Medications for chronic obstructive pulmonary disease: a historical non-interventional cohort study with validation...

Area of Science:

• Pulmonary Medicine
• Clinical Epidemiology
• Health Informatics

Background:

• Randomized controlled trials (RCTs) inform chronic obstructive pulmonary disease (COPD) treatment, but generalizability to excluded populations is unclear.
• The TOwards a Revolution in COPD Health (TORCH) trial provides benchmark data for COPD treatment effectiveness.
• Non-interventional methods using electronic health records (EHRs) offer a potential alternative for assessing treatment effects in broader patient groups.

Purpose of the Study:

• To validate non-interventional methods using UK Clinical Practice Research Datalink (CPRD) data against TORCH trial results for COPD treatment effectiveness.
• To apply validated methods to estimate COPD treatment effects in patients who would have been excluded from the TORCH trial.
• To compare the effectiveness of fluticasone propionate/salmeterol (FP-SAL) versus salmeterol alone or no FP-SAL in diverse COPD populations.

Main Methods:

• A historical non-interventional cohort design utilizing CPRD GOLD data (2000-2017).
• Validation cohort matched to TORCH trial participants based on inclusion/exclusion criteria.
• Analysis cohorts created for patients ineligible for TORCH due to age, asthma, comorbidity, or disease severity.
• Comparison of outcomes including exacerbations, mortality, and pneumonia between treatment groups.

Main Results:

• Exacerbation rate ratios for FP-SAL versus salmeterol were comparable between CPRD and TORCH.
• FP-SAL versus no FP-SAL showed a higher exacerbation rate in CPRD than in TORCH.
• Mortality and pneumonia results between active comparators were consistent with TORCH findings.
• In milder COPD, FP-SAL showed a stronger protective effect on exacerbations than observed in TORCH.
• An increased mortality risk was observed for FP-SAL in patients with prior asthma, not seen in the TORCH trial.

Conclusions:

• Routinely collected EHR data can effectively measure COPD treatment effects when comparing active treatments but not active versus no treatment.
• Treatment effects for FP-SAL in patients excluded from trials are generally similar to trial findings.
• Further investigation is warranted for the observed increased mortality risk in COPD patients with concomitant asthma using FP-SAL.