To assess whether rosacea is associated with subsequent irritable bowel syndrome (IBS), and whether IBS is associated with subsequent rosacea.
Approach:
Design and data: Retrospective population-based cohort study using the Korean National Health Insurance Service–National Sample Cohort, covering 2002–2013 and adults aged 20 years or older.
Comparison groups: Compared 1,085 patients with rosacea with 4,340 matched controls, and 244,156 patients with IBS with 244,156 matched controls. Matching was by age, sex, and index year.
Analysis: Cox proportional hazards models adjusted for sociodemographic factors and comorbidities. Sensitivity analyses assessed medication exposure, including isotretinoin, doxycycline, and minocycline.
Key Findings:
Over a median 6.25 years, IBS incidence was 56.3 per 1,000 person-years among patients with rosacea and 36.9 among controls; adjusted hazard was 30% higher with rosacea.
Over a median 6.24 years, rosacea incidence was 0.37 per 1,000 person-years among patients with IBS and 0.21 among controls; adjusted hazard was 61% higher with IBS. Absolute rosacea incidence remained low.
The rosacea–IBS association was statistically significant among men but not women. In the reverse analysis, adjusted rosacea hazard was approximately twice as high among men with IBS and 41% higher among women with IBS.
Associations persisted in medication-adjusted sensitivity analyses, although estimates were somewhat attenuated.
Interpretation:
The study found a reciprocal association between rosacea and IBS. The authors proposed shared microbial, inflammatory, and neuroimmune pathways involving the gut–skin axis, but the study did not directly assess these mechanisms and cannot establish causality.
Limitations:
Administrative diagnostic codes provided no information on disease severity.
The database lacked information on diet, stress, and microbiome composition, and analyses did not distinguish IBS subtypes.
More frequent health care encounters among affected patients could increase the likelihood of additional diagnoses.
Excluding patients diagnosed during the first study year may have introduced selection bias.
The observational design cannot establish causal or treatment-related effects.
Conclusion:
The authors called for further mechanistic and longitudinal research and stated that clinicians managing either condition should be aware of the potential coexistence of the other.
