Baseline characteristics of the study participants stratified by 25(OH)D quartiles
A total of 7061 subjects were included in the study, comprising 2012 males(28.5%) and 5049 females (71.5%).The mean age of the study population was 57.6 years (57.6±8.4 years) and the median (interquartile range) serum 25(OH)D levels was 15.91(11.56,19.98) ng/mL. The prevalence of vitamin D deficiency [25(OH)D<20 ng/mL] is 75.1%, merely 2.0% of those with vitamin D sufficiency[25(OH)D≥30 ng/mL]. Table 1 shows the baseline clinical and biochemical parameters according to the quartiles groups for the 25(OH)D. Overall, participants in lower quartile of 25(OH)D were more likely to be older, female, to had a lower smoking and drinking rate and had a higher prevalence of hypertension. Besides, the BMI, FPG, and 2h-PG levels of the study population decreased with the increase of 25(OH)D level (all P<0.05). However, no statistically significant differences were detected between in 25(OH)D quartiles group for levels of HbA1c, HDL-C, LDL-C, TC and TG (P>0.05).From first quartile to fourth quartile, the prevalence of IGR showed a significant downward trend (P<0.05), and the prevalence of diabetes also showed a weak downward trend, but there was no statistical difference (P>0.05).
Table 1 Baseline characteristics of the study participants stratified by 25(OH)D quartiles
Characteristics
|
Total
|
Quartile 1
|
Quartile 2
|
Quartile 3
|
Quartile 4
|
cut points(ng/mL)
|
1.35-65.69
|
≤11.56
|
11.57-15.91
|
15.92-19.98
|
≥19.99
|
no. of participants
|
7061
|
1767
|
1764
|
1765
|
1765
|
male vs. female
|
2012/5049
|
307/1460
|
447/1317a
|
528/1237ab
|
730/1035abc
|
age (years)
|
57.6±8.4
|
58.9±8.8
|
57.5±8.6a
|
57.2±8.0a
|
56.9±8.1a
|
hypertension(%)
|
2525(35.8)
|
702(39.7)
|
614(34.8)a
|
608(34.4)a
|
601(34.1)a
|
smoking(%)
|
1151(16.3)
|
218(12.3)
|
276(15.6)a
|
285(16.1)a
|
372(21.1)abc
|
drinking(%)
|
2008(28.4)
|
389(22.0)
|
484(27.4)a
|
532(30.1)a
|
603(34.2)ab
|
glucose metabolism status
|
|
|
|
|
|
NGT(%)
|
3272(46.3)
|
757(42.8)
|
818(46.4)
|
827(46.9)
|
870(49.3)a
|
IGR(%)
|
2001(28.3)
|
558(31.6)
|
499(28.3)
|
491(27.8)
|
453(25.7)a
|
T2DM(%)
|
1788(25.3)
|
452(25.6)
|
447(25.3)
|
447(25.3)
|
442(25.0)
|
BMI(kg/m2)
|
24.17±3.30
|
24.38±3.60
|
24.21±3.42
|
24.06±3.10a
|
24.01±3.04a
|
FPG(mmol/L)
|
6.00±1.71
|
6.09±1.84
|
6.03±1.76
|
5.95±1.57a
|
5.94±1.65a
|
2h-PG(mmol/L)*
|
7.60(6.10,10.00)
|
7.90(6.28,10.50)
|
7.60(6.13,9.90)
|
7.50(6.04,9.86)a
|
7.41(6.00,9.70)a
|
HbA1c(%)
|
6.16±1.03
|
6.18±1.08
|
6.18±1.06
|
6.16±1.02
|
6.10±0.97
|
HDL-C(mmol/L)
|
1.23±0.30
|
1.21±0.31
|
1.23±0.31
|
1.23±0.29
|
1.23±0.30
|
LDL-C(mmol/L)
|
2.56±0.77
|
2.55±0.82
|
2.56±0.76
|
2.59±0.77
|
2.56±0.75
|
TC(mmol/L)
|
4.57±1.05
|
4.57±1.12
|
4.58±1.07
|
4.58±1.04
|
4.53±0.97
|
TG(mmol/L)*
|
1.50(1.07,2.14)
|
1.53(1.09,2.19)
|
1.51(1.06,2.15)
|
1.52(1.05,2.14)
|
1.45(1.06,2.07)
|
25(OH)D(ng/ml)*
|
15.91(11.56,19.98)
|
9.86(8.66,10.64)
|
13.81(12.76,14.85)a
|
17.89(16.84,18.93)ab
|
22.70(21.19,24.86)abc
|
*:non-normally distributed variables were presented as medians (interquartile ranges);aP<0.05 Quartile 2,Quartile 3 or Quartile 4 versus Quartile 1; b P<0.05 Quartile 3 or Quartile 4 versus Quartile 2;c P<0.05 Quartile 4 versus Quartile 3
NGT, normal glucose tolerance; IGR, impaired glucose regulation; T2DM, type 2 diabetes mellitus; BMI, body mass index; FPG, fasting plasma glucose; 2h-PG, 2h postprandial plasma glucose; HbA1c, glycated haemoglobin; HDL-C, high density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; TC, total cholesterol; TG, triglyceride; 25(OH)D, 25-hydroxyvitamin D
Correlation between serum 25(OH)D concentration and glucolipid metabolism index
The associations between 25(OH)D concentration and glucolipid metabolism index were examined by the study. The partial correlation analysis showed that serum 25(OH)D concentration was negatively correlated with the levels of FPG, 2h-PG, HbA1c and TG when controlling gender, age, smoking, drinking, BMI, hypertension and diabetes, the partial correlation coefficients were -0.047, -0.056, -0.035, -0.042, respectively (all P<0.05). Meanwhile, serum 25(OH)D was positively correlated with HDL-C and LDL-C (P<0.05), but we did not find a correlation between serum levels of 25(OH)D and TC(Table 2).
Table 2 Correlation between serum 25(OH)D concentration and glucolipid metabolism index
Index
|
Unadjusted
|
Model 1
|
Model 2
|
Partial correlation
coefficient
|
P value
|
Partial correlation
coefficient
|
P value
|
Partial correlation
coefficient
|
P value
|
FPG(mmol/L)
|
-0.035
|
0.003
|
-0.048
|
0.000
|
-0.047
|
0.000
|
2h-PG(mmol/L)
|
-0.056
|
0.000
|
-0.053
|
0.000
|
-0.056
|
0.000
|
HbA1c(%)
|
-0.038
|
0.001
|
-0.040
|
0.001
|
-0.035
|
0.003
|
HDL-C(mmol/L)
|
0.013
|
0.274
|
0.061
|
0.000
|
0.050
|
0.000
|
LDL-C(mmol/L)
|
0.004
|
0.743
|
0.031
|
0.009
|
0.035
|
0.003
|
TC(mmol/L)
|
-0.022
|
0.069
|
0.018
|
0.126
|
0.023
|
0.058
|
TG(mmol/L)
|
-0.047
|
0.000
|
-0.055
|
0.000
|
-0.042
|
0.000
|
Model 1: adjusted for age, gender.
Model 2: adjusted for age, gender, smoking, drinking, body mass index, and history of hypertension and diabetes
FPG, fasting plasma glucose; 2h-PG, 2h postprandial plasma glucose; HbA1c, glycated haemoglobin; HDL-C, high density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; TC, total cholesterol; TG, triglyceride; 25(OH)D, 25-hydroxyvitamin D
Incidence of ASCVD among groups with 25(OH)D quartiles
After a median follow-up of 3.3 years, a total of 216 patients (3.1%) experienced ASCVD events, and the ASCVD cumulative incidence in quartiles of 25(OH)D groups were 4.1%, 3.0%, 3.1%, and 2.0%, respectively. There was a linear trend toward decreasing ASCVD events as serum 25(OH)D increased. This trend was statistically significant using Log-Rank test( c2=11.676, P=0.009) Figure 2.
Incidence of ASCVD among groups with 25(OH)D quartiles
After a median follow-up of 3.3 years, a total of 216 patients (3.1%) experienced ASCVD events, and the ASCVD cumulative incidence in quartiles of 25(OH)D groups were 4.1%, 3.0%, 3.1%, and 2.0%, respectively. There was a linear trend toward decreasing ASCVD events as serum 25(OH)D increased. This trend was statistically significant using Log-Rank test( c2=11.676, P=0.009) Figure 2.
Table 3 Hazard ratios for ASCVD events in participants stratified by 25(OH)D quartiles
Model
|
Quartile 1
|
Quartile 2
|
Quartile 3
|
Quartile 4
|
Unadjusted
|
1.972(1.322-2.942)
P=0.001
|
1.430(0.937-2.184)
P=0.098
|
1.522(1.000-2.317)
P=0.050
|
Ref.
|
Model 1
|
1.902(1.258-2.877)
P=0.002
|
1.458(0.951-2.235)
P=0.084
|
1.561(1.024-2.381)
P=0.039
|
Ref.
|
Model 2
|
1.765(1.165-2.675)
P=0.007
|
1.391(0.907-2.133)
P=0.130
|
1.510(0.989-2.304)
P=0.056
|
Ref.
|
Model 3
|
1.748(1.150-2.656)
P=0.009
|
1.378(0.897-2.118)
P=0.143
|
1.512(0.989-2.312)
P=0.056
|
Ref.
|
Model 1:adjusted for age and gender
Model 2:adjusted for age, gender, smoking, drinking, diabetes, hypertension and body mass index
Model 3: additionally adjusted for fasting plasma glucose, 2h postprandial plasma glucose, glycated haemoglobin, high density lipoprotein cholesterol, low-density lipoprotein cholesterol, total cholesterol, triglyceride
Hazard ratios for ASCVD in participants with 10 ng/ml per increase of 25(OH)D
Using serum 25(OH)D as a continuous variable(10 ng/ml per increase),Cox regression analyses showed that inverse association of the 25(OH)D concentration with the risk of ASCVD(HR:0.696, 95%CI:0.545-0.890, P=0.004).This association remained statistically significant after adjusted for confounders including age, gender, smoking, drinking, BMI, diabetes, hypertension, FPG, 2hPG, HbA1c, HDL-C, LDL-C, TC and TG (HR:0.761,95% CI:0.590-0.980,P = 0.034)(Table 4).
Table 4 Hazard ratios for ASCVD in participants with 10 ng/ml per increase of 25(OH)D
Model
|
β
|
SE
|
Wald
|
P value
|
HR (95% CI)
|
Unadjusted
|
-0.362
|
0.125
|
8.389
|
0.004
|
0.696(0.545-0.890)
|
Model 1
|
-0.341
|
0.128
|
7.134
|
0.008
|
0.711(0.554-0.913)
|
Model 2
|
-0.284
|
0.129
|
4.866
|
0.027
|
0.753(0.585-0.969)
|
Model 3
|
-0.274
|
0.129
|
4.481
|
0.034
|
0.761(0.590-0.980)
|
Model 1:adjusted for age and gender
Model 2:adjusted for age, gender, smoking, drinking, diabetes, hypertension, and body mass index
Model 3: additionally adjusted for fasting plasma glucose, 2h postprandial plasma glucose, glycated haemoglobin, high density lipoprotein cholesterol, low-density lipoprotein cholesterol, total cholesterol, triglyceride
Incidence of ASCVD among populations of different glucose metabolism status within 25(OH)D quartiles
According to different glucose metabolism status, 7061 individuals were classified into NGT group (n = 3272), IGR group (n = 2001) and T2DM group (n = 1788). A median follow-up of 3.3 years, the cumulative incidence of ASCVD in the three groups was 1.7%, 2.7%, and 5.9%, respectively. There was no statistically difference in ASCVD incidence among 25(OH)D quartiles in NGT and IGR groups. Interestingly, among the T2DM population, the incidence rate of ASCVD were 8.8%, 5.1%, 6.0%, and 3.6%, respectively, from the first quartile of serum 25(OH)D to fourth. Overall,the incidence of ASCVD showed a downward trend with the increase of 25(OH)D level in patients with T2DM(P=0.003)Figure 3.
Hazard ratios for ASCVD events in T2DM patients according to 25(OH)D quartiles
Compared with the fourth quartile in T2DM patients, those in first quartile had a higher risk of ASCVD incidence(P=0.002). After adjusted for all confounding factors, the trend was still significant, and the hazard ratio of ASCVD events (95% CI) was 2.284(1.242-4.202)in the lowest quartile than those in the highest (P=0.008)Table 5.
Table 5 Hazard ratios for ASCVD in T2DM patients according to 25(OH)D quartiles
Model
|
Quartile 1
|
Quartile 2
|
Quartile 3
|
Quartile 4
|
Unadjusted
|
2.481(1.389-4.430)
P=0.002
|
1.373(0.725-2.599)
P=0.330
|
1.666(0.897-3.091)
P=0.106
|
Ref.
|
Model 1
|
2.614(1.434-4.765)
P=0.002
|
1.455(0.765-2.7668)
P=0.253
|
1.759(0.945-3.275)
P=0.075
|
Ref.
|
Model 2
|
2.317(1.269-4.232)
P=0.006
|
1.370(0.720-2.607)
P=0.338
|
1.738(0.933-3.239)
P=0.082
|
Ref.
|
Model 3
|
2.284(1.242-4.202)
P=0.008
|
1.338(0.699-2.560)
P=0.379
|
1.736(0.928-3.247)
P=0.084
|
Ref.
|
Model 1:adjusted for age and gender
Model 2:adjusted for age, gender, smoking, drinking, hypertension, and body mass index
Model 3: additionally adjusted for fasting plasma glucose, 2h postprandial plasma glucose, glycated haemoglobin, high density lipoprotein cholesterol, low-density lipoprotein cholesterol, total cholesterol, triglyceride
Hazard ratios for ASCVD in T2DM patients with per 10 ng/ml increase of 25(OH)D
The results showed that for per 10ng/ml increase in basline serum 25(OH)D concentrations, the risk of onset ASCVD decreased by 40 percent approximately in T2DM patients(HR:0.597, 95% CI:0.415-0.857,P=0.005) when without adjusted for relevant risk factors. After further adjusted for confounding factors, we found that the risk of ASCVD still decreased gradually. In other words, the risk of ASCVD is reduced by 35% for each 10 ng/ml elevation of 25(OH)D(HR:0.649, 95% CI:0.444-0.948,P=0.025)Table 6.
Table 6 Hazard ratios for ASCVD in T2DM patients with per 10 ng/ml increase of 25(OH)D
Model
|
β
|
SE
|
Wald
|
P value
|
HR (95% CI)
|
Unadjusted
|
-0.516
|
0.185
|
7.814
|
0.005
|
0.597(0.415-0.857)
|
Model 1
|
-0.548
|
0.192
|
8.127
|
0.004
|
0.578(0.397-0.843)
|
Model 2
|
-0.450
|
0.192
|
5.512
|
0.019
|
0.637(0.438-0.928)
|
Model 3
|
-0.433
|
0.194
|
4.995
|
0.025
|
0.649(0.444-0.948)
|
Model 1:adjusted for age and gender
Model 2:adjusted for age, gender, smoking, drinking, hypertension, and body mass index
Model 3: additionally adjusted for fasting plasma glucose, 2h postprandial plasma glucose, glycated haemoglobin, high density lipoprotein cholesterol, low-density lipoprotein cholesterol, total cholesterol, triglyceride