Withdrawal of inhaled corticosteroids versus continuation of triple therapy in patients with COPD in real life: observational comparative effectiveness study.

<bold>Background: </bold>Inhaled corticosteroids (ICS) are indicated for prevention of exacerbations in patients with COPD, but they are frequently overprescribed. ICS withdrawal has been recommended by international guidelines in order to prevent side effects in patients in whom ICS are not indicated.<bold>Method: </bold>Observational comparative effectiveness study aimed to evaluate the effect of ICS withdrawal versus continuation of triple therapy (TT) in COPD patients in primary care. Data were obtained from the Optimum Patient Care Research Database (OPCRD) in the UK.<bold>Results: </bold>A total of 1046 patients who withdrew ICS were matched 1:4 by time on TT to 4184 patients who continued with TT. Up to 76.1% of the total population had 0 or 1 exacerbation the previous year. After controlling for confounders, patients who discontinued ICS did not have an increased risk of moderate or severe exacerbations (adjusted HR: 1.04, 95% confidence interval (CI) 0.94-1.15; p = 0.441). However, rates of exacerbations managed in primary care (incidence rate ratio (IRR) 1.33, 95% CI 1.10-1.60; p = 0.003) or in hospital (IRR 1.72, 95% CI 1.03-2.86; p = 0.036) were higher in the cessation group. Unsuccessful ICS withdrawal was significantly and independently associated with more frequent courses of oral corticosteroids the previous year and with a blood eosinophil count ≥ 300 cells/μL.<bold>Conclusions: </bold>In this primary care population of patients with COPD, composed mostly of infrequent exacerbators, discontinuation of ICS from TT was not associated with an increased risk of exacerbation; however, the subgroup of patients with more frequent courses of oral corticosteroids and high blood eosinophil counts should not be withdrawn from ICS. Trial registration European Network of Centres for Pharmacoepidemiology and Pharmacovigilance (EUPAS30851). [ABSTRACT FROM AUTHOR]