Bronchodilators, inhaled steroids, oxygen, and lung volume reduction.
What the evidence shows
Long-acting bronchodilators are the foundation. Where a LAMA and a LABA have been compared head to head, the LAMA prevented more exacerbations (POET-COPD), and dual bronchodilation outperformed an inhaled steroid combination in patients selected for exacerbations (FLAME).
Inhaled corticosteroids buy exacerbation reduction and pay for it in pneumonia. TORCH established both halves of that bargain and famously missed a mortality benefit at p=0.052. Adding a steroid to dual bronchodilation reduces exacerbations further (IMPACT, ETHOS, TRIBUTE), though IMPACT withdrew inhaled steroids at randomisation in most participants, which flatters the triple arm. ETHOS also found the lower budesonide dose worked as well as the higher. Going the other way, most patients without frequent exacerbations can have the steroid withdrawn without penalty — at the cost of a small fall in lung function, which was that trial's primary endpoint — and not at all if their blood eosinophils are raised (SUNSET).
The eosinophil has become a treatment target rather than merely a steroid-response marker. Dupilumab cut exacerbations by 30% in eosinophilic COPD already on triple therapy (BOREAS), and mepolizumab by 21% (MATINEE) — though neither improved quality of life as much as the exacerbation numbers suggest. Benralizumab given during an acute eosinophilic exacerbation outperformed prednisolone (ABRA).
Oral and inhaled add-ons occupy the space beyond triple therapy. Roflumilast's primary analysis missed significance and its sensitivity analysis met it (REACT); tolerability limits it in practice. Ensifentrine, the first new inhaled class in decades, improved lung function in both replicate trials with placebo-level side effects (ENHANCE). Daily azithromycin for a year reduces exacerbations at the cost of hearing loss and resistance (MACRO).
Oxygen is the oldest evidence here and the most misapplied. Long-term oxygen improves survival in severe resting hypoxaemia (NOTT, MRC oxygen) and does nothing at all for moderate desaturation (LOTT). More hours are not better: 24 a day was no better than 15 (REDOX).
No inhaled drug alters the trajectory. Four years of tiotropium produced a sustained improvement in FEV1, quality of life and exacerbations without changing the rate of decline (UPLIFT) — parallel curves shifted upward, which is the distinction between treating COPD and modifying it.
Exacerbations respond to antibiotics when the three Anthonisen criteria are met — increased dyspnoea, sputum volume and purulence (Anthonisen 1987), and to systemic steroids, which cut treatment failure from 33% to 23% at 30 days with no mortality benefit and no advantage beyond two weeks (SCCOPE). For volume reduction, surgery offers no overall survival benefit but a real one in upper-lobe emphysema with low exercise capacity, and harm in the wrong phenotype (NETT). Endobronchial valves achieve the same bronchoscopically, but only where the target lobe has no collateral ventilation (LIBERATE).