2-year infant and maternal outcomes investigating nicotine-replacement therapy for smoking cessation in pregnancy from the SNAP trial: a randomised controlled trial

Abstract

BackgroundNicotine-replacement therapy (NRT) is widely prescribed for smoking cessation in pregnancy, despite little evidence for its effectiveness and safety; the expert consensus is that NRT should be safer than smoking for the unborn infant. SNAP, a placebo randomised controlled trial comparing nicotine patches with placebo in pregnant smokers, found that NRT doubled cessation rates initially, but at delivery, no statistically significant effects were seen on either maternal smoking or birth outcomes. We report secondary outcomes from SNAP, comparing effects of NRT and placebo on infant development and respiratory problems and on maternal smoking 2 years after delivery. MethodsParticipants were aged 16–50 years, of 12–24 weeks' gestation, smoked five or more cigarettes daily, and were recruited in seven English antenatal hospital settings. Internet randomisation ensured 15 mg/16 h nicotine patch or placebo were randomly allocated in a 1:1 ratio. Apart from those with documented fetal deaths, all women were eligible for follow-up until their infants were aged 2 years; participants and those undertaking follow-up were masked to treatment allocation throughout the study. Outcomes were obtained by questionnaire (PQ2) sent to participants, which asked about maternal smoking and infant health and used items from the ages and stages questionnaire (ASQ-3) to assess infants' development. If participants did not respond, their general practitioner was sent a questionnaire (HPQ) to assess infant development instead. Of the secondary outcomes recorded at 2 years, the primary outcome was infant survival with no impairment, a composite of normal scores for all ASQ-3 domains plus no reported problems in remaining PQ2 items or HPQ responses. Singleton livebirth data were used for primary analysis of infant outcomes. Intention-to-treat analyses were done; for smoking outcomes, non-responders were assumed to be smokers, and for infant outcomes, complete case analyses were done with multiple imputations to explore the effect of missing data. Findings1050 women (mean age 26 years [SD 6·2], mean gestation 16 weeks [3·5]) were randomly assigned but 14 had fetal deaths and no birth details were obtained for 14. From the remaining 1022 women, we obtained questionnaires at 2 years for 900 (88%; 448 NRT, 452 placebo). 12 women had twins, leaving 1010 singleton infants (mean birthweight 3·2 kg [SD 0·6]), of which we obtained outcome data for 891 (445 NRT, 446 placebo). 323 (73%) singleton infants in the NRT group survived with no impairment compared with 290 (65%) in the placebo group (odds ratio [OR] 1·4, 95% CI 1·05–1·86; p=0·023). There was no difference in infants' reported respiratory symptoms between NRT and placebo groups: 132 (30%) in the NRT group and 111 (25%) in the placebo group (OR 1·3, 95% CI 0·97–1·74; p=0·083). Of the 1050 trial participants, 15 (3%) of 521 allocated NRT and nine (2%) of 529 allocated placebo abstained from smoking since a quit date set in pregnancy (OR 1·7, 95% CI 0·74–3·94; p=0·20). InterpretationFindings suggest that, at age 2 years, NRT used for smoking cessation in pregnancy results in better infant outcomes than with placebo. Thus, a smoking cessation intervention delivered in pregnancy could affect infant outcomes. We speculate that reported effects were caused by small changes in maternal smoking behaviour. FundingThe National Institute for Health Research Health Technology Assessment.