Guidelines for the Pharmacologic Treatment of COPD 2023: Canada versus GOLD

The 2023 Canadian Thoracic Society (CTS) guidelines for the pharmacological treatment of patients with chronic obstructive pulmonary disease (COPD) were recently published in two journals [1, 2]. The authors claim that these “guideline recommendations are consistent to other recent guidelines such as … the Global Initiative for Chronic Obstructive Lung Disease (GOLD) report 2023.” A careful reading of these Canadian guidelines find that several key recommendations stand in rather stark contrast with those of the GOLD 2023 report [3]. Note that I use the term “guidelines” for convenience of comparison, and because it is the term used by the CTS authors, recognizing that some consider them rather as strategies or recommendations. A comparison between the two reports is thus warranted, particularly in terms of the evidence behind the differing recommendations and the potential for unnecessary harm to patients.

Initial treatment recommendations

For the initial treatment of patients at low risk of exacerbation (0–1 prior moderate exacerbation), both the GOLD and CTS guidelines recommend to start with a long-acting β₂-agonist (LABA) or a long-acting muscarinic antagonist (LAMA) for those with low dyspnea and health status impairment scores (mMRC 0–1, CAT < 10). For those with higher dyspnea and health status impairment scores, the recommendation is to start with a LAMA + LABA combined bronchodilator (Figure 1).

Figure 1. Pharmacotherapy recommendations for the initial treatment of COPD comparing GOLD 2023 and CTS 2023.

For the initial treatment of patients at high risk of exacerbation (≥ 2 moderate or 1 severe exacerbation), GOLD recommends starting with a LAMA + LABA combined bronchodilator, irrespective of the degree of dyspnea and health status impairment scores (GOLD E patients), except for those with blood eosinophils ≥300 cells/µL, where a LAMA + LABA + ICS triple inhaler could be considered. In contrast, the CTS recommends that patients at high risk of exacerbation start with a LAMA + LABA + ICS triple inhaler if they have higher dyspnea and health status impairment scores, irrespective of eosinophil counts. No recommendation is given for patients at high risk of exacerbation with low dyspnea and health status impairment scores (Figure 1).

Follow-up treatment recommendations

For the follow-up treatment of COPD, when response to the initial treatment is inadequate, the GOLD guidelines recommend classifying patients according to two key treatable traits, dyspnea or exacerbations. For patients with the dyspnea trait, who are at low risk of exacerbation, the recommendation is to increase their treatment to a LAMA + LABA combined bronchodilator. Subsequently, GOLD recommends considering switching inhaler device or molecules, but there is no mention of introducing an ICS for these patients. On the other hand, the CTS recommends that these patients with dyspnea at low risk of exacerbation increase their treatment to a LAMA + LABA + ICS triple inhaler (Figure 2A).

Figure 2. Pharmacotherapy recommendations for the follow-up treatment of COPD comparing GOLD 2023 and CTS 2023, when response to the initial treatment is inadequate, for patients with: (A) a dyspnea profile, and (B) An exacerbation profile. Note that the labels “dyspnea profile” and “exacerbation profile” used here for comparative purposes may differ somewhat in definition between the two guidelines.

For the follow-up treatment of those patients with the high risk of exacerbation treatable trait, GOLD recommends that treatment be increased to a LAMA + LABA + ICS triple inhaler only if blood eosinophils ≥100 or 300 cells/µL, else to add roflumilast or azithromycin to the LAMA + LABA combined bronchodilator. In contrast, the CTS recommends that these patients at high risk of exacerbation, who already received a LAMA + LABA + ICS triple inhaler as initial treatment, should add oral therapies including macrolides, PDE-4 inhibitors, or mucolytic agents (Figure 2B).

Comments on the evidence

The CTS guidelines claim to “provide clinical guidance with evidence-based recommendations” by including “only randomized controlled trials (RCTs)” [1, 2]. There are, however, three sources of divergence between the evidence and the corresponding recommendations.

The first issue is that all trials used as evidence to recommend “initial” treatments for COPD in the CTS report were based on patients with longstanding COPD, who had been receiving maintenance therapy for a long time. For example, for the recommendation to start with a LAMA + LABA dual bronchodilator rather than monotherapy, 95% of the subjects enrolled in the largest trial (DYNAGITO) used as evidence, comparing a LAMA-LABA with a LAMA inhaler, were already treated at the time of randomization. Similarly, over 99% of the subjects in the ETHOS trial, used as evidence to initiate treatment of patients at high risk of exacerbation with a LAMA + LABA + ICS triple inhaler, were already treated at the time of randomization. It is unclear whether or to what extent the findings from a trial of patients with longstanding treated disease, patients who survived COPD up to that point, can be extrapolated to the initial treatment of treatment-naïve patients. On the other hand, while the GOLD guidelines also make recommendations for initial treatments, the authors are careful to avoid the “evidence-based” label, but rather makes these suggestions as “a practical recommendation; direct evidence is not available to guide therapy in naïve individuals” [3].

The second issue is the use of blood eosinophil counts to guide treatment, which is prominent and explicit in the GOLD guidelines but wholly absent in the CTS guidelines. For example, for the initial treatment of patients at high risk of exacerbation, GOLD recommends starting with a LAMA + LABA bronchodilator but suggests considering a LAMA + LABA + ICS triple inhaler for those with blood eosinophils ≥ 300 cells/µL. A more intricate use of blood eosinophil counts is used in the GOLD recommendations for the follow-up treatment of patients at high risk of exacerbation, with thresholds of 100 and 300 cells/µL guiding the choice of the next treatment. One recommendation involves the patients at high risk of exacerbation treated with a LAMA + LABA dual bronchodilator but with blood eosinophils < 100 cells/µL, for whom roflumilast or azithromycin is added to LAMA + LABA, rather than a LAMA + LABA + ICS triple inhaler.

This approach of using blood eosinophil count in the GOLD 2023 guidelines is in stark contrast with the CTS guidelines which do not consider the use of this important biomarker in their evaluations or recommendations. The reason given is that the guideline “is based on systematic reviews of RCTs” and that “most of the data on blood eosinophil count are derived from observational studies or post hoc analyses.” However, it was precisely post hoc analyses of large trials that were fundamental in establishing the link between blood eosinophil count and ICS effectiveness in COPD [4, 5]. A recent review finds solid and increasing evidence for using blood eosinophil counts “as part of a precision medicine strategy to identify the most suitable patients for inhaled corticosteroids treatment” [6].

The third issue relates to the recommendations for the use of a LAMA + LABA + ICS triple inhaler based on the reported mortality benefit from both the IMPACT and ETHOS trials. The CTS’s treatment chart specifically adds the label “reduces mortality” to the LAMA + LABA + ICS recommendation that patients at high risk of exacerbation with dyspnea should be initiated on a triple inhaler (their Figure 3) [1, 2]. This recommendation is based simply on the reported HRs of mortality comparing triple therapy with dual bronchodilators, namely 0.72 (95% CI: 0.53–0.99) in IMPACT and 0.51 (95% CI: 0.33–0.80) in ETHOS [7, 8]. The GOLD 2023 guideline also cites these two trials as evidence supporting a reduction in mortality with triple therapy (their Figure 6) [3].

It is, however, doubtful that IMPACT and ETHOS demonstrate a reduction in mortality with triple therapy. Indeed, it is unclear whether the observed reduction in all-cause mortality with triple therapy in these trials is the direct effect of the triple therapy or an effect of the abrupt discontinuation of ICS-based treatments prior to randomization. Indeed, this discontinuation, affecting 70–80% of the subjects, can result in a sudden increase in acute exacerbations and death among those allocated to LAMA-LABA [9–11]. Such an early adverse effect of abrupt ICS discontinuation has been previously shown on asthma mortality, relevant to the IMPACT and ETHOS trial populations that included patients with a history of asthma and other “asthma-like” patients [12].

Thus, rather than assessing the effect of initiating triple therapy in patients with COPD, these two randomized trials produced a confounded effect, namely a combined effect of discontinuing prior ICS treatment and of starting triple therapy. A proposed solution could be a sufficiently long run-in period that would tease out the effects of prior treatment discontinuation [9]. However, the choice of the common treatment during the run-in can possibly also confound the effect [13].

A cleaner untainted solution is to restrict the analysis to the subset of patients who were not previously on ICS. This allows to eliminate this bias by providing an untainted group, unaffected by the confounding effect of the abrupt ICS discontinuation [14]. Among these ICS-naïve patients, the HRs of mortality comparing triple therapy with dual bronchodilators become 1.25 (95% CI: 0.60–2.59) in IMPACT and 1.49 (95% CI: 0.49–4.55) in ETHOS, no longer suggesting a reduction in mortality [7, 8, 14]. This argument was supported and used in the FDA’s evaluations of IMPACT and ETHOS on mortality, and by their Pulmonary-Allergy Drugs Advisory Committee (PADAC), which concluded that there is no evidence that these triple therapies reduce all-cause mortality, in part because of the effect of abrupt withdrawal of prior ICS [15, 16].

Conclusion

The new 2023 Canadian Thoracic Society guidelines for the pharmacological treatment of patients with COPD make several “evidence-based” recommendations that diverge significantly from those in the GOLD 2023 report. Generally, this divergence arises from the Canadian Thoracic Society’s recommendations for broader and more aggressive use of LAMA + LABA + ICS triple inhalers. Practically, the Canadian Thoracic Society recommends a triple inhaler as first-line treatment for patients at higher risk of exacerbation with dyspnea, and as second-line for all patients with COPD, even those with low risk of exacerbation. However, the recommendation to initiate treatment with triple therapy is not evidence-based since no trial enrolled treatment-naïve patients. Moreover, all triple therapy trials were affected by the abrupt withdrawal of ICS at randomization, resulting in confounded effectiveness on exacerbations and mortality when compared with the LAMA + LABA inhaler. Consequently, such unsupported use of triple therapy can increase the risks of pneumonia, with no proven benefit. In contrast, the GOLD 2023 guidelines recommend a more targeted and judicious use of triple therapy, namely for patients at high risk of exacerbations and blood eosinophil count ≥300 cells/µL. This targeted recommendation is supported by a recent large observational study, using a design emulating a trial, that showed that triple therapy significantly reduced the incidence of exacerbation by 18% in this eosinophilic group, but not in those with count <300 cells/µL, as compared with a LAMA+LABA inhaler [17].

Financial disclosures

The author attended scientific advisory committee meetings or received speaking fees from AstraZeneca, Atara, Boehringer-Ingelheim, Bristol-Myers-Squibb, Merck, Novartis, Panalgo, Pfizer and Seqirus.

Funding information

This work is not funded. Pr. Suissa is the recipient of the Distinguished James McGill Professorship award.

Declaration of interest

The authors declare there is no Complete of Interest at this study.

Disclosure statement

No potential conflict of interest was reported by the author(s).