Prematurity accounts for the majority of neonatal morbidity and mortality in the developed world. The process of labour resembles inflammation, with prostaglandin and cytokine production both before and during labour. Anti-inflammatory drugs therefore have the potential to prevent preterm delivery. Indomethacin is the only tocolytic drug proven to delay delivery beyond 37 weeks and to reduce the incidence of low birth weight (<2500 g). There are, however, fetal side-effects such as ductal constriction and impaired renal function associated with its use. It is the type 2 isoform of cyclo-oxygenase (COX-2), which is important for the production of prostaglandins within intrauterine tissues and that up-regulation of COX-2 is associated with labour. Although indomethacin is currently the most common non-steroidal anti-inflammatory drug (NSAID) used in the treatment of preterm labour, it was hoped that COX-2-selective drugs, used as tocolytics, would target COX-2 activity and potentially spare COX-1-specific fetal side-effects. Experience with sulindac and nimesulide has been linked with both constriction of the ductus arteriosus and oligohydramnios. It is unclear whether this is due to COX-2-dependent side-effects, or due to accumulation of drug in the fetal circulation leading to levels that would cause COX-1 inhibition. Currently, the use of COX-2-selective drugs should therefore be confined to randomized controlled trials.