Data source and participants
The National Health and Nutrition Examination Survey (NHANES), conducted by the National Center for Health Statistics under the Centers for Disease Control and Prevention (Atlanta, Georgia, USA), is a cross-sectional study focusing on a nationally representative sample of the non-institutionalized US population. Employing a complex, stratified, multistage, probability cluster design, it selects non-institutionalized individuals from the US. All participants gave written informed consent, as mandated by the Ethics Review Committee of the National Center for Health Statistics. Our current analysis zeroes in on participants aged ≥20 years from three NHANES cycles, all of whom completed the intestinal health and dietary assessment questionnaire. From 2005-2010, out of the 31,034 NHANES participants, 13,531 filled out the dedicated questionnaire on intestinal health. After excluding pregnant individuals (n=383), those diagnosed with colorectal cancer (n=95), and those with missing data (n=553), the study considered a final sample of 12,499 participants, split between 6,325 males and 6,174 females. Once weighted, this sample mirrors a whopping 513,372,953 individuals from the US population. The flow chart of the systematic selection process is shown in Fig. 1.
Measures
Adiposity measures.At the Mobile Examination Center, trained health technicians measure weight and height using standardized methods described in a procedural manual. Each measurement session involves a health technician and an accompanying recorder. The BMI is determined using the formula:BMI = weight(kg)/height(m2). Based on BMI values, the World Health Organization categorizes individuals as: underweight (BMI < 18.5), normal weight (BMI 18.5–24.9), overweight (BMI 25–29.9), or obese (BMI ≥ 30).
Constipation.Prior studies have suggested that stool consistency, as indicated by the BSFS, is a more effective constipation measure than stool frequency. This is because the BSFS types exhibit a stronger correlation with colonic transit time. During their interview, participants discussed their bowel functions, and the associated questionnaire gathered data about their stool characteristics over the past 30 days. Participants were presented with a color picture card detailing the seven BSFS types. They were then asked to select the number that best described their most common stool type. For the purposes of this study, constipation was defined by reports of usually having stools that were either BSFS Type 1 (hard, nut-like lumps) or Type 2 (lumpy, sausage-like). Regular stool consistency encompassed BSFS Type 3 (sausage-shaped with surface cracks), Type 4(smooth, soft, snake-like), and Type 5(distinctly edged soft blobs). Chronic diarrhea was characterized by BSFS Type 6(mushy stool with ragged, fluffy pieces) or Type 7(liquid, with no solid pieces).
Covariates.Several variables were evaluated as potential covariates due to their hypothesized or previously demonstrated association with constipation and high BMI. These variables included the intake of energy[8], fiber[9], fat[10], phosphorus[11], magnesium[12] and selenium[13], as well as factors like depression scores[14], sleep duration[15], and exercise levels[16].In our sociodemographic assessment, we considered age (treated as a categorical variable), gender, race/ethnicity (spanning categories from Mexican-American to non-Hispanic black), educational attainment, and PIR (with two distinct ranges).
For behavioral risk factors, we examined smoking and drinking habits, physical activity intensity, and sleep patterns. We categorized individuals into never smokers and regular smokers based on their lifetime cigarette consumption. Those consuming a minimum of 12 alcoholic
Table 1. Baseline Characteristics of the Study Population from National Health and Nutrition Examination Survey 2005–2010 (Using the Stool Consistency Definition of Constipation), weighted drinks annually were labeled as drinkers. Physical activities were bifurcated into intense (e.g., heavy lifting or construction work causing significant breath or heart rate increase) and non-intense types. Sleep duration was segmented into four categories, ranging from very short to extended sleep durations.
Characteristics
|
Total
|
<18kg/m2
|
[18,25)
|
[25–30)
|
≥30kg/m2
|
bP
|
Sex,%(95%CI)a
|
|
|
|
|
|
<0.01
|
Male
|
49.50
(48.68,50.31)
|
27.31
(18.58,38.21)
|
41.95
(40.06,43.87)
|
58.12
(56.34,59.88)
|
48.41
(46.37,50.45)
|
|
Female
|
50.50
(49.69,51.32)
|
72.69
(61.79,81.42)
|
58.05
(56.13,59.94)
|
41.88
(40.12,43.66)
|
51.59
(49.55,53.63)
|
|
Age,%(95%CI)a
|
|
|
|
|
|
<0.01
|
20-29
|
18.73
(17.58,19.94)
|
45.79
(36.49,55.39)
|
26.88
(24.36,29.55)
|
15.84
(14.68,17.08)
|
13.70
(12.37,15.15)
|
|
30-39
|
18.38
(17.30,19.50)
|
10.12
(4.89,19.78)
|
19.06
(17.41,20.83)
|
17.87
(16.16,19.73)
|
18.49
(17.12,19.94)
|
|
40–49
|
21.18
(19.97,22.44)
|
14.55
(8.47,23.85)
|
19.71
(18.14,21.38)
|
21.41
(19.49,23.46)
|
22.40
(20.43,24.51)
|
|
50–59
|
19.05
(17.93,20.23)
|
10.95
(6.31,18.34)
|
15.73
(14.20,17.39)
|
19.88
(18.18,21.71)
|
21.35
(19.59,23.23)
|
|
60–69
|
12.11
(11.24,13.04)
|
5.96
(3.66,9.54)
|
8.66
(7.50,9.97)
|
12.81
(11.43,14.34)
|
14.60
(13.31,16.00)
|
|
70–79
|
7.15
(6.48,7.89)
|
6.87
(3.65,12.56)
|
5.93
(4.95,7.08)
|
7.98
(7.08,8.99)
|
7.44
(6.49,8.50)
|
|
>80
|
3.40
(2.97,3.89)
|
5.77
(3.14,10.35)
|
4.04
(3.34,4.89)
|
4.20
(3.55,4.95)
|
2.02
(1.68,2.42)
|
|
Ethnicity,%(95%CI)a
|
|
|
|
|
|
<0.01
|
Mexican American
|
7.72
(6.06,9.77)
|
3.90
(1.47,9.92)
|
5.65
(4.57,6.97)
|
9.18
(7.26,11.53)
|
8.22
(6.08,11.02)
|
|
Other Hispanic
|
4.09
(3.08,5.42)
|
2.43
(0.84,6.85)
|
3.62
(2.67,4.90)
|
4.52
(3.37,6.04)
|
4.13
(2.99,5.68)
|
|
Non-Hispanic White
|
72.29
(68.57,75.73)
|
76.15
(67.25,83.24)
|
74.84
(71.72,77.72)
|
72.55
(68.47,76.28)
|
69.75
(65.00,74.10)
|
|
Non-Hispanic Black
|
10.69
(8.93,12.74)
|
11.90
(7.60,18.18)
|
8.06
(6.69,9.68)
|
9.37
(7.82,11.19)
|
14.18
(11.56,17.28)
|
|
Other Race - Including Multi-Racial
|
5.21
(4.48,6.06)
|
5.62
(2.28,13.21)
|
7.83
(6.48,9.43)
|
4.38
(3.28,5.82)
|
3.73
(3.08,4.51)
|
|
Education,%(95%CI)a
|
|
|
|
|
|
<0.01
|
Below high school
|
17.36
(15.79,19.04)
|
24.14
(14.85,36.72)
|
15.34
(13.37,17.53)
|
17.82
(15.94,19.86)
|
18.49
(16.73,20.38)
|
|
High school
|
24.29
(22.95,25.69)
|
20.33
(13.84,28.83)
|
22.04
(20.21,23.99)
|
23.90
(21.93,25.99)
|
26.72
(24.66,28.88)
|
|
Above high school
|
58.35
(55.82,60.84)
|
55.54
(43.92,66.58)
|
62.62
(59.06,66.04)
|
58.29
(55.08,61.42)
|
54.79
(52.56,57.00)
|
|
Income-poverty Ratio,%
(95%CI)a
|
|
|
|
|
|
<0.01
|
<2
|
31.66
(29.68,33.71)
|
53.83
(42.85,64.46)
|
30.89
(28.55,33.34)
|
30.19
(27.55,32.96)
|
33.15
(30.43,35.99)
|
|
≥2
|
68.34
(66.29,70.32)
|
46.17
(35.54,57.15)
|
69.11
(66.66,71.45)
|
69.81
(67.04,72.45)
|
66.85
(64.01,69.57)
|
|
Diabetes,%(95%CI)a
|
|
|
|
|
|
<0.01
|
Yes
|
7.78
(7.14,8.47)
|
0.32
(0.04,2.34)
|
3.02
(2.50,3.66)
|
5.75
(5.05,6.55)
|
14.04
(12.62,15.59)
|
|
No
|
90.52
(89.74,91.24)
|
96.91
(88.43,99.23)
|
96.43
(95.71,97.03)
|
92.77
(91.90,93.55)
|
83.07
(81.49,84.54)
|
|
Borderline
|
1.70
(1.47,1.98)
|
2.77
(0.60,11.84)
|
0.55
(0.34,0.88)
|
1.48
(1.05,2.07)
|
2.89
(2.47,3.38)
|
|
Smoking,%(95%CI)a
|
|
|
|
|
|
<0.01
|
Yes
|
47.30
(45.44,49.17)
|
60.15
(47.89,71.26)
|
47.72
(44.80,50.65)
|
48.45
(46.10,50.81)
|
45.50
(43.34,47.68)
|
|
No
|
52.70
(50.83,54.56)
|
39.85
(28.74,52.11)
|
52.28
(49.35,55.20)
|
51.55
(49.19,53.90)
|
54.50
(52.32,56.66)
|
|
Alcohol use,%(95%CI)a
|
|
|
|
|
|
<0.01
|
Yes
|
76.65
(74.79,78.42)
|
73.61
(65.16,80.62)
|
79.04
(76.95,81.00)
|
79.12
(76.97,81.13)
|
72.30
(69.69,74.76)
|
|
No
|
23.35
(21.58,25.21)
|
26.39
(19.38,34.84)
|
20.96
(19.00,23.05)
|
20.88
(18.87,23.03)
|
27.70
(25.24,30.31)
|
|
Hypertension,%(95%CI)a
|
|
|
|
|
|
<0.01
|
Yes
|
30.24
(28.86,31.65)
|
13.75
(9.11,20.22)
|
17.22
(15.71,18.85)
|
28.81
(27.22,30.45)
|
43.34
(41.24,45.46)
|
|
No
|
69.76
(68.35,71.14)
|
86.25
(79.78,90.89)
|
82.78
(81.15,84.29)
|
71.19
(69.55,72.78)
|
56.66
(54.54,58.76)
|
|
Sleep duration,%(95%CI)a
|
|
|
|
|
|
<0.01
|
<5h/night
|
4.87
(4.35,5.46)
|
8.37
(3.96,16.81)
|
4.06
(3.27,5.04)
|
4.15
(3.58,4.82)
|
6.17
(5.28,7.20)
|
|
5–6h/night
|
31.48
(30.18,32.81)
|
22.18
(14.95,31.60)
|
28.04
(25.69,30.52)
|
31.68
(29.98,33.44)
|
34.52
(32.70,36.38)
|
|
Normal
|
57.11
(55.66,58.54)
|
56.62
(45.85,66.80)
|
59.94
(57.24,62.58)
|
58.39
(56.35,60.39)
|
53.45
(51.24,55.64)
|
|
≥9h/night
|
6.54
(6.03,7.09)
|
12.84
(7.08,22.17)
|
7.96
(7.11,8.89)
|
5.78
(4.92,6.77)
|
5.87
(5.05,6.82)
|
|
Depression,%(95%CI)a
|
|
|
|
|
|
<0.01
|
No
|
91.60
(90.73,92.40)
|
89.74
(82.56,94.17)
|
92.71
(91.50,93.76)
|
92.77
(91.60,93.78)
|
89.59
(88.17,90.86)
|
|
yes
|
8.40
(7.60,9.27)
|
10.26
(5.83,17.44)
|
7.29
(6.24,8.50)
|
7.23
(6.22,8.40)
|
10.41
(9.14,11.83)
|
|
Physical activity,%(95%CI)a
|
|
|
|
|
|
<0.01
|
Vigorous physical activity
|
26.76
(25.34,28.24)
|
28.22
(18.81,40.01)
|
28.77
(26.58,31.07)
|
27.99
(25.76,30.35)
|
23.81
(22.26,25.42)
|
|
No vigorous activity
|
73.24
(71.76,74.66)
|
71.78
(59.99,81.19)
|
71.23
(68.93,73.42)
|
72.01
(69.65,74.24)
|
76.19
(74.58,77.74)
|
|
Daily dietary intake (mean ± SD)a
|
|
|
|
|
|
|
Energy
(kcal/d)
|
2208.44
(2176.14,2240.75)
|
2105.43
(1901.50,2309.37)
|
2202.00
(2153.26,2250.73)
|
2246.70
(2203.30,2290.11)
|
2180.32
(2135.63,2225.01)
|
0.02
|
Protein
(gm/d)
|
84.79
(83.46,86.13)
|
75.57
(67.49,83.64)
|
82.41
(80.40,84.42)
|
86.50
(84.48,88.51)
|
85.49
(83.67,87.31)
|
<0.01
|
Carbohydrate
(gm/d)
|
263.43
(259.50,267.36)
|
265.50
(234.83,296.17)
|
267.82
(261.75,273.89)
|
267.37
(262.06,272.69)
|
255.81
(250.08,261.54)
|
<0.01
|
Dietary fiber
(gm/d)
|
16.25
(15.85,16.65)
|
14.62
(12.40,16.85)
|
16.63
(15.96,17.30)
|
16.64
(16.07,17.21)
|
15.59
(15.24,15.95)
|
<0.01
|
Total fat
(gm/d)
|
83.95
(82.34,85.55)
|
78.05
(69.06,87.04)
|
81.17
(79.07,83.27)
|
85.09
(83.08,87.10)
|
85.42
(83.13,87.72)
|
<0.01
|
Total saturated fatty acids
(gm/d)
|
27.77
(27.17,28.36)
|
26.02
(21.81,30.22)
|
26.87
(26.12,27.61)
|
28.08
(27.29,28.87)
|
28.30
(27.44,29.15)
|
<0.01
|
Total monounsaturated fatty acids
(gm/d)
|
30.81
(30.22,31.39)
|
28.80
(25.60,32.00)
|
29.75
(28.97,30.53)
|
31.17
(30.44,31.90)
|
31.43
(30.55,32.30)
|
<0.01
|
Total polyunsaturated fatty acids
(gm/d)
|
18.03
(17.67,18.40)
|
16.38
(14.52,18.24)
|
17.50
(16.96,18.04)
|
18.42
(17.91,18.93)
|
18.18
(17.68,18.68)
|
<0.01
|
Cholesterol
(mg/d)
|
295.88
(289.54,302.23)
|
225.50
(183.40,267.61)
|
275.93
(264.43,287.43)
|
302.18
(293.43,310.93)
|
309.08
(298.79,319.37)
|
<0.01
|
Phosphorus
(mg/d)
|
1394.47
(1373.73,1415.21)
|
1250.08
(1093.14,1407.01)
|
1369.79
(1337.05,1402.54)
|
1430.41
(1401.46,1459.35)
|
1385.48
(1356.02,1414.95)
|
<0.01
|
Magnesium
(mg/d)
|
304.64
(299.25,310.03)
|
279.02
(246.25,311.79)
|
309.70
(301.86,317.54)
|
313.39
(305.79,320.98)
|
292.60
(286.37,298.82)
|
<0.01
|
Selenium
(mcg/d)
|
113.80
(111.77,115.84)
|
97.68
(86.22,109.13)
|
112.04
(108.66,115.42)
|
115.56
(112.36,118.75)
|
114.10
(111.53,116.66)
|
<0.01
|
Caffeine
(mg/d)
|
192.94
(183.29,202.58)
|
200.35
(167.29,233.40)
|
184.65
(173.45,195.85)
|
208.20
(195.23,221.16)
|
185.37
(173.14,197.59)
|
<0.01
|
Moisture
(mg/d)
|
3104.85
(3047.73,3161.97)
|
2530.69
(2251.07,2810.30)
|
2990.19
(2913.94,3066.44)
|
3150.72
(3073.22,3228.22)
|
3175.99
(3093.86,3258.12)
|
<0.01
|
aSurvey-weighted percentage
(95%CI);
bP-value:survey-weighted Chi-square test.
Table 2.Regression Analyses of the Association Between BMI and Constipation (stool consistency), weighted
|
Model1 OR(95%CI)
|
Model2 OR(95%CI)
|
Model3 OR(95%CI)
|
<18kg/m2
|
1.00 (ref.)
|
1.00 (ref.)
|
1.00 (ref.)
|
[18,25)
|
0.78(0.45,1.36)
|
0.97(0.56,1.67)
|
1.04(0.60,1.82)
|
[25–30)
|
1.30(1.048,1.61)*
|
1.11(0.89,1.40)
|
1.146(0.92,1.43)
|
≥30kg/m2
|
1.62(1.33,1.98)**
|
1.54(1.25,1.89)**
|
1.67(1.35,2.07)**
|
P trend
|
<0.001
|
<0.001
|
<0.001
|
Model 1 was not adjusted for any confounding variables
Model 2 was adjusted for gender, age and ethnicity
Model 3 was adjusted for gender, age, ethnicity, education, ratio, diabetes, smoking, alcohol use, hypertensio, sleep duration, depression, physical activity and dietary intake of energy, protein, carbohydrate, dietary fiber, total fat, total saturated fatty acids, total monounsaturated fatty acids, total polyunsaturated fatty acids, cholesterol, phosphorus, magnesium, selenium, caffeine and moisture
OR, odds ratio.
*P<0.05; **P<0.01
Medical conditions were also considered. Diabetes was identified based on a physician's diagnosis, while hypertension was identified through professional medical advice. Depression status was ascertained using the NHANES mental health questionnaire, specifically relying on the Patient Health Questionnaire 9 scale, a validated tool for depression assessment[17].
Statistical analyses
We conducted our data analysis using R (version 3.4.3, 2021-12-21, R Foundation; http://www.r-project.org),Python(version 3.8) and Stata (MP 17.0). Accounting for sampling weights, we estimated sampling errors through Taylor series linearization. Using NHANES census data from 2005-2010, we summarized the weighted prevalence of US adults in proportional units. We undertook stratified analysis based on BMI classifications. For categorical variables, we presented data as percentages with 95% confidence intervals, employing the chi-squared test for group comparisons. Quantitative data were shown as mean ± standard deviation or standard error, with group differences ascertained through analysis of variance. We constructed regression models, adjusting for multiple confounders. While Model 1 remained unadjusted, Model 2 factored in age, gender, race/ethnicity, educational background, and marital status. Model 3 added considerations like alcohol consumption, smoking habits, health conditions such as diabetes, hypertension, depression, physical activity, and dietary components. Using logistic regression, we assessed the link between BMI and constipation, producing odds ratios (ORs) accompanied by 95% confidence intervals (95% CI). A p-value less than 0.05 was deemed statistically significant. To examine the relationship between BMI and stool quality, we employed a restricted cubic spline regression model with four knots located at the 25th, 50th, and 75th percentiles. To minimize potential errors from outliers, this model excluded the most extreme 1% of values. Additionally, we utilized a likelihood ratio test to assess non-linearity.