Functional impairment in COPD can be predicted using genomic-derived data.

Abstract

Objective: Reduced functional capacity and muscle weakness are two major contributors to functional impairment in chronic obstructive pulmonary disease (COPD). The underlying causes of functional impairment are poorly understood and, therefore, we sought to investigate the contribution of genetic factors.

Methods: We conducted a cross-sectional analysis of sociodemographic, clinical and genetic information of people with COPD. Hierarchical clustering based on functional capacity (6-minute walk test and 1-minute sit-to-stand test) and muscle strength (quadriceps isometric muscle strength and handgrip muscle strength) was performed. A genome-wide association study (GWAS) was performed using cluster assignment as phenotype. Polygenic risk scores (PRSs) were calculated for each variable. Genomic-derived data was used to construct a model to predict functional impairment.

Results: Two clusters were identified among 245 individuals. Cluster 1 (n=104) was composed of younger, less symptomatic patients, with preserved functional capacity and muscle strength, whereas cluster 2 (n=141) included those older, more symptomatic, with reduced functional capacity and muscle weakness. GWAS identified two polymorphisms suggestively associated with functional impairment, mapped to xanthine dehydrogenase. Cluster 2 was enriched in individuals with risk alleles for rs1991541 and rs10524730, and lower PRSs for functional capacity and muscle strength. A prediction model using genomic-derived data was constructed (n=159) and tested (n=37), yielding an area under the curve of 0.87 (0.76-0.99).

Conclusion: Genetic factors are significantly associated with functional impairment in COPD. The incorporation of genetic information, particularly PRSs, into a predictive model offers a promising avenue for timely identifying individuals at greater risk of functional decline, potentially facilitating personalised and preventive interventions. Further studies on independent external cohorts are needed to validate our model.