Abstract
Objective To explore the causes of and contributors to gastrointestinal (GI) dysfunction in systemic sclerosis (SSc) in a phenome-wide association study (PheWAS), using real-world clinical records data.
Methods Twelve thousand five hundred thirty-five documented clinical assessments of 2058 consenting individuals with SSc at the Royal Free Hospital (UK) were available for detailed phenotyping. Diagnoses and drugs were mapped to structured dictionaries of terms (Disease Ontology project and DrugBank Open Data, respectively). A PheWAS model was used to explore links between 6 important SSc-GI domains (constipation, diarrhea, dysmotility, incontinence, gastroesophageal reflux, and small intestinal bacterial overgrowth [SIBO]) and exposure to various comorbidities and drugs. "Hits" from the PheWAS model were confirmed and explored in a subcohort reporting quantitative GI symptom scores from the University of California Los Angeles Scleroderma Clinical Trials Consortium Gastrointestinal Tract Instrument 2.0 (GIT 2.0) questionnaire.
Results One thousand five hundred forty-six individuals were entered into the PheWAS analysis. Six hundred seventy-three distinct diagnoses and 634 distinct drugs were identified in the dataset, as well as SSc-specific phenotypes such as antinuclear antibodies (ANA). PheWAS analysis revealed associations between drugs, diagnoses, and ANAs with 6 important SSc-GI outcomes: constipation, diarrhea, dysmotility, incontinence, reflux, and SIBO. Subsequently, using GIT 2.0 symptom scores links with SSc-GI were confirmed for 22 drugs, 4 diagnoses, and 3 ANAs.
Conclusion Using a hypothesis-free PheWAS approach, we replicated known, and revealed potential novel, risk factors for SSc-GI dysfunction, including drug classes such as opioid, antimuscarinic, and endothelin receptor antagonist, and ANA subgroup.
