eGFR; and Model 4: further included serum phosphate.
Multi-variable models were constructed using a stepwise approach, with covariates selected a priori based on established CV and kidney-related risk factors. To assess whether T50 was associated with the presence of CAC, binary logistic regression was performed (CAC > 0 versus CAC = 0) using the same four nested models and covariates. In these models, odds ratios (ORs) and 95% confidence intervals (CIs) were calculated to estimate the change in odds of having any CAC per 10-min increase in T50, modelled as a continuous variable.
CAC was additionally analysed using proportional-odds ordinal logistic regression across ordered categories of 0, 1–99, 100–399, and at least 400 Agatston units, with T50 modelled per 10-min increase. The proportional-odds assumption for T50 was assessed using threshold-specific binary logistic models and an approximate heterogeneity test. CAC was also analysed continuously as log(CAC + 1), and non-linearity was assessed using a natural cubic spline with three degrees of freedom, compared with the linear model using a nested-model ANOVA F-test.
Stratified subgroup analyses were conducted across predefined subgroups based on established CV and kidney-related risk factors. Subgroups were defined by age (<65 versus ≥65 years), sex, BMI category (normal, overweight, obese), diabetes status, systolic blood pressure (<140 versus ≥140 mmHg), LDL cholesterol (<3.0 versus ≥3.0 mmol/l), medication use (yes/no), eGFR (<60 versus ≥60 ml/min/1.73 m²), and phosphate (<0.9 versus ≥0.9 mmol/l). Associations were estimated separately within each subgroup using fully adjusted models (Model 4). Results were summarized using forest plots.
Finally, sensitivity analyses were performed to assess the robustness of the main findings across different adjustment models. These included crude models, models adjusted only for eGFR, models only for serum phosphate, and models including single interaction terms between T50 and selected covariates to assess potential effect modification.
All data cleaning, statistical modelling, and analyses were conducted in RStudio (version 4.4.0; 2024-04-24, UCRT build) using the following R packages: broom (v1.0.8), dplyr (v1.1.4), foreign (v0.8.86), gtsummary (v2.2.0), ggplot2 (v3.5.1), haven (v2.5.4), MASS (v7.3-65), mice (v3.17.0), nnet (v7.3-20), nortest (v1.0-4), readr (v2.1.5), tidyr (v1.3.1), and VIM (v6.2.2). Missing data were handled using multiple imputation, and all analyses were performed on the imputed datasets. Imputation methods were chosen based on variable type and distribution using the mice package. Continuous variables were imputed using predictive mean matching, while categorical variables were imputed using logistic or polytomous regression as appropriate. BMI was imputed using the mean. Potential sources of bias include selection bias related to the ImaLife imaging subcohort and residual confounding. These were addressed by multi-variable adjustment and the use of uniform imaging and laboratory procedures.
RESULTS
Of the 11 678 participants who underwent CAC imaging within the ImaLife study, we selected the 1945 individuals with the lowest eGFR values after exclusion of participants with missing sex data or prior cardiac interventions, resulting in a study population with a mean eGFR of 76.9 ± 7.1 ml/min/1.73 m² (range 37.5–87.6) (Fig. 1).
Baseline characteristics
Baseline characteristics of the 1945 participants are shown in Table 1. Of these, 1099 (56.5%) had any CAC, including 601 (30.9%) with CAC 1–99, 281 (14.4%) with CAC 100–399, and 217 (11.2%) with CAC ≥400. Mean BMI increased across CAC categories, from 25.5 ± 3.7 kg/m² in the CAC = 0 group to 26.8 ± 3.3 kg/m² in the CAC ≥ 400 group (P < .001). Mean age increased from 49.1 ± 8.1 years in the CAC = 0 group to 63.5 ± 8.0 years in the CAC ≥ 400 group (P < .001). The proportion of male participants rose from 14% in the CAC = 0 group to 64% in the highest CAC category. The prevalence of diabetes was higher in participants with higher CAC categories, ranging from <1% in the CAC = 0 group to 5.5% in the CAC ≥ 400 group (P < .001). Participants with higher CAC scores had higher blood pressure, triglycerides, LDL cholesterol, and total cholesterol levels (all P < .001). Smoking patterns were different, with more ex-smokers in the higher CAC categories and fewer never-smokers. Medication use was more common among participants with higher CAC categories. Mean eGFR decreased across CAC categories (P < .05), whereas serum phosphate levels remained similar across groups.
Table 1:
Baseline characteristics of 1945 individuals grouped by coronary artery calcium score category (Agatston units: 0, 1–99, 100–399, ≥400).
| Variable | Overall (N = 1945) | CAC = 0 (n = 846) | CAC = 1–99 (n = 601) | CAC = 100–399 (n = 281) | CAC ≥ 400 (n = 217) | P-value |
|---|---|---|---|---|---|---|
| Age (years) | 54.0 ± 9.8 | 49.1 ± 8.1 | 54.8 ± 8.7 | 60.1 ± 8.6 | 63.5 ± 8.0 | < .001 |
| BMI (kg/m²) | 26.2 ± 3.8 | 25.5 ± 3.7 | 26.7 ± 3.9 | 26.8 ± 3.7 | 26.8 ± 3.3 | < .001 |
| Sex | < .001 | |||||
| • Female | 1390 (71%) | 728 (86%) | 420 (70%) | 163 (58%) | 79 (36%) | |
| • Male | 555 (29%) | 118 (14%) | 181 (30%) | 118 (42%) | 138 (64%) | |
| Diabetes | 51 (2.6%) | <10 (< 1%) | 17 (2.8%) | 14 (5.0%) | 12 (5.5%) | < .001 |
| Systolic blood pressure (mmHg) | 128.1 ± 15.7 | 124.0 ± 14.6 | 128.8 ± 15.0 | 134.3 ± 15.7 | 134.6 ± 16.7 | < .001 |
| Triglycerides | 1.2 ± 0.7 | 1.1 ± 0.6 | 1.3 ± 0.7 | 1.3 ± 0.7 | 1.4 ± 0.8 | < .001 |
| LDL cholesterol (mmol/l) | 3.4 ± 0.9 | 3.2 ± 0.9 | 3.6 ± 0.8 | 3.7 ± 0.9 | 3.6 ± 1.0 | < .001 |
| Total cholesterol | 5.4 ± 1.0 | 5.1 ± 1.0 | 5.5 ± 1.0 | 5.6 ± 1.0 | 5.5 ± 1.1 | < .001 |
| Education | < .001 | |||||
| • Low | 685 (35%) | 231 (27%) | 240 (40%) | 123 (44%) | 91 (42%) | |
| • Medium | 668 (34%) | 326 (39%) | 202 (34%) | 83 (30%) | 57 (26%) | |
| • High | 592 (30%) | 289 (34%) | 159 (26%) | 75 (27%) | 69 (32%) | |
| Smoking | < .001 | |||||
| • Never | 765 (39%) | 392 (46%) | 232 (39%) | 92 (33%) | 49 (23%) | |
| • Ex-smoker | 951 (49%) | 351 (41%) | 292 (49%) | 165 (59%) | 143 (66%) | |
| • Current smoker | 229 (12%) | 103 (12%) | 77 (13%) | 24 (9%) | 25 (12%) | |
| Medication use | 622 (32%) | 192 (23%) | 197 (33%) | 122 (43%) | 111 (51%) | < .001 |
| eGFR | 76.9 ± 7.1 | 78.5 ± 6.4 | 76.7 ± 6.7 | 74.9 ± 7.4 | 73.3 ± 8.0 | < .001 |
| Serum phosphate | 0.94 ± 0.17 | 0.94 ± 0.17 | 0.94 ± 0.17 | 0.93 ± 0.16 | 0.91 ± 0.17 | .07 |
| T50 (min) | 429.2 ± 95.8 | 419.8 ± 91.7 | 432.7 ± 98.8 | 433.8 ± 98.9 | 450.2 ± 95.0 | < .001 |
Continuous variables are presented as mean ± SD; categorical variables as number (%). P-values were derived using the Kruskal–Wallis test for continuous variables and Pearson’s chi-squared test for categorical variables. CAC, coronary artery calcification; BMI, body mass index; LDL, low-density lipoprotein cholesterol; eGFR, estimated glomerular filtration rate; T50, serum calcification propensity.
Serum T50 values ranged from 165 to 570 min, with a mean ± SD of 429.2 ± 95.8 min. Notably, the mean T50 value increased with higher CAC categories (P < .001).
T50 is not associated with CAC categories after multi-variable adjustment
Table 2 shows the results of the multinomial logistic regression models assessing the association between T50 and CAC categories. In the crude models, higher T50 values were significantly associated with higher odds of having CAC scores in the 1–99, 100–399, and ≥400 categories compared with CAC = 0, per 10-min increase in T50. For example, the crude OR for the highest CAC category (≥400) was 1.034 (95% CI: 1.018–1.050, P < .001) per 10-min increase in T50. However, after adjusting for age, sex, and BMI (Model 1), the associations lost statistical significance for all CAC categories. This lack of association remained consistent after additional adjustment for other CV risk factors (Model 2), kidney function (Model 3), and serum phosphate (Model 4).
Table 2:
Multinomial logistic regression models for T50 across CAC with the following categories: 0 (n = 846), 1–99 (n = 601), 100–399 (n = 281), and ≥400 (n = 217).
| CAC category | Crude model OR (95% CI) | Crude P-value | Model 1 OR (95% CI) | Model 1 P-value | ||
| 0 (n = 846) | Ref | |||||
| 1–99 (n = 601) | 1.01 (1.00–1.03) | .01 | 1.00 (0.99–1.01) | .86 | ||
| 100–399 (n = 281) | 1.02 (1.00–1.03) | .03 | 0.99 (0.97–1.00) | .11 | ||
| ≥400 (n = 217) | 1.03 (1.02–1.05) | <.001 | 0.99 (0.97–1.01) | .28 | ||
| CAC category | Crude model OR (95% CI) | Crude P-value | Model 1 OR (95% CI) | Model 1 P-value | ||
| Model 2 OR (95% CI) | Model 2P-value | Model 3 OR (95% CI) | Model 3P-value | Model 4 OR (95% CI) | Model 4P-value | |
| 0 (n = 846) | Ref | |||||
| 1–99 (n = 601) | 1.00 (0.99–1.01) | .88 | 1.00 (0.99–1.01) | .90 | 1.01 (0.99–1.02) | .31 |
| 100–399 (n = 281) | 0.99 (0.97–1.00) | .10 | 0.99 (0.97–1.00) | .09 | 0.98 (0.98–1.01) | .39 |
| ≥400 (n = 217) | 0.97 (0.97–1.01) | .25 | 0.99 (0.97–1.01) | .25 | 1.00 (0.98–1.02) | .81 |
OR represents the change per 10-min increase in T50. CAC = 0 was used as the reference category. OR, odds ratio; CI, confidence interval; CAC, coronary artery calcification; T50, serum calcification propensity. Model definitions are described in the Materials and methods section.
T50 is not associated with the presence of CAC after multi-variable adjustment
We subsequently explored whether there was a relationship between T50 and the presence or absence of CAC using binary logistic regression. Consistent with the multinomial models, the crude association between T50 and CAC presence was no longer significant after multi-variable adjustment (Table 3).
Table 3:
Binary logistic regression models for T50 and the presence of CAC.
| CAC presence (CAC > 0 versus 0) | Per 10-min increase in T50 OR (95% CI) | P-value |
|---|---|---|
| Crude model | 1.02 (1.01–1.03) | <.001 |
| Model 1 | 1.00 (0.98–1.01) | .47 |
| Model 2 | 1.00 (0.98–1.01) | .47 |
| Model 3 | 1.00 (0.98–1.01) | .47 |
| Model 4 | 1.00 (0.99–1.02) | .64 |
OR represents the change per 10-min increase in T50. CAC presence was defined as CAC > 0 versus CAC = 0. OR, odds ratio; CI, confidence interval; CAC, coronary artery calcification; T50, serum calcification propensity. Model definitions are described in the Materials and methods section.
T50 is not associated with CAC in ordinal or continuous analyses
We further examined the association between T50 and CAC using ordinal logistic regression, accounting for the ordered CAC categories. Higher T50 was associated with greater odds of being in a higher CAC category in the crude model, but this association was no longer significant after multi-variable adjustment (Table 4). Similarly, when CAC was analysed continuously as log(CAC + 1), the crude association with T50 was attenuated and no longer significant after multi-variable adjustment (Table 5). Modelling T50 using a natural cubic spline provided no evidence of a non-linear association with CAC (P = .145) (Fig. S2).
Table 4:
Ordinal regression analysis of the association between T50 and CAC.
| Model | Per 10-min increase in T50 OR (95% CI) | P-value |
|---|---|---|
| Crude model | 1.02 (1.01–1.03) | <.001 |
| Model 1 | 1.00 (0.99–1.00) | .35 |
| Model 2 | 1.00 (0.99–1.01) | .30 |
| Model 3 | 1.00 (0.99–1.01) | .32 |
| Model 4 | 1.00 (0.99–1.01) | .96 |
OR represents the change in the odds of being in a higher CAC category per 10-min increase in T50. CAC categories were 0, 1–99, 100–399, and ≥400 Agatston units. OR, odds ratio; CI, confidence interval; CAC, coronary artery calcification; T50, serum calcification propensity. Model definitions are described in the Materials and methods section.
Table 5:
Continuous regression analysis of the association between T50 and CAC.
| Model | Per 10-min increase in T50 β (95% CI) | P-value |
|---|---|---|
| Crude model | 0.02 (0.01–0.03) | <.001 |
| Model 1 | −0.01 (−0.02–0.00) | .08 |
| Model 2 | −0.01 (−0.02–0.00) | .07 |
| Model 3 | −0.01 (−0.02–0.00) | .07 |
| Model 4 | −0.01 (−0.02–0.01) | .35 |
β represents the change in log(CAC + 1) per 10-min increase in T50. CAC was analysed continuously as log(CAC + 1). β, regression coefficient; CI, confidence interval; CAC, coronary artery calcification; T50, serum calcification propensity. Model definitions are described in the Materials and methods section.
Adjustment for eGFR or phosphate alone does not alter the crude association between T50 and CAC
To evaluate whether kidney function or serum phosphate influenced the association between T50 and CAC categories, we ran two additional sensitivity models: one adjusting only for eGFR and one adjusting only for phosphate. In the crude model, higher T50 values were associated with higher CAC categories. When eGFR was added, the estimates for T50 remained similar to the crude model across all CAC groups, and the associations stayed statistically significant across CAC categories (Table 6). eGFR itself was inversely associated with higher CAC categories, with ORs between 0.91 and 0.96 and P-values <.001. In the model adjusted for phosphate, the association between T50 and higher CAC categories remained present, with effect estimates similar in magnitude to those seen in the crude and eGFR-adjusted models. Phosphate itself was not significantly associated with CAC, and its inclusion did not attenuate the T50 estimates. Together, these findings show that the crude association between T50 and CAC appears largely unchanged when adjusting for either eGFR or phosphate alone, and that phosphate does not change this association.
Table 6:
Sensitivity analyses for T50 across CAC categories using minimally adjusted multinomial logistic regression models.
| Model | Predictor | CAC 1–99 OR (95% CI) | P-value | CAC 100–399 OR (95% CI) | P-value | CAC ≥ 400 OR (95% CI) | P-value |
|---|---|---|---|---|---|---|---|
| Crude model | T50 | 1.015 (1.003–1.025) | .01 | 1.015 (1.001–1.030) | .03 | 1.034 (1.018–1.050) | <.001 |
| Crude + eGFR | T50 | 1.012 (1.001–1.023) | .01 | 1.011 (0.997–1.026) | .14 | 1.023 (1.006–1.041) | .01 |
| eGFR | 0.96 (0.94–0.98) | <.001 | 0.93 (0.91–0.95) | <.001 | 0.91 (0.89–0.93) | <.001 | |
| Crude + phosphate | T50 | 1.02 (1.001–1.029) | < .001 | 1.015 (1.000–1.028) | .04 | 1.032 (1.017–1.046) | <.001 |
| Phosphate | 1.67 (0.93–2.99) | .08 | 0.90 (0.45–1.79) | .76 | 0.73 (0.40–1.33) | .31 |
ORs were derived from multinomial logistic regression models with CAC = 0 as the reference category. ORs for T50 represent the change per 10-min increase in T50. OR, odds ratio; CI, confidence interval; CAC, coronary artery calcification; eGFR, estimated glomerular filtration rate.
No consistent associations between T50 and CAC are observed across subgroups
Because the main analyses showed no independent association between T50 and CAC, we examined whether this relationship differed across clinically relevant subgroups. As no specific hypotheses for effect modification were defined beforehand, these analyses were exploratory. Subgroup analyses were performed for age, BMI, sex, diabetes, systolic blood pressure, LDL cholesterol, medication use, eGFR, and serum phosphate and are presented in the Supplementary material (Fig. S1 and Table S1).
Based on biological plausibility, we focused our interpretation on diabetes status and serum phosphate. In exploratory subgroup analyses, small statistically significant associations between T50 and severe CAC were observed in both participants with and without diabetes (per 10-min increase in T50), with effect estimates close to unity. A similar pattern was observed for serum phosphate, with a modest association between T50 and severe CAC in participants with lower phosphate levels, while no significant association was observed in those with higher phosphate levels.
DISCUSSION
We hypothesized that lower T50 values would be associated with the presence and severity of CAC in this cross-sectional study; however, this hypothesis was not confirmed. These findings suggest that established CV risk factors, particularly age, may outweigh any potential contribution of T50 to CAC in this population. We therefore conclude that serum T50 was not independently associated with CAC in early-stage CKD. In our population-based cohort, participants had predominantly mildly to moderately reduced kidney function, were relatively young, and predominantly female. Notably, previous studies were conducted in populations with more advanced kidney disease, where vascular calcification is often more established.
In CKD stages 2–4, increases in fibroblast growth factor 23 may precede overt hyperphosphataemia and secondary hyperparathyroidism, indicating that compensatory changes in mineral metabolism occur before the development of an advanced chronic kidney disease–mineral and bone disorder (CKD-MBD) phenotype [29]. In this setting, prevalent CAC may reflect cumulative atherosclerotic burden and conventional CV risk factors more strongly than current serum calcification propensity. Age, male sex, hypertension, BMI, and diabetes have been associated with the incidence and progression of CAC in population-based studies [30]. Endothelial dysfunction and other non-mineral pathways may also contribute to early atherosclerosis, although these mechanisms were not assessed in the present study. The marked attenuation of the T50–CAC association after adjustment for age, sex, and BMI is consistent with confounding by conventional CV risk factors, but does not establish the mechanism underlying CAC.
T50 was comparatively long in the present cohort, with a mean of 429 min. Mean T50 values of 313 and 329 min have been reported in participants with CKD stages 2–4 in CRIC and in the PREVEND general-population cohort, respectively [13, 16]. A mean T50 of 335 min was reported in stages 3b–4 CKD, whereas baseline values of 233 and 247 min were observed in a haemodialysis trial [31, 32]. The longer T50 in the present cohort is consistent with lower serum calcification propensity.
Previous studies focusing on the association between serum T50 and CAC score in CKD populations have yielded variable results [16, 33–35]. The absence of an independent association with prevalent CAC does not exclude an association between T50 and subsequent CAC progression. CAC measured at a single time point reflects calcification accumulated over several years and may have developed before measurable loss of kidney function. In contrast, T50 reflects serum calcification propensity at the time of sampling and may therefore be more closely related to ongoing calcification than to established calcification. In the CRIC study, lower T50 was not independently associated with CAC prevalence or incident CAC, but was associated with greater CAC severity among participants with prevalent CAC and with subsequent CAC progression [16]. Longitudinal studies with repeated CAC measurements are required to determine whether T50 predicts calcification progression in individuals with early CKD.
Moreover, T50 was not associated with the presence of vascular calcification in a small cross-sectional study of patients with stages 4–5 CKD, in which calcification was assessed using radiographic scoring methods rather than CAC scoring [33].
A subsequent study from the same group assessed 400 incident haemodialysis patients [34]. In this cohort, T50 was also not associated with either the presence or the burden of calcification at baseline. In contrast, a cross-sectional study in kidney transplant patients identified an association between lower T50 and higher baseline CAC in the TransplantLines study [35].
Differences in study populations may further contribute to the discrepant findings. The TransplantLines cohort consisted of kidney transplant recipients ∼1 year post-transplantation, a population characterized by prior advanced kidney disease, altered mineral metabolism, and a relatively high burden of established vascular calcification [35]. In addition, cardiometabolic risk profiles differed substantially: diabetes mellitus was present in 38% of participants in the lowest T50 tertile compared with 20% in the highest tertile (P = .01). In contrast, our cohort comprised individuals with mildly to moderately reduced kidney function from a population-based setting, with lower diabetes prevalence and lower CV risk. Such differences in underlying CV risk may have contributed to the absence of significant results in our cohort.
Supporting the notion that T50 is only representative of CAC in high-risk populations, lower T50 values were associated with greater abdominal aortic calcification in primary aldosteronism [36]. Similarly, in a large cohort of individuals with type 2 diabetes mellitus, lower T50 was independently associated with higher CAC [37]. Of note, CAC scoring does not allow differentiation between intimal and medial calcification, and therefore provides limited insight into the underlying calcification phenotype [38]. CAC may therefore predominantly reflect atherosclerotic (intimal) calcification in individuals with early stage CKD [39].
Adjustment for age, sex, and BMI markedly attenuated the association between T50 and CAC. Previous studies suggest that the associations of these characteristics with T50 vary across populations. In the PREVEND cohort, older age was independently associated with shorter T50 [13]. Among individuals with type 2 diabetes, women had shorter T50 than men in one cohort, whereas age was not associated with T50; in a subsequent cohort, older age, but not BMI, was independently associated with shorter T50 [40, 41]. By contrast, studies in predialysis CKD and haemodialysis populations identified phosphate, calcium, fetuin-A, pyrophosphate, albumin, and bicarbonate as more prominent determinants of T50 [17, 19]. These findings indicate that age and sex may influence confounding of the T50–CAC association in a population-specific manner, whereas BMI is not an established independent determinant of T50.
Although we did not observe an independent association with CAC, it is important to note that our study does not contradict prior evidence linking lower T50 to adverse CV outcomes. In several cohorts, including kidney transplant recipients, patients receiving haemodialysis, and individuals from the general population, lower T50 values have been associated with CV and all-cause mortality, as well as progression of arterial stiffness [13, 14, 18, 8, 42]. These findings suggest that T50 may reflect vascular dysfunction and other CV risk factors, rather than CAC per se.
A major strength of our study is the large, well-characterized, population-based cohort allowed for detailed multi-variable analyses and subgroup exploration. Studying mildly to moderately reduced kidney function could provide useful insight into vascular calcification in an earlier phase of kidney decline, before advanced irreversible calcification develops.
Several limitations should be considered. Kidney function was assessed at a single time point without inclusion of albuminuria, which may have led to misclassification of CKD status. Finally, the study population was derived from a cohort in the Northern Netherlands and was predominantly female, which may limit generalizability. These findings are likely generalizable to populations with mildly to moderately reduced kidney function in similar population-based settings. Residual confounding cannot be excluded, and the cross-sectional design precludes assessment of longitudinal associations. The relevance of T50 for CAC progression therefore remains to be clarified in longitudinal studies.
In this population-based cohort of individuals with mildly to moderately reduced kidney function, serum T50 was not independently associated with the presence or severity of CAC after multi-variable adjustment. In conclusion, serum T50 is not independently associated with CAC in early CKD.
Supplementary Material
ACKNOWLEDGEMENTS
The authors wish to acknowledge the services of the Lifelines Cohort Study, the contributing research centres delivering data to Lifelines, and all the study participants. The Lifelines study, including the ImaLife subcohort, was approved by the Medical Ethical Review Committee of the University Medical Center Groningen (METC UMCG 2007/152). All participants provided written informed consent prior to participation.
Contributor Information
Guido de La Roij, Department of Medical BioSciences, Radboud University Medical Center, GA Nijmegen, The Netherlands.
Kevin Ocran, Department of Medical BioSciences, Radboud University Medical Center, GA Nijmegen, The Netherlands.
Marc Vervloet, Department of Nephrology, Radboud University Medical Center, GA Nijmegen, The Netherlands.
Andreas Pasch, Institute for Physiology and Pathophysiology, Johannes Kepler University Linz, Linz, Austria; Calciscon AG, Biel, Switzerland.
Rozemarijn Vliegenthart, Department of Radiology, University Medical Center Groningen, University of Groningen, AV Groningen, The Netherlands.
Gerjan Navis, Department of Internal Medicine, Division of Nephrology, University Medical Center Groningen, University of Groningen, AV Groningen, The Netherlands.
Martin H de Borst, Department of Internal Medicine, Division of Nephrology, University Medical Center Groningen, University of Groningen, AV Groningen, The Netherlands.
Joost G J Hoenderop, Department of Medical BioSciences, Radboud University Medical Center, GA Nijmegen, The Netherlands.
Jeroen H F de Baaij, Department of Medical BioSciences, Radboud University Medical Center, GA Nijmegen, The Netherlands.
CONFLICT OF INTEREST STATEMENT
A.P. is an inventor of the T50 test, and president and stock holder of Calciscon AG, Biel, Switzerland, a company commercializing the T50 test. M.H.de.B. received consultancy fees and grant/other support from Boehringer, Ingelheim, CSL Vifor, and Eli Lilly, all to institution. R.V. is supported by institutional research grants from Siemens Healthineers, and reports honorarium from invited lectures from Keya Medical, Wiley, Bayer Healthcare and Siemens Healthineers.
FUNDING
This work was funded by the Dutch Kidney Foundation (21OK022, 23OK2065). J.D.B. is supported by the European Research Council (101040682, IN-THE-KIDNEY) and the Netherlands Organization for Scientific Research (VIDI 09150172110040, IMAGE-THE-KIDNEY). The ImaLife study is supported by an institutional research grant from Siemens Healthineers and by the Ministry of Economic Affairs and Climate Policy by means of the PPP Allowance. The Lifelines initiative has been made possible by subsidy from the Dutch Ministry of Health, Welfare and Sport, the Dutch Ministry of Economic Affairs, the University Medical Center Groningen (UMCG), University of Groningen, and the Provinces in the North of the Netherlands (Drenthe, Friesland, Groningen). The work is also supported by Health Holland Stichting LSH-TKI project TREATABLE. The funders had no role in the study design, data collection, analysis, interpretation of the data, or in the decision to submit the manuscript for publication.
DATA AVAILABILITY STATEMENT
Data may be obtained from a third party and are not publicly available. Researchers can apply to use the Lifelines data used in this study. More information about how to request Lifelines data and the conditions of use can be found on their website (https://www.lifelines-biobank.com/researchers/working-with-us).
REFERENCES
- 1. Kovesdy CP. Epidemiology of chronic kidney disease: an update 2022. Kidney Int Suppl. 2022;12:7–11. 10.1016/j.kisu.2021.11.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Palit S, Kendrick J. Vascular calcification in chronic kidney disease: role of disordered mineral metabolism. Curr Pharm Des. 2014;20:5829–33. 10.2174/1381612820666140212194926. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Wang XR, Zhang JJ, Xu XX et al. Prevalence of coronary artery calcification and its association with mortality, cardiovascular events in patients with chronic kidney disease: a systematic review and meta-analysis. Ren Fail. 2019;41:244–56. 10.1080/0886022X.2019.1595646. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Kim JS, Hwang HS. Vascular calcification in chronic kidney disease: distinct features of pathogenesis and clinical implication. Korean Circ J. 2021;51:961–82. 10.4070/kcj.2021.0995. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Hutcheson JD, Goettsch C. Cardiovascular calcification heterogeneity in chronic kidney disease. Circ Res. 2023;132:993–1012. 10.1161/CIRCRESAHA.123.321760. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Miki T, Sakoda T, Yamamoto K et al. Development and validation of a prediction model for people with mild chronic kidney disease in Japanese individuals. BMC Nephrol. 2024;25:339. 10.1186/s12882-024-03786-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Pasch A, Farese S, Gräber S et al. Nanoparticle-based test measures overall propensity for calcification in serum. J Am Soc Nephrol. 2012;23:1744–52. 10.1681/ASN.2012030240. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Dahle DO, Åsberg A, Hartmann A et al. Serum calcification propensity is a strong and independent determinant of cardiac and all-cause mortality in kidney transplant recipients. Am J Transplant. 2016;16:204–12. 10.1111/ajt.13443. [DOI] [PubMed] [Google Scholar]
- 9. Bielesz B, Reiter T, Marculescu R et al. Calcification propensity of serum is independent of excretory renal function. Sci Rep. 2017;7:17941. 10.1038/s41598-017-18336-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Greenland P, Blaha MJ, Budoff MJ et al. Coronary calcium score and cardiovascular risk. J Am Coll Cardiol. 2018;72:434–47. 10.1016/j.jacc.2018.05.027. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Haq A, Veerati T, Walser-Kuntz E et al. Coronary artery calcium and the risk of cardiovascular events and mortality in younger adults: a meta-analysis. Eur J Prev Cardiol. 2024;31:1061–9. 10.1093/eurjpc/zwad399. [DOI] [PubMed] [Google Scholar]
- 12. Detrano R, Guerci AD, Carr JJ et al. Coronary calcium as a predictor of coronary events in four racial or ethnic groups. N Engl J Med. 2008;358:1336–45. 10.1056/NEJMoa072100. [DOI] [PubMed] [Google Scholar]
- 13. Eelderink C, te Velde-Keyzer CA, Frenay A-RS et al. Serum calcification propensity and the risk of cardiovascular and all-cause mortality in the general population: the PREVEND study. Arterioscler Thromb Vasc Biol. 2020;40:1942–51. 10.1161/ATVBAHA.120.314187. [DOI] [PubMed] [Google Scholar]
- 14. Pluquet M, Kamel S, Choukroun G et al. Serum calcification propensity represents a good biomarker of vascular calcification: a systematic review. Toxins. 2022;14:637. 10.3390/toxins14090637. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Mencke R, Al Ali L, de Koning M-SLY et al. Serum calcification propensity is increased in myocardial infarction and hints at a pathophysiological role independent of classical cardiovascular risk factors. Arterioscler Thromb Vasc Biol. 2024;44:1884–94. 10.1161/ATVBAHA.124.320974. [DOI] [PubMed] [Google Scholar]
- 16. Bundy JD, Cai X, Scialla JJ et al. Serum calcification propensity and coronary artery calcification among patients with CKD: the CRIC (Chronic Renal Insufficiency Cohort) study. Am J Kidney Dis. 2019;73:806–14. 10.1053/j.ajkd.2019.01.024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Smith ER, Ford ML, Tomlinson LA et al. Serum calcification propensity predicts all-cause mortality in predialysis CKD. J Am Soc Nephrol. 2014;25:339–48. 10.1681/ASN.2013060635. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Keyzer CA, de Borst MH, van den Berg E et al. Calcification propensity and survival among renal transplant recipients. J Am Soc Nephrol. 2016;27:239–48. 10.1681/ASN.2014070670. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Pasch A, Block GA, Bachtler M et al. Blood calcification propensity, cardiovascular events, and survival in patients receiving hemodialysis in the EVOLVE trial. Clin J Am Soc Nephrol. 2017;12:315–22. 10.2215/CJN.04720416. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Klijs B, Scholtens S, Mandemakers JJ et al. Representativeness of the LifeLines cohort study. PLoS One. 2015;10:e0137203. 10.1371/journal.pone.0137203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Stolk RP, Rosmalen JGM, Postma DS et al. Universal risk factors for multifactorial diseases: LifeLines: a three-generation population-based study. Eur J Epidemiol. 2008;23:67–74. 10.1007/s10654-007-9204-4. [DOI] [PubMed] [Google Scholar]
- 22. Scholtens S, Smidt N, Swertz M et al. LifeLines, a three-generation cohort study and biobank. Int J Epidemiol. 2015;44:1172–8. 10.1093/ije/dyu229 [DOI] [PubMed] [Google Scholar]
- 23. Sijtsma A, Rienks J, van der Harst P et al. Cohort Profile Update: Lifelines, a three-generation cohort study and biobank. Int J Epidemiol. 2022;51:e295–e302. 10.1093/ije/dyab257. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Xia C, Rook M, Pelgrim GJ et al. Early imaging biomarkers of lung cancer, COPD and coronary artery disease in the general population: rationale and design of the ImaLife (Imaging in Lifelines) Study. Eur J Epidemiol. 2020;35:75–86. 10.1007/s10654-019-00519-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Xiao H, Wang X, Yang P et al. Coronary artery calcium scoring assessment in ultra-low-dose chest computed tomography. Clin Imaging. 2024;106:110045. 10.1016/j.clinimag.2023.110045. [DOI] [PubMed] [Google Scholar]
- 26. Boersma HE, Xia C, van der Klauw MM et al. Association between skin autofluorescence and coronary calcification in the general population. PLoS One. 2024;19:e0309059. 10.1371/journal.pone.0309059. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Agatston AS, Janowitz WR, Hildner FJ et al. Quantification of coronary artery calcium using ultrafast computed tomography. J Am Coll Cardiol. 1990;15:827–32. 10.1016/0735-1097(90)90282-T. [DOI] [PubMed] [Google Scholar]
- 28. Inker LA, Eneanya ND, Coresh J et al. New creatinine- and cystatin C-based equations to estimate GFR without race. N Engl J Med. 2021;385:1737–49. 10.1056/NEJMoa2102953. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Isakova T, Wahl P, Vargas GS et al. Fibroblast growth factor 23 is elevated before parathyroid hormone and phosphate in chronic kidney disease. Kidney Int. 2011;79:1370–78. 10.1038/ki.2011.47 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Kronmal RA, McClelland RL, Detrano R et al. Risk factors for the progression of coronary artery calcification in asymptomatic subjects: results from the Multi-Ethnic Study of Atherosclerosis (MESA). Circulation. 2007;115:2722–30. 10.1161/CIRCULATIONAHA.106.674143 [DOI] [PubMed] [Google Scholar]
- 31. Pluquet M, Laville SM, Brazier F et al. The effect of sevelamer on serum calcification propensity in patients with chronic kidney disease: the results of a multicentre, double-blind, placebo-controlled, randomized clinical trial. Clin Kidney J. 2025;18:sfae343. 10.1093/ckj/sfae343 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Bressendorff I, Hansen D, Schou M et al. The effect of increasing dialysate magnesium on serum calcification propensity in subjects with end stage kidney disease: a randomized, controlled clinical trial. Clin J Am Soc Nephrol. 2018;13:1373–80. 10.2215/CJN.13921217 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Chen W, Anokhina V, Dieudonne G et al. Patients with advanced chronic kidney disease and vascular calcification have a large hydrodynamic radius of secondary calciprotein particles. Nephrol Dial Transplant. 2019;34:992–1000. 10.1093/ndt/gfy117. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Chen W, Fitzpatrick J, Monroy-Trujillo JM et al. Associations of serum calciprotein particle size and transformation time with arterial calcification, arterial stiffness, and mortality in incident hemodialysis patients. Am J Kidney Dis. 2021;77:346–54. 10.1053/j.ajkd.2020.05.031. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. van der Vaart A, Kremer D, Smith ER et al. Posttransplant dynamics of serum T50 and coronary artery calcium scores among kidney transplant recipients. Kidney Int Rep. 2025;108:4069–4072. 10.1016/j.ekir.2025.08.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Kantauskaite M, Bolten K, Boschheidgen M et al. Serum calcification propensity and calcification of the abdominal aorta in patients with primary aldosteronism. Front Cardiovasc Med. 2022;9:771096. 10.3389/fcvm.2022.771096. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Meer R, Hoek AG, Dal Canto E et al. Calciprotein crystallization time (T50) and its association with surrogate cardiovascular disease risk markers in individuals with type 2 diabetes mellitus: the cross-sectional EARLY-HFpEF study. Cardiovasc Diabetol. 2025;25:10. 10.1186/s12933-025-03016-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Yoshida K, Sugawara H, Mizuyama H et al. Association between ankle-brachial index and coronary artery calcification score in patients beginning hemodialysis. Cureus. 2025;17:e87749. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Nakamura S, Ishibashi-Ueda H, Niizuma S et al. Coronary calcification in patients with chronic kidney disease and coronary artery disease. Clin J Am Soc Nephrol. 2009;4:1892–900. 10.2215/CJN.04320709. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Mencke R, van der Vaart A, Pasch A Serum calcification propensity is associated with HbA1c in type 2 diabetes mellitus. BMJ Open Diabetes Res Care. 2021;9:e002016. 10.1136/bmjdrc-2020-002016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. van der Vaart A, Eelderink C, van Goor H et al. Serum T50 predicts cardiovascular mortality in individuals with type 2 diabetes: a prospective cohort study. J Intern Med. 2024;295:748–58. 10.1111/joim.13781 [DOI] [PubMed] [Google Scholar]
- 42. Zawada AM, Wolf M, Rincon Bello A et al. Assessment of a serum calcification propensity test for the prediction of all-cause mortality among hemodialysis patients. BMC Nephrol. 2023;24:35. 10.1186/s12882-023-03069-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data may be obtained from a third party and are not publicly available. Researchers can apply to use the Lifelines data used in this study. More information about how to request Lifelines data and the conditions of use can be found on their website (https://www.lifelines-biobank.com/researchers/working-with-us).