Did Triple Inhaler Therapy Really Lower COPD Deaths?

Two large trials, IMPACT and ETHOS, each reported that single-inhaler triple therapy lowered all-cause mortality in COPD, and inside those trials as they were built, the finding is real. The honest answer is narrower than the headline. The death benefit shows up when triple therapy is compared with a long-acting bronchodilator pair that contains no inhaled steroid, and it fades when the comparator already contains a steroid, or when patients were not taking a steroid before the trial began. So the signal is better read as evidence about inhaled corticosteroids, and about how these studies were assembled, than as proof that a three-drug canister rescues everyone.

The two trials and what they claimed

IMPACT (Lipson and colleagues, New England Journal of Medicine, 2018) randomized more than 10,000 people with symptomatic COPD and a history of exacerbations to a single inhaler holding fluticasone furoate, umeclidinium, and vilanterol, or to one of two dual therapies: fluticasone furoate plus vilanterol (a steroid with a long-acting beta agonist), or umeclidinium plus vilanterol (two long-acting bronchodilators, no steroid). ETHOS (Rabe and colleagues, New England Journal of Medicine, 2020) tested budesonide, glycopyrrolate, and formoterol at two budesonide doses against a matching bronchodilator-only pair and against a steroid-plus-bronchodilator pair, in roughly 8,500 patients. The primary endpoint in both was the rate of moderate or severe exacerbations, which triple therapy reduced. The result that traveled furthest, though, was a secondary one: fewer deaths.

What the death numbers actually were

In the IMPACT mortality analysis (Lipson and colleagues, American Journal of Respiratory and Critical Care Medicine, 2020, with vital status recovered for 99.6% of participants), about 2.4% of the triple-therapy group died versus about 3.2% of the bronchodilator-only group. The hazard ratio against umeclidinium plus vilanterol was 0.72 (95% CI 0.53 to 0.99). Against the steroid-containing pair, fluticasone furoate plus vilanterol, the hazard ratio was 0.89 (95% CI 0.67 to 1.16), which did not reach significance.

ETHOS told the same story more sharply. In its dedicated mortality analysis (American Journal of Respiratory and Critical Care Medicine, 2021), the higher-dose triple combination cut all-cause mortality against the bronchodilator-only pair with a hazard ratio of 0.51 (95% CI 0.33 to 0.80), close to a halving of risk. Against the steroid-containing comparator, the difference was not significant. In both trials, the mortality benefit lived in the comparison with the arm that carried no inhaled steroid.

The missing-data problem, and why it was fixed

The first reports drew fair criticism because vital status was incomplete. When patients withdraw from a trial and are not followed, no one knows whether they later lived or died, and a few percent of unknown outcomes can swamp a small mortality difference. Regulators pressed the sponsors to chase down those outcomes. In ETHOS, the proportion of patients with unknown week-52 vital status fell from roughly 4.5% to about 0.4% after investigators contacted sites and patients and searched public records and death registries. The mortality benefit held after the gap closed. That is a genuine point in the finding's favor, and it is why the signal cannot be waved away as a data artifact alone.

The confounding that will not go away

A separate and harder objection concerns who entered these trials. As Suissa has detailed (CHEST, 2022), roughly 70% to 80% of enrollees across IMPACT and ETHOS were already taking an inhaled steroid when they joined. Anyone randomized to a bronchodilator-only arm therefore had that steroid stopped abruptly on day one. Withdrawing an inhaled steroid from a steroid-dependent patient can trigger early exacerbations, so some of the deaths in the comparator arm may reflect harm from stopping a drug rather than benefit from adding two more.

The timing fits that reading. The separation in survival curves clustered in the first months, the window when withdrawal harm would be expected, a pattern epidemiologists call depletion of susceptibles. The mortality benefit was also concentrated in patients who had been on an inhaled steroid before randomization. In the smaller subgroup who were steroid-naive at entry, and who therefore had nothing withdrawn, the survival advantage was no longer evident and the confidence intervals crossed no difference. If triple therapy itself saved lives, the benefit should not track so closely with whether a steroid had just been taken away.

Why the comparator decides the answer

Put the pieces together and the direction of the evidence converges. Against a no-steroid comparator, triple therapy looks life-saving. Against a steroid-containing comparator, the difference is not significant. Among patients who were never on a steroid to begin with, no clear benefit appears. The common thread is the inhaled steroid, not the number of drugs in the device. Part of the apparent gain is a real steroid effect in patients who respond to steroids, and part is harm avoided in the comparator arm simply by not withdrawing a steroid.

That reframes the clinical question. The useful decision is which patients gain from an inhaled steroid at all, rather than whether a three-drug inhaler beats a two-drug one. Higher blood eosinophil counts and a history of frequent exacerbations mark the people most likely to benefit, while steroid exposure carries its own cost, including a raised risk of pneumonia. A blanket claim that triple therapy lowers COPD deaths overshoots what the trials support. A narrower claim, that adding an inhaled steroid reduces mortality in steroid-responsive patients relative to no steroid at all, sits closer to what the data show.

So, did triple inhaler therapy really lower COPD deaths? Within the design of IMPACT and ETHOS, and measured against a bronchodilator pair with no steroid, it did. As a general promise, the answer is qualified by the comparator, by who was enrolled, and by what was taken away at randomization. This article is educational and is not medical advice, and treatment decisions belong with a patient and their own clinician.