In total, 2,417 men and 4,568 women were included in this study (Figure 1). Table 1 shows the baseline characteristics of the study population. Among the study subjects, there were 57 men and 54 women newly diagnosed CRC after the entry of this cohort and their median follow-up years were 4.76 (IQR, 2.91-7.79) for men and 5.55 (IQR, 3.06-7.53) for women, respectively. The mean age of the study subjects was approximately 60 years old (data not shown), which is older than the general Korean population. CRC cases were significantly older than non-cancer controls. In men, alcohol consumption was higher in CRC cases, i.e., there was a high percentage of non-drinkers in non-cancer controls. On the other hand, CRC cases for women took in more beef or pork per week than non-cancer controls. In this study, 524 men (21.7%) and 1,297 (28.4%) women had at least three components of MetS at the entry of a cohort.
Table 1. Baseline Characteristics of Study Participants
Characteristics
|
Men
|
|
Women
|
Colorectal cancer
|
P-value
|
Colorectal cancer
|
P-value
|
No (N=2,360)
|
Yes (N=57)
|
No (N=4,514)
|
Yes (N=54)
|
Follow-up (years, Median (IQR))
|
10.43 (8.5-12.48)
|
4.76 (2.91-7.79)
|
<0.001
|
|
10.44 (9.48-12.88)
|
5.55 (3.06-7.53)
|
<0.001
|
Age (years, Mean±SD)
|
59.66±10.93
|
64.18±8.67
|
0.002
|
|
59.95±11.2
|
64.72±9.19
|
0.001
|
Physical activity (days/week, Mean±SD)
|
3.97±2.86
|
4.26±2.7
|
0.754
|
|
2.99±2.96
|
3.2±3.02
|
0.60
|
Intake of fruits or vegetables (days/week, Mean±SD)
|
5.43±1.44
|
5.19±1.57
|
0.28
|
|
5.51±1.46
|
5.04±1.8
|
0.073
|
Intake of beef or pork (days/week, Mean±SD)
|
2.55±1.4
|
2.44±1.49
|
0.732
|
|
1.75±1.33
|
1.31±1.33
|
0.012
|
Alcohol consumption [N(%)]
|
|
|
|
|
|
|
|
Non-drinkers
|
640 (27.12)
|
7 (12.28)
|
0.03
|
|
3,598 (79.71)
|
45 (83.33)
|
0.796
|
Moderate drinkers (<24 g/day)
|
797 (33.77)
|
26 (45.61)
|
|
|
797 (17.66)
|
8 (14.81)
|
|
Heavy drinkers (≥24 g/day)
|
923 (39.11)
|
24 (42.11)
|
|
|
119 (2.64)
|
1 (1.85)
|
|
Smoking status [N(%)]
|
|
|
|
|
|
|
|
Non-smokers
|
483 (20.47)
|
12 (21.05)
|
0.952
|
|
4,186 (92.73)
|
52 (96.3)
|
0.894
|
Moderate smokers (<20 pack-year)
|
669 (28.35)
|
17 (29.82)
|
|
|
263 (5.83)
|
2 (3.7)
|
|
Heavy smokers (≥20 pack-year)
|
1,208 (51.19)
|
28 (49.12)
|
|
|
65 (1.44)
|
0 (0)
|
|
Education level [N(%)]
|
|
|
|
|
|
|
|
Illiterate
|
222 (9.41)
|
6 (10.53)
|
0.448
|
|
1,445 (32.01)
|
24 (44.44)
|
0.24
|
Middle school or less
|
1,443 (61.14)
|
39 (68.42)
|
|
|
2,505 (55.49)
|
26 (48.15)
|
|
High school
|
484 (20.51)
|
10 (17.54)
|
|
|
438 (9.7)
|
3 (5.56)
|
|
College or more
|
211 (8.94)
|
2 (3.51)
|
|
|
126 (2.79)
|
1 (1.85)
|
|
Residential area [N(%)]
|
|
|
|
|
|
|
|
Sancheong-gun
|
1,270 (53.81)
|
23 (40.35)
|
0.209
|
|
2,380 (52.72)
|
21 (38.89)
|
0.072
|
Changwon-si
|
485 (20.55)
|
16 (28.07)
|
|
|
867 (19.21)
|
13 (24.07)
|
|
Chooncheon-si
|
167 (7.08)
|
7 (12.28)
|
|
|
438 (9.7)
|
3 (5.56)
|
|
Choongjoo-si
|
281 (11.91)
|
8 (14.04)
|
|
|
558 (12.36)
|
11 (20.37)
|
|
Haman-gun
|
157 (6.65)
|
3 (5.26)
|
|
|
271 (6)
|
6 (11.11)
|
|
Metabolic syndrome [N(%)]
|
|
|
|
|
|
|
|
No (No. of components of MetS <3)
|
1,848 (78.31)
|
45 (78.95)
|
0.907
|
|
3,243 (71.84)
|
28 (51.85)
|
0.001
|
Yes (No. of components of MetS ≥3)
|
512 (21.69)
|
12 (21.05)
|
|
|
1,271 (28.16)
|
26 (48.15)
|
|
Blood pressure [N(%)]
|
|
|
|
|
|
|
|
SBP <130 mmHg and DBP <85 mmHg
|
996 (42.2)
|
21 (36.84)
|
0.418
|
|
2,017 (44.68)
|
19 (35.19)
|
0.163
|
SBP ≥130 mmHg or DBP ≥85 mmHg
|
1,364 (57.8)
|
36 (63.16)
|
|
|
2,497 (55.32)
|
35 (64.81)
|
|
BMI [N(%)]
|
|
|
|
|
|
|
|
<25 kg/m2
|
1,667 (70.64)
|
40 (70.18)
|
0.94
|
|
2,840 (62.92)
|
33 (61.11)
|
0.785
|
≥25 kg/m2
|
693 (29.36)
|
17 (29.82)
|
|
|
1,674 (37.08)
|
21 (38.89)
|
|
HDL cholesterol [N(%)]
|
|
|
|
|
|
|
|
≥40 mg/dL for men, ≥50 mg/dL for women
|
1,913 (81.06)
|
54 (94.74)
|
0.009
|
|
2,309 (51.15)
|
26 (48.15)
|
0.661
|
<40 mg/dL for men, <50 mg/dL for women
|
447 (18.94)
|
3 (5.26)
|
|
|
2,205 (48.85)
|
28 (51.85)
|
|
Triglyceride level [N(%)]
|
|
|
|
|
|
|
|
<150 mg/dL
|
1,552 (65.76)
|
40 (70.18)
|
0.488
|
|
3,148 (69.74)
|
27 (50)
|
0.002
|
≥150 mg/dL
|
808 (34.24)
|
17 (29.82)
|
|
|
1,366 (30.26)
|
27 (50)
|
|
FBS [N(%)]
|
|
|
|
|
|
|
|
<110 mg/dL
|
2,047 (86.74)
|
46 (80.7)
|
0.186
|
|
4,042 (89.54)
|
49 (90.74)
|
0.775
|
≥110 mg/dL
|
313 (13.26)
|
11 (19.3)
|
|
|
472 (10.46)
|
5 (9.26)
|
|
IQR, interquartile range; SD, standard deviation; SBP, systolic blood pressure; DBP, diastolic blood pressure; BMI, body mass index; HDL, high-density lipoproteins; FBS, fasting blood sugar.
Tables 2, 3 and Figure S1~S4 show participants’ characteristics according to the presence of MetS and abnormal TG levels as well as the degree of imbalance among covariates. A statistically negligible difference in covariates was found in the following datasets: the matched, stratified, and IPTW datasets for MetS and abnormal TG levels in men; the matched, and IPTW datasets for MetS in women; the matched, stratified and IPTW datasets for increased TG level in women.
Table 2. Comparison of baseline characteristics according to metabolic syndrome and abnormal triglyceride (TG) levels and differences in baseline characteristics with different propensity score-based methods in men
Characteristics
|
Metabolic syndrome (N)
|
Standardized mean differencesa) (N)
|
Abnormal TG levels (N)
|
Standardized mean differences a) (N)
|
|
No
(1,893)
|
Yes
(524)
|
Crude
(2,417)
|
Matched
(1,038)
|
Strata5
(2,402)
|
IPTW
(2,417)
|
SMRW
(2,417)
|
No
(1,592)
|
Yes
(825)
|
Crude
(2,417)
|
Matched
(1,038)
|
Strata5
(2,402)
|
IPTW
(2,417)
|
SMRW
(2,417)
|
Age [years, Mean (SD)]
|
60.31 (11.03)
|
57.80 (10.17)
|
0.237
|
0.032
|
<0.001
|
0.029
|
0.385
|
60.58 (11.18)
|
58.19 (10.15)
|
0.224
|
0.006
|
0.021
|
0.007
|
0.308
|
Physical activity [days/week, Mean (SD)]
|
4.09 (2.84)
|
3.59 (2.88)
|
0.175
|
0.019
|
0.019
|
0.019
|
0.267
|
4.01 (2.85)
|
3.92 (2.88)
|
0.243
|
0.037
|
0.020
|
0.018
|
0.331
|
Intake of fruits or vegetables [days/week, Mean (SD)]
|
5.47 (1.42)
|
5.24 (1.48)
|
0.164
|
0.041
|
0.011
|
0.002
|
0.253
|
5.46 (1.42)
|
5.36 (1.48)
|
0.123
|
0.024
|
0.029
|
0.013
|
0.164
|
Intake of beef or pork [days/week, Mean (SD)]
|
2.52 (1.42)
|
2.62 (1.31)
|
0.073
|
0.052
|
0.003
|
0.005
|
0.127
|
2.51 (1.40)
|
2.62 (1.38)
|
0.030
|
0.013
|
0.020
|
0.010
|
0.034
|
Alcohol consumption [N(%)]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Non-drinkers
|
525 (27.7)
|
122 (23.3)
|
0.105
|
0.029
|
0.007
|
0.031
|
0.182
|
469 (29.46)
|
178 (21.58)
|
0.071
|
0.013
|
0.008
|
0.004
|
0.091
|
Moderate drinkers (<24 g/day)
|
640 (33.8)
|
183 (34.9)
|
|
|
|
|
|
560 (35.18)
|
263 (31.88)
|
|
|
|
|
|
Heavy drinkers (≥24 g/day)
|
728 (38.5)
|
219 (41.8)
|
|
|
|
|
|
563 (35.36)
|
384 (46.55)
|
|
|
|
|
|
Smoking status [N(%)]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Non-smokers
|
381 (20.1)
|
114 (21.8)
|
0.092
|
0.042
|
0.008
|
0.031
|
0.161
|
338 (21.23)
|
157 (19.03)
|
0.078
|
0.038
|
0.004
|
0.012
|
0.113
|
Moderate smokers (<20 pack-year)
|
554 (29.3)
|
132 (25.2)
|
|
|
|
|
|
473 (29.71)
|
213 (25.82)
|
|
|
|
|
|
Heavy smokers (≥20 pack-year)
|
958 (50.6)
|
278 (53.1)
|
|
|
|
|
|
781 (49.06)
|
455 (55.15)
|
|
|
|
|
|
Education level [N(%)]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Illiterate
|
194 (10.2)
|
34 ( 6.5)
|
0.211
|
0.049
|
0.016
|
0.029
|
0.337
|
165 (10.36)
|
63 (7.64)
|
0.118
|
0.017
|
0.011
|
0.014
|
0.168
|
Middle school or less
|
1180 (62.3)
|
302 (57.6)
|
|
|
|
|
|
978 (61.43)
|
504 (61.09)
|
|
|
|
|
|
High school
|
359 (19.0)
|
135 (25.8)
|
|
|
|
|
|
307 (19.28)
|
187 (22.67)
|
|
|
|
|
|
College or more
|
160 ( 8.5)
|
53 (10.1)
|
|
|
|
|
|
142 (8.92)
|
71 (8.61)
|
|
|
|
|
|
Residential area [N(%)]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Sancheong-gun
|
1046 (55.3)
|
247 (47.1)
|
0.312
|
0.027
|
0.068
|
0.011
|
0.448
|
867 (54.46)
|
426 (51.64)
|
0.141
|
0.030
|
0.005
|
0.012
|
0.185
|
Changwon-si
|
410 (21.7)
|
91 (17.4)
|
|
|
|
|
|
340 (21.36)
|
161 (19.52)
|
|
|
|
|
|
Chooncheon-si
|
130 ( 6.9)
|
44 ( 8.4)
|
|
|
|
|
|
109 (6.85)
|
65 (7.88)
|
|
|
|
|
|
Choongjoo-si
|
185 ( 9.8)
|
104 (19.8)
|
|
|
|
|
|
167 (10.49)
|
122 (14.79)
|
|
|
|
|
|
Haman-gun
|
122 ( 6.4)
|
38 ( 7.3)
|
|
|
|
|
|
109 (6.85)
|
51 (6.18)
|
|
|
|
|
|
SD, standard deviation; Crude, whole dataset; Matched, 1:1 matched dataset; Strata5, dataset stratified with 5 strata; IPTW, inverse probability-of-treatment weighted dataset; SMRW, standardized mortality ratio weighted dataset. a)The values highlighted in bold are significant imbalances according to metabolic syndrome (standardized mean differences > 0.1).
Table 3. Comparison of baseline characteristics according to metabolic syndrome and abnormal triglyceride (TG) levels and differences in baseline characteristics with different propensity score-based methods in women
Characteristics
|
Metabolic syndrome (N)
|
Standardized mean differencesa) (N)
|
Abnormal TG levels (N)
|
Standardized mean differences a) (N)
|
|
No
(3,271)
|
Yes
(1,397)
|
Crude
(4,568)
|
Matched
(2,580)
|
Strata5
(4,558)
|
IPTW
(4,568)
|
SMRW
(4,568)
|
No
(3,175)
|
Yes
(1,393)
|
Crude
(4,568)
|
Matched
(2,786)
|
Strata5
(4,554)
|
IPTW
(4,568)
|
SMRW
(4,568)
|
Age [years, Mean (SD)]
|
59.38 (11.72)
|
61.59 (9.54)
|
0.206
|
0.016
|
0.039
|
0.036
|
0.428
|
59.09 (11.74)
|
62.10 (9.47)
|
0.282
|
0.008
|
0.021
|
0.021
|
0.472
|
Physical activity [days/week, Mean (SD)]
|
3.13 (2.96)
|
2.65 (2.92)
|
0.164
|
0.02
|
0.001
|
0.006
|
0.199
|
3.08 (2.96)
|
2.78 (2.94)
|
0.102
|
0.042
|
0.017
|
0.003
|
0.121
|
Intake of fruits or vegetables [days/week, Mean (SD)]
|
5.55 (1.43)
|
5.37 (1.53)
|
0.127
|
0.004
|
0.002
|
0.004
|
0.173
|
5.51 (1.45)
|
5.48 (1.50)
|
0.021
|
0.011
|
0.011
|
0.008
|
0.028
|
Intake of beef or pork [days/week, Mean (SD)]
|
1.77 (1.34)
|
1.67 (1.31)
|
0.071
|
0.029
|
0.029
|
0.007
|
0.166
|
1.81 (1.34)
|
1.60 (1.31)
|
0.158
|
0.021
|
0.006
|
0.005
|
0.263
|
Alcohol consumption [N(%)]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Non-drinkers
|
2560 (78.3)
|
1083 (83.5)
|
0.137
|
0.027
|
0.011
|
0.007
|
0.255
|
2473 (77.9)
|
1170 (84.0)
|
0.156
|
0.017
|
0.008
|
0.009
|
0.261
|
Moderate drinkers (<24 g/day)
|
615 (18.8)
|
190 (14.6)
|
|
|
|
|
|
612 (19.3)
|
193 (13.9)
|
|
|
|
|
|
Heavy drinkers (≥24 g/day)
|
96 ( 2.9)
|
24 ( 1.9)
|
|
|
|
|
|
90 ( 2.8)
|
30 ( 2.2)
|
|
|
|
|
|
Smoking status [N(%)]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Non-smokers
|
3028 (92.6)
|
1210 (93.3)
|
0.034
|
0.016
|
0.011
|
0.008
|
0.056
|
2958 (93.2)
|
1280 (91.9)
|
0.059
|
0.037
|
0.016
|
0.004
|
0.084
|
Moderate smokers (<20 pack-year)
|
197 ( 6.0)
|
68 ( 5.2)
|
|
|
|
|
|
178 ( 5.6)
|
87 ( 6.2)
|
|
|
|
|
|
Heavy smokers (≥20 pack-year)
|
46 ( 1.4)
|
19 ( 1.5)
|
|
|
|
|
|
39 ( 1.2)
|
26 ( 1.9)
|
|
|
|
|
|
Education level [N(%)]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Illiterate
|
1040 (31.8)
|
429 (33.1)
|
0.263
|
0.056
|
0.111
|
0.023
|
0.534
|
993 (31.3)
|
476 (34.2)
|
0.246
|
0.034
|
0.052
|
0.012
|
0.441
|
Middle school or less
|
1757 (53.7)
|
774 (59.7)
|
|
|
|
|
|
1716 (54.0)
|
815 (58.5)
|
|
|
|
|
|
High school
|
356 (10.9)
|
85 ( 6.6)
|
|
|
|
|
|
354 (11.1)
|
87 ( 6.2)
|
|
|
|
|
|
College or more
|
118 ( 3.6)
|
9 ( 0.7)
|
|
|
|
|
|
112 ( 3.5)
|
15 ( 1.1)
|
|
|
|
|
|
Residential area [N(%)]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Sancheong-gun
|
1830 (55.9)
|
571 (44.0)
|
0.348
|
0.018
|
0.078
|
0.01
|
0.459
|
1724 (54.3)
|
677 (48.6)
|
0.159
|
0.031
|
0.057
|
0.007
|
0.214
|
Changwon-si
|
653 (20.0)
|
227 (17.5)
|
|
|
|
|
|
619 (19.5)
|
261 (18.7)
|
|
|
|
|
|
Chooncheon-si
|
304 ( 9.3)
|
137 (10.6)
|
|
|
|
|
|
291 ( 9.2)
|
150 (10.8)
|
|
|
|
|
|
Choongjoo-si
|
311 ( 9.5)
|
258 (19.9)
|
|
|
|
|
|
351 (11.1)
|
218 (15.6)
|
|
|
|
|
|
Haman-gun
|
173 ( 5.3)
|
104 ( 8.0)
|
|
|
|
|
|
190 ( 6.0)
|
87 ( 6.2)
|
|
|
|
|
|
SD, standard deviation; Crude, whole dataset; Matched, 1:1 matched dataset; Strata5, dataset stratified with 5 strata; IPTW, inverse probability-of-treatment weighted dataset; SMRW, standardized mortality ratio weighted dataset. a)The values highlighted in bold are significant imbalances according to metabolic syndrome (standardized mean differences > 0.1).
Table 4 showed the associations of MetS and abnormal TG levels on CRC risk according to various analytical methods. In women, unadjusted and adjusted HRs between abnormal MetS and CRC risk were 2.33 (95% CI: 1.37, 3.97) and 2.12 (95% CI: 1.22, 3.68), respectively. The HR from the PS 1:1 matching analysis was 2.19 (95% CI: 1.10, 4.33) and that from the IPTW analysis was 2.03 (95% CI: 1.40, 2.95). Besides, unadjusted and adjusted HRs between abnormal TG levels and CRC were 2.27 (95% CI: 1.33, 3.87) and 2.06 (95% CI: 1.20, 3.55), respectively. The HRs from the PS 1:1 matched, stratified, and IPTW datasets were 2.08 (95% CI: 1.07, 4.02), 2.26 (95% CI: 1.32, 3.84), and 1.98 (95% CI: 1.36, 2.89), respectively. On the other hand, in men, all associations between MetS and abnormal TG levels and CRC risk were not significant.
Additionally, associations between 4 metabolic components except abnormal TG levels (i.e., hypertension, obesity, abnormal high-density lipoprotein cholesterol levels, and abnormal fasting blood sugar levels) and CRC risk are summarized in Table S1. In women, All HRs between 4 metabolic components (hypertension, obesity, abnormal high-density lipoprotein cholesterol levels, and abnormal fasting blood sugar levels) and CRC risk according to were not consistently significant by various analyses. On the other hand, in men, only abnormal high-density lipoprotein cholesterol levels had a significant inverse association with CRC risk.
Besides, when associations between metabolic syndrome and the incidence of (a) colon cancer and (b) rectum cancer were evaluated, in women, there was an association between metabolic syndrome and rectum cancer (Table S2). The HRs from the PS 1:1 matched and IPTW datasets were 3.67 (95% CI: 1.03, 13.17) and 3.29 (95% CI: 1.29, 8.36), respectively, which were the only HRs estimated from datasets with balanced covariates (the degree of imbalance among covariates was not shown).
Table 4. Association between (a) Metabolic Syndrome, (b) Triglyceride Level and Colorectal Cancer risk
|
Total
|
Men
|
Women
|
Methods
|
Cases
(N)
|
Controls
(N)
|
HR
(95% CI)
|
P-value
|
Cases (N)
|
Controls (N)
|
HR
(95% CI)
|
P-value
|
Cases
(N)
|
Controls (N)
|
HR
(95% CI)
|
P-value
|
(a) Metabolic syndrome
|
|
|
|
|
|
|
|
|
|
|
|
|
General Cox's hazard regression
|
|
|
|
|
|
|
|
|
|
|
|
|
Unadjusted
|
111
|
6,874
|
1.46 (0.99, 2.16)
|
0.060
|
57
|
2,360
|
0.94 (0.50, 1.78)
|
0.856
|
54
|
4,514
|
2.33 (1.37, 3.97)
|
0.002
|
Multivariablea)
|
111
|
6,874
|
1.55 (1.04, 2.33)
|
0.033
|
57
|
2,360
|
1.04 (0.54, 1.99)
|
0.908
|
54
|
4,514
|
2.12 (1.22, 3.68)
|
0.008
|
PS-based Cox's hazard regression
|
|
|
|
|
|
|
|
|
|
|
|
|
Matched for PS
|
67
|
3,573
|
1.32 (0.81, 2.13)
|
0.266
|
25
|
1,013
|
0.93 (0.42, 2.03)
|
0.847
|
38
|
2,542
|
2.19 (1.10, 4.33)
|
0.025
|
Stratification into 5 strata by PS
|
111
|
6,862
|
1.48 (1.00, 2.19)
|
0.050
|
57
|
2,345
|
1.06 (0.56, 2.02)
|
0.859
|
53
|
4,505
|
2.23 (1.32, 3.77)
|
0.003
|
Regression adjusted with PS
|
|
|
|
|
|
|
|
|
|
|
|
|
as a continuous term
|
111
|
6,874
|
1.45 (0.97, 2.16)
|
0.071
|
57
|
2,360
|
1.05 (0.55, 2.01)
|
0.878
|
54
|
4,514
|
2.03 (1.17, 3.53)
|
0.012
|
as a quintile term
|
111
|
6,874
|
1.48 (0.99, 2.22)
|
0.054
|
57
|
2,360
|
1.02 (0.54, 1.95)
|
0.947
|
54
|
4,514
|
2.07 (1.20, 3.58)
|
0.009
|
Weighted models
|
|
|
|
|
|
|
|
|
|
|
|
|
IPTW model
|
111
|
6,874
|
1.43 (1.11, 1.85)
|
0.007
|
57
|
2,360
|
1.06 (0.73, 1.52)
|
0.772
|
54
|
4,514
|
2.03 (1.40, 2.95)
|
<0.001
|
SMRW model
|
111
|
6,874
|
1.44 (1.04, 2.01)
|
0.031
|
57
|
2,360
|
0.92 (0.52, 1.64)
|
0.780
|
54
|
4,514
|
2.48 (1.63, 3.75)
|
<0.001
|
(b) Triglyceride level
|
|
|
|
|
|
|
|
|
|
|
|
|
General Cox's hazard regression
|
|
|
|
|
|
|
|
|
|
|
|
|
Unadjusted
|
111
|
6874
|
1.39 (0.95, 2.03)
|
0.090
|
57
|
2,360
|
0.79 (0.45, 1.39)
|
0.416
|
54
|
4,514
|
2.27 (1.33, 3.87)
|
0.003
|
Multivariablea)
|
111
|
6874
|
1.33 (0.91, 1.95)
|
0.145
|
57
|
2,360
|
0.84 (0.47, 1.5)
|
0.557
|
54
|
4,514
|
2.06 (1.2, 3.55)
|
0.009
|
PS-based Cox's hazard regression
|
|
|
|
|
|
|
|
|
|
|
|
|
Matched for PS
|
76
|
4360
|
1.36 (0.86, 2.14)
|
0.191
|
32
|
1,606
|
1.11 (0.55, 2.21)
|
0.777
|
40
|
2,746
|
2.08 (1.07, 4.02)
|
0.031
|
Stratification into 5 strata by PS
|
111
|
6866
|
1.3 (0.89, 1.9)
|
0.168
|
57
|
2,354
|
0.82 (0.47, 1.45)
|
0.498
|
53
|
4,501
|
2.26 (1.32, 3.84)
|
0.003
|
Regression adjusted with PS
|
|
|
|
|
|
|
|
|
|
|
|
|
as a continuous term
|
111
|
6874
|
1.28 (0.87, 1.88)
|
0.210
|
57
|
2,360
|
0.85 (0.48, 1.51)
|
0.570
|
54
|
4,514
|
2.02 (1.18, 3.48)
|
0.011
|
as a quintile term
|
111
|
6874
|
1.28 (0.87, 1.88)
|
0.202
|
57
|
2,360
|
0.82 (0.46, 1.46)
|
0.504
|
54
|
4,514
|
2.03 (1.18, 3.50
|
0.010
|
Weighted models
|
|
|
|
|
|
|
|
|
|
|
|
|
IPTW model
|
111
|
6874
|
1.28 (0.99, 1.66)
|
0.063
|
57
|
2,360
|
0.85 (0.58, 1.23)
|
0.380
|
54
|
4,514
|
1.98 (1.36, 2.89)
|
<0.001
|
SMRW model
|
111
|
6874
|
1.44 (1.05, 1.97)
|
0.025
|
57
|
2,360
|
0.79 (0.48, 1.30
|
0.350
|
54
|
4,514
|
2.42 (1.6, 3.66)
|
<0.001
|
HR, hazard ratio; CI, confidence interval; PS, propensity score; IPTW, inverse probability-of-treatment weighted; SMRW, standardized mortality ratio weighted. a)Adjusted by age, sex (in case of total) education, smoking status, alcohol consumption, physical activity, frequency of intake of fruits or vegetables, frequency of intake of red meats, and residential area.