Derivation of cutoff values
|
Group
|
No
|
Age (mean(SD))
| BMI ≧24 |
Hyper- TC
|
Hyper-LDL-c
|
Hyper-TG
|
Hyper- FPG
| FS≧9 |
---|
Joinpoint regression
|
TSH < 2.9
|
2284
|
35.7 (9.1)
|
37 %
|
22 %
|
23 %
|
14 %
|
7 %
|
20 %
|
TSH≧2.9 |
572
|
36.9 (9.5)
|
35 %
|
27 %
|
27 %
|
18 %
|
8 %
|
23 %
|
P
| |
0.009
|
0.49
|
0.021*
|
0.07
|
0.030*
|
0.46
|
0.18
|
ROC curve
|
TSH < 2.6
|
2138
|
35.7 (9.1)
|
37 %
|
22 %
|
23 %
|
14 %
|
7 %
|
20 %
|
TSH≧2.6 |
718
|
36.8 (9.5)
|
36 %
|
28 %
|
27 %
|
17 %
|
8 %
|
23 %
|
P
| |
0.008
|
0.44
|
<0.001*
|
0.007*
|
0.024*
|
0.21
|
0.034*
|
Conventional accepted
|
TSH < 4.5
|
2714
|
35.9 (9.1)
|
37 %
|
23 %
|
24 %
|
15 %
|
7 %
|
20 %
|
TSH≧4.5 |
142
|
37.7 (10.2)
|
38 %
|
26 %
|
26 %
|
20 %
|
9 %
|
31 %
|
P
| |
0.046
|
0.77
|
0.008*
|
0.50
|
0.09
|
0.39
|
0.003*
|
- Values are means(SD) or probability. Comparison between any two groups by unpaired t-test, or Pearson Chi-square test*. Data in bold are statistically significant
-
FS Framingham score, LDL-c low-density lipoprotein cholesterol, TG triglyceride, FPG fasting plasma glucose