Systematic review of non-RCT studies to compare the safety of nicotine replacement therapy (NRT) use and smoking in pregnancy

Talk Code: 
Lauren Taylor
Tim Coleman, Tom Coleman-Haynes, Katarzyna Campbell, Sue Cooper
Author institutions: 
University of Nottingham


Smoking in pregnancy can lead to adverse health effects for mothers and their offspring. Stopping smoking in pregnancy is therefore an important public health target. Nicotine replacement therapy (NRT) is recommended for women who are unable to stop smoking with non-pharmacological interventions. Systematic review of RCT data has found no evidence that NRT is harmful in pregnancy. In this study we systematically reviewed non-RCT data to further investigate the safety of NRT use in pregnancy compared to smoking. Using non-RCT studies with larger sample sizes gives higher statistical power, reducing the risk of type 1 error that might be experienced when studying rare outcomes in RCTs.


A study protocol has been registered on PROSPERO. Systematic searches were performed using the MEDLINE, CINAHL, Embase, PsycINFO, CAB Abstracts, Social Sciences Citation Index and Economic and Social Research Council databases. Non-RCT studies of any design describing NRT use in pregnancy and reporting maternal, fetal or infant health outcomes were included. Once articles had been screened by two reviewers, the data was extracted and studies were quality assessed using a modified Newcastle-Ottawa scale. Data was requested from authors where there was insufficient detail to include the study or where further data was required for analysis. Analysis is ongoing; data will be meta-analysed where possible and other results will be described narratively.


After duplicate removal and full text screening was completed, 21 studies were included. Nine were large epidemiological cohorts of women using NRT as part of routine healthcare; 12 were smaller cohorts with interventional administration of NRT prior to outcome measurements. Quality of epidemiological studies was better (median score of 7/8; 4/7 for interventional cohorts); generally interventional studies did not describe the selection of their cohorts very well. The epidemiological studies with lower quality scores failed to validate self-reported outcomes by cross-referencing with medical records. Outcome measures were preterm birth, gestational age at birth, birthweight, delivery complications, stillbirth, infantile colic, congenital anomalies, strabismus, maternal side effects from NRT and fetal or maternal observations during nicotine administration. The interventional cohorts’ methods including the route and dose of NRT used were too dissimilar to pool the outcome data. Meta-analysis of data from epidemiological studies will be attempted for the following outcomes: congenital anomalies, low birthweight, preterm birth and stillbirth. Full results of these will be presented.


With ongoing analysis of our review data we hope to establish whether or not NRT is associated with any adverse maternal or infant health outcomes when it is used in pregnancy. Gaining further knowledge on the safety of NRT in pregnancy is likely to improve maternal confidence in, and adherence to treatment, which could increase smoking cessation rates.

Submitted by: 
Lauren Taylor
Funding acknowledgement: