This study aimed to analyse UK pharmacovigilance data to quantify adverse events (AEs) associated with the non-steroidal anti-inflammatory drug (NSAID) molecules found in veterinary medicines authorised for use in dogs and cats. It was hypothesised that the frequency of AEs would be lower when associated with cyclo-oxygenase-2 selective (coxib), compared to non-selective (non-coxib) NSAIDs. The UK Veterinary Medicines Directorate (VMD) supplied frequencies of AEs derived from Periodic Safety Update Reports subdivided by formulation and species for each NSAID molecule. Frequencies of AEs were similar between species. The five most reported AEs were emesis, death, anorexia, lethargy, and diarrhoea. Reported frequency of emesis, renal insufficiency and death was higher with injectable compared to oral NSAIDs (P = 0.043). Reported frequency of emesis, lethargy and death was higher with coxib, compared to non-coxib NSAIDs (P = 0.029). Median (range) interval since authorisation was shorter for coxibs at 5 (2.5-9) years compared to non-coxibs at 15 (12-25) years. A negative correlation between time elapsed since authorisation and the frequency of AEs was identified (rs = -0.11 to -0.94). Higher frequency of reported AEs with injectable NSAIDs may be related to perioperative administration. The AE frequency associated with coxib and non-coxib NSAIDs may be confounded by changes in reporting habits over time. This study highlights the value of interrogating passive surveillance data to identify low frequency AEs and the need to facilitate improvement in recording and collecting AEs in small animal practice.