Shared genetic liability between opioid analgesic use and lung cancer subtypes revealed by genome wide correlation...

Observational studies link opioid use to lung cancer risk, but findings are inconsistent due to potential confounding and reverse causality. Whether these associations reflect shared genetic liability with histologic lung cancer (LC) subtypes is unknown. This study quantified genome-wide genetic overlap and modeled their latent shared architecture.

This analysis used European-ancestry genome-wide association study (GWAS) summary statistics for opioid use traits (codeine/tramadol and dihydrocodeine) and lung cancer outcomes (overall lung cancer, non-small cell lung cancer, adenocarcinoma, and squamous cell carcinoma). Bivariate linkage disequilibrium score regression (LDSC) estimated heritability and genetic correlations. Genomic structural equation modeling (Genomic SEM) tested latent factor models, and multi-marker analysis of genomic annotation (MAGMA) performed gene, pathway, and tissue enrichment analyses. An exploratory Mendelian randomization (MR) analysis was additionally conducted when adequate instrumental variants were available.

LDSC indicated uniformly positive genetic correlations between opioid traits and lung cancer outcomes, strongest for CT-NSCLC (rg = 1.1017, p = 6.86 × 10^- 4) and DHC-NSCLC (rg = 0.9773, p = 4.46 × 10^- 2); other positive pairs included DHC-LC (rg = 0.5938, p = 1.59 × 10^- 5) and DHC-SCC-L (rg = 0.6366, p = 9.03 × 10^- 4), with remaining correlations smaller but positive (CT-LC rg = 0.2702; CT-LAC rg = 0.2055; CT-SCC-L rg = 0.3358; DHC-LAC rg = 0.3377). Genomic SEM supported a two-factor model (CFI > 0.99; SRMR = 0.0602) separating cancer outcomes (LAC β = 0.64, LC β = 1.10, SCC-L β = 0.83) from opioid traits (DHC β = 1.22; CT β = 0.61). MAGMA identified enrichment for oxidative phosphorylation/mitochondrial electron transport chain, Notch, cell-cycle, DNA damage response-p53/TP53, and immune pathways. Exploratory MR was feasible only for CT under the prespecified instrument-selection criteria, whereas DHC did not yield sufficient instruments and therefore could not be evaluated by MR. The CT-based MR analysis did not provide robust evidence for a causal effect of codeine/tramadol use liability on lung cancer outcomes, indicating that the observed LDSC associations are more appropriately interpreted as shared genetic liability rather than confirmed causality.

Common-variant liability is broadly shared between opioid medication use and lung cancer, particularly CT-NSCLC, with a correlated two-factor structure separating cancer susceptibility from medication use. Enrichment analyses highlighted mitochondrial energetics, DNA damage response/TP53, immune signaling, and subtype-specific pathways. Exploratory MR was feasible only for CT and did not support a definitive causal interpretation, reinforcing the need to view the findings primarily as evidence of shared genetic liability.