Characteristics of participants
Table 1 shows the characteristics of the study participants (6,281 men and 8,263 women). To analyze the differences between the sexes, men and women were separately classified into groups according to their hs-CRP levels: low hs-CRP (≤1.0 mg/L) and high hs-CRP (>1.0 mg/L). Among all participants, the percentage of those with very poor to poor SRH was higher in the high hs-CRP group (22.47%) than in the low hs-CRP group (17.66%). Similarly, for both men and women, the percentages of participants with very poor to poor SRH were higher in the high hs-CRP groups than in the low hs-CRP groups, whereas the percentages of participants with good to very good SRH were lower in the low hs-CRP groups than in the high hs-CRP groups (P<0.001) (Fig. 2).
Table 1 Baseline characteristics of the study population in the 2015-2017 KNHANES
Characteristic
|
Men
|
Women
|
Low hs-CRP (≤1.0 mg/L) (n=4,495)
|
High hs-CRP (>1.0 mg/L) (n=1,786)
|
P-value†
|
Low hs-CRP (≤1.0 mg/L) (n=6,306)
|
High hs-CRP (>1.0 mg/L) (n=1,957)
|
P-value†
|
Age (years), mean (SD)
|
49.8 (0.249)
|
53.3(0.391)
|
.000*
|
49.9 (0.203)
|
54.2 (0.373)
|
.000*
|
Educational level, n (%)
|
|
|
.000*
|
|
|
.000*
|
Elementary school or lower
|
580 (12.9)
|
333 (18.6)
|
|
1,424 (22.6)
|
653 (33.4)
|
|
Middle school
|
446 (9.9)
|
205 (11.5)
|
|
651 (10.3)
|
203 (10.4)
|
|
High school
|
1,575 (35.0)
|
568 (31.8)
|
|
2,001 (31.7)
|
547 (28.0)
|
|
College or higher
|
1,894 (42.1)
|
680 (38.1)
|
|
2,230 (35.4)
|
554 (28.3)
|
|
Marital status, n (%)
|
|
|
.002*
|
|
|
.000*
|
Married
|
3,536 (78.7)
|
1,487 (83.3)
|
|
5,396 (85.6)
|
1,772 (90.5)
|
|
Unmarried
|
959 (21.3)
|
299 (16.7)
|
|
910 (14.4)
|
185 (9.5)
|
|
Household income level, n (%)
|
|
|
.101
|
|
|
.009*
|
Low
|
1,034 (23.0)
|
457 (25.6)
|
|
1,444 (22.9)
|
522 (26.7)
|
|
Lower middle
|
1,077 (24.0)
|
464 (26.0)
|
|
1,591 (25.2)
|
519 (26.5)
|
|
Upper middle
|
1,156 (25.7)
|
435 (24.4)
|
|
1,609 (25.5)
|
470 (24.0)
|
|
High
|
1,228 (27.3)
|
430 (24.1)
|
|
1,662 (26.4)
|
446 (22.8)
|
|
Weight (kg), mean (SD)
|
70.0 (0.158)
|
73.4 (0.351)
|
.000*
|
56.8 (0.103)
|
62.1 (0.253)
|
.000*
|
Height (cm), mean (SD)
|
170.5 (0.101)
|
169.6 (0.161)
|
.000*
|
157.5 (0.081)
|
156.2 (0.149)
|
.000*
|
BMI (kg/m2), mean (SD)
|
24.0 (0.044)
|
25.4 (0.087)
|
.000*
|
22.9 (0.040)
|
25.4 (0.091)
|
.000*
|
Smoking, n (%)
|
|
|
.123
|
|
|
.483
|
Nonsmoker
|
1,130 (25.1)
|
389 (21.8)
|
|
5,710 (90.5)
|
1,742 (89.0)
|
|
Past smoker
|
1,926 (42.8)
|
777 (43.5)
|
|
341 (5.4)
|
118 (6.0)
|
|
Current smoker
|
1,439 (32.0)
|
620 (34.7)
|
|
255 (4.0)
|
97 (5.0)
|
|
Alcohol consumption, n (%)
|
|
|
.001*
|
|
|
.000*
|
Never drink
|
196 (4.4)
|
85 (4.8)
|
|
951 (15.1)
|
392 (20.0)
|
|
<Once/month
|
1,004 (22.3)
|
457 (25.6)
|
|
2,634 (41.8)
|
864 (44.1)
|
|
<Five times/month
|
1,691 (37.6)
|
577 (32.3)
|
|
2,020 (32.0)
|
496 (25.3)
|
|
≥Five times/month
|
1,604 (35.7)
|
667 (37.3)
|
|
701 (11.1)
|
205 (10.5)
|
|
Chronic diseases**, n (%)
|
|
|
.000*
|
|
|
.000*
|
None
|
2,579 (57.4)
|
789 (44.2)
|
|
4,000 (63.4)
|
966 (49.4)
|
|
One
|
1,186 (26.4)
|
590 (33.0)
|
|
1,274 (20.2)
|
536 (27.4)
|
|
Two or more
|
730 (16.2)
|
407 (22.8)
|
|
1,032 (16.4)
|
455 (23.2)
|
|
Self-rated health, n (%)
|
|
|
.000*
|
|
|
.000*
|
Very poor
|
80 (1.8)
|
59 (3.3)
|
|
270 (4.3)
|
108 (5.5)
|
|
Poor
|
532 (11.8)
|
264 (14.8)
|
|
1,026 (16.3)
|
410 (21.0)
|
|
Fair
|
2,271 (50.5)
|
943 (52.8)
|
|
3,328 (52.8)
|
1,009 (51.6)
|
|
Good
|
1,317 (29.3)
|
431 (24.1)
|
|
1,431 (22.7)
|
370 (18.9)
|
|
Very good
|
295 (6.6)
|
89 (5.0)
|
|
251 (4.0)
|
60 (3.1)
|
|
Menopause, n (%)
|
|
|
|
|
|
.000*
|
No
|
|
|
|
3,211 (50.9)
|
752 (38.4)
|
|
Yes
|
|
|
|
3,072 (48.7)
|
1,189 (60.8)
|
|
* p ≤ .05
** Chronic diseases include hypertension, diabetes, dyslipidemia, coronary heart disease (myocardial infarction or angina pectoris), stroke, and rheumatoid arthritis.
† P-values of t-test or the Mann-Whitney U test for continuous variables and Pearson’s chi-squared test for categorical variables to determine differences between groups according to high-sensitivity C-reactive protein (hs-CRP) levels.
KNHANES, Korea National Health and Nutrition Examination Survey; SE, standard error; SD, standard deviation; BMI, body mass index.
Correlation between SRH and high hs-CRP levels (overall)
Table 2 shows the overall correlations between SRH and high hs-CRP levels along with their respective ORs and 95% CIs. In the crude model (Model 1), the poorer the SRH, the higher was the risk in relation to the reference group (very good SRH). After adjustment for age, educational level, marital status, and household income levels (Model 2), the associations between fair to very poor SRH and high hs-CRP levels remained significant. However, after adjustment for BMI, smoking, and alcohol consumption (Model 3) and after additional adjustment for chronic diseases (Model 4), the associations only remained significant for poor and fair SRH.
Table 2 Association between high hs-CRP level and SRH in all subjects
SRH
|
Model 1
|
Model 2
|
Model 3
|
Model 4
|
OR
|
95% CI
|
P-value
|
OR
|
95% CI
|
P-value
|
OR
|
95% CI
|
P-value
|
OR
|
95% CI
|
P-value
|
Very poor
|
2.02
|
1.47
|
2.78
|
.000*
|
1.54
|
1.11
|
2.12
|
.010*
|
1.22
|
0.87
|
1.72
|
.245
|
1.22
|
0.87
|
1.71
|
.259
|
Poor
|
1.82
|
1.44
|
2.29
|
.000*
|
1.68
|
1.32
|
2.13
|
.000*
|
1.40
|
1.09
|
1.79
|
.008*
|
1.38
|
1.08
|
1.77
|
.011*
|
Fair
|
1.51
|
1.22
|
1.87
|
.000*
|
1.46
|
1.18
|
1.82
|
.001*
|
1.34
|
1.06
|
1.68
|
.013
|
1.32
|
1.05
|
1.66
|
.016*
|
Good
|
1.16
|
0.92
|
1.46
|
.211
|
1.17
|
0.93
|
1.47
|
.183
|
1.12
|
0.88
|
1.42
|
.375
|
1.12
|
0.87
|
1.42
|
.377
|
Very good (ref.)
|
1.00
|
|
|
|
1.00
|
|
|
|
1.00
|
|
|
|
1.00
|
|
|
|
P for trend
|
<0.001*
|
<0.001*
|
<0.001*
|
<0.001*
|
Odds ratios with adjustments using logistic regression models.
* p ≤ .05
Model 1: unadjusted.
Model 2: adjusted for age, educational level, marital status, and household income level.
Model 3: adjusted for Model 2 confounders + body mass index, smoking, and alcohol consumption.
Model 4: adjusted for Model 3 confounders + chronic diseases.
hs-CRP, high-sensitivity C-reactive protein; SRH, self-rated health; OR, odds ratio; CI, confidence interval.
Correlation between SRH and high hs-CRP levels (male vs. female participants)
When we analyzed male participants, those with fair to very poor SRH had a significantly higher risk of being in the high hs-CRP group than those with very good SRH in both the crude model (Model 1) and Model 2 (adjustment for age, educational level, marital status, and household income level). After adjustment for BMI, smoking status, and alcohol consumption status (Model 3), the associations between fair to very poor SRH and high hs-CRP levels remained significant. However, after adjustment for chronic diseases (Model 4), the association only remained significant for very poor and fair SRH (Table 3).
Table 3 Association between high hs-CRP level and SRH in men
SRH
|
Model 1
|
Model 2
|
Model 3
|
Model 4
|
OR
|
95% CI
|
P-value
|
OR
|
95% CI
|
P-value
|
OR
|
95% CI
|
P-value
|
OR
|
95% CI
|
P-value
|
Very poor
|
2.69
|
1.68
|
4.32
|
.000*
|
2.19
|
1.37
|
3.51
|
.001*
|
1.78
|
1.07
|
2.96
|
.026*
|
1.74
|
1.04
|
2.90
|
.034*
|
Poor
|
1.79
|
1.30
|
2.44
|
.000*
|
1.70
|
1.23
|
2.35
|
.001*
|
1.43
|
1.02
|
2.00
|
.041*
|
1.40
|
0.99
|
1.97
|
.057*
|
Fair
|
1.63
|
1.22
|
2.16
|
.001*
|
1.61
|
1.20
|
2.14
|
.001*
|
1.44
|
1.06
|
1.96
|
.019*
|
1.42
|
1.04
|
1.93
|
.025*
|
Good
|
1.12
|
0.83
|
1.51
|
.462
|
1.14
|
0.84
|
1.54
|
.396
|
1.08
|
0.78
|
1.49
|
.637
|
1.08
|
0.78
|
1.49
|
.637
|
Very good (ref.)
|
1.00
|
|
|
|
1.00
|
|
|
|
1.00
|
|
|
|
1.00
|
|
|
|
P for trend
|
<0.001*
|
<0.001*
|
<0.001*
|
<0.001*
|
Odds ratios with adjustments using logistic regression models.
* p ≤ .05
Model 1: unadjusted.
Model 2: adjusted for age, educational level, marital status, and household income level.
Model 3: adjusted for Model 2 confounders + body mass index, smoking, and alcohol consumption.
Model 4: adjusted for Model 3 confounders + chronic diseases.
hs-CRP, high-sensitivity C-reactive protein; SRH, self-rated health; OR, odds ratio; CI, confidence interval.
When we analyzed female participants, those with fair to very poor SRH had a significantly higher risk of being in the high hs-CRP group than those with very good SRH in the crude model (Model 1). After adjustment for age, educational level, marital status, and household income levels (Model 2), the association only remained significant for poor SRH. However, in contrast with that observed for male participants, no significant correlations remained after adjustment for BMI, smoking, and alcohol consumption (Model 3) and after additional adjustment for chronic diseases (Model 4) (Table 4).
Table 4 Association between high hs-CRP level and SRH in women
SRH
|
Model 1
|
Model 2
|
Model 3
|
Model 4
|
OR
|
95% CI
|
P-value
|
OR
|
95% CI
|
P-value
|
OR
|
95% CI
|
P-value
|
OR
|
95% CI
|
P-value
|
Very poor
|
1.97
|
1.28
|
3.01
|
.002*
|
1.35
|
0.87
|
2.10
|
.182
|
0.99
|
0.62
|
1.59
|
.980
|
1.01
|
0.63
|
1.62
|
.968
|
Poor
|
2.03
|
1.43
|
2.90
|
.000*
|
1.77
|
1.24
|
2.53
|
.002*
|
1.39
|
0.95
|
2.03
|
.086
|
1.40
|
0.96
|
2.03
|
.083
|
Fair
|
1.50
|
1.06
|
2.11
|
.021*
|
1.38
|
0.98
|
1.95
|
.063
|
1.27
|
0.89
|
1.80
|
.186
|
1.27
|
0.89
|
1.80
|
.187
|
Good
|
1.25
|
0.88
|
1.78
|
.213
|
1.24
|
0.87
|
1.76
|
.233
|
1.21
|
0.84
|
1.76
|
.301
|
1.22
|
0.84
|
1.76
|
.297
|
Very good (ref.)
|
1.00
|
|
|
|
|
|
|
|
1.00
|
|
|
|
1.00
|
|
|
|
P for trend
|
<0.001*
|
<0.001*
|
<0.001*
|
<0.001*
|
Odds ratios with adjustments using logistic regression models.
* p ≤ .05
Model 1: unadjusted.
Model 2: adjusted for age, educational level, marital status, and household income level.
Model 3: adjusted for Model 2 confounders + body mass index, smoking, alcohol consumption, and menopause.
Model 4: adjusted for Model 3 confounders + chronic diseases.
hs-CRP, high-sensitivity C-reactive protein; SRH, self-rated health; OR, odds ratio; CI, confidence interval.