2025 Volume 32 Issue 6 Pages 763-774
Aim: Patients with type 2 diabetes mellitus (T2D) are prone to develop vascular calcification. Fetuin-A protects against vascular calcification but it increases insulin resistance. T50 calciprotein crystallization (also called serum calcification propensity) is a novel marker of calcification stress. This study examined whether T2D affects T50 and the potential role of fetuin-A in the relationship between T2D and T50.
Methods: This cross-sectional study included 101 individuals with T2D and 101 individuals without diabetes (controls). T50 and fetuin-A levels were measured using the established nephelometric method and an enzyme-linked immunosorbent assay, respectively.
Results: Although fetuin-A levels were higher in the T2D group, T50 was not significantly different between the T2D and control groups. In multivariable-adjusted analyses of the total population, T50 was not independently associated with the presence of T2D, fasting plasma glucose, or HbA1c, whereas T50 was significantly associated with fetuin-A, phosphate, and calcium levels. The association between T50 and fetuin-A was modified by the presence of T2D. A subgroup analysis revealed that the positive association between T50 and fetuin-A was significant but smaller in the T2D group, and that the associations of T50 with serum phosphate and calcium were more evident in the T2D group. Additional analyses showed that T50/fetuin-A ratio was lower in the T2D group and that T50/fetuin-A ratio was inversely correlated with fasting glucose and HbA1c levels.
Conclusions: T2D itself was not significantly associated with T50 but T2D modified the association between T50 and fetuin-A in favor of developing vascular calcification in T2D.
Individuals with type 2 diabetes (T2D) have a higher prevalence and extent of vascular calcification than those without diabetes1-3). However, patients with T2D have no apparent abnormalities in their serum calcium or phosphate levels, which are considered key factors for vascular calcification. Thus, more information is required to explain the link between T2D and advanced vascular calcification.
Serum calcification propensity (T50)4), which was recently renamed the T50 calciprotein crystallization test5), is an inverse biomarker of calcification stress6). Although plasma is hypersaturated with calcium and phosphate, precipitation of calcium phosphate crystals does not occur. This is inhibited by a liver-derived circulating glycoprotein fetuin-A, which adsorbs amorphous calcium phosphate clusters and prevents them from growing into larger crystalline precipitates7). Fetuin-A and calcium phosphate clusters form nano-sized particles called calciprotein particles (CPPs), which disperse as colloids8). Small and spherical CPPs (primary CPPs) containing amorphous calcium phosphate clusters spontaneously aggregate and transform into larger and irregular-shaped ones (secondary CPPs) containing crystallized calcium phosphate. Secondary CPPs have cytotoxic and pro-inflammatory properties, and can promote inflammation, atherosclerosis, and vascular calcification9). The time required to transform primary to secondary CPPs in vitro (T50) can be measured using the assay system developed by Pasch et al.4). A serum sample with a shorter T50 is interpreted as having a higher propensity for calciprotein crystallization or higher calcification stress4). In clinical studies of patients with chronic kidney disease (CKD), a shorter T50 has been shown to be associated with arterial calcification10) and a higher risk for cardiovascular events and all-cause mortality10-15). In addition to many studies on patients with CKD10, 16), recent studies have provided information regarding T50 in other populations, such as the general population17), systemic lupus erythematosus18), and primary aldosteronism19). Thus, T50 appears to be an established biomarker of calcification stress6). The addition of calcium or phosphate to the assay system decreased T50, whereas the addition of magnesium, fetuin-A, or albumin increased the T50 values4).
Mixed results have been reported regarding the possible changes in T50 associated with T2D and/or hyperglycemia. Mencke et al.20) showed an independent and inverse association between HbA1c and T50 levels in T2D patients. In contrast, Nakatani et al.21) showed an independent and positive association of T50 with fasting plasma glucose, but not with HbA1c, in T2D. According to Eelderink et al.17), plasma glucose showed an independent and positive association with T50 in a large sample of the general population. To date, no previous study has compared T50 levels between individuals with and without T2D. In addition, it is possible that the presence of T2D modifies the relationship between T50 and some factors known to affect T50, such that patients with T2D are more prone to vascular calcification than individuals without diabetes.
The purpose of this study was to examine whether T50 is altered in T2D and to explore whether the presence of T2D could modify the association between T50 and any factors, such as calcium, phosphate, magnesium, fetuin-A, and albumin, which are known to affect T50.
This was a cross-sectional study of individuals selected from the MedCity21 health examination registry, Osaka Metropolitan University, Osaka, Japan. Data available for analysis were extracted from the registry database. We measured T50, fetuin-A, and magnesium using freshly frozen serum samples that were collected at the same time as the health examination and stored at −80℃ because these data were not available from the database.
Ethical ConsiderationsThis study was conducted in accordance with the latest version of the Declaration of Helsinki and Ethical Guidelines for Clinical Studies by the Ministry of Health, Labor, and Welfare, Japan. The MedCity21 health examination registry protocol was reviewed and approved by the ethics committee of Osaka City University Graduate School of Medicine (Approval No. 2927, September 01, 2014). All participants provided written informed consent before participation. The protocol of the present analysis was also reviewed and approved by the ethics committee of Osaka City University Graduate School of Medicine (Approval No. 2021-197, December 15, 2021). Informed consent was obtained using the opt-out method because of the observational nature of the study.
Selection of ParticipantsEligible participants for this study were screened from a total of 1358 records of participants of health examinations, including vascular checkups, between March 2015 and September 2021 in the database of the MediCity21 health examination registry of Osaka Metropolitan University. No specific inclusion criteria were established for the study. The exclusion criteria were (1) liver dysfunction exceeding x3 of the upper limit of the reference range of aspartate transaminase or alanine aminotransferase; (2) chronic respiratory failure or chronic obstructive lung disease; (3) osteoporosis receiving medical treatment; and (4) diabetes other than T2D (type 1 diabetes, gestational diabetes, and/or specific types of diabetes due to other causes). These criteria were set to avoid patients with conditions that could affect the T50 values and to exclude patients with diabetes other than T2D. We excluded individuals for whom the presence of diabetes could not be defined because of missing fasting glucose and/or HbA1c values. We used data from the first visit if the participants had repeated measurements on different occasions. Eligible individuals were divided into two groups: T2D and control. We then selected the same number of subjects from each group who were similar in age and sex, without knowing other information. For this purpose, we first divided the records of eligible individuals without diabetes (N = 805) into age groups (30–39 years, 40–49 years, 50–59 years, 60–69 years, and ≥ 70 years) for men and women. Then, we sorted the participants into lists using the study ID. Finally, we selected the same number of individuals without diabetes as the T2D group for each age category from the top of the list for men and women separately.
Definition of Diabetes and Classification of T2DIn this study, diabetes was defined if any of the following criteria were met: (1) fasting glucose ≥ 126 mg/dL, (2) HbA1c ≥ 6.5%, and/or (3) current use of any medication for diabetes22). As we excluded patients with type 1 diabetes, gestational diabetes, and/or specific types of diabetes due to other causes, the individuals with diabetes in this analysis were classified as those with T2D.
Measurements of T50, Fetuin-A, and MagnesiumWe measured T50 in our laboratory16, 21) according to the method described by Pasch et al.4). Freshly frozen serum samples stored at −80℃ were carefully thawed on ice for 3 h and kept in iced water for 1 h. The test serum samples (80 µL per well) were placed in a 96-well microplate containing NaCl solution (20 µL per well). The microplate was placed in a nephelometric apparatus (Nephelostar Plus®, BGM Labtech, Ortenberg, Germany) equipped with a thermoconstant room (controlled at 36–37℃). Pre-warmed phosphate-containing solution (50 µL per well) and calcium-containing solution (50 µL per well) were added using an automated dispenser to start the measurement. Continuous nephelometric monitoring was performed for 600 min to detect changes in signals caused by the transformation of primary CPPs to secondary CPPs. T50 level was determined using the analysis software program MARS®, and the results were expressed in minutes. We used the average of duplicate measurements for statistical analysis. The T50 assay in our laboratory was validated against the gold standard measurement at Calciscon AG (Nidau, Switzerland)16). For quality control, each assay included two serum controls. The intra- and inter-assay coefficients of variation were less than 4.5%16).
Serum fetuin-A levels were determined by enzyme-linked immunoassay using a commercial assay kit (Human fetuin-A ELISA kit, BioVender, Brno, Czech Republic) as previously reported23, 24). Serum magnesium was measured by an enzymatic assay (L-type Wako Mg·N, FUJIFILM Wako Pure Chemical Co. Ltd., Osaka, Japan) using an automated analyzer.
Other VariablesWe extracted clinical data on age, sex, height, body weight, blood pressure, smoking status, past history, comorbidities, medication use, and fasting laboratory data from the MediCity21 Health Examination Registry database. In this study, serum calcium denotes serum total calcium, which was not adjusted for serum albumin. Body mass index (BMI) was calculated as the body weight (kg) divided by the squared height (m2). Estimated glomerular filtration rate (eGFR) was calculated from age, sex, and serum creatinine concentration using a formula for Japanese individuals25). Hypertension was defined as a systolic blood pressure ≥ 140 mmHg, a diastolic blood pressure ≥ 90 mmHg, and/or the current use of any medication for hypertension26). Dyslipidemia was defined as low-density lipoprotein cholesterol ≥ 140 mg/dL, triglyceride ≥ 150 mg/dL, high-density lipoprotein cholesterol ＜40 mg/dL in the fasting state, and/or the current use of any medication for dyslipidemia27,28).
Statistical AnalysisWe summarized the clinical characteristics of the T2D and control groups using medians [25th and 75th percentile levels] for continuous variables and numbers (percentages) for categorical variables. Comparisons between two groups were conducted using the Mann–Whitney U test or Fisher’s exact test.
To examine whether T50 was associated with the presence of T2D, we first compared T50 values between the two groups using the Mann–Whitney U test. A multiple regression model was used to examine whether T50 was independently associated with the presence of T2D when adjusted for potential confounders. Although we initially planned to adjust for eight factors including age, sex, eGFR, serum calcium, phosphate, magnesium, albumin, and fetuin-A in the total study population, the model was further adjusted for four factors that were significantly different between the T2D and control groups (BMI, current smoking, hypertension, and dyslipidemia). In these multiple regression analyses, the log-transformed value for T50 level was included in the model to meet the assumption of a normal distribution of the residual. Using the same covariates, we examined the role of hyperglycemia by replacing T2D with fasting plasma glucose or HbA1c levels. Finally, we explored the possible effect modification by the presence of T2D on the association between T50 and serum calcium, phosphate, magnesium, albumin, and fetuin-A by inserting the interaction term between T50 and each of these candidates. If the effect modification was significant, then a subgroup analysis was conducted in the T2D and control groups.
Because we found a significant effect modification by the presence of T2D on the relationship between T50 and fetuin-A, we conducted additional analysis using the ratio of T50 to fetuin-A (T50/fetuin-A ratio) as a simple index for the protective function of fetuin-A per unit against calciprotein crystallization. We compared T50/fetuin-A ratio between the T2D and control groups using the Mann–Whitney U test, and we examined the correlations of T50/fetuin-A ratio with fasting plasma glucose and HbA1c levels using Spearman’s rank correlation.
These statistical calculations were conducted using JMP version 14.0 (SAS Institute Japan, Tokyo, Japan) and the free software EZR (with R version 4.1.2), which was developed by Professor Kanda at Jichi Medical University Saitama Medical Center29). A two-sided P value of less than 0.05 was to be considered statistically significant.
Fig.1 shows the selection of the participants for this analysis. From a total of 1358 records, we excluded records of the second and later visits for those who had repeated measurements. Further, we excluded 56 individuals because of the exclusion criteria and one individual because of missing data that were needed to define the presence of diabetes. Thus, we identified 908 eligible participants, including 103 with diabetes. Because serum samples of two persons with diabetes were missing, we finally selected 101 individuals with T2D (T2D group) and 101 individuals without diabetes (control group), who were similar in age and sex.
Eligible individuals were divided into two groups: T2D and control. We then selected the same number of subjects from each group who were similar in age and sex, without knowing other information. The method for selecting the control group is explained in detail in the text.
Abbreviations: COPD, chronic obstructive pulmonary disease; CHF, congestive heart failure; T2D, type 2 diabetes; HbA1c, hemoglobin A1c.
Table 1 summarizes the clinical characteristics of the individuals in the T2D and control groups. In addition to the plasma glucose and HbA1c levels, the two groups showed significant differences in BMI and the prevalence of dyslipidemia, hypertension, and current smokers. Regarding the factors potentially affecting T50, no difference was found in age, sex, calcium, phosphate, magnesium, or albumin levels between the two groups, whereas fetuin-A levels were significantly higher in the T2D group.
| Characteristics | Control group | T2D group | P value |
|---|---|---|---|
| Number of participants (N, %) | 101 (100%) | 101 (100%) | --- |
| Male sex (N, %) | 70 (69.3%) | 71 (70.3%) | 0.880 |
| Female sex (N, %) | 31 (30.7%) | 30 (29.7%) | |
| Age (year) | 63 [53, 70] | 64 [55, 71] | 0.650 |
| BMI (kg/m2) | 24.1 [22.1, 25.4] | 25.1 [23.1, 27.6] | 0.011 |
| Prior CVD (N, %) | 5 (4.9%) | 6 (5.9%) | 0.759 |
| T2D (N, %) | 0 (0%) | 101 (100%) | ＜0.001 |
| Fasting plasma glucose (mg/dL) | 102 [98, 108] | 134 [126, 154] | ＜0.001 |
| HbA1c (%) | 5.7 [5.5, 5.9] | 6.7 [6.4, 7.4] | ＜0.001 |
| Use of medication for diabetes (N, %) | 0 (0%) | 48 (47.5%) | ＜0.001 |
| Dyslipidemia (N, %) | 44 (43.5%) | 68 (67.3%) | ＜0.001 |
| Total cholesterol (mg/dL) | 199 [174, 224] | 189 [174, 214] | 0.114 |
| Triglycerides (mg/dL) | 87 [69, 115] | 113 [82, 172] | ＜0.001 |
| HDL-C (mg/dL) | 58 [50, 70] | 50 [43, 62] | ＜0.001 |
| LDL-C (mg/dL) | 111 [95, 132] | 113 [92, 124] | 0.324 |
| Use of medication for dyslipidemia | 19 (18.8%) | 43 (42.6%) | ＜0.001 |
| Hypertension (N, %) | 39 (38.6%) | 61 (60.4%) | 0.001 |
| Systolic blood pressure (mmHg) | 124 [114, 132] | 126 [116, 138] | 0.187 |
| Diastolic blood pressure (mmHg) | 76 [69, 81] | 76 [70, 84] | 0.815 |
| Use of medication for hypertension | 29 (28.7%) | 48 (47.5%) | 0.006 |
| Current smoking (N, %) | 14 (13.8%) | 33 (32.7%) | 0.003 |
| eGFR (mL/min/1.73 m2) | 71.1 [62.9, 80.0] | 74.8 [61.6, 84.8] | 0.261 |
| Serum calcium (mg/dL) | 9.3 [9.2, 9.6] | 9.4 [9.2, 9.6] | 0.646 |
| Serum phosphate (mg/dL) | 3.3 [3.1, 3.6] | 3.4 [3.1, 3.6] | 0.232 |
| Serum magnesium (mg/dL) | 2.1 [2.0, 2.1] | 2.0 [1.9, 2.1] | 0.057 |
| Serum fetuin-A (μg/mL) | 221 [190, 252] | 240 [212, 273] | 0.001 |
| Serum albumin (g/dL) | 4.2 [4.1, 4.3] | 4.3 [4.1, 4.4] | 0.125 |
The table provides medians [interquartile ranges] for continuous variables and numbers (percentages) for categorical variables. P-values for the between-group comparison were calculated using the Mann–Whitney U test or Fisher’s exact test.
Abbreviations: T2D, type 2diabetes; BMI, body mass index; CVD, cardiovascular disease; HbA1c, hemoglobin A1c; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; eGFR, estimated glomerular filtration rate.
The left panel of Fig.2 shows a comparison of T50 between the T2D and control groups. The T50 of the T2D group (192 [168, 211] min) was not significantly different from the T50 of the control group (183 [158, 211] min, P = 0.414). The right panel of Fig.2 shows the correlation between T50, fasting plasma glucose, and HbA1c levels. T50 was not significantly correlated with fasting plasma glucose or HbA1c levels in the total subjects. No significant association was found when the T2D and control groups were analyzed separately.
The left panel shows that the T50 values followed a log-normal distribution in the T2D and control groups, as indicated by the red curves. The median [interquartile range] values and P values by the Mann–Whitney U test are shown in the figure. The right panel shows Spearman’s correlation of T50 with fasting plasma glucose and HbA1c levels.
Abbreviations: T50, T50 calciprotein crystallization (serum calcification propensity); T2D, type 2 diabetes; HbA1c, hemoglobin A1c.
The independent association between T50 and T2D in the total population was examined using a multiple regression analysis (Table 2). Model 1 showed that T2D was not significantly associated with T50. When T2D was replaced with fasting plasma glucose (model 2) or HbA1c (model 3), these glycemic parameters showed no significant association with T50. T50 was inversely associated with serum phosphate levels and positively correlated with calcium and fetuin-A levels in all three models.
| Variables | Model 1 | Model 2 | Model 3 | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Std. coeff. | [95% CI] | P value | Std. coeff | [95% CI] | P value | Std. coeff | [95% CI] | P value | |
| Age | 0.001 | [−0.140, 0.143] | 0.984 | −0.008 | [−0.150, 0.134] | 0.913 | −0.012 | [−0.155, 0.130] | 0.867 |
| Male sex (yes = 1, no = 0) | −0.104 | [−0.233, 0.025] | 0.115 | −0.108 | [−0.239, 0.024] | 0.108 | −0.109 | [−0.239, 0.021] | 0.100 |
| Body mass index | 0.027 | [−0.105, 0.159] | 0.686 | 0.020 | [−0.112, 0.151] | 0.766 | 0.017 | [−0.115, 0.149] | 0.798 |
| Currernt smoking (yes = 1, no = 0) | −0.100 | [−0.225, 0.025] | 0.116 | −0.112 | [−0.236, 0.012] | 0.077 | −0.115 | [−0.239, 0.009] | 0.069 |
| Hypertension (yes = 1, no = 0) | −0.049 | [−0.187, 0.090] | 0.489 | −0.056 | [−0.194, 0.083] | 0.429 | −0.058 | [−0.197, 0.08] | 0.408 |
| Dyslipidemia (yes = 1, no = 0) | 0.146 | [0.018, 0.273] | 0.025 | 0.137 | [0.011, 0.263] | 0.034 | 0.134 | [0.008, 0.261] | 0.037 |
| eGFR | 0.033 | [−0.103, 0.168] | 0.633 | 0.027 | [−0.108, 0.162] | 0.695 | 0.025 | [−0.111, 0.160] | 0.717 |
| Serum calcium | 0.209 | [0.049, 0.369] | 0.011 | 0.216 | [0.056, 0.376] | 0.008 | 0.218 | [0.058, 0.378] | 0.008 |
| Serum phosphate | −0.173 | [−0.308, −0.039] | 0.012 | −0.181 | [−0.317, −0.045] | 0.010 | −0.188 | [−0.327, −0.049] | 0.009 |
| Serum magnesium | 0.094 | [−0.034, 0.223] | 0.147 | 0.100 | [−0.030, 0.23] | 0.133 | 0.100 | [−0.028, 0.229] | 0.126 |
| Serum fetuin−A | 0.494 | [0.370, 0.618] | ＜0.001 | 0.486 | [0.362, 0.609] | ＜0.001 | 0.484 | [0.361, 0.608] | ＜0.001 |
| Serum albumin | −0.036 | [−0.198, 0.125] | 0.658 | −0.048 | [−0.209, 0.112] | 0.552 | −0.049 | [−0.208, 0.111] | 0.550 |
| T2D (yes = 1, no = 0) | −0.054 | [−0.183, 0.075] | 0.411 | −−− | −−− | ||||
| Fasting plasma glucose | −−− | 0.010 | [−0.117, 0.137] | 0.875 | −−− | ||||
| HbA1c | −−− | −−− | 0.030 | [−0.098, 0.158] | 0.643 | ||||
| Model summary | R2 = 0.362, p＜0.001 | R2 = 0.360, p＜0.001 | R2 = 0.360, p＜0.001 | ||||||
The table provides the results of the multiple regression analysis of the factors associated with T50 in the total study population (N = 202). T50 was included in the model after log-transformation to meet the assumption of a normal distribution of residuals.
Abbreviations: eGFR, estimated glomerular filtration rate; T2D, type 2 diabetes; HbA1c, hemoglobin A1c; Std coeff, standardized coefficient of regression (β); CI, confidence interval; R2, coefficient of determination.
We found that the presence of T2D significantly modified the association between T50 and fetuin-A concentration (P for interaction = 0.005). We then conducted a subgroup analysis in the T2D and control groups (Table 3). The results showed that T50 was strongly and positively associated with fetuin-A in the control group (β = 0.591, P＜0.001), whereas the positive association between T50 and fetuin-A was significant, but less in the T2D group (β = 0.345, P = 0.001). The positive association between T50 and calcium and inverse association between T50 and phosphate were significant in the T2D group, whereas these associations were not significant in the control group.
| Exposure variables | Control group (N = 101) | T2D group (N = 101) | ||||
|---|---|---|---|---|---|---|
| Std. coeff. | [95% CI] | P value | Std. coeff. | [95% CI] | P value | |
| Age | 0.025 | [−0.150, 0.200] | 0.776 | −0.091 | [−0.325, 0.144] | 0.444 |
| Male sex (yes = 1, no = 0) | −0.082 | [−0.248, 0.084] | 0.330 | −0.127 | [−0.333, 0.079] | 0.224 |
| Body mass index | 0.069 | [−0.093, 0.232] | 0.399 | −0.024 | [−0.235, 0.187] | 0.823 |
| Current smoking (yes = 1, no = 0) | −0.072 | [−0.233, 0.090] | 0.380 | −0.011 | [−0.208, 0.186] | 0.913 |
| Hypertension (yes = 1, no = 0) | −0.078 | [−0.249, 0.092] | 0.364 | 0.019 | [−0.197, 0.234] | 0.864 |
| Dyslipidemia (yes = 1, no = 0) | 0.183 | [0.021, 0.345] | 0.027 | 0.023 | [−0.173, 0.218] | 0.818 |
| eGFR | −0.048 | [−0.217, 0.120] | 0.570 | 0.107 | [−0.110, 0.324] | 0.329 |
| Serum calcium | 0.069 | [−0.127, 0.266] | 0.486 | 0.437 | [0.168, 0.706] | 0.002 |
| Serum phosphate | −0.099 | [−0.263, 0.066] | 0.236 | −0.279 | [−0.496, −0.062] | 0.012 |
| Serum magnesium | 0.108 | [−0.070, 0.286] | 0.231 | 0.053 | [−0.141, 0.246] | 0.589 |
| Serum fetuin−A | 0.591 | [0.424, 0.758] | ＜0.001 | 0.345 | [0.155, 0.534] | 0.001 |
| Serum albumin | −0.022 | [−0.222, 0.177] | 0.826 | −0.149 | [−0.415, 0.116] | 0.267 |
| Model summary | R2 = 0.541, P＜0.001 | R2 = 0.289, P = 0.002 | ||||
As there was a significant effect modification by the presence of T2D on the association between fetuin-A and T50 (P for interaction = 0.005), we conducted subgroup analysis in the control and T2D subgroups using a multiple regression model. T50 was included in the model after log- transformation to meet the assumption of a normal distribution of residuals.
Abbreviations: T2D, type 2 diabetes; eGFR, estimated glomerular filtration rate; Std coeff, standardized coefficient of regression (β); CI, confidence interval; R2, coefficient of determination.
We conducted an additional analysis using the ratio of T50 to fetuin-A (T50/fetuin-A ratio) as a simple index for the protective function of fetuin-A per unit against calciprotein crystallization. T50/fetuin-A ratio was significantly lower in the T2D group than in the control group and was significantly and inversely associated with fasting plasma glucose and HbA1c levels (Fig.3).
The ratio of T50 to fetuin-A (T50/fetuin-A ratio) was calculated as a simple index of the protective function of fetuin-A per unit against calciprotein crystallization. The ratio was lower in the T2D group than in the control group (left panel), and the ratio was inversely associated with fasting plasma glucose and HbA1c levels (right panel). Box-and-whisker plots indicate the 10th, 25th, 50th, 75th, and 90th percentiles.
Abbreviations: T50, T50 calciprotein crystallization (serum calcification propensity); T2D, type 2 diabetes; HbA1c, hemoglobin A1c.
Studies have shown that patients with T2D have a higher prevalence and extent of vascular calcification1-3) than those without diabetes, although patients with T2D have no apparent abnormalities in serum calcium or phosphate. Since a lower T50 value indicates a higher propensity for calciprotein crystallization or higher calcification stress4), we hypothesized that T50 would be lower in the T2D group than in the control group in the present study. Contrary to our hypothesis, T50 was not significantly different between the two groups, and T50 was not significantly associated with fasting plasma glucose or HbA1c in the total subjects. However, we found that T2D significantly modified the association between T50 and fetuin-A. A subgroup analysis revealed that the positive association between T50 and fetuin-A was significant but weaker in the T2D group and that the associations of T50 with serum calcium and phosphate were more apparent in the T2D group than in the control group. An additional analysis revealed that T50/fetuin-A ratio was significantly lower in the T2D group than in the control group, and that T50/fetuin-A ratio was inversely associated with fasting plasma glucose and HbA1c levels. These results suggest that the resistive function of fetuin-A against calcification stress was decreased in the presence of T2D, which could explain why patients with T2D are prone to vascular calcification without apparent abnormalities in the serum calcium of phosphate.
Regarding the possible influences of T2D and/or hyperglycemia on T50, a previous study of 932 patients with T2D reported that T50 was inversely associated with HbA1c, whereas another study found a positive association between T50 and fasting plasma glucose in 132 patients with T2D21). A report from the PREVEND study (N = 6231) showed that plasma glucose was positively associated with T50 in the general population17). Our study showed no significant association between T50 and the fasting plasma glucose or HbA1c levels. Our study is the first to compare T50 values between T2D and control groups without diabetes and found no significant difference in T50 between the groups. Based on these studies, including ours, the influence of hyperglycemia and the presence of T2D on T50 appears to be neutral or modest.
The observed neutral results can be explained by the altered profile of mineralization regulators in T2D. Fetuin-A is a multi-functional liver-derived glycoprotein that inhibits the transformation of primary to secondary CPPs and suppresses ectopic calcification on one hand7, 30), while fetuin-A inhibits insulin signaling and induces insulin resistance31, 32). In this study, the T2D group had significantly higher levels of serum fetuin-A, which would increase the T50 4). At the same time, the T2D group had insignificantly lower levels of serum magnesium (P = 0.057), which was in favor of lower T50 levels in the T2D group, because serum magnesium is a known factor that increase T50 levels33, 34). According to Kurstjens et al.35), hypomagnesemia is prevalent in T2D, and hyperglycemia is one of the determinants of plasma magnesium levels. In this study, we confirmed that serum magnesium was inversely correlated with fasting plasma glucose (Spearman’s r = −0.204, P = 0.004) and fetuin-A was positively correlated with HbA1c (Spearman’s r = 0.180, P = 0.010). Therefore, the alterations of these T50-modulating factor levels in T2D could cancel each other out, thus explaining the absence of a difference in T50 between the T2D and control groups. In addition, the balance between fetuin-A and magnesium levels may vary among study populations, which could explain the above-mentioned discrepancy between studies.
This study showed that the presence of dyslipidemia was positively associated with T50 in the total participants and in the control group, but not in the T2D group. The positive association of T50 with the presence of dyslipidemia or serum lipid levels is consistent with other studies17, 20), although the precise mechanisms are unknown. As the prevalence of dyslipidemia was significantly higher in the T2D group than in the control group in this study, this difference may explain the lack of reduction in T50 in the T2D group. However, the presence of dyslipidemia does not fully explain the result of T50 because no significant association was found between T50 and the presence of T2D, fasting glucose, or HbA1c, even when adjusted for multiple factors, including the presence of dyslipidemia.
It is a novel finding that the association between T50 and fetuin-A was significantly modified by the presence of T2D and that the degree of association was smaller in the T2D group than in the control group when assessed by standardized regression coefficients in the subgroup analysis. In addition, the subgroup analysis indicated that the positive association of serum calcium and the inverse association of serum phosphate with T50 were more apparent in the T2D group than in the control group. We interpret these results to indicate that the ability of fetuin-A to protect against the transformation of primary CPPs to secondary CPPs is impaired in the presence of T2D, and that the same degree of change in serum calcium or phosphate would result in a greater effect on T50 in the presence of T2D. In other words, the influence of serum calcium and phosphate on T50 might be exaggerated in the presence of T2D, presumably due to the impaired function of fetuin-A in T2D.
To explore this possibility, we conducted an additional analysis using the ratio of T50 to fetuin-A as a simple index for the protective function of fetuin-A per unit against calciprotein crystallization. The results showed that T50/fetuin-A ratio was significantly lower in the T2D group than in the control group, and the ratio was inversely correlated with fasting plasma glucose and HbA1c levels, thus supporting the notion that the protective function of fetuin-A could be impaired in the presence of T2D and hyperglycemia. Fetuin-A is fully phosphorylated and active when secreted from hepatocytes, whereas a large part (approximately 90%) of fetuin-A in serum or plasma is in its inactive form36). Although we did not separately measure the active and inactive forms of fetuin-A in the serum, the presence of T2D and/or hyperglycemia might affect the proportion of the active form of fetuin-A, resulting in decreased T50/fetuin-A, as shown in this study. Further studies are therefore required to confirm this possibility.
It may appear strange that serum calcium was positively associated with T50 in this study, firstly because the study by Pasch et al. reported that calcium was a factor that decreased the T50 value when added to the assay system4), and secondly because hypercalcemia is generally believed to promote vascular calcification. Regarding the first point, however, we found no study in the literature that reported an inverse association between T50 and serum calcium concentration. Instead, among the four studies that examined the independent association between T50 and serum calcium levels by multiple regression analysis, the association of calcium with T50 was not significant in one study of patients with CKD37), whereas the other three studies showed a significant and positive association between T50 and serum calcium in patients with T2D21), pediatric patients on hemodialysis38), and the general population17). Thus, this is the fourth study to show an independent and positive association between serum calcium and T50. This was not because we used serum total calcium instead of calcium levels which were adjusted for serum albumin, because we confirmed the positive association of T50 with albumin-corrected serum calcium, which was calculated using the equation by Payne et al.39) and by the equation recommended by K/DOQI clinical practice guidelines40). Regarding the second point, although previous studies identified a higher serum phosphate level as an independent factor associated with vascular calcification41), there is insufficient evidence that a higher serum calcium level is significantly associated with vascular calcification. Kinugasa et al. reported that serum phosphate, but not serum calcium, was a significant factor associated with vascular calcification in non-dialysis patients undergoing cardiovascular surgery42). In addition, London et al. showed that hyperphosphatemia, but not hypercalcemia, was significantly associated with higher odds of having vascular calcification in patients undergoing hemodialysis43). In addition, according to Koubaity et al., although coronary artery disease was more severe in patients with primary hyperparathyroidism (PHP) than in the control group, the patients in the mild normocalcemic PHP group had 5-folds higher odds of having a coronary calcification score ＞100 than those in the classical hypercalcemic PHP group44). Although further clarification is needed at these points, we speculate that the in vitro effect of adding calcium to the assay system of T50 is not necessarily the same as the association of serum calcium concentration with T50 and presumably with vascular calcification in vivo.
This study has several limitations. First, we could not relate T50 to vascular calcification because data on vascular calcification were not available in the database. Second, this study did not address the biochemical mechanisms underlying the above findings, including the altered association between T50 and fetuin-A in the presence of T2D. Third, we did not measure the concentration of CPPs. Thus, this study does not exclude the possibility that circulating levels of primary and secondary CPPs are affected by T2D and/or hyperglycemia. Fourth, we had no data on albuminuria, which may be associated with hyperfiltration (eGFR), fetuin-A level, and T50. Fifth, the sample size of this study (N = 101 for each group) was smaller than that of other studies, which may limit the statistical power to detect the true association. Sixth, because of the observational nature of this cross-sectional study, causality was unknown for the observed associations. To the best of our knowledge, this study is the first to show that the association between T50 and fetuin-A is modified by the presence of T2D. This finding is novel and one of the strengths of this study.
This study showed that T50 was not significantly different between the T2D and control groups, and that T50 was not significantly associated with fasting plasma glucose or HbA1c levels. However, T2D was a significant effect modifier of the relationship between T50 and fetuin-A. Namely, the degree of positive association of T50 with fetuin-A was smaller, but the associations of T50 with calcium and phosphate were more apparent in the T2D group than in the control group. T50/fetuin-A ratio was lower in the T2D group, suggesting an impaired protective function of fetuin-A in T2D. Although these results are in line with the fact that patients with T2D are prone to vascular calcification without apparent abnormalities in serum calcium and phosphate levels, further studies are needed to confirm the findings of this study in other clinical settings and to clarify the mechanisms underlying the observed associations.
The authors acknowledge the valuable contributions of Mayumi Taniguchi-Seko and Chiaki Murakami at the Department of Metabolism, Endocrinology, and Metabolism, Osaka Metropolitan University, for their excellent technical assistance. The kind support by Yasuko Okui and Wataru Yamashita at the Department of Central Clinical Laboratory, Osaka Metropolitan University Hospital, is also acknowledged. Part of this study was presented at the Young Investigators Award session of the 56th Annual Scientific Meeting of the Japan Atherosclerosis Society (Kobe, Japan; July 6-7, 2024), and the abstract in Japanese was published. No external funding was received for this study.
All authors report that they have nothing to disclose for this study.
