A total of 4236 participants aged between 60 and 85 were included in our analysis. The weighted characteristics of the participants were subdivided according to the tertiles of serum total cholesterol (low: ≥75 mg/dL to 191 mg/dL; medium: ≥191 mg/dL to < 224 mg/dL; high: ≥224 mg/dL to < 704 mg/dL). As shown in Table 1, there were significant differences in baseline characteristics between the tertiles of total cholesterol, except for race / ethnicity. Compared with other subgroups, participants in the highest tertile of total cholesterol were more likely to be female, 60–69 years old, and sedentary. Participants in the top tertile of total cholesterol had lower income to poverty ratio, BUN, serum uric acid, and BMI levels but higher total protein and serum calcium levels.
Table 1
Baseline characteristics of participants (N = 4236).
Total Cholesterol (mg/dL)
|
Total
|
Low
(≥ 75 to 191)
|
Middle
(≥ 191 to < 224)
|
High
(≥ 224 to < 704)
|
P-value
|
Sex (%)
|
|
|
|
|
< 0.0001
|
Male
|
43.67
|
58.55
|
44.3
|
29.59
|
|
Female
|
56.33
|
41.45
|
55.7
|
70.41
|
|
Age (%)
|
|
|
|
|
0.0298
|
60–69 years
|
60.29
|
58.3
|
59.39
|
62.93
|
|
>= 70 years
|
39.71
|
41.7
|
40.61
|
37.07
|
|
Race/ethnicity (%)
|
|
|
|
|
0.3112
|
Non-Hispanic White
|
83.16
|
81.85
|
84.77
|
82.86
|
|
Non-Hispanic Black
|
8.62
|
9.29
|
7.78
|
8.80
|
|
Mexican American
|
4.2
|
4.1
|
4.29
|
4.20
|
|
Other races/ethnicities
|
4.02
|
4.76
|
3.16
|
4.14
|
|
Physical activity (%)
|
|
|
|
|
0.0258
|
Sedentary
|
27.31
|
26.54
|
25.67
|
29.50
|
|
Low
|
26.64
|
26.97
|
26.29
|
26.66
|
|
Moderate
|
17.57
|
16.53
|
17.37
|
18.67
|
|
High
|
28.49
|
29.96
|
30.67
|
25.17
|
|
Income to poverty ratio
|
2.82 ± 1.52
|
2.90 ± 1.50
|
2.85 ± 1.53
|
2.71 ± 1.51
|
0.0044
|
Blood urea nitrogen (mg/dL)
|
16.24 ± 6.80
|
16.81 ± 7.91
|
15.99 ± 6.39
|
15.96 ± 5.99
|
0.0009
|
Total protein (mg/dL)
|
7.24 ± 0.50
|
7.17 ± 0.52
|
7.25 ± 0.49
|
7.30 ± 0.48
|
< 0.0001
|
Serum uric acid (mg/dL)
|
5.63 ± 1.44
|
5.83 ± 1.49
|
5.58 ± 1.41
|
5.50 ± 1.41
|
< 0.0001
|
Serum calcium (mg/dL)
|
9.51 ± 0.41
|
9.45 ± 0.43
|
9.50 ± 0.39
|
9.56 ± 0.39
|
< 0.0001
|
Body mass index
|
28.65 ± 5.73
|
29.50 ± 6.24
|
28.26 ± 5.55
|
28.24 ± 5.32
|
< 0.0001
|
Lumber spine BMD (g/cm2)
|
1.01 ± 0.18
|
1.05 ± 0.19
|
1.02 ± 0.18
|
0.98 ± 0.17
|
< 0.0001
|
Note: Mean +/- standard deviation for: age, income to poverty, ratio blood urea nitrogen, total protein, serum uric acid, serum calcium, BMI, total days drink year, total BMD, head BMD, lumber spine BMD, lumber pelvis BMD. P value was calculated using weighted linear regression model. % For: age, sex, race/ethnicity, education, physical activity. P value was calculated using weighted chi-square test. Weighted by: Full sample mobile examination center exam weight.
BMD, Bone mineral density; BMI, Body mass index.
Table 2 shows a stratified analysis of total cholesterol versus lumbar spine BMD. Except for non-Hispanic Black and Mexican American race / ethnicity, with the lowest tertile of total cholesterol as the reference, the β items in the highest tertile of total cholesterol were negative, and P < 0.01, a statistically significant difference.
Table 2
Stratified analysis between total cholesterol and lumbar bone mineral density.
|
Total cholesterol (tertiles)
|
Low
|
Middle
β (95%CI) P value
|
High
β (95%CI) P value
|
Age
|
|
|
|
60–69 years
|
Reference
|
-0.0271 (-0.0450, -0.0091) 0.0032
|
-0.0552 (-0.0727, -0.0378) < 0.0001
|
>= 70 years
|
Reference
|
-0.0236 (-0.0442, -0.0030) 0.0248
|
-0.0887 (-0.1095, -0.0680) < 0.0001
|
Sex
|
|
|
|
Male
|
Reference
|
-0.0074 (-0.0260, 0.0112) 0.4376
|
-0.0359 (-0.0564, -0.0153) 0.0006
|
Female
|
Reference
|
-0.0115 (-0.0300, 0.0069) 0.2211
|
-0.0343 (-0.0518, -0.0169) 0.0001
|
Race ethnicity
|
|
|
|
Non-Hispanic White
|
Reference
|
-0.0269 (-0.0422, -0.0117) 0.0005
|
-0.0737 (-0.0887, -0.0586) < 0.0001
|
Non-Hispanic Black
|
Reference
|
-0.0218 (-0.0580, 0.0143) 0.2370
|
-0.0319 (-0.0663, 0.0025) 0.0699
|
Mexican American
|
Reference
|
0.0205 (-0.0442, 0.0853) 0.5354
|
0.0183 (-0.0457, 0.0823) 0.5757
|
Other races/ethnicities
|
Reference
|
-0.0372 (-0.1150, 0.0406) 0.3501
|
-0.0861 (-0.1570, -0.0153) 0.0188
|
Physical activity
|
|
|
|
Sedentary
|
Reference
|
-0.0458 (-0.0710, -0.0207) 0.0004
|
-0.0753 (-0.0992, -0.0515) < 0.0001
|
Low
|
Reference
|
-0.0564 (-0.0841, -0.0287) < 0.0001
|
-0.0818 (-0.1089, -0.0548) < 0.0001
|
Moderate
|
Reference
|
-0.0124 (-0.0490, 0.0242) 0.5057
|
-0.0859 (-0.1212, -0.0507) < 0.0001
|
High
|
Reference
|
0.0043 (-0.0217, 0.0303) 0.7443
|
-0.0360 (-0.0628, -0.0093) 0.0084
|
Income to poverty ratio (tertiles)
|
|
0–1.38
|
Reference
|
-0.0281 (-0.0537, -0.0025) 0.0318
|
-0.0515 (-0.0758, -0.0272) < 0.0001
|
1.39–2.99
|
Reference
|
-0.0381 (-0.0636, -0.0126) 0.0034
|
-0.0784 (-0.1033, -0.0536) < 0.0001
|
3.01–5.00
|
Reference
|
-0.0236 (-0.0466, -0.0007) 0.0440
|
-0.0572 (-0.0803, -0.0342) < 0.0001
|
Blood urea nitrogen (tertiles, mg/dL)
|
|
2.00–12.00
|
Reference
|
-0.0356 (-0.0601, -0.0110) 0.0046
|
-0.0527 (-0.0768, -0.0286) < 0.0001
|
13.00–16.00
|
Reference
|
-0.0261 (-0.0507, -0.0016) 0.0369
|
-0.0627 (-0.0867, -0.0386) < 0.0001
|
17.00–122.00
|
Reference
|
-0.0182 (-0.0400, 0.0036) 0.1015
|
-0.0805 (-0.1020, -0.0590) < 0.0001
|
Total protein (tertiles, mg/dL)
|
|
|
5.40–7.00
|
Reference
|
-0.0177 (-0.0425, 0.0070) 0.1601
|
-0.0631 (-0.0887, -0.0374) < 0.0001
|
7.10–7.40
|
Reference
|
-0.0363 (-0.0602, -0.0124) 0.0030
|
-0.0529 (-0.0757, -0.0301) < 0.0001
|
7.50–11.00
|
Reference
|
-0.0253 (-0.0480, -0.0026) 0.0293
|
-0.0834 (-0.1055, -0.0613) < 0.0001
|
Serum uric acid (tertiles, mg/dL)
|
|
|
1.50–4.80
|
Reference
|
-0.0054 (-0.0287, 0.0179) 0.6491
|
-0.0559 (-0.0783, -0.0334) < 0.0001
|
4.90–6.00
|
Reference
|
-0.0086 (-0.0323, 0.0151) 0.4757
|
-0.0466 (-0.0700, -0.0233) < 0.0001
|
6.10–13.70
|
Reference
|
-0.0399 (-0.0624, -0.0175) 0.0005
|
-0.0656 (-0.0882, -0.0430) < 0.0001
|
Serum calcium (tertiles, mg/dL)
|
|
|
6.70–9.20
|
Reference
|
-0.0441 (-0.0693, -0.0188) 0.0007
|
-0.0611 (-0.0878, -0.0343) < 0.0001
|
9.30–9.50
|
Reference
|
-0.0015 (-0.0259, 0.0229) 0.9040
|
-0.0673 (-0.0909, -0.0438) < 0.0001
|
9.60–12.50
|
Reference
|
-0.0286 (-0.0502, -0.0070) 0.0096
|
-0.0638 (-0.0847, -0.0429) < 0.0001
|
Body mass index (tertiles)
|
|
|
15.18–25.91
|
Reference
|
-0.0008 (-0.0233, 0.0218) 0.9476
|
-0.0532 (-0.0757, -0.0307) < 0.0001
|
25.93–30.04
|
Reference
|
-0.0135 (-0.0376, 0.0106) 0.2733
|
-0.0630 (-0.0864, -0.0395) < 0.0001
|
30.05–58.58
|
Reference
|
-0.0360 (-0.0587, -0.0133) 0.0019
|
-0.0604 (-0.0826, -0.0382) < 0.0001
|
We performed a univariate analysis of lumbar BMD (Table 3). Compared with men, women’s lumbar spine BMD was reduced by 0.15, a relatively large difference that was statistically significant (P < 0.001). Other results are shown in Table 3.
Table 3
Univariate analysis for lumber spine bone mineral density.
|
Statistics
|
β (95%CI)
|
P-value
|
Total cholesterol
|
209.81 ± 42.95
|
-0.00 (-0.00, -0.00)
|
< 0.0001
|
Total cholesterol quartiles
|
|
Low
|
1405 (33.17%)
|
Reference
|
|
Middle
|
1388 (32.77%)
|
-0.02 (-0.04, -0.01)
|
0.0008
|
High
|
1443 (34.07%)
|
-0.06 (-0.08, -0.05)
|
< 0.0001
|
Sex
|
|
|
|
Male
|
2062 (48.68%)
|
Reference
|
|
Female
|
2174 (51.32%)
|
-0.15 (-0.16, -0.15)
|
< 0.0001
|
Race/ethnicity
|
|
|
Non-Hispanic white
|
3231 (76.27%)
|
Reference
|
|
Non-Hispanic black
|
740 (17.47%)
|
0.08 (0.07, 0.09)
|
< 0.0001
|
Mexican American
|
144 (3.40%)
|
-0.04 (-0.07, -0.02)
|
0.0002
|
Other races/ethnicities
|
121 (2.86%)
|
-0.01 (-0.03, 0.02)
|
0.4821
|
Age
|
|
|
60–69 years
|
2294 (54.15%)
|
Reference
|
|
>= 70 years
|
1942 (45.85%)
|
-0.02 (-0.03, -0.00)
|
0.0109
|
Income to poverty ratio
|
2.46 ± 1.51
|
0.02 (0.02, 0.02)
|
< 0.0001
|
Income to poverty ratio (tertile)
|
|
Low
|
1277 (33.31%)
|
Reference
|
|
Middle
|
1279 (33.36%)
|
0.03 (0.02, 0.04)
|
< 0.0001
|
High
|
1278 (33.33%)
|
0.06 (0.05, 0.07)
|
< 0.0001
|
Blood urea nitrogen
|
16.46 ± 7.46
|
-0.00 (-0.00, 0.00)
|
< 0.0001
|
Blood urea nitrogen (tertile)
|
|
Low
|
1178 (27.92%)
|
Reference
|
|
Middle
|
1348 (31.95%)
|
-0.00 (-0.01, 0.01)
|
0.9751
|
High
|
1693 (40.13%)
|
-0.01 (-0.02, 0.00)
|
0.0886
|
Total protein
|
7.33 ± 0.52
|
-0.01 (-0.01, 0.00)
|
0.1642
|
Total protein (tertile)
|
|
Low
|
1249 (29.62%)
|
Reference
|
|
Middle
|
1345 (31.89%)
|
-0.00 (-0.01, 0.01)
|
0.8438
|
High
|
1623 (38.49%)
|
-0.01 (-0.02, -0.00)
|
0.0254
|
Serum uric acid
|
5.66 ± 1.47
|
0.02 (0.02, 0.03)
|
< 0.0001
|
Serum uric acid (tertile)
|
|
|
|
Low
|
1304 (30.91%)
|
Reference
|
|
Middle
|
1412 (33.47%)
|
0.05 (0.04, 0.06)
|
< 0.0001
|
High
|
1503 (35.62%)
|
0.08 (0.07, 0.09)
|
< 0.0001
|
Continued Table 3.
|
Statistics
|
β (95%CI)
|
P-value
|
Serum calcium
|
9.48 ± 0.42
|
-0.01 (-0.02, -0.00)
|
0.0087
|
Serum calcium (tertile)
|
|
|
|
Low
|
1172 (27.78%)
|
Reference
|
|
Middle
|
1301 (30.84%)
|
0.01 (-0.00, 0.02)
|
0.1108
|
High
|
1746 (41.38%)
|
-0.00 (-0.01, 0.01)
|
0.5408
|
Physical activity
|
|
|
Sedentary
|
1328 (31.35%)
|
Reference
|
|
Low
|
1026 (24.22%)
|
0.04 (0.03, 0.05)
|
<0.0001
|
Moderate
|
619 (14.61%)
|
0.05 (0.04, 0.06)
|
<0.0001
|
High
|
997 (23.54%)
|
0.06 (0.05, 0.08)
|
<0.0001
|
Not recorded
|
266 (6.28%)
|
0.07 (0.05, 0.09)
|
<0.0001
|
BMI
|
28.60 ± 5.65
|
0.00 (0.00, 0.01)
|
<0.0001
|
BMI (tertile)
|
|
Low
|
1383 (33.32%)
|
Reference
|
|
Middle
|
1382 (33.29%)
|
0.05 (0.04, 0.06)
|
<0.0001
|
High
|
1386 (33.39%)
|
0.06 (0.05, 0.07)
|
<0.0001
|
Abbreviations: CI, confidence interval; BMI, Body mass index.
The results of the multiple regression equation analysis are shown in Table 4. In the unadjusted model, total cholesterol was negatively correlated with lumbar spine BMD (β=-0.0007, 95% confidence interval [CI]: -0.0008, -0.0005, P < 0.001). After adjusting for confounding factors, this correlation still existed in Model 2 (β=-0.0004, 95% CI: -0.0005, -0.0002, P < 0.001), and Model 3 (β=-0.0003, 95% CI: -0.0004, -0.0001, P < 0.001), with P for trend < 0.001. We also stratified by continuous variables (third quantiles) and categorical variables; The results are shown in Table 4.
Table 4
Relationship between total cholesterol and lumbar spine bone mineral density.
Outcome
|
Crude Model
|
Model Ⅰ
|
Model Ⅱ
|
β (95%CI)
|
P-value
|
β (95%CI)
|
P-value
|
β (95%CI)
|
P-value
|
Total cholesterol
|
-0.0007 (-0.0008, -0.0005)
|
< 0.0001
|
-0.0004 (-0.0005, -0.0002)
|
< 0.0001
|
-0.0003 (-0.0004, -0.0001)
|
< 0.0001
|
Total cholesterol (tertiles)
|
|
|
|
|
|
Low
|
Refrence
|
|
Refrence
|
|
Refrence
|
|
Middle
|
-0.0255 (-0.0390, -0.0119)
|
0.0002
|
-0.0079 (-0.0208, 0.0050)
|
0.2312
|
-0.0016 (-0.0151, 0.0118)
|
0.8147
|
High
|
-0.0669 (-0.0802, -0.0535)
|
< 0.0001
|
-0.0344 (-0.0474, -0.0214)
|
< 0.0001
|
-0.0195 (-0.0330, -0.0060)
|
0.0047
|
P for trend
|
< 0.001
|
|
< 0.001
|
|
< 0.001
|
|
Sex
|
|
|
|
|
|
|
Male
|
-0.0004 (-0.0006, -0.0002)
|
< 0.0001
|
-0.0004 (-0.0006, -0.0002)
|
0.0001
|
-0.0003 (-0.0005, -0.0001)
|
0.0013
|
Female
|
-0.0003 (-0.0005, -0.0002)
|
0.0001
|
-0.0003 (-0.0005, -0.0002)
|
< 0.0001
|
-0.0002 (-0.0003, -0.0000)
|
0.0460
|
Age
|
|
|
|
|
|
|
60–69 years
|
-0.0006 (-0.0007, -0.0004) <
|
< 0.0001
|
-0.0003 (-0.0005, -0.0002)
|
< 0.0001
|
-0.0002 (-0.0003, 0.0000)
|
0.0590
|
>= 70 years
|
-0.0008 (-0.0010, -0.0006)
|
< 0.0001
|
-0.0004 (-0.0006, -0.0002)
|
< 0.0001
|
-0.0003 (-0.0005, -0.0001)
|
0.0007
|
Race/ethnicity
|
|
|
|
|
|
|
Non-Hispanic White
|
-0.0007 (-0.0009, -0.0006)
|
< 0.0001
|
-0.0004 (-0.0005, -0.0003)
|
< 0.0001
|
-0.0003 (-0.0004, -0.0002)
|
< 0.0001
|
Non-Hispanic Black
|
-0.0004 (-0.0007, -0.0000)
|
0.0267
|
-0.0002 (-0.0005, 0.0002)
|
0.3208
|
-0.0001 (-0.0004, 0.0003)
|
0.6625
|
Mexican American
|
-0.0001 (-0.0008, 0.0005)
|
0.7457
|
-0.0000 (-0.0007, 0.0006)
|
0.9187
|
0.0004 (-0.0003, 0.0011)
|
0.3169
|
Other races/ethnicities
|
-0.0006 (-0.0013, 0.0001)
|
0.0902
|
-0.0004 (-0.0011, 0.0003)
|
0.2223
|
-0.0003 (-0.0010, 0.0004)
|
0.3774
|
Serum uric acid (tertiles)
|
|
|
|
|
|
Low
|
-0.0006 (-0.0008, -0.0004)
|
< 0.0001
|
-0.0004 (-0.0006, -0.0002)
|
0.0005
|
-0.0003 (-0.0005, -0.0001)
|
0.0093
|
Middle
|
-0.0007 (-0.0009, -0.0004)
|
< 0.0001
|
-0.0004 (-0.0007, -0.0002)
|
0.0001
|
-0.0004 (-0.0006, -0.0002)
|
0.0012
|
High
|
-0.0005 (-0.0007, -0.0003)
|
< 0.0001
|
-0.0003 (-0.0005, -0.0001)
|
0.1351
|
-0.0001 (-0.0003, 0.0001)
|
0.3774
|
Physical activity (tertiles)
|
|
|
|
|
|
|
Sedentary
|
-0.0007 (-0.0009, -0.0005)
|
< 0.0001
|
-0.0004 (-0.0006, -0.0002)
|
0.0003
|
-0.0003 (-0.0005, -0.0000)
|
0.0231
|
Low
|
-0.0007 (-0.0009, -0.0004)
|
< 0.0001
|
-0.0004 (-0.0007, -0.0002)
|
0.0007
|
-0.0002 (-0.0005, 0.0000)
|
0.0978
|
Moderate
|
-0.0010 (-0.0014, -0.0007)
|
< 0.0001
|
-0.0006 (-0.0010, -0.0003)
|
0.0006
|
-0.0004 (-0.0007, -0.0000)
|
0.0432
|
High
|
-0.0005 (-0.0008, -0.0003)
|
0.0001
|
-0.0003 (-0.0005, -0.0001)
|
0.0186
|
-0.0002 (-0.0005, 0.0000)
|
0.1155
|
Continued Table 4.
Outcome
|
Crude Model
|
Model Ⅰ
|
Model Ⅱ
|
β (95%CI)
|
P-value
|
β (95%CI)
|
P-value
|
β (95%CI)
|
P-value
|
Income to poverty ratio (tertiles)
|
|
|
|
|
|
Low
|
-0.0004 (-0.0006, -0.0002)
|
0.0003
|
-0.0002 (-0.0004, -0.0000)
|
0.0252
|
-0.0001 (-0.0003, 0.0001)
|
0.2724
|
Middle
|
-0.0007 (-0.0010, -0.0005)
|
<0.0001
|
-0.0004 (-0.0007, -0.0002)
|
0.0002
|
-0.0003 (-0.0005, -0.0001)
|
0.0121
|
High
|
-0.0007 (-0.0009, -0.0004)
|
<0.0001
|
-0.0004 (-0.0006, -0.0001)
|
0.0034
|
-0.0003 (-0.0006, -0.0001)
|
0.0062
|
Blood urea nitrogen (tertiles)
|
|
|
|
|
|
Low
|
-0.0005 (-0.0008, -0.0003)
|
<0.0001
|
-0.0002 (-0.0004, 0.0000)
|
0.0584
|
-0.0002 (-0.0004, 0.0000)
|
0.1151
|
Middle
|
-0.0007 (-0.0009, -0.0004)
|
<0.0001
|
-0.0004 (-0.0006, -0.0002)
|
0.0010
|
-0.0003 (-0.0005, -0.0001)
|
0.0181
|
High
|
-0.0007 (-0.0009, -0.0005)
|
<0.0001
|
-0.0005 (-0.0007, -0.0003)
|
<0.0001
|
-0.0003 (-0.0005, -0.0001)
|
0.0062
|
Total protein (tertiles)
|
|
|
|
|
|
|
Low
|
-0.0006 (-0.0008, -0.0003)
|
<0.0001
|
-0.0003 (-0.0005, -0.0001)
|
0.0103
|
-0.0002 (-0.0004, 0.0001)
|
0.1457
|
Middle
|
-0.0006 (-0.0008, -0.0004)
|
<0.0001
|
-0.0003 (-0.0005, -0.0001)
|
0.0060
|
-0.0003 (-0.0005, -0.0001)
|
0.0171
|
High
|
-0.0008 (-0.0010, -0.0006)
|
<0.0001
|
-0.0004 (-0.0006, -0.0002)
|
<0.0001
|
-0.0003 (-0.0005, -0.0001)
|
0.0018
|
Serum calcium (tertiles)
|
|
|
|
|
|
|
Low
|
-0.0007 (-0.0009, -0.0004)
|
<0.0001
|
-0.0004 (-0.0007, -0.0002)
|
0.0007
|
-0.0004 (-0.0007, -0.0001)
|
0.0026
|
Middle
|
-0.0005 (-0.0007, -0.0003)
|
<0.0001
|
-0.0002 (-0.0004, 0.0001)
|
0.1429
|
-0.0001 (-0.0003, 0.0001)
|
0.2873
|
High
|
-0.0007 (-0.0009, -0.0005)
|
<0.0001
|
-0.0005 (-0.0007, -0.0003)
|
<0.0001
|
-0.0003 (-0.0005, -0.0001)
|
0.0071
|
BMI (tertiles)
|
|
|
|
|
|
|
Low
|
-0.0006 (-0.0009, -0.0004)
|
<0.0001
|
-0.0004 (-0.0006, -0.0002)
|
0.0002
|
-0.0004 (-0.0007, -0.0002)
|
<0.0001
|
Middle
|
-0.0006 (-0.0008, -0.0003)
|
<0.0001
|
-0.0002 (-0.0004, 0.0000)
|
0.1046
|
-0.0001 (-0.0003, 0.0002)
|
0.6275
|
High
|
-0.0007 (-0.0009, -0.0004)
|
<0.0001
|
-0.0004 (-0.0006, -0.0001)
|
0.0012
|
-0.0004 (-0.0007, -0.0002)
|
0.0002
|
Outcome variable: Total BMD. Exposure variable: Total cholesterol.
The crude model is not adjusted. Model Ⅰ adjusted for age, sex, and serum uric acid. Model Ⅱ adjusted for age, sex, serum uric acid, race ethnicity, education, income to poverty ratio, blood urea nitrogen, total protein, serum uric acid, serum calcium, physical activity, BMI.
Abbreviations: CI, confidence interval; BMD, bone mineral density BMI, body mass index.
We also performed weighted generalized additive models and smooth curve fitting to assess their association (Fig. 2). Figure 2 shows that total cholesterol was linearly negatively correlated with lumbar BMD. We also performed smooth curve fitting in subgroups stratified by categorical variables (Fig. 3). Figure 3a shows a turning point in the fitting curve for elderly people 70 years or older, at a total cholesterol of 280 mg/dL (Table 5). Figure 3d shows that the fitting curve of older people with more than moderate physical activity had a turning point at a total cholesterol of 203 mg/dL (Table 5). Other turning points are shown in Fig. 4 and Table 5.
Table 5
Threshold effect and saturation effect analysis.
|
Model I
|
Model II
|
|
A straight-line effect
|
Fold points (K)
|
< K-segment effect 1
|
>K-segment Effect 2
|
Effect size difference of 2 versus 1
|
Equation predicted values at break points
|
Log likelihood ratio tests
|
Total
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0003
(-0.0004, -0.0002)
|
146
|
0.0003
(-0.0011, 0.0017)
|
-0.0003
(-0.0005, -0.0002)
|
-0.0006
(-0.0020, 0.0009)
|
1.0584
(1.0461, 1.0706)
|
0.430
|
P-value
|
< 0.0001
|
|
0.7028
|
< 0.0001
|
0.4308
|
|
|
Stratification variables
|
|
|
|
|
|
|
Age
|
|
|
|
|
|
|
|
60–69 years
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0002
(-0.0004, -0.0001)
|
256
|
-0.0001
(-0.0003, 0.0001)
|
-0.0007
(-0.0011, -0.0002)
|
-0.0006
(-0.0012, -0.0000)
|
1.0010
(0.9858, 1.0161)
|
0.034
|
P-value
|
0.0057
|
|
0.4673
|
0.0027
|
0.0350
|
|
|
>= 70 years
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0003
(-0.0006, -0.0001)
|
280
|
-0.0005
(-0.0007, -0.0002)
|
0.0006
(-0.0002, 0.0014)
|
0.0011
(0.0001, 0.0020)
|
0.9196
(0.8978, 0.9415)
|
0.023
|
P-value
|
0.0016
|
|
0.0001
|
0.1672
|
0.0236
|
|
|
Sex
|
|
|
|
|
|
|
|
Male
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0004
(-0.0006, -0.0002)
|
138
|
0.0012
(-0.0013, 0.0037)
|
-0.0004
(-0.0007, -0.0002)
|
-0.0017
(-0.0042, 0.0009)
|
1.1131
(1.0957, 1.1305)
|
0.196
|
P-value
|
0.0001
|
|
0.3336
|
< 0.0001
|
0.1985
|
|
|
Continued table 5.
|
Model I
|
Model II
|
|
A straight-line effect
|
Fold points (K)
|
< K-segment effect 1
|
> K-segment Effect 2
|
Effect size difference of 2 versus 1
|
Equation predicted values at break points
|
Log likelihood ratio tests
|
Female
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0002
(-0.0003, 0.0000)
|
207
|
0.0003
(-0.0002, 0.0007)
|
-0.0004
(-0.0007, -0.0001)
|
-0.0006
(-0.0012, -0.0001)
|
0.9638
(0.9511, 0.9765)
|
0.033
|
P-value
|
0.0658
|
|
0.2311
|
0.0050
|
0.0337
|
|
|
Race/ethnicity
|
|
|
|
|
|
|
|
Non-Hispanic White
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0004
(-0.0005, -0.0002)
|
146
|
0.0007
(-0.0010, 0.0023)
|
-0.0004
(-0.0006, -0.0002)
|
-0.0011
(-0.0027, 0.0006)
|
1.0505
(1.0371, 1.0639)
|
0.216
|
P-value
|
<0.0001
|
|
0.4327
|
<0.0001
|
0.2171
|
|
|
Non-Hispanic Black
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0001
(-0.0004, 0.0003)
|
154
|
-0.0009
(-0.0038, 0.0019)
|
-0.0000
(-0.0004, 0.0004)
|
0.0009
(-0.0021, 0.0039)
|
1.1161
(1.0867, 1.1455)
|
0.544
|
P-value
|
0.7282
|
|
0.5234
|
0.9422
|
0.5485
|
|
|
Mexican American
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
0.0001
(-0.0006, 0.0008)
|
181
|
-0.0033
(-0.0070, 0.0003)
|
0.0005
(-0.0003, 0.0013)
|
0.0038
(-0.0001, 0.0078)
|
0.9334
(0.8814, 0.9854)
|
0.045
|
P-value
|
0.8348
|
|
0.0730
|
0.2362
|
0.0621
|
|
|
Continued table 5.
|
Model I
|
Model II
|
|
A straight-line effect
|
Fold points (K)
|
< K-segment effect 1
|
>K-segment Effect 2
|
Effect size difference of 2 versus 1
|
Equation predicted values at break points
|
Log likelihood ratio tests
|
Other races/ethnicities
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0001
(-0.0009, 0.0007)
|
132
|
-0.0016
(-0.0100, 0.0068)
|
-0.0000
(-0.0009, 0.0008)
|
0.0016
(-0.0071, 0.0103)
|
1.0036
(0.9288, 1.0783)
|
0.701
|
P-value
|
0.8128
|
|
0.7060
|
0.9332
|
0.7219
|
|
|
Physical activity
|
|
|
|
|
|
|
|
Sedentary
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0003
(-0.0005, -0.0000)
|
146
|
0.0008
(-0.0017, 0.0032)
|
-0.0003
(-0.0005, -0.0001)
|
-0.0011
(-0.0035, 0.0014)
|
1.0338
(1.0119, 1.0556)
|
0.403
|
P-value
|
0.0238
|
|
0.5397
|
0.0161
|
0.4063
|
|
|
Low
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0003
(-0.0006, -0.0000)
|
272
|
-0.0005
(-0.0008, -0.0002)
|
0.0008
(-0.0000, 0.0017)
|
0.0013
(0.0004, 0.0023)
|
0.9424
(0.9163, 0.9684)
|
0.006
|
P-value
|
0.0271
|
|
0.0010
|
0.0614
|
0.0069
|
|
|
Moderate
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0002
(-0.0006, 0.0001)
|
203
|
0.0007
(-0.0001, 0.0015)
|
-0.0008
(-0.0014, -0.0002)
|
-0.0015
(-0.0027, -0.0003)
|
1.0491
(1.0237, 1.0745)
|
0.011
|
P-value
|
0.2225
|
|
0.0917
|
0.0063
|
0.0118
|
|
|
Continued table 5.
|
Model I
|
Model II
|
|
A straight-line effect
|
Fold points (K)
|
< K-segment effect 1
|
>K-segment Effect 2
|
Effect size difference of 2 versus 1
|
Equation predicted values at break points
|
Log likelihood ratio tests
|
High
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0003
(-0.0006, -0.0001)
|
256
|
-0.0002
(-0.0005, 0.0002)
|
-0.0008
(-0.0014, -0.0001)
|
-0.0006
(-0.0014, 0.0002)
|
1.0021
(0.9797, 1.0245)
|
0.133
|
P-value
|
0.0115
|
|
0.3555
|
0.0163
|
0.1358
|
|
|
Income to poverty ratio (tertiles)
|
|
|
|
|
|
|
Low
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0002
(-0.0004, -0.0000)
|
208
|
-0.0000
(-0.0005, 0.0004)
|
-0.0003
(-0.0007, -0.0000)
|
-0.0003
(-0.0010, 0.0003)
|
0.9847
(0.9684, 1.0010)
|
0.329
|
P-value
|
0.0307
|
|
0.9188
|
0.0309
|
0.3320
|
|
|
Middle
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0004
(-0.0006, -0.0002)
|
145
|
0.0008
(-0.0022, 0.0039)
|
-0.0004
(-0.0007, -0.0002)
|
-0.0013
(-0.0044, 0.0018)
|
1.0795
(1.0571, 1.1019)
|
0.416
|
P-value
|
0.0010
|
|
0.5864
|
0.0008
|
0.4196
|
|
|
High
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0003
(-0.0005, -0.0000)
|
150
|
-0.0016
(-0.0040, 0.0009)
|
-0.0002
(-0.0005, 0.0001)
|
0.0013
(-0.0012, 0.0039)
|
1.0692
(1.0483, 1.0901)
|
0.294
|
P-value
|
0.0369
|
|
0.2090
|
0.1192
|
0.2978
|
|
|
Continued table 5.
|
Model I
|
Model II
|
|
A straight-line effect
|
Fold points (K)
|
< K-segment effect 1
|
>K-segment Effect 2
|
Effect size difference of 2 versus 1
|
Equation predicted values at break points
|
Log likelihood ratio tests
|
Blood urea nitrogen (tertiles)
|
|
|
|
|
|
|
Low
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0003
(-0.0006, -0.0000)
|
146
|
0.0011
(-0.0013, 0.0035)
|
-0.0004
(-0.0006, -0.0001)
|
-0.0015
(-0.0040, 0.0010)
|
1.0490
(1.0263, 1.0717)
|
0.246
|
P-value
|
0.0197
|
|
0.3696
|
0.0096
|
0.2495
|
|
|
Middle
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0003
(-0.0005, -0.0000)
|
281
|
-0.0003
(-0.0006, -0.0001)
|
0.0004
(-0.0008, 0.0016)
|
0.0007
(-0.0006, 0.0020)
|
0.9554
(0.9299, 0.9808)
|
0.285
|
P-value
|
0.0357
|
|
0.0188
|
0.5471
|
0.2890
|
|
|
High
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0003
(-0.0005, -0.0001)
|
143
|
0.0016
(-0.0009, 0.0040)
|
-0.0003
(-0.0005, -0.0001)
|
-0.0019
(-0.0044, 0.0006)
|
1.0752
(1.0557, 1.0946)
|
0.139
|
P-value
|
0.0072
|
|
0.2138
|
0.0025
|
0.1413
|
|
|
Total protein (tertiles)
|
|
|
|
|
|
|
Low
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0003
(-0.0005, -0.0001)
|
142
|
0.0023
(-0.0001, 0.0047)
|
-0.0004
(-0.0006, -0.0001)
|
-0.0027
(-0.0052, -0.0002)
|
1.0742
(1.0524, 1.0960)
|
0.031
|
P-value
|
0.0138
|
|
0.0586
|
0.0024
|
0.0321
|
|
|
Continued table 5.
|
Model I
|
Model II
|
|
A straight-line effect
|
Fold points (K)
|
< K-segment effect 1
|
>K-segment Effect 2
|
Effect size difference of 2 versus 1
|
Equation predicted values at break points
|
Log likelihood ratio tests
|
Middle
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0003
(-0.0005, -0.0000)
|
187
|
-0.0009
(-0.0017, -0.0001)
|
-0.0001
(-0.0004, 0.0002)
|
0.0008
(-0.0002, 0.0018)
|
1.0134
(0.9952, 1.0316)
|
0.105
|
P-value
|
0.0216
|
|
0.0243
|
0.5513
|
0.1076
|
|
|
High
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0003
(-0.0005, -0.0001)
|
152
|
-0.0008
(-0.0028, 0.0012)
|
-0.0002
(-0.0005, -0.0000)
|
0.0006
(-0.0015, 0.0027)
|
1.0448
(1.0249, 1.0648)
|
0.588
|
P-value
|
0.0152
|
|
0.4247
|
0.0408
|
0.5907
|
|
|
Serum uric acid (tertiles)
|
|
|
|
|
|
|
Low
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0003
(-0.0005, -0.0000)
|
218
|
0.0000
(-0.0004, 0.0005)
|
-0.0006
(-0.0010, -0.0002)
|
-0.0006
(-0.0014, 0.0001)
|
0.9576
(0.9406, 0.9745)
|
0.089
|
P-value
|
0.0198
|
|
0.8597
|
0.0072
|
0.0912
|
|
|
Middle
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0003
(-0.0006, -0.0001)
|
278
|
-0.0002
(-0.0005, 0.0001)
|
-0.0009
(-0.0016, -0.0002)
|
-0.0007
(-0.0015, 0.0002)
|
0.9699
(0.9460, 0.9939)
|
0.113
|
P-value
|
0.0068
|
|
0.1472
|
0.0170
|
0.1157
|
|
|
Continued table 5.
|
Model I
|
Model II
|
|
A straight-line effect
|
Fold points (K)
|
< K-segment effect 1
|
>K-segment Effect 2
|
Effect size difference of 2 versus 1
|
Equation predicted values at break points
|
Log likelihood ratio tests
|
High
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0003
(-0.0005, -0.0000)
|
279
|
-0.0004
(-0.0007, -0.0001)
|
0.0005
(-0.0003, 0.0012)
|
0.0008
(-0.0000, 0.0017)
|
1.0041
(0.9806, 1.0276)
|
0.055
|
P-value
|
0.0223
|
|
0.0033
|
0.2482
|
0.0568
|
|
|
Serum calcium (tertiles)
|
|
|
|
|
|
|
|
Low
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0006
(-0.0009, -0.0003)
|
141
|
0.0005
(-0.0020, 0.0029)
|
-0.0006
(-0.0009, -0.0004)
|
-0.0011
(-0.0037, 0.0014)
|
1.0856
(1.0619, 1.1092)
|
0.384
|
P-value
|
<0.0001
|
|
0.7064
|
<0.0001
|
0.3882
|
|
|
Middle
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0001
(-0.0003, 0.0001)
|
268
|
-0.0002
(-0.0005, 0.0001)
|
0.0005
(-0.0003, 0.0012)
|
0.0007
(-0.0002, 0.0016)
|
0.9699
(0.9467, 0.9930)
|
0.114
|
P-value
|
0.4314
|
|
0.1298
|
0.2171
|
0.1170
|
|
|
High
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0002
(-0.0004, -0.0000)
|
208
|
0.0003
(-0.0002, 0.0007)
|
-0.0005
(-0.0008, -0.0002)
|
-0.0007
(-0.0014, -0.0001)
|
1.0152
(0.9998, 1.0305)
|
0.028
|
P-value
|
0.0408
|
|
0.2825
|
0.0029
|
0.0294
|
|
|
Continued table 5.
|
Model I
|
Model II
|
|
A straight-line effect
|
Fold points (K)
|
< K-segment effect 1
|
>K-segment Effect 2
|
Effect size difference of 2 versus 1
|
Equation predicted values at break points
|
Log likelihood ratio tests
|
Body mass index (tertiles)
|
|
|
|
|
|
|
Low
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0004
(-0.0006, -0.0002)
|
259
|
-0.0003
(-0.0006, 0.0000)
|
-0.0007
(-0.0013, -0.0001)
|
-0.0004
(-0.0011, 0.0003)
|
0.9217
(0.9020, 0.9414)
|
0.261
|
P-value
|
0.0010
|
|
0.0597
|
0.0198
|
0.2648
|
|
|
Middle
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0002
(-0.0005, 0.0000)
|
203
|
0.0001
(-0.0005, 0.0007)
|
-0.0004
(-0.0008, -0.0000)
|
-0.0005
(-0.0013, 0.0003)
|
1.0288
(1.0116, 1.0461)
|
0.238
|
P-value
|
0.0651
|
|
0.7904
|
0.0358
|
0.2419
|
|
|
High
|
|
|
|
|
|
|
|
Adjusted β
(95% CI)
|
-0.0004
(-0.0006, -0.0002)
|
282
|
-0.0004
(-0.0007, -0.0002)
|
0.0001
(-0.0011, 0.0013)
|
0.0005
(-0.0007, 0.0018)
|
1.0071
(0.9833, 1.0310)
|
0.406
|
P-value
|
0.0009
|
|
0.0010
|
0.8715
|
0.4095
|
|
|
Note: Adjusted β (95% CI), P-value. Outcome variables: lumber spine BMD. exposure variables: total cholesterol. Adjustment variables: age, sex, race/ethnicity, income to poverty ratio, blood urea nitrogen, total protein, serum uric acid, serum calcium, physical activity and body mass.