Pro-Inflammatory Diets Linked to Bowel Control Problems, But Not to New Cases

A large new analysis spanning three continents has found that people whose diets score higher on a measure of inflammatory potential are more likely to report existing bowel control problems, including fecal incontinence. But in a twist that complicates the story, the same dietary scores did not predict who would go on to develop these symptoms in the future. The findings, published in Food Science & Nutrition, suggest that the relationship between diet and continence may be more tangled than a simple cause-and-effect narrative allows, and that reverse causation—people changing what they eat because of their symptoms—may explain part of the association.

The research team, led by Mengmeng Jiang and Honggang Wang of Huai’an First People’s Hospital, drew on three major population studies: the United States National Health and Nutrition Examination Survey (NHANES), the English Longitudinal Study of Ageing (ELSA), and the Chinese Longitudinal Healthy Longevity Survey (CLHLS). Each dataset played a distinct role. NHANES provided the primary cross-sectional evidence, ELSA offered a partially harmonized comparison in a different country, and CLHLS extended the question to older Chinese adults with a food-based score rather than a nutrient-based index. The investigators followed the STROBE reporting guidelines for observational studies and used survey-weighted statistical models designed to account for the complex sampling schemes behind each survey.

At the heart of the study is the Dietary Inflammatory Index, or DII, a literature-derived score that summarizes the inflammatory potential of a person’s diet. Each of 25 dietary components—ranging from fiber, magnesium, and omega-type fats on the anti-inflammatory side to saturated fat, cholesterol, and energy intake on the pro-inflammatory side—is standardized against global reference values, mapped through a normal distribution to a centered percentile, and multiplied by a published inflammatory effect score. The component scores are then summed, so higher values represent a more pro-inflammatory dietary pattern. In the NHANES analysis, the researchers reconstructed the index from one or two 24-hour dietary recalls collected between 2005 and 2010, the only cycles containing the relevant bowel-health questions.

The NHANES results were striking in their consistency. Among 13,385 adults with complete data, 1,213 reported at least monthly leakage of mucus, liquid stool, or solid stool during the previous 30 days—a weighted prevalence of 8.45 percent. Each one-unit increase in the DII was associated with roughly a 7 percent higher odds of leakage in the core-adjusted model, an odds ratio of 1.075. When the researchers additionally adjusted for body mass index, diabetes, and depressive symptoms, the estimate softened slightly to 1.056 but remained statistically significant. People in the highest quartile of dietary inflammatory potential had about 1.5 times the odds of leakage compared with those in the lowest quartile. Restricted cubic spline analyses found no evidence that the relationship was nonlinear; the dose-response curve appeared essentially linear across the observed range of scores.

Perhaps more impressive was the temporal reproducibility. Splitting the NHANES data into three survey periods produced nearly identical estimates—1.081, 1.071, and 1.071 per unit—with no statistical heterogeneity between periods, and leave-one-period-out analyses ranged only from 1.072 to 1.077. Sensitivity analyses using two-day average recalls, restricted plausible energy intakes, and an alternative 24-component score all preserved the association. Of the three leakage types, only liquid-stool leakage survived correction for multiple comparisons, a detail the authors say motivates attention to stool properties such as water content and consistency, though they caution it does not establish a mechanism.

To test whether the finding traveled beyond American shores, the team built a common 15-component DII using nutrients available in both NHANES and ELSA, where Wave 9 intake came from the online Oxford WebQ questionnaire. The concurrent association was remarkably similar in both cohorts: an odds ratio of 1.128 in NHANES and 1.136 in ELSA for self-reported bowel-control problems during the preceding 12 months. Among ELSA participants whose problems had persisted for more than a month, the estimate was even stronger at 1.183. Yet the authors are careful to label this only a partially harmonized comparison. The instruments, recall windows, and outcome definitions differ, and an uncertain mapping of ELSA carotene to the beta-carotene scoring parameter adds another layer of caution. Weighting for nutrition-substudy participation also attenuated the ELSA estimate to 1.088, with a confidence interval that just included the null.

The prospective analyses are where the story turns. Among ELSA participants free of bowel-control problems at baseline, the dietary score showed no clear association with new problems reported at Wave 10, Wave 11, or either follow-up, with cumulative incident odds ratios hovering near 1. The CLHLS extension told a similar story. At baseline in 2008, among more than 16,000 adults aged 65 and older, a standardized 13-food directional score—built from food-frequency items in which infrequent consumption of fruits, vegetables, fish, eggs, legumes, tea, garlic, dairy, nuts, and mushrooms or algae pushed the score upward—was associated with combined urinary and bowel continence impairment, with an odds ratio of 1.44 per standard deviation. But adjusting for cognitive function attenuated the estimate, and baseline scores did not predict incident impairment across the 2011, 2014, and 2018 waves.

The authors argue that reverse causation is a plausible explanation for the concurrent findings. Qualitative research has shown that older women with fecal incontinence actively avoid certain foods, change preparation methods, and modify meal size and frequency to manage their symptoms. Avoiding nutrient-rich foods could raise an inflammatory score, while cutting out fatty foods or alcohol could lower it—meaning diet recorded after symptom onset may partly reflect symptom management rather than a cause of symptoms. Alternative biological pathways also deserve attention. A randomized trial found that psyllium, a gel-forming fiber with strong water-holding capacity, reduced leakage episodes in adults with loose stools, while carboxymethylcellulose increased them. Fermentable carbohydrates and individual food tolerances can alter luminal water, gas production, and transit—processes that are distinct from the systemic inflammatory construct the DII represents.

The study has real strengths, including complex-survey analysis, period-specific reproducibility checks, and prospective follow-up in two additional countries, but the limitations are equally instructive. The CLHLS food score has not been validated as a measure of inflammation, its outcome combined urinary and bowel control, and the survey oversamples the oldest-old, introducing survival and response selection. ELSA lacked nutrition-specific calibration weights, and neither weighting nor imputation can remove unmeasured confounding or exposure error. The authors conclude that higher dietary inflammatory potential is consistently associated with existing continence-related symptoms across settings, but that baseline scores showed no clear association with incident outcomes—findings that extend the cross-sectional evidence while limiting its interpretation as proof of prevention. Whether deliberately lowering an inflammatory diet score can improve or prevent continence problems remains a question for intervention studies, not observational data.

Subject of Research: Association between dietary inflammatory potential and continence-related symptoms across three population studies

Article Title: Dietary Inflammatory Potential and Continence‐Related Symptoms: Cross‐Sectional and Longitudinal Analyses of Three Population Studies

Article References: Jiang, M., Zhong, M., Yu, J., Zhang, J., & Wang, H. (2026). Dietary Inflammatory Potential and Continence‐Related Symptoms: Cross‐Sectional and Longitudinal Analyses of Three Population Studies. Food Science & Nutrition, 14(10), Article e72477. https://doi.org/10.1002/fsn3.72477

Image Credits: AI Generated

DOI: 10.1002/fsn3.72477

Keywords: dietary inflammatory index, fecal incontinence, NHANES, ELSA, CLHLS, nutritional epidemiology, dietary fiber, bowel control, reverse causation, cross-sectional analysis, longitudinal study, inflammation