To the Editor:
Clinical trials in chronic obstructive pulmonary disease (COPD) patients have shown that the long-term use of inhaled corticosteroids (ICS) in COPD patients reduced the number of exacerbations per patient per year and improved health status [1]. Early studies have suggested increased ICS efficacy in patients with low lung function and frequent exacerbations [2]. The efficacy was reinforced when ICS was used in conjunction with long-acting β2-agonists (LABA) [3]. In most countries, health authorities approved ICS/LABA combinations in COPD patients with severe airflow impairment and frequent exacerbations, as also recommended in the Global Initiative for Obstructive Lung Disease (GOLD) 2007 document [4]. However, several surveys found poor adherence to this proposal among primary care physicians and pulmonologists in “real life”, ICS being often prescribed at a milder stage of the disease.
The GOLD 2011 document proposed a new multidimensional system for the assessment and management of patients with COPD [5]. This system classifies COPD patients into four categories (A, B, C and D) based on the level of symptoms (dyspnoea or global clinical impact) and the risk of future exacerbations, as assessed using the severity of airflow limitation and the past history of exacerbations [5]. The GOLD 2011 proposal, largely based on expert opinions, was challenged by studies investigating the association of COPD categories with future risk of exacerbation, hospitalisation and mortality [6, 7]. Notably, some authors found that subgroups of categories C and D (named C1, C2, C3, D1, D2 and D3) had different risks of exacerbation depending on whether a patient enters these categories because of low forced expiratory volume in 1 s (FEV1) only, past exacerbations only or both criteria combined [6]. Importantly, GOLD 2011 also proposed substantial changes in therapeutic options, ICS/LABA combinations being proposed as the first-line treatment option in GOLD categories C and D [5]. Thus, some patients with FEV1 >50% predicted or without repeated exacerbations could now be eligible for ICS/LABA therapy [5]. Consequences of this change in the indication of ICS/LABA combinations between GOLD 2007 and GOLD 2011 have not been specifically addressed in any clinical study to date.
Here, we investigated ICS use in real-life COPD patients and compared it to GOLD 2007 and GOLD 2011 proposals. Data were issued from the French COPD longitudinal cohort INITIATIVES BPCO, enrolling COPD subjects in 17 university hospitals [8]. For this analysis, data were extracted between June 2006 (when the long-acting muscarinic antagonist (LAMA) tiotropium was released in France) and June 2012 (before dissemination of the GOLD 2011 document). Classification of these patients (n=421) according to GOLD 2011 is presented in table 1. 253 (60%) patients were using ICS as single therapy (n=9, 2%), double therapy (n=107, 25%; including eight patients using ICS/LAMA and 99 using ICS/LABA) or triple therapy (ICS/LABA/LAMA; n=137, 33%). Based on the GOLD 2007 proposal [4], ICS/LABA was inappropriately prescribed in patients with FEV1 ≥50% pred (n=116, 46%) and in those with FEV1 <50% pred but with fewer than two exacerbations in the previous year (n=62, 25%); ICS monotherapy (n=4) or ICS/LAMA combination (n=1) were also considered inappropriate. Thus, according to GOLD 2007, 183 (72%) out of 253 patients were inappropriately receiving ICS therapy. Because the ICS/LABA combination salmeterol/fluticasone is approved in France in patients with FEV1 <60% pred and frequent exacerbations, we further examined ICS prescription in patients with FEV1 ≥50% to <60% pred: only 13% (n=34) of ICS patients had FEV1 in this range, of whom only half (n=17) had more than two exacerbations per year. Next, we compared ICS prescription to the GOLD 2011 proposal: ICS were inappropriately prescribed in GOLD A patients (n=44, 17%) and GOLD B patients (n=28, 11%); ICS prescription was also considered inappropriate in GOLD C and D patients receiving ICS alone (n=5, 2%) or ICS/LAMA (n=4, 2%). Thus, according to GOLD 2011, only 81 (32%) out of 253 patients were inappropriately receiving ICS therapy.
These results indicate that the real-life prescription of ICS in COPD patients in France was closer to GOLD 2011 than to GOLD 2007. Changes in proposals on ICS use between GOLD 2011 and GOLD 2007 are not based on new findings or randomised controlled studies (RCTs) [9], but result mostly from differences in interpretation of previously available data by experts. In our cohort, approximately half of the patients receiving ICS had FEV1 >50% pred. In this group of patients, evidence supporting the prescription of ICS/LABA is limited, except for salmeterol/fluticasone in patients with FEV1 between ≥50% and <60% pred and frequent exacerbations [3]. Furthermore, the efficacy of ICS/LABA in patients with severe airflow limitation but without frequent exacerbations remains unclear.
In summary, real-life treatment of COPD patients in France anticipated the new GOLD 2011 proposal. This observation questions the development and dissemination of recommendations for chronic diseases, including COPD. Most of the guidelines try to grade recommendations based on evidence, relying on the results of RCTs. This approach is highly acknowledged and well used, but most COPD patients are not eligible for RCTs for many reasons [10]. Clinical trials cannot answer all real-life questions, which may in part explain the marked discrepancies between GOLD 2007 proposals, based on RCT results, and daily practice. By contrast, GOLD 2011 appeared to be a paradigm shift by providing a more flexible expert interpretation of published evidence, leading to proposals reflecting more closely the attitude of clinicians. The GOLD 2011 document presents itself as a research-stimulating set of proposals that should be prospectively validated [5, 6]. We suggest that the proposal to use ICS/LABA outside indications validated by registration RCTs and, in many countries, outside indications approved by regulatory agencies should prompt new academic-based clinical trials to investigate whether the benefit–risk ratio of ICS/LABA remains favourable under these circumstances.
Footnotes
Support statement: The INITIATIVES BPCO cohort is supported by unrestricted grants from Boehringer Ingelheim France and Pfizer France.
Conflict of interest: Disclosures can be found alongside the online version of this article at www.erj.ersjournals.com
- Received September 16, 2013.
- Accepted October 14, 2013.
- ©ERS 2014