Alexander Mathioudakis, MD, PhD, from the biology department at The University of Manchester.

Alexander Mathioudakis, MD, PhD

chronic obstructive pulmonary disease (COPD)

, with real-world adherence to inhaled maintenance therapies estimated at just 20–30%. With COPD affecting an estimated 480 million adults globally in 2020 and projected to reach 600 million by 2050, the clinical consequences of treatment discontinuation warrant careful evaluation.

A post-hoc analysis of the 52-week FLAME trial, which included 3362 patients with moderate-to-severe COPD and a history of exacerbations, suggests discontinuation is not clinically neutral.

The analysis showed that stopping long-acting muscarinic antagonists (LAMA) was associated with a transient increase in moderate-to-severe exacerbations during the first quarter after withdrawal, with rate ratios as high as 2.2 in certain subgroups. Inhaled corticosteroid (ICS) discontinuation was linked to an earlier rise in severe exacerbations, with effects observed regardless of baseline blood eosinophil counts.

These withdrawal-related peaks were most pronounced within the first 3 months and appeared to attenuate over time, supporting a time-limited vulnerability window. The study also identified potential LAMA withdrawal effects with rate ratios approaching 3.7 in patients least influenced by concomitant therapy.

In this Q&A, Alexander Mathioudakis, MD, PhD, from the biology department at The University of Manchester, discusses the clinical interpretation of these findings and how they should inform adherence counseling, deprescribing decisions, and post-withdrawal monitoring in routine COPD care.

What prompted your team to investigate the effects of discontinuing ICS and (AMA in COPD patients?

 Treatment withdrawal effects are more common than we often acknowledge and have been described across many drug classes, including corticosteroids, beta-blockers, statins, anti-depressants, and anti-epileptics. In COPD, where adherence to inhaled therapy is frequently poor, periods of unintentional discontinuation are common in real-world practice.

While ICS withdrawal effects have been suggested previously, data on LAMA withdrawal were lacking. This prompted us to systematically explore whether discontinuing these treatments is associated with short-term increases in exacerbation risk. The findings from FLAME were revealing and clinically relevant.

 Your post-hoc analysis of the FLAME trial identified early increases in exacerbation risk after discontinuing ICS or LAMA. How should clinicians interpret these transient withdrawal effects in the context of routine COPD care?

 The first implication is the importance of supporting and reinforcing treatment adherence. Patients should understand that intermittent use or unplanned gaps in maintenance therapy may carry a short-term risk of exacerbations.

Second, when treatments are intentionally discontinued or switched, clinicians should anticipate a possible transient increase in risk. This is particularly relevant for ICS, where inappropriate initiation can make later withdrawal challenging. Aligning initiation and discontinuation decisions with GOLD recommendations remains essential to minimize avoidable harm.

 The study shows differences between moderate-to-severe and severe exacerbations. Could you clarify how these risks vary depending on prior therapy and baseline eosinophil counts?

 FLAME included relatively few severe exacerbations, meaning these analyses were underpowered to detect clear differences by severity. We do not believe this reflects a true biological difference in withdrawal effects between moderate and severe events.

Importantly, our findings suggest that ICS withdrawal effects may not be confined to patients with elevated blood eosinophil counts. Even patients who may derive limited benefit from ICS could still experience withdrawal-related risk. For LAMA, there is no biological rationale for an eosinophil-dependent effect. Apparent differences reflect trial design, as ICS initiation likely masked withdrawal effects in patients with higher eosinophil counts.

 In practice, many patients discontinue inhaled therapy intermittently. How might real-world patterns of non-adherence affect the risk of early exacerbations compared with what you observed in the trial?

 In real-world practice, many patients are non-adherent and intermittently stop and restart inhaled therapies. It is likely that such patterns of non-adherence could trigger withdrawal effects similar to those observed in FLAME.

Further evaluation using real-world data or studies with detailed adherence monitoring is needed to quantify the frequency and magnitude of these effects. Our analysis underscores the importance of treating treatment discontinuation as an active exposure state, rather than assuming non-adherence is clinically neutral.

 For patients being considered for ICS or LAMA withdrawal, what clinical factors or baseline characteristics should guide decision-making to minimize exacerbation risk?

 Current GOLD and international guidelines provide clear indications for ICS withdrawal, and LAMA discontinuation is generally rare, usually limited to intolerance or adverse effects. These recommendations remain appropriate.

Our findings reinforce the importance of avoiding unnecessary initiation of ICS, as inappropriate use increases the likelihood of difficult or risky withdrawal later. Decisions to discontinue therapy should be deliberate, guideline-aligned, and accompanied by appropriate clinical follow-up.

 Did your analysis suggest an optimal monitoring strategy after discontinuation to detect and manage early exacerbations effectively?

Our data suggest that the first 3 months after discontinuing ICS or LAMA represent a period of increased vulnerability to exacerbations. Heightened clinical awareness during this window is therefore reasonable. Exacerbations should, as always, be recognized and treated promptly.

 How might these findings inform step-down or de-escalation strategies, particularly for patients with moderate-to-severe COPD who are stable on combination therapy?

 Guidelines recommend considering ICS step-down in patients who are stable, have low exacerbation risk and low blood eosinophil counts, or in those experiencing recurrent pneumonia or infective exacerbations. Our findings do not challenge this approach.

However, they suggest that early exacerbations after ICS withdrawal may reflect a transient withdrawal effect rather than immediate treatment failure. Clinicians may therefore wish to assess exacerbation characteristics and eosinophil levels before rapidly re-escalating therapy during the first three months, where clinically appropriate.

Derman C. Stopping Inhaled Corticosteroids or LAMA in COPD Linked to Early Exacerbations. HCPLive. Published on January 15, 2026. Accessed on January 27, 2026. 

https://www.hcplive.com/view/stopping-inhaled-corticosteroids-lama--copd-linked-early-exacerbations

Mathioudakis AG, Bate S, Chatzimavridou-Grigoriadou V, et al. Disproportionate increase in COPD exacerbation risk for 3 months after discontinuing LAMA or ICS: insights from the FLAME trial. Thorax. Published online December 15, 2025. 

Articles

Chronic obstructive pulmonary disease (COPD) treatment discontinuation is prevalent, with adherence to inhaled therapies at 20-30%. A post-hoc analysis of the FLAME trial revealed that stopping long-acting muscarinic antagonists (LAMA) and inhaled corticosteroids (ICS) leads to a transient increase in exacerbations, particularly within the first three months. These findings emphasize the importance of adherence, careful deprescribing, and monitoring post-withdrawal. The study suggests that withdrawal effects are not limited to patients with high eosinophil counts, highlighting the need for guideline-aligned decision-making in COPD management.

In this Q&A, Mathioudaki discusses the post-hoc FLAME trial, which links ICS and LAMA discontinuation to transient exacerbation risk in COPD.