Evaluating Customized Antibiotic Duration (CDA) Strategy in Acute Exacerbations of COPD

Background: Antibiotics are routinely given to adults hospitalised with an acute exacerbation of COPD (AECOPD). International guidance (GOLD) advises that treatment should normally last about 5 days, but in many hospitals in Pakistan courses of 14 days remain usual practice. Longer courses increase antibiotic exposure, adverse effects and the risk of antimicrobial resistance, yet there is no randomised evidence from Pakistan on whether a shorter course is as effective for hospitalised patients. Intervention: The experimental arm receives a 7-day systemic antibiotic course and the active comparator arm the conventional 14-day course. The treating physician chooses the agent, dose and route according to institutional practice and available microbiology, so only the planned duration differs between arms. Duration is counted from the first antibiotic dose for the current exacerbation. All other care, including bronchodilators, corticosteroids, oxygen and ventilatory support, follows usual practice in both arms. The intervention was developed through a three-round Delphi consensus process with prescribers before the trial began. Procedures: Blood and sputum samples are taken at baseline for culture and susceptibility testing; results may guide a change of agent but do not change the allocated duration. Participants are assessed during treatment (Day 3-5), at Day 7, at Day 14 and at Day 30, in hospital or by telephone. Allocation uses a computer-generated sequence with permuted blocks of variable size, concealed through a web-based randomisation system. Sample size: Assuming 80% clinical success in both arms, a non-inferiority margin of 10 percentage points, a one-sided alpha of 0.025 and 80% power, 252 evaluable participants per arm are required; allowing for 10% attrition, 560 participants will be enrolled. Analysis: The primary analysis estimates the risk difference in clinical success at Day 30 (7-day minus 14-day) with a two-sided 95% Miettinen-Nurminen confidence interval. Non-inferiority will be concluded only if the lower confidence limit exceeds -10 percentage points in both the intention-to-treat and per-protocol populations. Full details are given in the study protocol (version 2.0) and statistical analysis plan (version 1.1).