Abstract
The gut-kidney axis is increasingly recognized in chronic kidney disease (CKD) pathophysiology. Although stool frequency and consistency reflect gut microbiome status, their association with CKD remains underexplored. This study aimed to examine the relationship between bowel habits and CKD risk. In this cross-sectional analysis of 11,760 adults from National Health and Nutrition Examination Survey 2005 to 2010, stool frequency (weekly) and consistency (Bristol Stool Form Scale) were assessed. CKD was defined as estimated glomerular filtration rate (eGFR) < 60 mL/min/1.73m2 and/or urine albumin-to-creatinine ratio (uACR) ≥ 30 mg/g. Weighted logistic regression adjusted for multiple confounders was used; subgroup and mediation analyses were performed. Diarrhea-like stool (Bristol Stool Form Scale types 6/7) was independently associated with higher CKD risk (odds ratio [OR] = 1.57, 95% confidence interval [CI]: 1.14–2.27), consistent across eGFR < 60 (OR = 1.42, 95% CI: 1.01–1.99) and uACR ≥ 30 mg/g (OR = 1.55, 95% CI: 1.03–2.23). Mediation analysis indicated that systemic inflammation (C-reactive protein) explained 10.1%, 8.5%, and 15.3% of these associations for eGFR decline, elevated uACR, and overall CKD, respectively. Subgroup analyses supported robustness across demographic and clinical strata. Bowel movement frequency was not significantly associated with CKD. Diarrhea-like stool consistency, partly mediated by systemic inflammation, is associated with increased CKD risk.
Keywords: bowel movements, chronic kidney disease, glomerular filtration rate, stool consistency, urine albumin-creatinine ratio
1. Introduction
Chronic kidney disease (CKD) is a common public health problem with increasing prevalence and is strongly associated with cardiovascular disease (CVD) and mortality.[1] Several risk factors for CKD have been identified, such as diabetes and hypertension. However, it is critical to identify additional and simpler risk markers for primary intervention. Constipation or diarrhea, the prototype of functional gastrointestinal disorders, is one of the most common disorders in primary care.[2] These conditions can disrupt the intestinal microbial environment, leading to ecological and metabolic dysfunction, abnormal immune function and chronic inflammation in the body. In addition, abnormalities in the gut microbiota may influence the onset and progression of chronic diseases.[3,4]
An accumulating number of studies have emphasized the synergistic relationship of the microbiota-gut-kidney axis.[5] The incidence of constipation or diarrhea due to gut microbial imbalance is also higher in patients with CKD than in the general population.[6] A study revealed that the severity of constipation was linked to a high risk of CKD and end-stage renal disease events, indicating a potential connection between defecation and CKD.[7] In clinical practice, stool frequency is a common criterion for diagnosing constipation or diarrhea. However, the inconsistency of the demarcation criteria often affects the conclusions from these observations. Recent research has shown that fecal character was closely linked to the abundance of the intestinal microbiota, colonic transit time, stool type, and bacterial growth rate. Stool consistency was considered to be a representative indicator of intestinal bacterial metabolism and intestinal function.[8,9] Moreover, it was included in the Rome III criteria and Bristol Stool Form Scale (BSFS) entries.
Currently, little has been reported on the relationship between stool and CKD risk. Therefore, we aimed to evaluate the association between bowel movements and stool types with estimated glomerular filtration rate (eGFR) and albuminuria (indicators of CKD) using data representative of the United States (U.S.) adult population.
2. Methods
2.1. Data and participants
The National Health and Nutrition Examination Survey (NHANES) is a national nutrition and health-related survey study of adults and children. Data were obtained from a representative sample of the population in stratified, multistage probability aggregation, which was obtained through structured questionnaires, health examinations, and laboratory tests.[10] Detailed information is available from the Centers for Disease Control and Prevention (https://www.cdc.gov/nchs/nhanes/). On this occasion, we used continuous data from NHANES 2005 to 2010, which provides complete information on stool types and frequency. The NHANES protocol was approved by the Research Ethics Review Board of the National Center for Health Statistics, and all participants provided written informed consent.
We included 31,034 initial subjects at 6 years (3 cycles). Participants aged < 20 years (n = 13,902), pregnant (n = 461), with chronic gastrointestinal disease (n = 1254) or gastrointestinal cancer (n = 128) or missing stool information (n = 2432), or lacking laboratory data related to renal function (n = 1097) were excluded by itemized exclusion criteria, and data from 11,760 participants were ultimately used for analysis (Fig. 1A)
Figure 1.
Research Flowchart. (A) Participants screening flowchart; (B) Classification scheme.
2.2. Exposure variables-bowel movements and stool types
Bowel movements (frequency) and stool types (consistency), which are also commonly used to define constipation or diarrhea, were self-reported during the NHANES 2005 to 2010 cycle using the Bowel Health questionnaire. Bowel movements were counted using weekly data, and the stool consistency scale was measured using the BSFS. BSFS has been applied in a series of studies. These studies have shown that a subject’s rating of his/her own individual stools is robustly correlated with measures of how long it takes radiopaque markers to pass through the gastrointestinal tract. The specific typing was: type 1 (separate hard lumps, like nuts), type 2 (sausage-like, but lumpy), type 3 (like a sausage but with cracks in the surface), type 4 (like a sausage or snake, smooth and soft), type 5 (soft blobs with clear-cut edges), type 6 (fluffy pieces with ragged edges, a mushy stool) and type 7 (watery, no solid pieces).[11] When the number of bowel movements was used as a reference, typically ≤ 3 per week was considered constipation, 3 to 21 was normal, and ≥ 21 was diarrhea. Correspondingly, when stool characteristics were used as a reference, BSFS types 1 to 2 were considered constipation, BSFS types 3 to 5 were normal, and BSFS types 6 to 7 were defined as diarrhea.[12]
2.3. Outcome variable-CKD
CKD was identified when the eGFR < 60 mL/min/1.73 m2 and/or the urine albumin-to-creatinine ratio (uACR) ≥ 30 mg/g. The eGFR was calculated using the CKD Epidemiology Collaborative Study formula (a gender-, creatinine-, and ethnicity-specific formula), while the uACR was obtained from the urine albumin/creatinine ratio.[13] In our research, CKD was the primary study outcome variable, while eGFR and uACR were secondary.
2.4. Covariates
Baseline information on demographic characteristics, smoking status, alcohol consumption, physical activity (PA), mental health, disease history (CVD, hypertension or diabetes mellitus), diet (dietary energy, carbohydrates, protein, fat, fiber and moisture) and drugs were obtained through a standardized questionnaire. Body mass index (BMI) was calculated based on measured weight/height2 (kg/m2). Smoking status was categorized as: nonsmoker (smoking < 100 cigarettes to date), former smoker (smoking more than 100 cigarettes to date but not currently a smoker) or current smoker (smoking more than 100 cigarettes and currently a smoker). Drinking status was divided as: nondrinkers, low-to moderate drinkers (< 2 drinks per day for men and < 1 drink per day for women) or heavy drinkers (≥ 2 drinks per day for men and ≥ 1 drink per day for women). PA was calculated as a sum according to the metabolic equivalent estimate (MET) corresponding to the specific type of activity, with individual PA (MET-min/week) = MET × weekly frequency × duration. We calculated the total PA of the subjects, which included the sum of metabolic equivalents for work, walking and bicycling, muscular exercise, household tasks, and personal recreational activities. And in terms of the total PA values, they were divided into 3 groups: none or < 600 (MET-min/week), 600 to 3999 (MET-min/week), and ≥ 4000 (MET-min/week).[14] Mental Health referred to as the Depression Screener Questionnaire, all 9 relevant questions were derived from the Patient Health Questionnaire-9 (PHQ-9) and were classified into 3 categories depending on the total PHQ-9 score: no depression < 5, mild depression 5 to 9, and moderate to severe depression ≥ 10.[15] Dietary fiber and water intake were then obtained based on calculations from 2 24-hour dietary reviews and averaged. The definition of CVD included individuals who self-reported or had been diagnosed with congestive heart failure, coronary heart disease, angina pectoris, heart attack, or stroke. Hypertension was defined as a systolic blood pressure ≥ 130 mm Hg and/or diastolic blood pressure ≥ 80 mm Hg (mean of all valid measurements), or self-reported diagnosis of hypertension or the use of antihypertensive medications. Diabetes status was defined based on a self-reported physician diagnosis of diabetes mellitus; the use of insulin or oral hypoglycemic medications; a fasting blood glucose level ≥ 7.0 mmol/L; a 2-hour postprandial blood glucose level ≥ 11.1 mmol/L from an oral glucose tolerance test; or a glycosylated hemoglobin A1c (HbA1c) level ≥ 6.5% (48 mmol/mol).[16] The history of other diseases was determined based on self-reported medical history.
Additional laboratory data relevant to this study were obtained, including hemoglobin, albumin, lipids, urea nitrogen, HbA1c and C-reactive protein (CRP). More details of the tests are available in the NHANES Laboratory Medical Technician Procedures Manual.
2.5. Statistical analyses
All analyses incorporated NHANES sampling weights to account for the complex survey design. Data are presented as weighted mean ± standard deviation (normal continuous variables), weighted median (interquartile range) (non‑normal continuous variables), or unweighted counts with weighted percentages (categorical variables). Participants were categorized by bowel frequency (≤ 3/week [group a, constipation], 3–21/week [group b, normal], ≥ 21/week [group c, diarrhea]) and stool consistency using the BSFS (types 1–2 [group A, hard], 3–5 [group B, normal], 6–7 [group C, diarrhea‑like]). Group differences were assessed with 1‑way ANOVA, Kruskal–Wallis, or χ2 tests, as appropriate.
Weighted multivariable logistic regression was used to estimate odds ratios (ORs) and 95% confidence intervals (CI) for the association of stool patterns with CKD. Three sequential models were fitted: Model 1 adjusted for stool classifications; Model 2 added age, sex, race, BMI, smoking, and alcohol intake; Model 3 further adjusted for CVD, hypertension, diabetes, PHQ-9 score, PA (MET-minutes/week), medication use (number of prescriptions, diuretics, antacids, opioids, calcium-channel blockers), dietary components (energy, protein, carbohydrate, fat, fiber, moisture), HbA1c, and CRP.
To assess robustness, we repeated analyses using eGFR < 60 mL/min/1.73 m2 and uACR ≥ 30 mg/g as separate outcomes and performed subgroup analyses across demographic and clinical strata. We also examined joint categories of stool frequency and consistency. In addition, causal mediation analysis was conducted to estimate the proportion of the stool–CKD association explained by systemic inflammation (log‑transformed CRP), using linear regression for eGFR and uACR, and logistic regression for binary CKD. All analyses were performed in R version 4.3.1 (R Foundation for Statistical Computing, Vienna, Austria), with 2‑sided P < .05 considered statistically significant.
3. Results
3.1. Basic characteristics of the participants
A total of 11,760 (Weighted to represent about 155 million U.S. population) participants were included in the analysis (mean age, 47.2 ± 16.4 years; male, 51.4%; Non-Hispanic Black, 10.1%; weighted). Among the 2091 participants with CKD, 1332 had albuminuria (uACR ≥ 30 mg/g) and 1105 had reduced eGFR (< 60 mL/min/1.73m2), with an overlap of 350 individuals. As previously described, we triple-categorized the bowel movements according to the frequency of ≤ 3 bowel movements per week (group a), 3 to 21 per week (group b), ≥ 21 per week (group c), and BSFS type 1/2 (group A), type 3 to 5 (group B), and type 6/7 (group C). The weighted basic characteristics of all participants according to these 2 classifications were shown in Table 1. For the CKD or not, there was a significant difference among the 3 groups categorized according to stool types (P < .001), while the difference between the groups classified according to bowel movements was close to being statistically significant (P = .051). Variables such as age, sex, race, hypertension status, diabetes status, BMI, smoking status, drinking status, PHQ-9 score, MET, drugs, diet, hemoglobin, albumin, triglyceride, HbA1c, CRP, serum creatinine and uACR exhibited significant differences between groups in both classifications (P < .05). Conversely, CVD, urea nitrogen, and eGFR demonstrated distinct trends of differentiation between the 2 classifications. When dividing all subjects by CKD status, there was a compelling discrepancy in the weighted percentages of the stool types but not the bowel movements between the 2 groups. The CKD group had a significantly higher proportion of abnormal fecal consistency compared to the non-CKD group. Further details of other characteristics can be found in Table S1 and S2.
Table 1.
Baseline characteristics of participants classified by bowel movements and stool types, weighted.
| Characteristics† | Overall (n = 11,760) |
Classified by bowel movements |
P
value |
Classified by stool types | P value | ||||
|---|---|---|---|---|---|---|---|---|---|
| Group a (n = 891) |
Group b (n = 10,107) |
Group c (n = 762) |
Group A (n = 853) |
Group B (n = 10,050) |
Group C (n = 857) |
||||
| Age, yrs | 47.2 ± 16.4 | 43.6 ± 17.0 | 47.5 ± 16.4 | 47.7 ± 15.2 | < .001 | 46.5 ± 17.3 | 47.1 ± 17.3 | 50.5 ± 15.6 | < .001 |
| Male, n (%) | 6059 (51.4) | 289 (27.5) | 5317 (51.5) | 453 (62.4) | < .001 | 248 (29.1) | 5266 (52.4) | 388 (45.3) | < .001 |
| Race, n (%) | < .001 | < .001 | |||||||
| Non-Hispanic Black | 2238 (10.1) | 271 (18.6) | 1805 (9.2) | 162 (12.6) | 186 (13.3) | 1873 (9.8) | 179 (12.0) | ||
| Non-Hispanic White | 5919 (72.3) | 430 (68.8) | 5168 (73.1) | 321 (65.0) | 379 (66.4) | 5187 (73.2) | 353 (66.4) | ||
| Mexican American | 2129 (7.9) | 95 (5.1) | 1840 (7.8) | 194 (12.0) | 162 (9.1) | 1773 (7.6) | 194 (10.4) | ||
| Others | 1474 (9.7) | 95 (7.5) | 1294 (9.8) | 85 (10.4) | 126 (11.2) | 1271 (9.4) | 131 (11.3) | ||
| CVD, n (%) | 1242 (7.8) | 112 (8.8) | 1058 (7.7) | 72 (7.6) | .538 | 92 (7.4) | 1031 (7.6) | 119 (10.5) | .018 |
| Hypertension, n (%) | 4545 (35.9) | 304 (30.7) | 3919(35.9) | 322 (41.5) | .002 | 298 (31.1) | 3884(36.0) | 363 (39.1) | .012 |
| Diabetes, n (%) | 2047 (12.3) | 134 (10.3) | 1748 (12.2) | 165 (15.5) | .011 | 138 (10.9) | 1691 (12.0) | 218 (18.5) | < .001 |
| BMI, kg/m2 | 28.6 ± 6.6 | 27.2 ± 5.9 | 28.6 ± 6.5 | 30.6 ± 7.0 | < .001 | 27.5 ± 6.6 | 28.6 ± 6.5 | 30.0 ± 7.9 | < .001 |
| Smoking, n (%) | < .001 | .003 | |||||||
| No | 6195 (53.1) | 495 (55.5) | 5313(52.8) | 387 (53.7) | 508 (60.3) | 5255 (52.8) | 432 (48.4) | ||
| Former | 3068 (25.5) | 175 (18.3) | 2660 (25.8) | 233 (30.0) | 190 (21.2) | 2630(25.6) | 248 (28.9) | ||
| Current | 2493 (21.4) | 221 (22.1) | 2130 (21.4) | 142 (16.3) | 155 (18.5) | 2161 (21.6) | 177 (22.7) | ||
| Drinking, n (%) | < .001 | < .001 | |||||||
| No | 3440 (27.5) | 318 (35.7) | 2908 (26.9) | 214 (26.6) | 342 (39.5) | 2790 (23.2) | 308 (33.0) | ||
| Low-to-moderate | 5953 (62.3) | 407 (58.0) | 5187 (62.7) | 359 (61.5) | 366 (53.3) | 5217(63.4) | 370 (55.4) | ||
| Heavy | 941 (10.1) | 46 (6.2) | 813 (10.5) | 82 (11.8) | 41 (7.2) | 830 (10.4) | 70 (11.4) | ||
| PHQ-9, n (%) | < .001 | < .001 | |||||||
| < 5 | 9223 (80.1) | 599 (70.0) | 8053 (81.2) | 571 (76.6) | 599 (72.5) | 8018 (87.3) | 606 (72.5) | ||
| 5–9 | 1676(14.1) | 163 (16.6) | 1391 (13.7) | 122 (16.7) | 157 (18.0) | 1383 (13.7) | 136 (16.1) | ||
| ≥ 10 | 842 (5.6) | 129 (13.4) | 644 (5.1) | 69 (6.6) | 95 (9.5) | 635 (5.1) | 112 (11.4) | ||
| MET, min/week | < .001 | < .001 | |||||||
| < 600 | 4282 (34.8) | 400 (46.3) | 3615 (34.0) | 267 (32.6) | 169 (47.4) | 3544 (38.7) | 369 (47.4) | ||
| 600–3999 | 4007 (41.8) | 228 (34.3) | 3491 (42.5) | 251 (39.3) | 251 (32.3) | 3529 (38.3) | 247 (31.7) | ||
| ≥ 4000 | 2433 (23.4) | 143 (19.3) | 2101 (23.4) | 189 (28.1) | 159 (20.4) | 2111 (23.0) | 163 (20.9) | ||
| Number of drugs | 2.0 (1.0, 4.0) | 3.0 (1.0, 5.0) | 2.0 (1.0, 4.0) | 3.0 (1.0, 5.0) | .023 | 2.5 (1.0, 4.0) | 2.0 (1.0, 4.0) | 3.0 (2.0, 6.0) | < .001 |
| Drugs, n (%) | |||||||||
| Diuretics | 1491 (10.1) | 96 (7.7) | 1294 (10.2) | 101 (12.5) | .013 | 101 (9.0) | 1260 (10.1) | 130 (12.4) | .065 |
| Antacids | 1305 (10.2) | 106 (10.2) | 1103 (10.1) | 96 (11.6) | .513 | 86 (9.9) | 1089 (9.9) | 130 (14.6) | .038 |
| Opioids | 221 (1.7) | 26 (2.9) | 184 (1.6) | 11 (1.1) | .019 | 17 (2.3) | 185 (1.6) | 19 (2.3) | .096 |
| CCB | 792 (5.0) | 69 (5.2) | 675 (4.9) | 48 (5.2) | .093 | 54 (4.1) | 656 (4.9) | 82 (6.9) | .032 |
| Total diet, per day | |||||||||
| Energy, kcal | 2147 ± 872 | 11925 ± 767 | 2150 ± 866 | 2377 ± 1005 | < .001 | 1909 ± 728 | 2168 ± 879 | 2108 ± 879 | < .001 |
| Protein, gm | 84.3 ± 36.8 | 71.2 ± 32.6 | 84.7 ± 36.4 | 94.6 ± 42.1 | < .001 | 73.3 ± 31.5 | 85.2 ± 36.9 | 82.3 ± 38.1 | < .001 |
| Carbohydrate, gm | 258.5 ± 111.4 | 242.5 ± 110.8 | 258.5 ± 101.7 | 278.9 ± 122.7 | < .001 | 240.2 ± 92.4 | 260.3 ± 112.2 | 254.1 ± 109.6 | < .001 |
| Fat, gm | 81.5 ± 40.7 | 72.9 ± 35.5 | 81.5 ± 40.2 | 91.9 ± 49.7 | < .001 | 70.7 ± 34.5 | 82.4 ± 40.8 | 79.9 ± 42.1 | < .001 |
| Fiber, mg | 16.7 ± 8.7 | 13.4 ± 7.0 | 16.9 ± 8.7 | 18.2 ± 9.4 | < .001 | 14.5 ± 7.3 | 16.9 ± 8.5 | 16.6 ± 10.7 | < .001 |
| Dietary moisture, mg/day | 2749 (2085-3618) |
2224 (1757-2884) |
2782 (2114-3642) |
3702 (2140-4029) |
< .001 | 2323 (1784-3124) |
2783 (2118-3655) |
2691 (1997-3594) |
< .001 |
| Hemoglobin, g/L | 14.4 ± 1.5 | 13.8 ± 1.5 | 14.5 ± 1.5 | 14.7 ± 1.4 | < .001 | 13.9 ± 1.4 | 14.4 ± 1.5 | 14.2 ± 1.5 | < .001 |
| Albumin, g/L | 42.7 ± 3.2 | 42.3 ± 3.4 | 42.8 ± 3.1 | 42.4 ± 3.4 | .003 | 42.3 ± 3.2 | 42.8 ± 3.2 | 42.3 ± 3.4 | < .001 |
| TC, mmol/L | 5.0 ± 1.1 | 4.9 ± 1.1 | 5.0 ± 1.1 | 5.2 ± 1.1 | .001 | 5.0 ± 1.1 | 5.0 ± 1.1 | 5.1 ± 1.1 | .836 |
| Triglyceride, mmol/L | 1.2 (0.9–1.8) | 1.2 (0.8–1.6) | 1.2 (0.9–1.8) | 1.4 (1.0–2.0) | < .001 | 1.2 (0.8–1.8) | 1.2 (0.9–1.8) | 1.4 (1.0–2.0) | .009 |
| HbA1c, % | 5.7 ± 1.0 | 5.6 ± 1.0 | 5.7 ± 1.0 | 5.8 ± 1.2 | < .001 | 5.7 ± 1.1 | 5.7 ± 1.0 | 5.8 ± 1.1 | < .001 |
| CRP, mg/dL | 0.17 (0.07–0.39) |
0.16 (0.06–0.39) |
0.17 (0.07–0.39) |
0.22 (0.09–0.47) |
.001 | 0.16 (0.07–0.38) |
0.17 (0.07–0.38) |
0.22 (0.08–0.50) |
.001 |
| Serum creatinine, mg/dL | 0.89 (0.75–1.00) |
0.80 (0.71–0.95) |
0.89 (0.75–1.00) |
0.90 (0.77–1.02) |
< .001 | 0.82 (0.70–0.92) |
0.90 (0.76–1.00) |
0.86 (0.72–1.00) |
< .001 |
| Urea nitrogen, mg/dL | 12 (10–15) | 11 (9–14) | 12 (10–15) | 13 (10–16) | < .001 | 12 (10–15) | 12 (10–15) | 12 (9–16) | .534 |
| eGFR,ml/min/1.73 m2 | 94 (79–108) | 98 (80–114) | 93 (78–107) | 94 (81–109) | .002 | 95 (78–112) | 94 (79–107) | 93 (78–107) | .082 |
| uACR, mg/g | 6.0 (4.0–10.7) | 6.4 (4.2–12.3) | 6.0 (4.0–10.6) | 6.0 (4.2–10.5) | .026 | 6.3 (4.3–12.1) | 5.9 (4.0–10.4) | 6.9 (4.3–15.0) | < .001 |
| CKD, n (%) | 2091 (13.2) | 168 (15.5) | 1780 (12.9) | 143 (15.0) | .051 | 156 (14.1) | 1729 (12.8) | 106 (18.7) | < .001 |
BMI = body mass index, CCB = calcium-channel blockers, CKD = chronic kidney disease, CRP = C-reactive protein, CVD = cardiovascular disease, eGFR = estimated glomerular filtration rate, HbA1c = glycosylated hemoglobin A1c, MET = metabolic equivalent, n = number of participants, PHQ = patient health questionnaire, TC = total cholesterol, uACR = urine albumin/creatinine ratio.
n (%): unweighted numbers (weighted percentage).
3.2. Associations of bowel movements and stool types with CKD
We analyzed in detail the potential association of stool classification with CKD using 3 outcome variables (eGFR < 60 ml/min/1.73m2, uACR ≥ 30 mg/g and CKD). As previously described in the Statistical analyses section, we adjusted for multiple covariates and 3 models. The groups of normal stool (groups b and B) as references. As shown in Table 2, group C (representing BSFS type 6/7, which predicts markedly unformed stools or diarrhea) was associated with a high risk of CKD, and this striking association was observed in the crude model [1.57 (1.25–1.97), P < .001], model 1 [1.24 (1.24–1.93), P < .001], model 2 [1.40 (1.04–1.87), P = .026] and model 3 [1.57 (1.14–2.57), P = .009], whereas there was no statistical association for group A (P = .979).
Table 2.
Weighted OR (95% CI) for CKD, eGFR < 60ml/min/1.73 m2 and uACR ≥ 30 mg/g classified by bowel movements and stool types.
| Unadjusted Model | Model 1 | Model 2 | Model 3 | |||||
|---|---|---|---|---|---|---|---|---|
| OR (95% CI) | P value | OR (95% CI) | P value | OR (95% CI) | P value | OR (95% CI) | P value | |
| CKD | ||||||||
| Classified by bowel movements | ||||||||
| Group a | 1.23 (1.04–1.46) | .014* | 1.22 (1.04–1.44) | .015* | 1.47 (1.15–1.89) | 0.003* | 1.36 (0.95–1.94) | .084 |
| Group b | 1.0 | - | 1.0 | - | 1.0 | - | 1.0 | - |
| Group c | 1.19 (0.92–1.54) | .180 | 1.12 (0.87–1.45) | .362 | 1.25 (0.92–1.71) | 0.142 | 1.24 (0.86–1.77) | .223 |
| Classified by stool types | ||||||||
| Group A | 1.12 (0.89–1.41) | .326 | 1.09 (0.87–1.36) | .461 | 1.01 (0.75–1.36) | 0.966 | 1.00 (0.70–1.41) | .979 |
| Group B | 1.0 | - | 1.0 | - | 1.0 | - | 1.0 | - |
| Group C | 1.57 (1.25–1.97) | < .001* | 1.24 (1.24–1.93) | < .001* | 1.40 (1.04–1.87) | 0.026* | 1.57 (1.14–2.17) | .009* |
| eGFR < 60ml/min/1.73 m2 | ||||||||
| Classified by bowel movements | ||||||||
| Group a | 1.19 (0.90–1.58) | .204 | 1.19 (0.90–1.57) | .217 | 1.53 (0.97–2.44) | 0.074 | 1.47 (0.83–2.61) | .171 |
| Group b | 1.0 | - | 1.0 | - | 1.0 | - | 1.0 | - |
| Group c | 0.79 (0.57–1.08) | .132 | 0.76 (0.55–1.05) | .090 | 0.87 (0.57–1.33) | 0.504 | 0.84 (0.49–1.42) | .491 |
| Classified by stool types | ||||||||
| Group A | 1.10 (0.81–1.50) | .529 | 1.07 (0.79–1.46) | .652 | 0.87 (0.57–1.33) | 0.518 | 0.82 (0.51–1.32) | .386 |
| Group B | 1.0 | - | 1.0 | - | 1.0 | - | 1.0 | - |
| Group C | 1.31 (1.03–1.66) | .028* | 1.35 (1.06–1.72) | .017* | 1.18 (0.87–1.62) | 0.274 | 1.42 (1.01–1.99) | .042* |
| uACR ≥ 30 mg/g | ||||||||
| Classified by bowel movements | ||||||||
| Group a | 1.24 (0.99–1.55) | .053 | 1.28 (0.99–1.54) | .059 | 1.46 (1.12–1.89) | 0.006* | 1.29 (0.91–1.82) | .127 |
| Group b | 1.0 | - | 1.0 | - | 1.0 | - | 1.0 | - |
| Group c | 1.48 (1.12–1.94) | .006* | 1.39 (1.05–1.82) | .020* | 1.44 (1.07–1.93) | 0.019* | 1.44 (1.01–2.06) | .044* |
| Classified by stool types | ||||||||
| Group A | 1.11 (0.86–1.44) | .430 | 1.08 (0.83–1.39) | .568 | 1.06 (0.74–1.53) | 0.749 | 1.07 (0.72–1.60) | .702 |
| Group B | 1.0 | - | 1.0 | - | 1.0 | - | 1.0 | - |
| Group C | 1.69 (1.26–2.26) | < .001* | 1.62 (1.21–2.17) | .002* | 1.44 (1.02–2.04) | 0.040* | 1.55 (1.03–2.33) | .038* |
Model 1 was adjusted for bowel movements and stool types, Model 2 was adjusted for Model 1 + age + sex + race + body mass index + smoking + drinking, and Model 3 was adjusted for Model 2 + cardiovascular disease + hypertension + diabetes + patient health questionnaire-9 + metabolic equivalent + drugs + diet + glycosylated hemoglobin A1c + C-reactive protein.
CI = confidence interval, CKD = chronic kidney disease, eGFR = estimated glomerular filtration rate, OR = weighted odds ratio, uACR = urine albumin/creatinine ratio.
P < .05.
Furthermore, we performed separate tests for the 2 sub-diagnostic criteria of eGFR < 60 ml/min/1.73m2 and uACR ≥ 30 mg/g, and consistent results were found. Group C continued to be a high-risk factor for the 2 associated adverse conditions, the weighted ORs were 1.42 (1.01–1.99) and 1.55 (1.03–2.33), respectively. Interestingly, after adjusting for model 3, classified by bowel movements (group a and c) were not positively associated with the risk of CKD and eGFR < 60 ml/min/1.73 m2 (P < .05). However, group c demonstrated a somewhat positive and statistically significant association with uACR ≥ 30 mg/g. The weighted ORs between other variables and outcomes were shown in Table S3.
Additionally, we analyzed the association between specific BSFS classification types (types 1–7) and CKD. It was consistent that both types that make up group C (types 6/7) showed a positive association with CKD. The results were shown in Table S4.
3.3. Subgroup analysis
The strong significance of the association between group C and CKD in the stool types subgroups was noticed in the aforementioned population-wide analysis, so we further analyzed the risk association between group C with CKD (group B as a control) in different subgroups. As shown in Figure 2, the OR of group C with CKD remained stable across most strata in the weighted subgroup interaction analysis. We only found an interaction between group C and sex (P value for interaction = .034). Notably, in the subgroup with CVD, we appeared to observe a significantly stronger positive risk association between group C and CKD (OR = 3.35, 95% CI: 1.22–9.15), although the the interaction between groups was not significant (P for interaction > .05).
Figure 2.
Forest plot of subgroup analysis of the risk between the group C and CKD, group B as the reference. Odds ratio was adjusted for adjusted for bowel movements + stool types + age + sex + race + BMI + smoking + drinking + CVD + hypertension + diabetes + PHQ-9 + MET + drugs + diet + HbA1c + CRP. BMI = body mass index, CKD = chronic kidney disease, CRP = C-reactive protein, CVD = cardiovascular disease, eGFR = estimated glomerular filtration rate, HbA1c = Glycosylated hemoglobin A1c, MET = metabolic equivalent, PHQ = patient health questionnaire, uACR = urine albumin/creatinine ratio..
3.4. Association of combined classification of stool frequency and traits with CKD
Stool frequency and characteristics are often assessed together in clinical practice. We combined pairs based on the 2 classifications, which formed 9 new subgroups (groups 1–9) (Fig. 1B). By related analysis, we found that stool frequency increased progressively with stool irregularity, as shown in Figure 3A. However, the subgroups of stool frequency and traits grouped on the same definitions did not overlap as well, as shown in Figure 3B. The weighted percentage distributions of CKD, eGFR < 60 ml/min/1.73m2, and uACR ≥ 30 mg/g did not perform consistently across the 3 groupings, as shown in Figure 3C. Notably, the high prevalence of CKD in groups C, groups 3, groups 6, and groups 9, which were grouped based on diarrhea-like stool traits, was well represented in Figure 3C. The characteristics of participants classified by groups 1 to 9 were provided in Table 3.
Figure 3.
Characteristics of different stool classifications. (A)Relationship between different stool types and the number of bowel movements; (B) Diagnostic concordance between 2 different stool classifications; (C) Weighted percentage distribution of CKD, eGFR < 60 ml/min/1.72 m2 and uACR ≥ 30 mg/g in 3 classifications. CKD = chronic kidney disease, eGFR = estimated glomerular filtration rate, uACR = urine albumin/creatinine ratio.
Table 3.
Characteristics of participants classified by bowel movements combined with stool types, weighted.
| Characteristics† | Group 1 (n = 174) |
Group 2 (n = 670) |
Group 3 (n = 47) |
Group 4 (n = 644) |
Group 5 (n = 8798) |
Group 6 (n = 665) |
Group 7 (n = 35) |
Group 8 (n = 582) |
Group 9 (n = 145) |
P value |
|---|---|---|---|---|---|---|---|---|---|---|
| Age, yrs | 47.6 ± 17.4 | 42.4 ± 16.6 | 48.4 ± 17.9 | 46.3 ± 17.4 | 47.4 ± 16.3 | 50.7 ± 15.6 | 46.1 ± 15.9 | 47.1 ± 15.1 | 50.5 ± 15.1 | < .001 |
| Male, n (%) | 39 (17.3) | 234 (29.6) | 16 (34.5) | 235 (31.3) | 4784(53.2) | 298 (44.8) | 17 (47.0) | 365 (65.6) | 71 (52.7) | < .001 |
| Race, n (%) | < .001 | |||||||||
| Non-Hispanic Black | 56 (19.7) | 204 (18.7) | 11 (13.0) | 124 (11.7) | 1540(8.9) | 141 (12.0) | 6 (10.7) | 129 (13.0) | 27 (11.1) | |
| Non-Hispanic White | 79 (66.2) | 327 (69.0) | 24 (76.2) | 285 (66.2) | 4614 (74.1) | 269 (65.4) | 15 (69.5) | 245 (64.2) | 61 (67.5) | |
| Mexican American | 18 (5.1) | 68 (4.9) | 9 (8.4) | 135 (10.2) | 1558(7.5) | 147 (10.3) | 9 (9.1) | 146 (12.2) | 39 (12.0) | |
| Others | 21 (9.0) | 71 (7.5) | 3 (2.4) | 100 (11.8) | 1084 (9.5) | 110 (12.3) | 5 (10.7) | 62 (10.6) | 18 (9.4) | |
| CVD, n (%) | 26 (11.6) | 74 (7.8) | 12 (14.9) | 61 (6.1) | 912 (7.7) | 85 (9.5) | 5 (10.0) | 45 (6.1) | 22 (13.8) | .035 |
| Hypertension, n (%) | 61 (31.1) | 221 (30.0) | 22 (40.5) | 220 (30.8) | 3416(36.1) | 283 (38.9) | 17 (36.7) | 247 (42.5) | 58 (38.3) | .010 |
| Diabetes, n (%) | 30 (11.6) | 91 (9.3) | 13 (20.6) | 101 (10.8) | 1490(12.1) | 157 (16.9) | 7 (9.8) | 110 (13.7) | 48 (25.1) | < .001 |
| BMI, kg/m2 | 26.6 ± 5.7 | 27.3 ± 5.9 | 29.5 ± 7.2 | 27.7 ± 6.7 | 28.6 ± 6.4 | 29.8 ± 8.0 | 28.9 ± 5.5 | 30.6 ± 6.9 | 31.4 ± 8.1 | < .001 |
| Smoking, n (%) | < .001 | |||||||||
| No | 100 (57.4) | 370 (55.7) | 25 (44.9) | 387 (61.0) | 4588 (52.6) | 338 (47.8) | 21 (61.2) | 296 (53.4) | 70 (52.7) | |
| Former | 38 (21.8) | 126 (17.1) | 11 (25.0) | 142 (20.7) | 2328(25.9) | 190 (29.2) | 10 (27.2) | 175 (30.4) | 48 (29.1) | |
| Current | 36 (20.8) | 174 (27.2) | 11 (30.1) | 115 (18.3) | 1876 (21.5) | 139 (23.0) | 4 (11.6) | 111 (16.2) | 27 (18.2) | |
| Drinking, n (%) | < .001 | |||||||||
| No | 69 (45.5) | 229 (32.9) | 20 (45.9) | 258 (37.9) | 2413 (25.8) | 237 (32.1) | 15 (42.6) | 147 (24.4) | 52 (32.7) | |
| Low-to-moderate | 66 (49.8) | 322 (60.5) | 19 (47.2) | 286 (54.2) | 4598 (63.6) | 303 (57.0) | 14 (52.2) | 296 (64.3) | 49 (50.5) | |
| Heavy | 5 (4.6) | 38 (6.6) | 3 (6.8) | 34 (7.9) | 728 (10.6) | 51 (10.9) | 2 (5.1) | 64 (11.3) | 16 (16.8) | |
| PHQ-9, n (%) | < .001 | |||||||||
| < 5 | 108 (65.1) | 467 (72.8) | 24 (44.1) | 468 (74.8) | 7101(82.0) | 484 (75.4) | 23 (68.4) | 449 (79.0) | 99 (68.2) | |
| 5–9 | 33 (18.7) | 118 (15.0) | 12 (34.2) | 114 (17.2) | 1175 (13.5) | 102 (13.9) | 10 (28.9) | 89 (15.1) | 23 (20.7) | |
| ≥ 10 | 33 (16.2) | 85 (12.2) | 11 (21.7) | 60 (8.0) | 506 (4.5) | 78 (10.7) | 2 (2.7) | 44 (5.9) | 23 (11.2) | |
| MET, min/week | < .001 | |||||||||
| < 600 | 83 (46.1) | 300 (46.7) | 17 (40.9) | 273 (41.3) | 3053 (33.2) | 289 (39.9) | 13 (29.4) | 190 (30.5) | 64 (43.4) | |
| 600–3999 | 48 (34.8) | 204 (33.9) | 13 (40.8) | 190 (36.6) | 3109(43.4) | 192 (35.7) | 13 (37.1) | 195 (40.0) | 43 (36.4) | |
| ≥ 4000 | 24 (19.0) | 110 (19.4) | 9 (18.2) | 127 (22.1) | 1847(23.5) | 127 (24.4) | 8 (33.5) | 154 (29.4) | 27 (20.2) | |
| Number of drugs | 3.0 (2.0,5.0) | 2.0 (1.0,5.0) | 3.0 (3.0,6.9) | 2.0 (1.0,4.0) | 2.0 (1.0,4.0) | 3.0 (1.0,5.7) | 3.0 (1.9,4.7) | 2.0 (1.0,4.0) | 5.0 (2.0,7.0) | < .001 |
| Drugs, n (%) Diuretics Antacids Opioids CCB |
23 (10.1) 21 (10.2) 6 (4.7) 13 (5.2) |
7 (7.2) 77 (9.7) 19 (2.5) 49 (5.2) |
6 (6.1) 8 (17.2) 1 (1.6) 7 (5.2) |
74 (8.6) 5 (9.6) 11 (1.9) 40 (5.2) |
1120 (10.2) 945 (9.9) 159 (1.5) 574 (5.2) |
100 (11.9) 99 (14.2) 14 (2.5) 61 (5.2) |
4 (12.1) 6 (14.0) 0 (0) 1 (5.2) |
72 (11.4) 66 (10.4) 7 (1.1) 33 (5.2) |
25 (17.8) 24 (16.3) 4 (1.5) 14 (5.2) |
.186 .008 .350 .037 |
| Total diet, per day | ||||||||||
| Energy, kcal | 1754 ± 645 | 1975.23 ± 790 | 1804 ± 745 | 1946 ± 751 | 2168.3 ± 872 | 2078 ± 856 | 1991 ± 613 | 2410 ± 1027 | 2343 ± 977 | < .001 |
| Protein, gm | 63.8 ± 23.3 | 73.7 ± 34.5 | 59.6 ± 26.0 | 75.2 ± 32.5 | 85.4 ± 36.5 | 82.5 ± 37.3 | 83.9 ± 36.3 | 96.5 ± 42.3 | 88.7 ± 42.1 | < .001 |
| Carbohydrate, gm | 226.3 ± 88.5 | 247.3 ± 104.6 | 226.8 ± 97.9 | 244.4 ± 93.8 | 260.0 ± 111.7 | 249.1 ± 107.9 | 233.7 ± 81.2 | 280.2 ± 126.1 | 286.3 ± 115.1 | < .001 |
| Fat, gm | 65.0 ± 29.2 | 74.9 ± 36.4 | 70.1 ± 38.0 | 71.9 ± 35.7 | 82.2 ± 40.4 | 78.7 ± 39.9 | 76.8 ± 33.7 | 93.8 ± 50.0 | 88.3 ± 51.3 | < .001 |
| Fiber, mg | 12.9 ± 6.7 | 13.6 ± 7.1 | 12.4 ± 6.0 | 14.8 ± 7.3 | 17.1 ± 8.6 | 16.5 ± 10.9 | 16.1 ± 8.8 | 18.5 ± 9.2 | 17.9 ± 10.7 | < .001 |
| Dietary moisture, mg/day | 2223 (1703, 3003) |
2234 (1767, 2874) |
1907 (1729, 2606) |
2329 (1795, 3076) |
2816 (2151, 3671) |
2687 (2000, 3614) |
3097 (1889, 4018) |
3094 (2135, 4124) |
2979 (2229, 3729) |
< .001 |
| Hemoglobin, g/L | 13.7 ± 1.3 | 13.9 ± 1.5 | 13.7 ± 1.5 | 14.0 ± 1.4 | 14.5 ± 1.5 | 14.2 ± 1.5 | 14.2 ± 1.9 | 14.8 ± 1.3 | 14.2 ± 1.4 | < .001 |
| Albumin, g/L | 41.9 ± 2.9 | 42.5 ± 3.5 | 41.6 ± 3.9 | 42.3 ± 3.3 | 42.8 ± 3.1 | 42.6 ± 3.2 | 42.9 ± 4.0 | 42.7 ± 3.2 | 41.2 ± 3.8 | < .001 |
| TC, mmol/L | 5.13 ± 1.2 | 4.9 ± 1.1 | 4.8 ± 1.0 | 5.1 ± 1.0 | 5.1 ± 1.1 | 5.2 ± 1.1 | 5.3 ± 1.1 | 5.3 ± 1.0 | 5.0 ± 1.1 | .002 |
| Triglyceride, mmol/L | 1.3 (0.8, 1.9) | 1.1 (0.8, 1.6) | 1.1 (0.7, 2.0) | 1.2 (0.9, 1.7) | 1.2 (0.9, 1.8) | 1.4 (1.0, 1.9) | 1.1 (0.9, 2.1) | 1.4 (1.0, 1.9) | 1.5 (1.0, 2.4) | < .001 |
| HbA1c, % | 5.5 ± 0.8 | 5.5 ± 0.8 | 5.6 ± 0.9 | 5.6 ± 1.0 | 5.5 ± 0.8 | 5.7 ± 0.9 | 5.7 ± 1.3 | 5.6 ± 0.9 | 5.7 ± 1.0 | < .001 |
| CRP, mg/dL | 0.18 (0.07, 0.47) |
0.16 (0.05, 0.38) |
0.21 (0.08, 0.42) |
0.16 (0.07, 0.41) |
0.17 (0.07, 0.38) |
0.19 (0.08, 0.46) |
0.12 (0.07, 0.41) |
0.21 (0.09, 0.43) |
0.32 (0.14, 0.70) |
.001 |
| Serum creatinine, mg/dL | 0.81 (0.70, 0.92) |
0.80 (0.71, 0.97) |
0.77 (0.68, 0.89) |
0.82 (0.71, 0.93) |
0.90 (0.76, 1.00) |
0.85 (0.72, 1.00) |
0.82 (0.71, 0.98) |
0.90 (0.80, 1.02) |
0.90 (0.73, 1.02) |
< .001 |
| Urea nitrogen, mg/dL | 12 (9, 16) | 11 (9, 14) | 11.00 (9, 13) | 12 (10, 15) | 12 (10, 15) | 12 (9, 16) | 15 (10, 17) | 13 (10, 16) | 13 (10, 16) | < .001 |
| eGFR,ml/min/1.73 m2 | 92 (76, 110) | 99 (81, 114) | 94 (83, 110) | 96 (78, 112) | 93 (79, 107) | 92 (77, 106) | 96 (90, 113) | 94 (81, 108) | 92 (75, 107) | .004 |
| uACR, mg/g | 6.2 (4.4, 13.2) | 6.6 (4.2, 11.8) | 7.2 (5.0, 18.8) | 6.3 (4.2, 12.1) | 5.9 (4.0, 10.4) | 6.7 (4.2, 13.1) | 7.7 (4.7, 9.6) | 5.6 (3.8, 9.3) | 7.4 (4.8, 25.5) | .001 |
| CKD, n (%) | 33 (15.8) | 122 (15.0) | 13 (21.5) | 116 (13.7) | 1506 (12.6) | 158 (17.5) | 7 (12.9) | 100 (13.1) | 36 (24.1) | .004 |
BMI = body mass index, CCB = calcium-channel blockers, CKD = chronic kidney disease, CRP = C-reactive protein, CVD = cardiovascular disease, eGFR = estimated glomerular filtration rate, HbA1c = glycosylated hemoglobin A1c, MET = metabolic equivalent, PHQ = patient health questionnaire, TC = total cholesterol, uACR = urine albumin/creatinine ratio.
n (%): unweighted numbers (weighted percentage).
In addition, for the 9 detailed joint subgroups, weighted logistic regression analyses were again performed, and as shown in Figure 4, the results showed that there was a significant association between group 6 (representing the stool trait subgroup-group C and the stool frequency subgroup-group b) and CKD, which predicted that even if the stool frequency was normal, a markedly softer or watery stool type still had a high risk of CKD [OR = 1.51 (1.02–2.22), P = .040]. Moreover, compared to the reference group with normal stools, group 9 had the largest OR [3.90 (1.93–7.86), P = .002], which represents the group with frequent bowel movements and significantly softened stools. The other subgroups did not show any statistical significance.
Figure 4.
Weighted adjusted ORs of combined stool groups 1–9 to CKD, eGFR < 60 ml/min/1.73m2 and uACR ≥ 30 mg/g. OR was adjusted for adjusted for bowel movements + stool types + age + sex + race + BMI + smoking + drinking + cardiovascular disease + hypertension + diabetes + PHQ-9 + metabolic equivalent + drugs + diet + glycosylated hemoglobin A1c + C-reactive protein. BMI = body mass index, CKD = chronic kidney disease, CRP = C-reactive protein, eGFR = estimated glomerular filtration rate, uACR = urine albumin/creatinine ratio, OR = odds ratio, PHQ-9 = patient health questionnaire-9.
3.5. Mediation analyses
We further performed parallel causal mediation analyses to examine whether systemic inflammation, indicated by high‑sensitivity C‑reactive protein (log‑CRP), mediated the association between diarrhea‑like stool types (BSFS types 6/7) and renal outcomes. Using linear regression models for continuous outcomes (eGFR and uACR) and logistic regression for the binary CKD outcome, we observed a consistent mediating role of CRP across all endpoints. The proportion of the total effect mediated by CRP was 10.1% (4.1%–25.6%) for eGFR, 8.5% (3.2%–18.1%) for uACR, and 15.3% (6.5%–32.7%) for CKD (Fig. 5).
Figure 5.
Mediation analysis of systemic inflammation in the association between diarrhea-like stool consistency and CKD risk. (A) For eGFR; (B) For uACR; (C) For overall CKD risk. CKD = chronic kidney disease, eGFR = estimated glomerular filtration rate, uACR = urine albumin-creatinine ratio.
4. Discussion
In this cross-sectional study of NHANES adults, we found that diarrhea-like stool traits were independently associated with an increased risk of CKD, especially among women. Mediation analysis further indicated that systemic inflammation (log_CRP) accounted for approximately 10.1%, 8.5%, and 15.3% of the associations with eGFR decline, elevated uACR, and overall CKD risk, respectively. These associations remained robust after comprehensive adjustment for diet, hydration, PA, depression, medication use, and traditional confounders. As stool consistency is an easily observable clinical sign, it may serve as an early indicator of kidney disease progression, highlighting its potential utility in public health screening and CKD prevention strategies.
In recent years, the communication between CKD and the intestinal environment has gained increasing attention. A healthy gut microbiota, characterized by high abundance and diversity, is essential for systemic homeostasis.[17] Abnormal bowel movements can disrupt this ecosystem, impairing the intestinal barrier (composed of the mucus layer, epithelial cells, tight junctions, immune cells, and microbiota) and promoting chronic inflammation. This breach enhances intestinal permeability, allowing inflammatory mediators, immune cells, and uremic toxins to enter systemic circulation, thereby accelerating CKD progression. Conversely, CKD-associated inflammation and immune dysregulation further perturb gut flora and barrier integrity, creating a vicious cycle. Notably, irritable bowel syndrome, a classic brain‑gut‑microbiome axis disorder, is more prevalent in CKD patients.[18] In summary, intestinal dysbiosis, barrier dysfunction, and sustained inflammation form key mechanistic pathways linking abnormal defecation to CKD progression.[19–22]
Recently, a study based on macro-genomics and untargeted metabolomics of gut microorganisms confirmed that predictive models using microbial species, fecal metabolites, and their combinations showed significantly higher efficacy in predicting mild CKD than traditional indicators such as blood creatinine, suggesting that the intestinal microenvironment of patients with CKD may be altered at an earlier age than serum toxin levels.[23] Therefore, more attention should be paid to abnormal changes in defecation in public health, especially in clinical practice related to CKD.
In the past, experts have characterized the healthy colonic microbiota, often using the number of bowel movements, especially some definitions of constipation or chronic diarrhea. However, a newer study on the relationship between stool traits and colon microbial composition and metabolic capacity observed that intestinal species richness declined with increasing BSFS type scores, reaching a minimum in dilute-pooped individuals, suggesting that stool consistency is a highly relevant signal of intestinal microenvironmental dysbiosis. It has been noted that fecal passage time is a key determinant of intestinal microbial habitat, influencing nutrient and water uptake in the gut and determining the rate at which luminal microbial populations are cleared during defecation. Fecal consistency as classified by the BSFS score serves as a strong proxy for colonic transit rate, integrating the effects of 2 major and correlated factors of the intestinal ecosystem: intestinal transit rate and water activity.[24] These findings may support our results on stool type and high risk of CKD.
Accumulating epidemiologic evidence indicates that constipation is independently associated with adverse renal outcomes. A large cohort of U.S. veterans with preserved kidney function demonstrated that constipation was linked to a higher risk of end-stage renal disease and eGFR decline over a median 7-year follow-up.[6] Another analysis of the same NHANES population also reported an association between reduced bowel frequency and poor CKD prognosis, though that study adjusted for fewer confounders and did not account for key dietary, inflammatory, and PA variables.[25] In our study, constipation defined by bowel frequency (group a) showed a significant association with CKD in crude and partially adjusted models, but this association was attenuated and lost statistical significance in the fully adjusted model 3. This attenuation likely reflects the inclusion in model 3 of several modifiable factors (such as dietary fiber, water intake, PA, and depressive symptoms) which are known correlates of constipation and may also exert protective effects against CKD progression.[26,27] On the other hand, baseline data showed that diarrhea-related stool groups consistently exhibited higher CRP levels compared to normal or constipation-related groups, aligning with the known inflammatory phenotype of CKD. Our mediation analysis further supports this link, indicating that systemic inflammation (CRP) mediates approximately 8.5–15.3% of the association between diarrhea-like stools and renal impairment. Therefore, diarrhea-related bowel abnormalities may be particularly relevant to CKD, possibly through inflammatory pathways. These findings highlight the need for further mechanistic and longitudinal studies to clarify the role of stool consistency and inflammation in CKD progression.
Several limitations should be considered when interpreting our findings. Firstly, the NHANES survey provides only a single measurement of renal function/albuminuria and non-repeated survey. Although our definition of CKD aligns with Kidney Disease: Improving Global Outcomes recommendations for epidemiological studies, we could not confirm persistence of abnormalities for ≥ 3 months, which is required for clinical diagnosis. This may result in some misclassification, such as the inclusion of individuals with acute kidney injury or transient albuminuria. However, this approach is consistent with other large-scale public health surveys and burden estimates (e.g., the Global Burden of Disease Study), enabling comparative assessments of CKD prevalence and risk associations at the population level. Additionally, the cross-sectional analysis could not establish a causal relationship between stool abnormalities and CKD. Furthermore, there is a known wide variability and ambiguity in the definition of constipation among the general public.[28] The information about feces in this study was based on subjective recall, which may create some bias when used as a diagnosis of constipation or diarrhea. More objective measurements, such as photographic documentation of stools or quantification of water content, would improve the accuracy of the measurements. Finally, beyond the prescription medications adjusted for in this study, future longitudinal investigations that incorporate data on the persistence and treatment response of bowel symptoms, along with primary kidney disease etiology, detailed dietary and sleep patterns, genetic factors, and specific diarrhea‑inducing drugs (which were unavailable in NHANES), would better elucidate the nature and direction of this association. In summary, further research should explore potential mechanisms linking stool consistency, gut microbiota status, and CKD development. The ability to readily assess and modify stool form through public health interventions merits additional investigation.
5. Conclusion
Overall, the study found that stool traits, but not bowel movements, were associated with CKD. Specifically, self-reported diarrhea-like stool traits (BSFS type 6/7) were significantly linked to a heightened risk of CKD, eGFR < 60 ml/min/1.73m2, and uACR ≥ 30 mg/g, partly mediated by systemic inflammation. These findings suggest a potential opportunity to identify or control CKD progression. By looking dynamically at changes in defecation and linking it to impaired renal function, causality can be better determined.
Acknowledgments
The authors thank the NHANES participants and staff for their valuable contributions to data collection and for their efforts to make the data publicly available.
Author contributions
Conceptualization: Dao Wang, Fengping Zhang.
Data curation: Dao Wang.
Formal analysis: Dao Wang.
Writing – original draft: Fengping Zhang.
Writing – review & editing: Fengping Zhang.
Abbreviations:
- BMI
- body mass index
- BSFS
- Bristol Stool Form Scale
- CI
- confidence intervals
- CKD
- chronic kidney disease
- CRP
- C-reactive protein
- CVD
- cardiovascular disease
- eGFR
- estimated glomerular filtration rate
- HbA1c
- glycosylated hemoglobin A1c
- MET
- metabolic equivalent estimate
- NHANES
- National Health and Nutrition Examination Survey
- OR
- odds ratio
- PA
- physical activity
- PHQ-9
- Patient Health Questionnaire-9
- uACR
- urine albumin-to-creatinine ratio
- U.S.
- United States
The authors’ organization did not conduct an ethical review of this study, as the data were obtained from publicly available databases. All participants provided written informed consent at the time of the survey. All procedures were conducted in accordance with relevant NHANSE guidelines and regulations.
The authors have no funding and conflicts of interest to disclose.
The datasets generated during and/or analyzed during the current study are publicly available.
Supplemental Digital Content is available in the online version of this article (http://dx.doi.org/10.1097/MD.0000000000048888).
How to cite this article: Wang D, Zhang F. Associations of bowel movements and stool types with risk of chronic kidney disease. Medicine 2026;105:21(e48888).
References
- [1].Chen TK, Knicely DH, Grams ME. Chronic kidney disease diagnosis and management: a review. JAMA. 2019;322:1294–304. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [2].Fan Y, Pedersen O. Gut microbiota in human metabolic health and disease. Nat Rev Microbiol. 2021;19:55–71. [DOI] [PubMed] [Google Scholar]
- [3].de Vos WM, Tilg H, Van Hul M, Cani PD. Gut microbiome and health: mechanistic insights. Gut. 2022;71:1020–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Huang G, Khan R, Zheng Y, Lee P-C, Li Q, Khan I. Exploring the role of gut microbiota in advancing personalized medicine. Front Microbiol. 2023;14:1274925. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [5].Müller DN, Wilck N, Haase S, Kleinewietfeld M, Linker RA. Sodium in the microenvironment regulates immune responses and tissue homeostasis. Nat Rev Immunol. 2019;19:243–54. [DOI] [PubMed] [Google Scholar]
- [6].Sumida K, Yamagata K, Kovesdy CP. Constipation in CKD. Kidney Int Rep. 2019;5:121–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [7].Sumida K, Molnar MZ, Potukuchi PK, et al. Constipation and Incident CKD. J Am Soc Nephrol. 2017;28:1248–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Vandeputte D, Falony G, Vieira-Silva S, Tito RY, Joossens M, Raes J. Stool consistency is strongly associated with gut microbiota richness and composition, enterotypes and bacterial growth rates. Gut. 2016;65:57–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Lewis SJ, Heaton KW. Stool form scale as a useful guide to intestinal transit time. Scand J Gastroenterol. 1997;32:920–4. [DOI] [PubMed] [Google Scholar]
- [10].Wang M, Huang ZH, Zhu YH, He P, Fan Q-L. Association between the composite dietary antioxidant index and chronic kidney disease: evidence from NHANES 2011-2018. Food Funct. 2023;14:9279–86. [DOI] [PubMed] [Google Scholar]
- [11].Blake MR, Raker JM, Whelan K. Validity and reliability of the Bristol Stool Form Scale in healthy adults and patients with diarrhoea-predominant irritable bowel syndrome. Aliment Pharmacol Ther. 2016;44:693–703. [DOI] [PubMed] [Google Scholar]
- [12].Nordin E, Hellström PM, Brunius C, Landberg R. Modest conformity between self-reporting of bristol stool form and fecal consistency measured by stool water content in irritable bowel syndrome and a FODMAP and gluten trial. Am J Gastroenterol. 2022;117:1668–74. [DOI] [PubMed] [Google Scholar]
- [13].Kidney Disease: Improving Global Outcomes (KDIGO) Glomerular Diseases Work Group. KDIGO 2021 clinical practice guideline for the management of glomerular diseases. Kidney Int. 2021;100:S1–S276. [DOI] [PubMed] [Google Scholar]
- [14].Diao X, Ling Y, Zeng Y, et al. Physical activity and cancer risk: a dose-response analysis for the Global Burden of Disease Study 2019. Cancer Commun (Lond). 2023;43:1229–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [15].Zhang Z, Jackson SL, Gillespie C, Merritt R, Yang Q. Depressive symptoms and mortality among US adults. JAMA Netw Open. 2023;6:e2337011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [16].Qiu Z, Chen X, Geng T, et al. Associations of serum carotenoids with risk of cardiovascular mortality among individuals with type 2 diabetes: results from NHANES. Diabetes Care. 2022;45:1453–61. [DOI] [PubMed] [Google Scholar]
- [17].Wang X, Yang S, Li S, et al. Aberrant gut microbiota alters host metabolome and impacts renal failure in humans and rodents. Gut. 2020;69:2131–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [18].Lozupone CA, Stombaugh JI, Gordon JI, Jansson JK, Knight R. Diversity, stability and resilience of the human gut microbiota. Nature. 2012;489:220–30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Özkul D, Güney I, Saçkan F, Coşkun Yavuz Y, Selcuk NY, Tonbul HZ. Irritable bowel syndrome frequency and related factors in hemodialysis patients. Hemodial Int. 2020;24:359–66. [DOI] [PubMed] [Google Scholar]
- [20].Cano AE, Neil AK, Kang JY, et al. Gastrointestinal symptoms in patients with end-stage renal disease undergoing treatment by hemodialysis or peritoneal dialysis. Am J Gastroenterol. 2007;102:1990–7. [DOI] [PubMed] [Google Scholar]
- [21].Ng QX, Yau CE, Yaow CYL, et al. What has longitudinal “Omics” studies taught us about irritable bowel syndrome? A systematic review. Metabolites. 2023;13:484. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [22].Wu IW, Lin C-Y, Chang L-C, et al. Gut microbiota as diagnostic tools for mirroring disease progression and circulating nephrotoxin levels in chronic kidney disease: discovery and validation study. Int J Biol Sci. 2020;16:420–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [23].Wang H, Ainiwaer A, Song Y, et al. Perturbed gut microbiome and fecal and serum metabolomes are associated with chronic kidney disease severity. Microbiome. 2023;11:3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [24].Saad RJ, Rao SSC, Koch KL, et al. Do stool form and frequency correlate with whole-gut and colonic transit? Results from a multicenter study in constipated individuals and healthy controls. Am J Gastroenterol. 2010;105:403–11. [DOI] [PubMed] [Google Scholar]
- [25].He M, Ding G, Yang Y, Zhong J. Bowel habits were associated with mortality in chronic kidney disease: results from a nationwide prospective cohort study. Ren Fail. 2023;45:2292150. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [26].Wang Y, Shen X, Wang P. Constipation is associated with depression of any severity, but not with suicidal ideation: insights from a large cross-sectional study. Int J Colorectal Dis. 2023;38:231. [DOI] [PubMed] [Google Scholar]
- [27].Wang J, Kong W, Liu M, Wang Y, Zheng Y, Zhou Y. Association between dietary carotenoids intake and chronic constipation in American men and women adults: a cross-sectional study. BMC Public Health. 2023;23:1597. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [28].Soh AYS, Kang JY, Siah KTH, Scarpignato C, Gwee K-A. Searching for a definition for pharmacologically refractory constipation: a systematic review. J Gastroenterol Hepatol. 2018;33:564–75. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.