ObjectiveTo investigate relationships between maternal smoking status in pregnancy and infant development. The largest randomised controlled trial of nicotine replacement therapy (NRT) for smoking cessation in pregnancy, the smoking, nicotine and pregnancy (SNAP) trial, found that at 1 month after randomisation, smoking cessation rates were doubled in the NRT group compared with the placebo group. At delivery, there was no significant difference in cessation rates between groups. Surprisingly, infants born to women randomised to NRT were more likely to have unimpaired development at 2 years. We hypothesised that this apparently protective effect was due to smoking cessation caused by NRT and so, investigate this relationship using the same cohort.DesignSecondary analysis of a randomised controlled trial.SettingSeven antenatal hospitals in the Midlands and North-West England.ParticipantsEight hundred and eighty-four pregnant smokers randomised to receive either NRT patches or visually-identical placebo in the SNAP trial. Participants’ smoking behaviour were recorded at randomisation, 1 month after their target quit date and at delivery.MethodsUsing logistic regression models, we investigated associations between participants’ smoking measures and infant development (assessed using the Ages and Stages questionnaire) at 2 years.Main outcome measures2 year infant development.ResultsDevelopmental impairment was reported for 12.7% of study 2 year olds. Maternal heaviness of smoking at randomisation (OR: 1.26, 95% CI: 0.82 to 1.96, p=0.091), validated smoking abstinence recorded at 1 month after a quit date (OR: 1.02, 95% CI: 0.60 to 1.74, p=0.914) and validated smoking abstinence recorded at both 1 month after a quit date and at the end of pregnancy (OR: 1.52, 95% CI: 0.81 to 2.85, p=0.795) were not independently associated with infant developmental impairment at 2 years.ConclusionWe found no evidence that NRT treatment improved infants' developmental outcomes through smoking cessation.Trial registration numberCTA03057/0002/001-0001; Post-results