A Phenome-Wide Comorbidity Atlas of Age-Related Hearing Loss, Speech-in-Noise Deficits, and Tinnitus: Distinguishing Causal Signals from Correlation.

Purpose: The present study employed a data-driven and hypothesis-free approach to identify comorbidities associated with age-related hearing loss (ARHL), speech-in-noise (SIN) deficits, and tinnitus.

Methods: The study performed phenome-wide co-occurrence association analyses using the UK Biobank cohort to identify comorbidities associated with ARHL (N = 429,318), SIN deficits (N = 437,155), tinnitus (N = 172,527), and tinnitus severity (N = 57,657). Medical health records were accessed to obtain ICD-10 codes, which were converted into phecodes reflecting a modern disease classification. The statistical analysis was conducted to identify comorbidities associated with ARHL, SIN deficits, tinnitus, and tinnitus severity while statistically controlling for age, sex, ethnicity, and genetic ethnicity. Phenotype risk scores (PheRS) for hearing conditions were calculated. A complementary phenome-wide genetic correlation analysis was conducted to identify genetic comorbidities associated with these conditions. We utilized the summary statistics of recent genome-wide association studies of ARHL (N = 723,266), SIN deficits (N = 443,482), tinnitus (N = 132,438), and tinnitus severity (N = 132,438). The results of the phenome-wide association analyses were subjected to enrichment analysis to identify trait categories involved in hearing conditions. A complementary phenome-wide latent causal variant (LCV) analysis was employed to obtain causal inference by distinguishing between horizontal pleiotropy and true causality.

Results: The phenome-wide co-occurrence association analysis identified 383, 449, 283, and 216 medical conditions associated (FDR p < 0.05) with ARHL, SIN deficits, tinnitus, and tinnitus severity, respectively. Gastrointestinal conditions revealed significant enrichment across all traits. Respiratory, genitourinary, and sense organs showed significant enrichment with ARHL, SIN deficits, and tinnitus. SIN deficits and tinnitus severity showed significant enrichment with mental Health and neurological conditions. Elevated PheRS significantly increased the risk of expressing their respective phenotypes. A phenome-wide genetic correlation analysis identified 376, 254, 97, and 188 health traits associated with ARHL, SIN deficits, tinnitus, and tinnitus severity, respectively. Mental health and medical symptoms were significantly enriched for all hearing conditions in the genetic correlation analyses. The results of LCV analyses revealed widespread horizontal pleiotropy driving most genetic correlations. In contrast, only a few traits demonstrated a true causal relationship.

Conclusion: This study mapped phenotypic and genotypic comorbidity profiles of ARHL, SIN deficits, tinnitus, and tinnitus severity. We observed a robust enrichment of gastrointestinal traits with all hearing conditions, suggesting a potential role of gut dysbiosis in their pathogenesis. The associations between mental health and hearing conditions suggest a complex interplay between auditory and psychological health. Genetic analyses provided compelling evidence that most comorbidities are driven by a shared genetic architecture, rather than true causality.