This is a study comparing the effects of ICS containing treatments in patients with chronic obstructive pulmonary disease (COPD) in a real world setting, using the UK Clinical Practice Research Datalink (CPRD) linked with Hospital Episode Statistics (HES). The main outcome to be assessed is exacerbation rates.

Background/Rationale:

For people with chronic obstructive pulmonary disease (COPD), standard maintenance inhaler treatments consist of inhaled corticosteroids (ICS) and long acting bronchodilators (principal classes include long-acting beta-2-agonists (LABAs) and long-acting muscarinic antagonists (LAMAs)). Clinical trials indicate that adding ICS to treatment combinations may provide rapid and sustained improvements (1). However, ICS may be associated with adverse effects, notably pneumonia.

There is a need for real-world effectiveness data regarding COPD treatment in order to demonstrate that improvements in lung function translate into reductions in exacerbations, hospitalizations or morbidity.

Objectives and Hypotheses: We hypothesize that ICS containing treatment regimens (including triple therapy - ICS/LABA/LAMA) are more effective at preventing AECOPD than non-ICS containing regimens. The main objective is to assess the impact of ICS therapy on exacerbation outcomes in a COPD population and identify which patient subgroups may achieve the greatest benefit.

Methods:

Study design: Cohort study examining comparative effectiveness

Data Source(s): Clinical Practice Research Datalink (CPRD) GOLD, Hospital Episode Statistics (HES) and Office for National Statistics (ONS) mortality data

Study Population: Patients ≥40 years-old, with a validated diagnosis of COPD registered between the 1st of January 2006 and the 29 February 2016. Eligible patients must have a smoking history, data recorded at least 12 months prior to the study index date and have Up-To-Standard (UTS) data as defined by CPRD.

Exposure(s): ICS containing (LAMA/LABA/ICS and LABA/ICS) regimens and non-ICS containing regimens (LABA/LAMA, LAMA monotherapy).

Outcome(s): Exacerbationsmof COPD, defined using a published algorithm, both GP treated and hospitalised(2), hospitalised pneumonias.

Sample Size Estimations:

A two-sample log-rank test for power indicated that a total of n=2,858 patients (1,429 patients per group) will be required to detect a 15% risk reduction for exacerbation in the triple therapy population compared to the dual-therapy group (comparison which requires the maximum number of patients). This is based on a 60% annual hazard rate for exacerbation in the dual-therapy population and a 51% annual hazard rate for exacerbation in the triple therapy group; assuming power of 90%, alpha=0.05, and a conservative accrual period of one year and a minimum follow-up time of one year.

Statistical Analysis: Descriptive statistics will be used to characterize patients according to baseline demographic, clinical and treatment factors and to study treatment patterns during follow-up.

Analytical statistics will estimate time to exacerbation events per treatment type during the follow-up period. These statistics will include extended Cox regression models, marginal structural models (MSM) and inverse probability weighting (IPW).

Sensitivity analyses will assess follow-up time in incremental periods in order to examine the impact of drug type and drug changes over time. We will also analyse associations between demographic and clinical factors and treatment received.

Inclusion Criteria:

- All adult COPD patients aged ≥40, with a smoking history, incident prescription of a study therapy, and research acceptable data.

Exclusion criteria:

- Any subjects with less than 12 months of UTS data prior to index date.

We will include all patients ≥40 years-old, with a validated diagnosis of COPD (29) recorded between the 1st of January 2006 and 29 Feb 2016 and incident use of one of the therapies under investigation during the study period. Eligible patients must have a smoking history, data recorded at least 12 months prior to the study index date and have up to standard (UTS) data as defined by CPRD (Appendix A lists Read codes for identification and characterization of the COPD population).

The COPD definition has been validated in the CPRD against a reference standard of physician review of patient notes by the Imperial College research team. These definitions have high Positive Predictive Value (PPV) >85%, and will be used in this analysis.

All eligible CPRD GOLD patients will be included in the descriptive analysis, regardless of their eligibility for data linkage to HES, ONS and IMD data. The follow-up and analytical statistics will only be applied to patients eligible for linkage.