Patients in clinical trials on COPD triple therapy compared to real world populations

Introduction: Randomized clinical trials (RCTs) regularly include populations defined by narrow inclusion/exclusion criteria. To guide general treatment decisions it is therefore crucial to understand the validity of a specific trial in the general COPD population. Recently, large RCTs on the effectiveness of triple vs. dual therapies in COPD have been published. Here we report the frequency of patients matching inclusion criteria of TRIBUTE, IMPACT and KRONOS in the real-life DACCORD cohort.

Methods: DACCORD is a prospective, non-interventional, ‘real-life’ study conducted in Germany. Patients were required to fulfill criteria for the German COPD disease management program (DMP) including spirometry and to either initiate or change COPD maintenance medication prior to study entry. Criterion for exclusion was participation in the Asthma DMP and participation in a RCT.

CAT total score and spirometry were evaluated annually and exacerbations were captured every 3 months during follow-up. To allow for a high quality of prospectively assessed data in DACCORD, inclusion criteria of the 3 RCTs were evaluated at the end of the 2 year follow-up period. [Table 1] shows the selection of inclusion criteria evaluated.

Table 1 Inclusion criteria for TRIBUTE, IMPACT and KRONOS.

	TRIBUTE	IMPACT	KRONOS
FEV1/FVC	< 0.7	< 0.7	< 0.7
Lung function	Post-bronchodilator FEV1 < 50% pred.	≥ 1 moderate or severe exacerbation if post-bronchodilator FEV1 < 50% pred. or ≥ 2 moderate or ≥ 1 severe COPD exacerbation if post-bronchodilator FEV1 50 – 80% pred. in the prior year	Post-bronchodilator FEV1 < 80% and ≥ 25 pred.
Exacerbation history	≥ 1 moderate or severe exacerbation in previous 12 months		Patients were not required to have an exacerbation history
Medication prior to screening	LAMA or dual therapy (i.e. ICS+LABA, ICS+LAMA or LABA+LAMA) for ≥ 2 months prior to screening	Existing COPD maintenance treatment for ≥ 3 months prior to screening	LABA/LAMA, LABA/ICS, LAMA/ICS or LABA/LAMA/ICS prior to baseline
Age	≥ 40	≥ 40	≥ 40 and ≤ 80
Smoking history	Smokers or ex-smokers ≥ 10 pack-years	Smokers or ex-smokers ≥ 10 pack-years	Smokers or ex-smokers ≥ 10 pack-years
CAT	≥ 10	≥ 10	≥ 10

Results: All of the 6611 analyzed patients were > 40 years. 19.4% of patients were non-smokers and 17.5% had a CAT score < 10. 20.4% of patients had an exacerbation in the prior year, of whom 13.8% had 1 exacerbation, 6.6% had ≥ 2 exacerbations and 2.9% had severe exacerbations. Mean FEV₁ predicted was 63.2%. In summary 1.8%, 5.4% and 24.0% of patients in DACCORD would qualify for TRIBUTE, IMPACT or KRONOS, respectively.

Conclusion: A high proportion of patients in the DACCORD real-life study would not meet the selective inclusion criteria for the above RCTs. Hence, it is questionable whether results of these RCTs can be extrapolated to the DACCORD cohort. While the selection of the right population is crucial for RCTs to generate valid answers to specific research questions, our results strongly support the need to understand how RCT findings apply in more pragmatic settings.