Association between psychiatric disorders and constipation: population-based cross-sectional study (NHANES 2009-2010)
Participants using prescribed antipsychotics had a constipation prevalence nearly twice that of those without psychiatric disorder (9.6% vs 4.9%), with a positive association persisting after multivariable adjustment.
| Population | Adults ≥20 years from the NHANES 2009-2010 sample (n=5,092), representative of the US non-institutionalized civilian population |
|---|---|
| Exposure | Presence of psychiatric disorder, operationally defined as use of a prescribed antipsychotic medication in the prior 30 days |
| Comparator | Absence of psychiatric disorder (no prescribed antipsychotic) |
| Outcome | Constipation prevalence (PD vs non-PD group); PD-constipation association in crude model (Model 1); PD-constipation association adjusted for age, sex, race (Model 2); Fully adjusted PD-constipation association (Model 3); Subgroup analyses consistency (effect by sex, age, race) |
Summary of findings
| Outcome | Effect | 95% CI | Certainty | Clinical relevance | Notes |
|---|---|---|---|---|---|
| Constipation prevalence (PD vs non-PD group) | OR 1.84 (95% CI not explicitly reported); crude prevalence 9.6% vs 4.9%, absolute difference 4.7pp, p<0.001 | — | Low | — | 1 studies |
| PD-constipation association in crude model (Model 1) | OR ~2.07 (derived from 9.6% vs 4.9%; 95% CI not reported) | — | Low | — | 1 studies |
| PD-constipation association adjusted for age, sex, race (Model 2) | OR not explicitly transcribed in provided text; 95% CI not reported | — | Low | — | 1 studies |
| Fully adjusted PD-constipation association (Model 3) | OR 1.84 (95% CI not explicitly reported in provided text) | — | Low | — | 1 studies |
| Subgroup analyses consistency (effect by sex, age, race) | Direction consistently positive across subgroups; individual OR and 95% CI not available in provided text | — | Low | — | 1 studies |
Context
Constipation affects approximately 15% of the global population and accounts for 8 million annual medical visits in the US. Psychiatric disorders share pathophysiological mechanisms with constipation, including neurotransmitter dysregulation and gut microbiota alterations. Robust quantification of this association in representative samples is lacking.
What the study showed
Crude constipation prevalence was 9.6% in the psychiatric disorder group vs 4.9% in controls (absolute difference 4.7 percentage points; p<0.001). In the fully adjusted multivariable model (Model 3), psychiatric disorder was associated with higher odds of constipation (OR=1.84; 95% CI not explicitly reported in the provided text). The direction of effect consistently favors higher constipation risk in the psychiatric disorder group across all models. Subgroup analyses maintained the positive direction, but disaggregated 95% CIs were not provided in the available text.
How it was done
Population-based cross-sectional study using NHANES 2009-2010 cycle data. From 20,905 original participants, 5,092 were included after exclusions for age (<20 years), incomplete prescription medication questionnaire data (n=10,376), incomplete BHQ data (n=5,301), and missing covariates (n=136). Multivariable logistic regression with three progressive adjustment models was applied with complex survey weights (WTMEC2YR).
Effect magnitude
Crude OR approximately 2.07 (derived from prevalences of 9.6% vs 4.9%); adjusted OR reported as 1.84 in Model 3, without an explicitly transcribed 95% CI in the provided text — a critical limitation for assessing effect precision.
Risk of bias
Cross-sectional design precludes establishing causality or temporality. Psychiatric disorder was defined solely by antipsychotic prescription use, conflating disease effect with medication effect (confounding by indication) and excluding unmedicated patients. Constipation was self-reported without clinical or manometric confirmation. Authors did not apply a formal risk of bias tool (ROBINS-I). Exclusion of over 15,000 participants due to incomplete data introduces substantial selection bias risk.
What this study does NOT prove
This study does not prove causality between psychiatric disorder and constipation; the direction of the relationship may be reversed or mediated by unmeasured confounders. Results are not generalizable to non-US populations or to institutionalized or unmedicated patients.
In clinical practice
Clinicians managing patients on antipsychotics should actively screen for constipation, given that prevalence in this population is approximately double that of the general population. The operational definition used (antipsychotic use = psychiatric disorder) does not allow distinction between disease-attributable and drug-attributable risk — a clinically relevant distinction for therapeutic decision-making.
Limitations
Cross-sectional design precludes establishing causality or temporality. Psychiatric disorder was defined solely by antipsychotic prescription use, conflating disease effect with medication effect (confounding by indication) and excluding unmedicated patients. Constipation was self-reported without clinical or manometric confirmation. Authors did not apply a formal risk of bias tool (ROBINS-I). Exclusion of over 15,000 participants due to incomplete data introduces substantial selection bias risk.
What is still missing
Prospective studies separating the effect of the psychiatric disorder itself from the specific effect of each antipsychotic class on colonic transit and constipation are needed. Clinical trials evaluating constipation interventions in this specific population are also absent.
Technical appendix
Version history
- 1.0 · 2026-11-13 — Auto-generated under Evidence Standard v1.0