3.1 Baseline characteristics
We compared the characteristics of individuals with constipation, diarrhea, and normal bowel movements (Table 1). Among these participants, 594 reported constipation, 7,622 had normal bowel movements, and 613 had diarrhea. The prevalence of constipation was higher than diarrhea among those with lower education. Women, individuals with low BMI, diabetes, and hypertension had a lower prevalence of constipation, while individuals with high BMI and diarrhea had a higher prevalence of diabetes and hypertension. In the drinking module, we observed that individuals with constipation had a significantly lower drinking frequency compared to those with diarrhea and normal subjects (P < 0.001). Individuals who consumed more than 5 drinks per day were significantly more likely to experience diarrhea (P < 0.001). Among individuals who consumed more than 12 drinks per year, the normal group (91.5%) had a significantly higher likelihood of experiencing diarrhea compared to the constipated group (85.2%, P <0.001). There was no significant difference between the three groups in terms of the amount of alcohol consumed per drink.
Table 1: Characteristics of participants in constipation group, diarrhea group and normal group
Variable
|
Constipation, N = 594
|
Healthy bowel, N = 7622
|
Diarrhea, N = 613
|
P-valuea
|
P-valueb
|
Sex
|
|
|
|
<0.001
|
<0.001
|
female
|
66.0%
|
44.6%
|
52.0%
|
|
|
male
|
34.0%
|
55.4%
|
48.0%
|
|
|
age
|
43.559 (17.237)
|
46.363 (17.122)
|
49.853 (15.909)
|
<0.001
|
<0.001
|
race
|
|
|
|
0.004
|
<0.001
|
Mexican American
|
18.0%
|
16.4%
|
22.2%
|
|
|
Other Hispanic
|
10.1%
|
7.4%
|
8.5%
|
|
|
Non-Hispanic White
|
46.6%
|
54.2%
|
45.2%
|
|
|
Non-Hispanic Black
|
21.5%
|
18.2%
|
20.1%
|
|
|
Other Race - Including Multi-Racial
|
3.7%
|
3.9%
|
4.1%
|
|
|
Education level
|
|
|
|
<0.001
|
<0.001
|
Less Than 9th Grade
|
10.4%
|
6.8%
|
16.8%
|
|
|
9-11th Grade (Includes 12th grade with no diploma)
|
16.5%
|
14.1%
|
18.4%
|
|
|
High School Grad/GED or Equivalent
|
26.1%
|
23.1%
|
21.9%
|
|
|
Some College or AA degree
|
27.4%
|
30.9%
|
25.0%
|
|
|
College Graduate or above
|
19.5%
|
25.1%
|
17.9%
|
|
|
BMI(mean ± SD)
|
27.952 (6.425)
|
28.632 (6.452)
|
30.365 (8.516)
|
0.013
|
<0.001
|
Family PIR(mean ± SD)
|
2.996±1.628
|
2.847 (1.646)
|
2.472 (1.613)
|
<0.001
|
<0.001
|
diabetes
|
|
|
|
0.044
|
<0.001
|
Yes
|
6.4%
|
7.6%
|
11.7%
|
|
|
No
|
93.1%
|
90.7%
|
85.2%
|
|
|
Borderline
|
0.5%
|
1.7%
|
3.1%
|
|
|
hypertension
|
|
|
|
0.008
|
<0.001
|
Yes
|
24.4%
|
29.5%
|
37.8%
|
|
|
No
|
75.6%
|
70.5%
|
62.2%
|
|
|
ALQ101
|
|
|
|
<0.001
|
0.295
|
No
|
15.3%
|
10.1%
|
11.4%
|
|
|
Yes
|
84.7%
|
89.9%
|
88.6%
|
|
|
ALQ150
|
|
|
|
0.226
|
<0.001
|
Yes
|
13.6%
|
15.5%
|
22.2%
|
|
|
No
|
86.4%
|
84.5%
|
77.8%
|
|
|
ALQ130(drinks, mean ± SD)
|
2.680 (2.527)
|
2.894 (2.995)
|
3.042 (2.963)
|
0.091
|
0.236
|
ALQ120(day/week, mean ± SD)
|
1.128 (1.691)
|
1.595 (2.034)
|
1.597 (2.175)
|
<0.001
|
0.978
|
PIR, poverty income ratio; BMI, body mass index; ALQ101, Whether more than 12 drinks were consumed in a year;ALQ120, Frequency of drinking (daily/weekly); ALQ130, The average number of drinks consumed per occasion in the past 12 months; ALQ150, Whether there was a period of drinking five or more drinks daily; P-valuea represents the difference between constipation group and normal group; P-valueb represents the difference between diarrhea group and normal group.
3.2 Univariate logistic regression analysis of constipation and diarrhea
Based on univariate logistic regression analysis (Tables 2 and 3), we found that gender, age, race, education, BMI, Family PIR, hypertension, blood glucose, and alcohol consumption were significantly associated with diarrhea and constipation (P < 0.05). Upon gender categorization, we observed that the female population with diarrhea was significantly associated with the male population with diarrhea in several common factors, including age, Non-Hispanic White race, education level, BMI, Family PIR, hypertension etc(P < 0.05). However, the factors significantly associated in the female population with constipation with the male population with constipation were not consistently observed. They were co-significantly correlated only in age and Family PIR (P < 0.05).
Table 2: Univariate logistic regression analysis of constipation population
Variable
|
Constipation-healthy bowel
|
|
Female
|
Male
|
Total
|
|
OR(95%CI)
|
P-value
|
OR(95%CI)
|
P-value
|
OR(95%CI)
|
P-value
|
Age
|
0.992 (0.985 -0.998)
|
0.011
|
0.991 (0.983 -1.000)
|
0.042
|
0.990 (0.985 - 0.995)
|
<.001
|
Race: ref. = Mexican American
|
|
|
|
|
|
|
Other Hispanic
|
1.74 (1.13 - 2.679)
|
0.012
|
0.686(0.388 - 1.213)
|
0.195
|
1.248 (0.896 - 1.739)
|
0.189
|
Non-Hispanic White
|
1.074 (0.776 - 1.486)
|
0.668
|
0.456(0.321 - 0.648)
|
<.001
|
0.783 (0.621 - 0.987)
|
0.039
|
Non-Hispanic Black
|
1.454 (1.005 - 2.104)
|
0.047
|
0.714 (0.476 - 1.072)
|
0.104
|
1.076 (0.823 - 1.407)
|
0.590
|
Other Race - Including Multi-Racial
|
1.31 (0.722 - 2.378)
|
0.374
|
0.459 (0.195 - 1.083)
|
0.075
|
0.873 (0.543 - 1.406)
|
0.578
|
Education level: ref. = 9-11th Grade (Includes 12th grade with no diploma)
|
|
|
|
|
|
|
Less Than 9th Grade
|
1.132 (0.688 - 1.86)
|
0.626
|
1.855 (1.144 - 3.008)
|
0.012
|
1.303 (0.932 - 1.821)
|
0.121
|
High School Grad/GED or Equivalent
|
0.838 (0.595 - 1.18)
|
0.311
|
1.197 (0.782 - 1.832)
|
0.408
|
0.966 (0.742 - 1.258)
|
0.877
|
Some College or AA degree
|
0.703 (0.509 - 0.972)
|
0.033
|
0.71 (0.451 - 1.118)
|
0.139
|
0.758 (0.584 - 0.983)
|
0.037
|
College Graduate or above
|
0.768 (0.549 -1.074)
|
0.123
|
0.339 (0.192 -0.599)
|
<.001
|
0.663 (0.502 - 0.877)
|
0.004
|
BMI
|
0.983 (0.968 - 0.999)
|
0.034
|
0.984 (0.959 - 1.009)
|
0.207
|
0.983 (0.969 - 0.996)
|
0.013
|
Family PIR
|
0.924(0.867 - 0.985)
|
0.015
|
0.813(0.743 -0.89)
|
<.001
|
0.885 (0.840 - 0.932)
|
<.001
|
Hypertension:Yes vs. No
|
0.855(0.672 - 1.088)
|
0.203
|
0.683(0.49 - 0.953)
|
0.025
|
0.771 (0.635 - 0.935)
|
0.008
|
Diabetes: ref. = Borderline
|
|
|
|
|
|
|
Yes
|
3.059 (0.7 -13.361)
|
0.137
|
2.935 (0.38 -22.653)
|
0.302
|
2.859 (0.869 - 9.404)
|
0.084
|
No
|
3.025(0.734 - 12.47)
|
0.126
|
3.869(0.535 -27.965)
|
0.180
|
3.465(1.099 - 10.919)
|
0.034
|
ALQ150:Yes vs. No
|
1.054(0.711- 1.563)
|
0.794
|
1.19(0.859- 1.648)
|
0.295
|
0.861 (0.676 - 1.097)
|
0.227
|
ALQ101:Yes vs. No
|
0.786 (0.602 - 1.027)
|
0.077
|
0.712 (0.414 - 1.225)
|
0.220
|
0.620 (0.490 - 0.785)
|
<.001
|
ALQ120
|
0.891(0.827 -0.96)
|
0.003
|
0.933(0.869 -1.001)
|
0.052
|
0.869 (0.826 - 0.915)
|
<.001
|
ALQ130
|
0.997 (0.939 - 1.059)
|
0.929
|
1.014 (0.98 - 1.049)
|
0.440
|
0.971 (0.939 - 1.004)
|
0.087
|
OR, odds ratio; CI, confidence interval; PIR, poverty income ratio; BMI, body mass index; ALQ101, Whether more than 12 drinks were consumed in a year;ALQ120, Frequency of drinking (daily/weekly); ALQ130, The average number of drinks consumed per occasion in the past 12 months; ALQ150, Whether there was a period of drinking five or more drinks daily.
Table 3: Univariate logistic regression analysis of diarrhea population:
Variable
|
Diarrhea-healthy bowel
|
|
Female
|
Male
|
Total
|
|
OR(95%CI)
|
P-value
|
OR(95%CI)
|
P-value
|
OR(95%CI)
|
P-value
|
Age
|
1.01 (1.003 - 1.017)
|
0.004
|
1.015 (1.008 - 1.022)
|
<.001
|
1.012 (1.007 - 1.017)
|
<.001
|
Race: ref. = Mexican American
|
|
|
|
|
|
|
Other Hispanic
|
0.91 (0.572 - 1.448)
|
0.691
|
0.77 (0.473 - 1.254)
|
0.293
|
0.851 (0.609 - 1.190)
|
0.346
|
Non-Hispanic White
|
0.641 (0.47- 0.875)
|
0.005
|
0.577 (0.428 - 0.779)
|
<.001
|
0.616 (0.497 - 0.764)
|
<.001
|
Non-Hispanic Black
|
0.893 (0.619 -1.288)
|
0.546
|
0.735(0.513 - 1.053)
|
0.093
|
0.814 (0.630 - 1.051)
|
0.114
|
Other Race - Including Multi-Racial
|
0.845(0.451 - 1.584)
|
0.599
|
0.717 (0.381 - 1.351)
|
0.304
|
0.781 (0.500 - 1.219)
|
0.276
|
Education level: ref. = 9-11th Grade (Includes 12th grade with no diploma)
|
|
|
|
|
|
|
Less Than 9th Grade
|
1.729 (1.123 -2.664)
|
0.013
|
2.335 (1.563 - 3.488)
|
<.001
|
1.877 (1.409 - 2.501)
|
<.001
|
High School Grad/GED or Equivalent
|
0.621 (0.435 - 0.886)
|
0.009
|
0.871 (0.59 - 1.288)
|
0.489
|
0.724 (0.558 - 0.941)
|
0.016
|
Some College or AA degree
|
0.511 (0.366 - 0.715)
|
<.001
|
0.73 (0.494 - 1.08)
|
0.115
|
0.617 (0.479 - 0.795)
|
<.001
|
College Graduate or above
|
0.408 (0.28 - 0.595)
|
<.001
|
0.738 (0.494 - 1.104)
|
0.139
|
0.545 (0.415 -0.717)
|
<.001
|
BMI
|
1.039 (1.025 -1.053)
|
<.001
|
1.025 (1.007- 1.042)
|
0.005
|
1.034 (1.023 - 1.045)
|
<.001
|
Family PIR
|
0.844 (0.786- 0.907)
|
<.001
|
0.896 (0.833 - 0.965)
|
0.003
|
0.868 (0.825 - 0.914)
|
<.001
|
Hypertension:Yes vs. No
|
1.547 (1.219 - 1.964)
|
<.001
|
1.387(1.086 - 1.771)
|
0.009
|
1.453 (1.225 - 1.723)
|
<.001
|
Diabetes: ref. = Borderline
|
|
|
|
|
|
|
Yes
|
0.841 (0.389 -1.817)
|
0.660
|
0.909 (0.423 -1.952)
|
0.806
|
0.855 (0.498 - 1.468)
|
0.570
|
No
|
0.456(0.229 - 0.908)
|
0.025
|
0.566(0.28 -1.141)
|
0.111
|
0.516 (0.317 - 0.843)
|
0.008
|
ALQ150:Yes vs. No
|
2.198 (1.571 - 3.074)
|
<.001
|
1.53 (1.18 - 1.982)
|
0.001
|
1.555 (1.273 - 1.900)
|
<.001
|
ALQ101:Yes vs. No
|
0.888(0.656 - 1.202)
|
0.442
|
1.142 (0.657 - 1.986)
|
0.637
|
0.870 (0.671 - 1.129)
|
0.295
|
ALQ120
|
1.012 (0.948 - 1.081)
|
0.715
|
1.02 (0.968 - 1.075)
|
0.463
|
1.001 (0.961 - 1.042)
|
0.977
|
ALQ130
|
1.061 (1.006 - 1.119)
|
0.028
|
1.017 (0.989 - 1.046)
|
0.241
|
1.015 (0.990 - 1.040)
|
0.236
|
OR, odds ratio; CI, confidence interval; PIR, poverty income ratio; BMI, body mass index; ALQ101, Whether more than 12 drinks were consumed in a year; ALQ120, Frequency of drinking (daily/weekly); ALQ130, The average number of drinks consumed per occasion in the past 12 months; ALQ150, Whether there was a period of drinking five or more drinks daily.
Among individuals with constipation, "Whether more than 12 drinks were consumed in a year" and "Frequency of drinking" showed significant associations (P < 0.05). When stratified by gender, only the female "Frequency of drinking" was significantly associated with constipation (P < 0.05). In male patients with constipation, no significant correlation was observed in the alcohol module. In the diarrhea group, "whether there was a period of drinking five or more drinks daily" showed a significant association with diarrhea. When analyzed by gender, "whether there was a period of drinking five or more drinks daily" was significantly associated with diarrhea in both males and females (P <0.05), with the exception of "the average number of drinks consumed per occasion in the past 12 months" in females, which was significantly associated with diarrhea (P <0.05). Additionally, "the average number of drinks consumed per occasion in the past 12 months" showed a significant association with diarrhea in females (P <0.05).
3.3 Multivariate logistic regression analysis of constipation, diarrhea and alcohol consumption
Among women with diarrhea, Model I indicated an increased risk of diarrhea associated with "whether there was a period of drinking five or more drinks daily" (OR: 2.128, 95% CI: 1.496-3.027, P < 0.001). This association was supported by Model II (OR: 2.126, 95% CI: 1.493 - 3.029, P < 0.001), Model III (OR: 1.888, 95% CI: 1.31 - 2.719, P < 0.001), and Model IV (OR: 1.870, 95% CI: 1.297 - 2.696, P = 0.001). Interestingly, only Model II (OR: 1.071, 95% CI: 1.011 - 1.134, P = 0.020) demonstrated a positive association between the average number of drinks consumed per occasion in the past 12 months and the risk of diarrhea. In the male population with diarrhea, only Model I (OR: 1.501, 95% CI: 1.148 - 1.963, P = 0.003) and Model II (OR: 1.467, 95% CI: 1.122 - 1.916, P = 0.005) indicated an increased risk of diarrhea associated with "whether there was a period of drinking five or more drinks daily" (Table 4).
Table 4: Multivariate logistic regression analysis of diarrhea population
|
Model Ⅰ
|
Model Ⅱ
|
Model Ⅲ
|
Model Ⅳ
|
|
OR(95%CI)
|
P-value
|
OR(95%CI)
|
P-value
|
OR(95%CI)
|
P-value
|
OR(95%CI)
|
P-value
|
Female
|
|
|
|
|
|
|
|
|
ALQ120
|
1(0.934 - 1.071)
|
0.998
|
0.973(0.908 - 1.044)
|
0.448
|
1.021(0.951 -1.096)
|
0.569
|
1.023(0.953 -1.099)
|
0.530
|
ALQ130
|
1.038(0.982 - 1.098)
|
0.186
|
1.071(1.011 - 1.134)
|
0.020
|
1.035(0.974 -1.101)
|
0.264
|
1.035(0.974 - 1.101)
|
0.265
|
ALQ101:
Yes vs. No
|
0.806(0.587- 1.109)
|
0.185
|
0.855(0.62- 1.177)
|
0.336
|
1.008(0.728 - 1.397)
|
0.961
|
1.005(0.725 - 1.393)
|
0.975
|
ALQ150:
Yes vs. No
|
2.128 (1.496- 3.027)
|
<0.001
|
2.126(1.493 - 3.029)
|
<0.001
|
1.888(1.31 - 2.719)
|
<0.001
|
1.870(1.297 - 2.696)
|
0.001
|
Male
|
|
|
|
|
|
|
|
|
ALQ120
|
1.003 (0.95 - 1.059)
|
0.901
|
0.98(0.928 - 1.035)
|
0.463
|
1.009(0.955 - 1.066)
|
0.753
|
1.01(0.956 - 1.068)
|
0.723
|
ALQ130
|
1.01 (0.98 - 1.041)
|
0.519
|
1.027(1 - 1.056)
|
0.052
|
1.007(0.977 - 1.038)
|
0.643
|
1.008(0.978 - 1.038)
|
0.626
|
ALQ101:
Yes vs. No
|
1.045 (0.594 - 1.841)
|
0.878
|
1.099(0.624 - 1.938)
|
0.743
|
1.22(0.685 - 2.171)
|
0.499
|
1.218(0.685 -2.168)
|
0.502
|
ALQ150:
Yes vs. No
|
1.501 (1.148 - 1.963)
|
0.003
|
1.467(1.122 -1.916)
|
0.005
|
1.309(0.993 - 1.726)
|
0.056
|
1.295(0.981 - 1.709)
|
0.068
|
OR, odds ratio; CI, confidence interval. ALQ101, Whether more than 12 drinks were consumed in a year; ALQ120, Frequency of drinking (daily/weekly); ALQ130, The average number of drinks consumed per occasion in the past 12 months; ALQ150, Whether there was a period of drinking five or more drinks daily. Model Ⅰ, without adjusting any covariates; Model II, adjusted for age; According to type Ⅱ and race/nationality, education level, Family PIR and BMI, the type Ⅲ model was adjusted. Type Ⅳ adjusted the covariate of type III with diabetes and hypertension. To test the stability of the association between drinking and defecation types.
In the constipated female population, models I (OR: 0.898, 95% CI: 0.831 - 0.969, P = 0.006), II (OR: 0.911, 95% CI: 0.842 - 0.985, P < 0.020), III (OR: 0.905, 95% CI: 0.834 - 0.981, P = 0.015), and IV (OR: 0.906, 95% CI: 0.835 - 0.982, P = 0.016) all showed that frequency of drinking was negatively associated with the risk of constipation. In contrast, in the constipated male population, all four models showed no significant correlation between alcohol consumption and constipation (Table 5).
Table 5: Multivariate logistic regression analysis of constipation population
|
Model Ⅰ
|
Model Ⅱ
|
Model Ⅲ
|
Model Ⅳ
|
|
OR(95%CI)
|
P-value
|
OR(95%CI)
|
P-value
|
OR(95%CI)
|
P-value
|
OR(95%CI)
|
P-value
|
Female
|
|
|
|
|
|
|
|
|
ALQ120
|
0.898(0.831 - 0.969)
|
0.006
|
0.911(0.842 - 0.985)
|
0.020
|
0.905(0.834 - 0.981)
|
0.015
|
0.906(0.835 - 0.982)
|
0.016
|
ALQ130
|
1.004(0.945 - 1.067)
|
0.892
|
0.982(0.919 - 1.049)
|
0.591
|
0.974(0.91 - 0.993)
|
0.458
|
0.975(0.91 - 1.044)
|
0.465
|
ALQ101:Yes vs. No
|
0.87(0.657 - 1.151)
|
0.329
|
0.845(0.637 - 1.119)
|
0.240
|
0.858(0.644 - 1.142)
|
0.294
|
0.857(0.643 - 1.141)
|
0.291
|
ALQ150:Yes vs. No
|
1.136 (0.757- 1.705)
|
0.537
|
1.157(0.771 - 1.736)
|
0.481
|
1.087(0.72 - 1.642)
|
0.692
|
1.087(0.719 - 1.642)
|
0.693
|
Male
|
|
|
|
|
|
|
|
|
ALQ120
|
0.931(0.866 - 1.001)
|
0.054
|
0.941(0.874 - 1.014)
|
0.110
|
0.961(0.892 - 1.035)
|
0.294
|
0.959(0.89 - 1.033)
|
0.266
|
ALQ130
|
1.012(0.977 - 1.048)
|
0.507
|
1.004(0.965 - 1.043)
|
0.857
|
0.971(0.927 - 1.017)
|
0.217
|
0.97(0.926 - 1.017)
|
0.205
|
ALQ101:Yes vs. No
|
0.76(0.434 - 1.33)
|
0.337
|
0.747(0.426 - 1.309)
|
0.308
|
0.823(0.466 - 1.452)
|
0.501
|
0.82(0.465 - 1.449)
|
0.495
|
ALQ150:Yes vs. No
|
1.262(0.903 - 1.762)
|
0.173
|
1.285(0.919 - 1.798)
|
0.143
|
1.117(0.792 - 1.576)
|
0.528
|
1.151(0.815 - 1.626)
|
0.426
|
OR, odds ratio; CI, confidence interval. ALQ101, Whether more than 12 drinks were consumed in a year; ALQ120, Frequency of drinking (daily/weekly); ALQ130, The average number of drinks consumed per occasion in the past 12 months; ALQ150, Whether there was a period of drinking five or more drinks daily. Model Ⅰ, without adjusting any covariates; Model II, adjusted for age; According to type Ⅱ and race/nationality, education level, Family PIR and BMI, the type Ⅲ model was adjusted. Type Ⅳ adjusted the covariate of type III with diabetes and hypertension. To test the stability of the association between drinking and defecation types.
3.4 Multilevel Linear Regression Analysis of Stool Shape and Alcohol Consumption
Next, we performed a multilevel linear regression with the covariates (age, gender, race/ethnicity, education level, BMI, Famliy PIR, diabetes, and hypertension as the control level, and "Frequency of drinking" and "the average number of drinks consumed per occasion in the past 12 months" as level 1) to analyze the effect of "Frequency of drinking" and "the average number of drinks consumed per occasion in the past 12 months" on stool shape (Bristol) and alcohol consumption (Bristol). Stratified regression was performed to analyze the degree of influence of "Frequency of drinking" and "the average number of drinks consumed per occasion in the past 12 months" on stool shape (Bristol type 1 to type 7). The results showed that based on the hierarchical model: the control level model was significant at P < 0.001, level of significance, the model was valid, and the model's goodness of fit R² was 0.033 The hierarchical 1 model was significant at P < 0.001, level of significance, the model was valid, and the model's goodness of fit R² was 0.035 (Table 6).
Table 6: Multi-level linear regression analysis of stool shape
|
Control level
|
Level 1
|
R²
|
0.028
|
0.03
|
adj.R²
|
0.028
|
0.03
|
F
|
F(5, 8829) =50.989,P=<0.001
|
F(7, 8828) =39.427,P=<0.001
|
△R²
|
0.028
|
0.002
|
△F
|
F(5, 8829) =50.989,P=<0.001
|
F(2, 8828) =10.253,P=<0.001
|
Control level includes age, gender, race/ethnicity, education level, BMI, Famliy PIR, diabetes, and hypertension. Level 1 includes Frequency of drinking and The average number of drinks consumed per occasion in the past 12 months.