Adverse effects of amoxicillin for acute lower respiratory tract infection in primary care: secondary and subgroup analysis of a randomised clinical trial

Abstract

A European placebo-controlled trial of antibiotic treatment for lower respiratory tract infection (LRTI) conducted in 16 primary care practices networks recruited participants between November 2007 and April 2010, and found adverse events (AEs) occurred more often in patients prescribed amoxicillin compared to placebo. This secondary analysis explores the causal relationship and estimates specific AEs (diarrhoea, nausea, rash) due to amoxicillin treatment for LRTI, and if any subgroup is at increased risk of any or a specific AE. A total of 2061 patients were randomly assigned to amoxicillin (1038) and placebo (1023); 595 (28%) were 60 and older. A significantly higher proportion of any AEs (diarrhoea or nausea or rash) (OR = 1.31, 95% CI 1.05–1.64, number needed to harm (NNH) = 24) and of diarrhoea (OR 1.43 95% CI 1.08–1.90, NNH = 29) was reported in the amoxicillin group during the first week after randomisation. Subgroup analysis showed rash was significantly more often reported in males prescribed amoxicillin (interaction term 3.72 95% CI 1.22–11.36; OR of amoxicillin in males 2.79 (95% CI 1.08–7.22). No other subgroup at higher risk was identified. Although the study was not powered for subgroup analysis, this analysis suggests that most patients are likely to be equally harmed when prescribed antibiotics.