Genes May Not Predict COPD Inhaler Response, Systematic Review Finds

People with chronic obstructive pulmonary disease (COPD) respond very differently to the inhaled medicines that form the backbone of their treatment, and a new systematic review concludes that, so far, genetics cannot explain those differences well enough to guide therapy. The review, published in Frontiers in Pharmacology, synthesized eight studies covering seven independent populations and found that while several genetic variants were linked to lung function, disease severity, or exacerbation risk, the evidence that any of them predict how a patient responds to long-acting bronchodilators or inhaled corticosteroids remains weak, inconsistent, and largely unreplicated.

COPD is a heterogeneous lung disease characterized by persistent airflow limitation, in which patients differ in symptoms, exacerbation frequency, and clinical course. Its mainstay pharmacological treatment for stable disease relies on two classes of long-acting bronchodilators: long-acting beta2-agonists (LABAs), which stimulate beta2-adrenergic receptors on airway smooth muscle, and long-acting muscarinic antagonists (LAMAs), which block the bronchoconstricting effects of acetylcholine on M3 receptors. Inhaled corticosteroids (ICS), usually added as triple therapy, are targeted at patients with an exacerbation-prone phenotype and elevated blood eosinophil counts. Despite these options, many patients remain symptomatic, which is why researchers have long hoped that pharmacogenetics could identify who will benefit from which drug.

The review team searched PubMed and Scopus along with gray literature sources for studies published between January 2015 and May 2026, following PRISMA 2020 guidance and registering the protocol prospectively in PROSPERO. From 51 identified records, eight studies met the eligibility criteria: adults with COPD receiving LABA and/or LAMA treatment, including combination regimens, with genetic variants evaluated in relation to therapeutic response or clinical outcomes. Study selection and data extraction were performed by two independent reviewers, and risk of bias was assessed with the Newcastle-Ottawa Scale for observational studies and the Cochrane Risk of Bias 2 tool for pharmacogenetic analyses derived from randomized trials. Because of substantial methodological, clinical, and statistical heterogeneity, no meta-analysis was attempted; the findings were synthesized narratively.

The included studies ranged in size from 100 to 8,439 patients and investigated variants in genes including ADRB2, FAM13A, SUMF1, MAPK8, AZU1, ARFGEF3, CSE1L, STAU1, and CCL2. Four publications reported pharmacogenetic analyses from randomized controlled trials, two prospective studies came from the same Ukrainian cohort of 100 patients treated with tiotropium/olodaterol, one was an observational cohort from the Rotterdam Study, and one was a North Indian case-control study. The review applied a strict framework, counting as pharmacogenetic evidence only cases where the effect of treatment differed by genotype, for example through a treatment-by-genotype interaction, and separating these from prognostic associations, disease susceptibility signals, and mechanistic findings.

The most scrutinized gene was ADRB2, which encodes the beta2-adrenergic receptor, the direct molecular target of LABAs. Three coding variants, Arg16Gly (rs1042713), Gln27Glu (rs1042714), and Thr164Ile (rs1800888), are known to affect receptor signaling after ligand stimulation. Yet the evidence was contradictory. In a pooled pharmacogenetic analysis of 17 randomized trials comprising 6,075 patients treated with umeclidinium, vilanterol, or both, none of the three variants was associated with changes in FEV1, FVC, or FEV1/FVC in response to treatment. In the two Ukrainian prospective studies, the ADRB2 variants were associated with aspects of disease course, including exacerbation-related hospital admissions and GOLD stage D classification, but after one year of tiotropium/olodaterol, exacerbations, hospitalizations, and symptom scores improved similarly across all genotype groups, indicating no differential treatment response.

The Rotterdam Study added a further layer of complexity. Among 1,053 COPD patients from this observational cohort who were current users of inhaled beta2-agonists, including short-acting and long-acting agents, each copy of the Arg allele of rs1042713 was associated with a 30% lower risk of moderate or severe exacerbations (HR 0.70; 95% CI 0.59–0.84), and each copy of the Gln allele of rs1042714 with a 20% lower risk. A sensitivity analysis restricted to LABA-only users showed a similar association for rs1042713 but not rs1042714. However, the study did not formally demonstrate an interaction between genotype and treatment response, and the reviewers caution that the observational design cannot exclude residual confounding by disease severity, treatment indication, adherence, or treatment intensity.

The most direct pharmacogenetic signal came from a substudy of the randomized HISTORIC trial, in which 165 Caucasian patients were genotyped for two SUMF1 variants. Among patients receiving LAMA/LABA plus the ICS budesonide, heterozygotes for rs793391 showed significantly greater improvement in FEV1 over 12 months compared with those receiving LAMA/LABA plus placebo (p = 0.002), along with a mean 5.5-point reduction in SGRQ score considered clinically relevant. No significant treatment-related improvement was seen for the TT or GG genotypes, and interpretation of the GG subgroup was limited by only nine carriers. The authors of that analysis considered the result exploratory and emphasized the need for validation in larger, independent cohorts, and the review likewise flags it as based on limited data lacking independent replication.

Large genome-wide efforts fared no better on replication. A post hoc GWAS of 8,439 patients with moderate-to-severe COPD from the IMPACT trial, which compared ICS/LAMA/LABA, ICS/LABA, and LAMA/LABA regimens, identified MAPK8 rs56195836 as associated with increased exacerbation rates in patients with baseline eosinophils of at least 150 cells/µL receiving fluticasone furoate-containing therapy, and AZU1 rs1962343 as associated with reduced exacerbation rates among frequent exacerbators on fluticasone furoate/vilanterol. Neither association was confirmed in independent follow-up analyses, leading the original investigators to conclude that common genetic variants are unlikely to play a major role in predicting exacerbation treatment response. Similarly, a post hoc GWAS of 2,005 patients on fluticasone propionate/salmeterol found seven correlated low-frequency variants near CSE1L associated with changes in SGRQ score, but associations with binary SGRQ response and FEV1 did not reach genome-wide significance.

The Indian case-control study illustrated the distinction between susceptibility and response. The CCL2 rs3760399 GG genotype was significantly more frequent among 150 COPD patients than 150 healthy controls (13.3% vs. 4.0%; OR 3.929), indicating an association with disease susceptibility. But in the subset of 46 patients assessed for response to budesonide/formoterol, genotype did not differ between responders and non-responders; instead, higher CCL2 expression, rather than the rs3760399 genotype itself, was independently associated with treatment non-response (OR 4.956; p = 0.031), hinting that gene expression measurements might hold more promise than single variants.

The overall verdict is sobering but carefully reasoned. Risk of bias assessment found one observational study of high quality, with the others rated moderate, while the trial-derived pharmacogenetic analyses showed low risk of bias in randomization and measurement but some concerns about selection of reported results given their exploratory, multiple-testing nature. The review concludes that current evidence is insufficient to support routine pharmacogenetic testing to guide LABA-, LAMA-, or ICS-containing therapy in COPD, and that prognostic variants, which may still inform risk stratification, should not be interpreted as predictors of treatment response without demonstrated treatment-by-genotype interaction. Before clinical implementation, candidate markers must be replicated in independent, diverse populations and validated in studies designed specifically to test their clinical utility. For now, the promise of genetically personalized inhaler choice in COPD remains exactly that: a promise awaiting proof.

Subject of Research: Genetic variants and treatment response to long-acting bronchodilators and inhaled corticosteroids in COPD

Article Title: Influence of genetic variants on the response to long-acting β2-agonists (LABAs) and long-acting muscarinic antagonists (LAMAs) and combination therapies in patients with chronic obstructive pulmonary disease (COPD): a systematic review

Article References: Torrecillas-Baena, B., Puentes-Osorio, Y., Ramírez, V., Salamanca-Miñarro, M. A., García-Alcaide, A., & Pajuelo, L. S. M. (2026). Influence of genetic variants on the response to long-acting β2-agonists (LABAs) and long-acting muscarinic antagonists (LAMAs) and combination therapies in patients with chronic obstructive pulmonary disease (COPD): a systematic review. Frontiers in Pharmacology, 17, Article 1973557. https://doi.org/10.3389/fphar.2026.1973557

Image Credits: AI Generated

DOI: 10.3389/fphar.2026.1973557

Keywords: COPD, pharmacogenetics, ADRB2, LABA, LAMA, inhaled corticosteroids, bronchodilators, GWAS, systematic review, SUMF1, genetic variants, treatment response