Helping pregnant smokers quit: a multicentre randomised controlled trial of electronic cigarettes versus nicotine replacement therapy

Article history

The research reported in this issue of the journal was funded by the HTA programme as project number 15/57/85. The contractual start date was in May 2017. The draft report began editorial review in October 2021 and was accepted for publication in September 2022. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HTA editors and publisher have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the draft document. However, they do not accept liability for damages or losses arising from material published in this report.

Copyright statement

Copyright © 2023 Przulj et al. This work was produced by Przulj et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This is an Open Access publication distributed under the terms of the Creative Commons Attribution CC BY 4.0 licence, which permits unrestricted use, distribution, reproduction and adaption in any medium and for any purpose provided that it is properly attributed. See: https://creativecommons.org/licenses/by/4.0/. For attribution the title, original author(s), the publication source – Journals Library, and the DOI of the publication must be cited.

2023 Przulj et al.

Overview of trial design

This was a pragmatic multi-centre RCT comparing ECs with NPs, with both interventions accompanied by the same brief behavioural support.

Objectives and hypothesis

The objective of the study was to compare ECs and NPs in effectiveness in helping pregnant smokers quit and in their safety. We hypothesised that ECs would be more effective than NPs. We did not formulate any hypotheses regarding comparative safety of the two products.

Changes to trial design

Table 1 provides a summary of the changes made to the study protocol.

Participants

Trial procedures

Smokers who were interested in the study were provided with the patient information sheet. If they decided to take part, they were invited to the baseline visit, arranged where possible to coincide with their hospital appointments.

Measures

The following were collected at baseline (for categories and units, e.g. age in years, ng/ml, see Table 3):

1. demographics: age, ethnicity, employment, whether living with a partner and level of education

2. smoking history: Fagerström Test for Cigarette Dependence (FTCD),38 cigarettes per day, previous stop-smoking product use, longest previous period of abstinence and whether they live with other smokers

3. saliva sample to measure nicotine intake via cotinine levels.

The following were collected weekly (from 1 to 4 weeks post TQD), and at EOP:

1. smoking status

2. allocated product use during the trial: on how many days per week; product type; reasons for stopping/switching (where applicable); for ECs only: e-liquid flavour and strength

3. non-allocated product use (same as above)

4. contact with local SSS

5. adverse events (AEs) and adverse reactions (ARs), and infant AEs at 3 months post partum

6. at EOP only: weeks of using products regularly (on ≥5 days/week) and weeks of using occasionally (at least once a week)

7. respiratory health questionnaire asking about shortness of breath, phlegm, wheezing and cough39

8. in participants reporting abstinence from smoking: salivary cotinine and anabasine or CO reading (in abstainers using nicotine products)

9. in participants reporting reduction of smoking by at least 50% and in those reporting both smoking and using nicotine products: salivary cotinine.

Birth outcomes were also collected by research staff at the participating site via hospital records.

Primary outcome

Prolonged abstinence from smoking from 2 weeks after the TQD until the EOP defined as per Russell Standard (up to five lapses allowed with no smoking at all during the previous week at the time of final follow-up)40 and validated by salivary cotinine (<10 ng/ml41) for those not reporting using any nicotine product or a CO of <8 parts per million or anabasine (<1 ng/ml) for those reporting other forms of nicotine use at the EOP follow-up.

We supplemented anabasine validation with CO validation when we noticed during an unrelated study that we were conducting at the time that salivary anabasine may be a less reliable indicator of smoking than thought. We used anabasine validation where CO readings were not available.

Participants with missing validation or missing information on smoking status were included as non-abstainers.

Secondary smoking status outcomes

Self-reported prolonged abstinence at EOP, 7-day and 2-month validated and self-reported abstinence at EOP and self-reported abstinence at 4 weeks. (Note: if participants were reached only post delivery and reported smoking now, but were abstinent at the time of delivery, they were included as self-reported abstainers.)

Proportion of participants who reduced their cigarette consumption by at least 50% at EOP compared to baseline, self-reported and also validated by a reduction of at least 50% in salivary cotinine levels (including abstainers; and only including non-abstainers).

Adherence outcomes

Number setting TQD, weekly support contacts completed, number receiving support from local SSS; number of weeks products were used regularly (≥5 days/week) and occasionally (<5 days/week), amount used per day, products used, product changes; and use of non-allocated products (same details as above).

Safety measures and health outcomes

Proportion of participants reporting AEs, serious adverse events (SAEs), ARs [using Medical Dictionary for Regulatory Activities (MedDRA) terminology] and new respiratory symptoms (phlegm, shortness of breath, cough and wheezing).

Data collection and entry

A web-based application, using an Oracle 11g database, was used to collect data. Oracle Database 11g Enterprise Edition Release 11.2.0.1.0 (Oracle Corporation, Austin, TX, USA) was used for building the database and Java programming language for the building of the study application. This was set up and hosted by the Centre for Cancer Prevention, QMUL. Data were kept in accordance with good clinical practice and data protection requirements.

Sample size

We estimated the quit rate at delivery in the NP arm at 8%12 and in the EC arm at 14%39 [odds ratio 1.87, risk ratio (RR) = 1.75]. Using the power command in Stata^® version 15 (StataCorp LP, College Station, TX, USA) indicated that 1140 participants would be needed for 90% power (570 in each condition; alpha = 0.05, two-tailed test) to detect this difference. This sample size would allow detecting smaller treatment effects with lower power, for example, 80% power to detect a difference between 8% and 13.1% (odds ratio 1.74, RR = 1.64).

Randomisation

Participants were randomised (1:1) via a pre-programmed list generated by an independent statistician using R Statistical Software (Foundation for Statistical Computing, Vienna, Austria). 44 The randomisation scheme consisted of a sequence of blocks of random size. The randomisation list was programmed into the study database application by the study programmer who was not involved in the recruitment of participants. Research midwives at participating sites enrolled participants and conducted the randomisation through the study database application. Study staff were unaware of the randomisation sequence.

Treatment blinding

Researchers and participants were blind to treatment allocation until the point of randomisation which was generated via the study database. Once randomised, participants were informed by the research midwife which treatment they would receive. Researchers conducting follow-up calls were blind to treatment allocation until condition-specific questions were reached. The trial statistician was blind to participant allocation until the analysis was complete.

Statistical methods

Participant withdrawal

Participants were able to withdraw from the study at any time. Participants who withdrew their consent for further data collection were not replaced. Those who withdrew prior to the EOP follow-up were included as non-abstainers. Data collected up to the point of the consent withdrawal were used.

Participants who were randomised in error and found ineligible by the site PIs, or during the first contact with the study team, or who were randomised twice by mistake, were withdrawn and replaced.

Patient and public involvement

Study design and protocol were discussed with smokers and stop-smoking advisers. A member of the public served on the Trial Steering Committee (TSC) and advised on participant engagement.

At our annual ‘Update and Supervision Days’ for SSS staff, we asked for input from advisers working with pregnant smokers. The advisors noted that EC use by pregnant smokers is becoming more common. There was a consensus that information on safety and efficacy of such use is urgently needed and that the trial results will have an important practical impact.

Study approvals and trial conduct

The study was sponsored by the QMUL Joint Management Research Office. Ethical approval was obtained from the National Research Ethics Service Committee London –South East (ref: 17/LO/0962) on 29 June 2017 and approvals were obtained from the Health Research Authority on 24 July 2017.

The study was approved by the Medicines and Healthcare Products Regulatory Agency on 5 June 2017 via the Clinical Trial of Investigational Medicinal Product (CTIMP) Notification Scheme. The study was pre-registered on ISRCTN, ref: ISRCTN62025374.

A Data Monitoring and Ethics Committee and a TSC were convened every 6–12 months during the recruitment and follow-up phase of the trial. All members were independent. A Trial Monitoring Group consisting of members of the study team also met regularly. Table 2 shows the members of the committees.

Quality control

A risk assessment was carried out in conjunction with the study sponsor and Barts CTU (the CTU at time of study commencement), which was used as a basis for the study monitoring plan. During the recruitment phase, a monitor from the co-ordinating site carried out a monitoring visit at each of the sites within 3 months of the first participant being enrolled. Monitoring visits were then carried out every 6 months. The monitoring was initially at the study site but later moved to remote working during the COVID-19 pandemic. The sponsor was responsible for oversight of the monitoring process and monitoring of the co-ordinating centre. Trial oversight was provided by the Barts CTU (between May 2017 and March 2019) and the King’s CTU (from April 2019).

Saliva analysis

ABS Laboratories was used for cotinine and anabasine analysis. ABS Laboratories is a good laboratory practice (GLP) accredited contract research organisation that specialises in the quantification of drugs, metabolites and biomarkers in biological and non-biological samples. Cotinine and anabasine in saliva samples were determined using high-performance liquid chromatography coupled to tandem mass spectrometry with multiple reaction monitoring.

Participant flow

Figure 1 shows the Consolidated Standards of Reporting Trials (CONSORT) flow diagram. 48

FIGURE 1.

CONSORT flow diagram. * Four in each arm.

We were able to establish self-reported smoking status in 531 (93%) and 516 (91%) participants in the EC and NP arms, respectively, either via direct contact or from hospital records.

Table 3 presents the sample characteristics. The participants in the two study arms had similar profiles.

Effectiveness outcomes

For the primary outcome, saliva samples were required from participants who reported abstinence at EOP. Obtaining the samples proved difficult, with only 108 of 196 self-reported abstainers (55.1%) providing useable samples. Availability of useable samples in participants reporting reduction in smoking (46.9%) and those reporting dual use (60.5%) was broadly similar. We checked for an effect of COVID-19 lockdown on data collection, but it was only minor (see Appendix 3).

Due to problems with saliva sampling, validated prolonged abstinence rates were low (6.8% vs. 4.4% in the EC and NP arms, respectively). The difference between the two study arms was not significant (Table 4). The BF = 2.69 which suggests the data are insensitive. 49

The results of ‘per-protocol’ and multiple-imputation analyses were similar. However, in the analysis excluding abstainers who regularly used the non-allocated product (0 ECs vs. 5 NRT), the difference between the two study arms (6.8% vs. 3.6%) was significant (Table 4). The BF = 10.0 which indicates a strong effect.

Regarding secondary outcomes, the EC arm had generally higher abstinence rates, but the difference only reached statistical significance for abstinence at 4 weeks and for self-reported abstinence at EOP. Regarding regular use of non-allocated products, six self-reported PP abstainers in the EC arm and 25 in the NP arm reported such use. With these participants excluded, all differences between the two study arms became significant (Table 5). Results for abstinence at 4 weeks and abstinence over 2-month pre-EOP are in Appendix 4.

We also conducted an exploratory post hoc sensitivity analysis that assumed that abstainers using non-allocated products would not succeed in stopping smoking without such use and instead of excluding them, included them as non-abstainers. This approach allows the inclusion of the whole sample and maintains randomisation and statistical power. The results are shown in Table 6.

The rates of at least a 50% reduction in cigarettes smoked per day accompanied by at least a 50% reduction in salivary cotinine compared to baseline in participants who did not achieve abstinence were very low (in part due to the problem with saliva sampling) and did not differ between the study arms. Self-reported smoking reduction favoured the EC arm, but the difference only became significant when reducers using non-allocated products were excluded (Table 7).

Treatment adherence

Regarding treatment adherence, about 30% of participants did not set a TQD, with rates similar in the two study arms (Table 8). The uptake of support phone calls was low in both study arms. Product use was initially also low, especially in the NP arm. More participants returned to using their products later in pregnancy, with product use being significantly higher in the EC arm across different indicators. Additional product use data are in Appendix 5. A third of participants in the EC arm were using ECs at EOP, compared to 6% using NRT in the NP arm. Use of non-allocated products mirrored the difference, with significantly more participants in the NP arm becoming EC users than the other way round.

Table 8 shows the product use among the full sample. We also looked at product use in participants reporting abstinence from smoking at EOP. Among this sample, 58 (49.2%; 57 allocated and one non-allocated) participants in the EC arm were using a nicotine product, either ECs or NRT. In the NP arm, 15 (19.2%; five allocated, eight non-allocated and two both) reported such use [χ² (df = 1) = 18.0; p < 0.001].

Among various NRT products, NPs, which were the starting product, remained in use by far the most frequently. Of the 238 participants in the NP arm who reported at EOP that they had used NRT during pregnancy, 236 (99.2%) used NPs, including 16 who used a combination of NPs with other NRT products; one used only a mouth spray and one used only an inhaler. Regarding patch strength, of 351 patch products dispensed by the study team, only 29 (8.3%) were 10 mg NPs, while the rest were 15 mg NPs.

Among available EC products, refillable ECs, which were the starting product, remained by far the most frequently used. Among 344 EC users in the EC arm, 94.2% used refillable ECs. E-liquids with a higher nicotine content (11–20 mg/ml) and tobacco and fruit flavours were popular (Table 9).

A total of 244 participants provided information on their products at 4 weeks and at EOP. Nicotine concentrations in their e-liquids decreased significantly over time [Bhapkar χ² (df = 2) = 32.0; p < 0.001].

Safety outcomes

There were 1095 singleton births and 13 pairs of twins (nine in the EC and four in the NP arm). Safety data were available from 1110 women (97.4% of the sample; 97.4% in each arm). A total of 39 participants (20 in the EC and 19 in the NP arm) delivered infants in non-study sites and no data were available on birthweight for 10 of them and on gestational age and birthweight for 29. Two women (one in each arm) had an elective termination and were excluded from the analyses.

Participants in the NP arm had more infants with low birthweight (<2.5 kg; see Table 10), BF = 10.3. In a post hoc exploratory analysis, we also compared the proportions of babies with birthweight below the 10th centile (2.38 kg). This too was more frequent in the NP (10.9%) than in the EC arm (7.1%); RR = 0.65, 95% CI 0.44 to 0.96.

The rates of other adverse birth outcomes and mean birthweight were similar in the two study arms. The analysis including twins provided similar results (see Appendix 6).

The two study arms were also similar in other AEs (Table 10). SAEs and AEs that occurred fewer than three times and that are marked as ‘Other’ in Table 11 are listed in Appendix 7. ARs deemed possibly related to study products consisted primarily of skin irritation and nausea in the NP arm, and cough and throat irritation in the EC arm.

There were no differences between the two study arms in overall number of SAEs and AEs (476 vs. 479 in the EC and NP arm, respectively) or in the number of participants experiencing any SAEs or AEs (285 vs. 292 in the EC and NP arm, respectively; RR = 0.97, 95% CI 0.87 to 1.09).

In the sample of participants who answered the questions regarding respiratory symptoms (n = 64 in the EC arm and n = 79 in the NP arm), 84.4% in the EC arm and 79.8% in the NP arm reported new symptoms since the start of treatment (RR = 1.06, 95% CI 0.91 to 1.23). Among those who rated the severity of their symptoms (n = 53 in the EC and n = 63 in the NP arm), 5.7% in the EC arm and 19.1% in the NP arm reported experiencing severe symptoms (symptoms stopped them from carrying out everyday activities ‘a lot’ vs. ‘a little’ or ‘had no effect’).

Effect of e-cigarettes on smoking cessation

The measure that comprised the primary outcome required a biochemical validation of abstinence from smoking via salivary sampling. This required participants to read the instructions on the sampling kit posted to their address, provide a sample in sufficient quantity and mail the sample back to the study team. The compliance with this request was limited, with only about half of the eligible participants providing useable samples. It is possible that the timing of the sampling was particularly inconvenient for women in late pregnancy or with newborn babies. Future studies should consider other approaches. Due to this issue, the validated sustained abstinence rates were low. The difference favoured the EC arm, but it was not statistically significant.

Six times as many participants in the NP arm used ECs regularly during pregnancy than participants in the EC arm who used any form of NRT (18% vs. 3%). A pre-specified sensitivity analysis of the primary outcome that excluded abstainers who regularly used the non-allocated product showed a significantly higher abstinence rate in the EC arm (RR = 1.93, 95% CI = 1.14 to 3.26). When using the same adjustment for other abstinence outcomes, ECs were almost twice as effective as NPs across different outcome measures (RRs ranging from 1.89 to 2.03).

The two treatments thus did not differ significantly in the primary outcome, but the effect of ECs appears to have been masked by EC use in the NP study arm. When this was controlled for, ECs were superior to NPs. The difference between the two treatments in biochemically validated abstinence (7% vs. 4%) is low because of the saliva sampling problem, but the difference in self-reported 7-day PP abstinence at EOP (21% vs. 14%, or 20% vs. 10% with adjustment for contamination) suggests a clinically important effect.

Behaviour change trials that compare an old and a new treatment are at a risk that some participants who did not benefit from the old treatment join the trial to access the new treatment and if allocated to the old one, are disappointed and drop out. In this trial, similar proportions of participants had previous experience with NRT and ECs (just under 50%). They all stopped using the products and continued to smoke. Familiarity with the two products and the rate of previous failure with them was thus similar. Reassuringly, the proportions of participants in the two study arms who never set up a TQD were also similar in the two study arms. However, more participants in the NP arm never tried their allocated product (32% vs. 15%). This, however, seems to have been due to participants being keen to try ECs, rather than not wanting NPs, because NP use was higher than in the previous large UK Smoking, Nicotine and Pregnancy (SNAP) trial of nicotine versus placebo patches that recruited pregnant smokers in the same way and from largely the same locations. 10,12,50–53

For any treatment to make an impact, clients must be willing to use it. In previous studies of pregnant smokers, treatment adherence was limited. 12 For instance, in the SNAP trial behavioural support was rarely used and only 14% (72 out of 521) of participants in the NP arm requested further patches after the first 4-week supply. 12 Here, the adherence to behavioural advice was also low. Almost a third of participants did not set up a TQD, and the response to support calls was low. In the NP arm, 23% of participants used their allocated products for at least 4 weeks. ECs had a higher initial appeal, with 40% of participants in the EC arm using their allocated product at 4 weeks. The two products further diverged during the follow-up period. At EOP, 6% and 34% were using their products in the NP and EC study arms, respectively. The contrast was even stronger among self-reported EOP abstainers, with 9% in the NP arm versus 57% in the EC arm using their allocated product at EOP. Interviews with trial participants suggested that positive beliefs about the necessity of vaping to achieve smoking cessation outweighed concerns about vaping. 54 The difference at EOP could have been in part at least due to participants expecting that NPs are supposed to be used for only 3 months, although in this study, participants in both study arms were encouraged to use their products for as long as needed. They also needed to ask for a prescription for NRT, while this was not needed for ECs. On the other hand, participants were able to obtain NPs and any other NRT they asked for free of charge, but had to source EC supplies and pay for them themselves. The number of participants who switched to regular use of the non-allocated product can be considered another indication of treatment attractiveness. This was much more frequent in the NP arm. As in previous studies with general cohorts of smokers,39,55 ECs seem to be a more attractive option than NPs in this client group as well.

The higher rate of ongoing use of ECs compared to ongoing use of NRT raises the question of what effects this may have over time on participants who stopped smoking, as well as on those who became dual users. Smokers who switch to dual use can be expected to reduce their smoking and toxicant intake and so gain a degree of harm reduction. 56–58 The effect on ex-smokers is less clear. It could be negative because EC use is likely to carry some health risks if used over an extended time,59 and for some vapers at least, the cost and/or compulsive nature of ongoing nicotine use is likely to be undesirable. Ongoing EC use, however, could also have some positive effects if it helps with reducing irritability39 and weight gain60 that can accompany cessation of nicotine use and possibly help with maintaining enjoyment that was previously derived from smoking. 61 Perhaps the key issue is whether extended EC use facilitates or prevents relapse back to smoking. Future studies should examine quality of life, relapse risk and health outcomes in comparable ex-smokers who did and did not switch to EC use.

Regarding the type of EC products used by participants, refillable ECs were used almost exclusively. In the USA, pod-based ECs, such as JUUL™ (Juul Labs Inc., San Francisco, CA, USA), are now the most popular EC product. 62 JUUL has a high nicotine content (59%) and provides nicotine in a way similar to cigarettes. 63 The UK, however, is currently subject to European Union (EU) regulations that ban ECs with nicotine content above 20%. This means that pod products on sale in the EU provide only low nicotine levels that are unlikely to be helpful for smokers trying to stop smoking and that indeed did not become popular in the UK. 63

As in previous studies,39,55 fruit flavour was the most popular EC flavour choice. The study also replicated an unexpected previous finding39 of EC users reducing nicotine content of their e-liquid over time. This could be the result of a conscious effort at weaning oneself off nicotine, but it could also be due to an improved vaping technique64 or using more effective EC devices.

Regarding treatment costs, it is worth noting that while ECs were more costly for participants, NRT was much more costly to treatment providers.

Safety of e-cigarettes compared to safety of nicotine patches

Existing data on safety of ECs in pregnancy come from observational studies. In two reports, infants of exclusive vapers had a higher risk of smallness for gestational age65 and a higher incidence of low birthweight and pre-term birth66 compared to non-smokers. The findings are difficult to interpret as most or all vapers had been smoking during early pregnancy, and no comparison is provided with ex-smokers who quit without using ECs. One study found birthweight of infants born to exclusive vapers matching that of infants of never-smokers, and significantly higher than in infants of smokers. 67 A study of Neonatal Behavioural Assessment Scale scores reported a greater number of abnormal reflexes in infants of both smokers and EC users compared to non-smokers. 68 This could be related to EC use, but as in the two studies above, the finding could reflect differences between smoking and non-smoking mothers, or tobacco exposure in early pregnancy.

In our study, AEs and pregnancy outcomes were similar in the two study arms, but there were more infants born with low birthweight (<2500 g) in the NP arm. The BF of 10.3 indicates strong evidence for the effect, though a possibility needs to be considered that as the two study arms were compared on a number of health outcomes, the finding could be an artefact of multiple testing. The most likely explanation of the finding is that participants in the EC arm reduced their smoking more than those randomised to NPs. The finding echoes the result from the SNAP trial mentioned earlier, where infants born to women randomised to NPs had better outcomes than those randomised to placebo patches. 69 The secondary analyses could not explain this finding in terms of smoking cessation70 but smoking reduction was not measured and it is plausible that women on NPs reduced their smoking more than those on placebo patches. The potentially important implication of both of these findings is that the risks of negative birth outcomes in smokers is due to other chemicals in cigarette smoke rather than nicotine.

It is important to note that these results only concern effects of nicotine in later pregnancy. All study participants were smoking for at least the first 3 months of pregnancy and so were exposed to both nicotine and to other tobacco chemicals early on. The results do not rule out possible detrimental effects of nicotine during these early stages. However, they provide reassuring evidence that providing pregnant smokers with alternative nicotine delivery devices such as NRT or ECs does not generate any additional risks and is likely in fact to reduce the risk of low birthweight. In addition, a recent study looking at women who stopped smoking at various stages of their pregnancy showed that smoking during the second trimester or through the entire pregnancy is associated with a higher incidence of low birthweight. 71 Our results suggest that nicotine is not implicated in these effects. If the effects of smoking in the first trimester are due to the same processes, nicotine on its own may not affect intrauterine growth.

Trial strengths

This was a large ‘real-world’ trial. The participants were representative of the population of pregnant smokers in the UK and the two interventions were delivered in a way that is economic and practicable and can be applied in routine care. The trial included detailed examinations of smoking behaviours as well as monitoring of product safety and of birth outcomes for both women and infants.

Trial limitations

Validation of smoking status via postal saliva sampling proved problematic. Almost half of eligible participants did not provide useable samples, which led to low primary outcome abstinence rates and reduced the study power. The trial results concern primarily NPs rather than NRT combinations. Although participants were encouraged to use additional NRT products, this was used only rarely. In non-pregnant smokers, combinations of patches with other NRT products were shown more effective than single NRT. 72 Participants could only access ECs with a maximum of 20 mg/ml nicotine because EU regulations do not allow higher nicotine concentrations. The results may not generalise to modern ‘pod’ EC products with higher nicotine delivery.

Implications for health care

Although nicotine in late pregnancy may not have any detrimental effects on pregnancy outcomes, given the question marks regarding possible effects of continuing nicotine use on quality of life, health outcomes and risk of relapse, stopping smoking without nicotine-containing aids is preferable to switching to such products. However, where the choice is between using nicotine products such as NRT or EC, or continuing to smoke, nicotine product use would be the recommended option. Specialist SSSs should include EC starter packs among the treatment options offered to pregnant smokers. Such an offer is likely to reach more smokers and generate better smoking reduction at lower cost than the offer of NPs.

Recommendations for research

As noted earlier, future studies should avoid postal saliva sampling. Regarding use of non-allocated products, this is likely to continue to occur in future studies. Researchers should pre-specify how this will be controlled for. We opted for excluding abstainers who used the non-allocated products, but re-classifying them as non-abstainers could be a better approach, as discussed above. Future studies may also consider partially randomised patient preference design. The inclusion of long-term follow-up of the offspring would address concerns about effects of nicotine on later offspring development. Regarding more general research recommendations, in this field, long-term health effects of EC use are the main current research priority. Studies are needed that compare biomarkers or risks and eventually also long-term health outcomes, quality of life and relapse rates in comparable ex-smokers who either stopped smoking without switching to EC use, or stopped smoking and use ECs.

Conclusion

In the unadjusted primary analysis, there was insufficient evidence to confidently demonstrate that ECs are more effective in helping pregnant smokers quit than NPs. EC effects appear to have been masked by EC use in the NP arm. When this was controlled for, ECs were more effective than patches. Regarding product safety, ECs do not seem to pose more risks to birth outcomes that were assessed in this study than NPs and may reduce the incidence of low birth weight.

Contributions of authors

Dunja Przulj (https://orcid.org/0000-0003-1133-8835) (Research Fellow) was the study manager. She co-wrote the original grant application, co-designed the trial, trained staff, delivered the interventions, contributed to data collection and interpretation of the trial findings and assisted with the drafting of the report.

Francesca Pesola (https://orcid.org/0000-0002-2054-7930) (Study Statistician) led the statistical analysis plan, analysed the trial data and assisted with drafting of the report.

Katie Myers Smith (https://orcid.org/0000-0003-1837-3924) (Senior Research Fellow) was the study manager and co-wrote the original grant application, co-designed the trial, co-wrote the statistical analysis plan, trained staff, delivered the interventions, contributed to data collection and assisted with interpretation of the trial findings and drafting of the report.

Hayden McRobbie (https://orcid.org/0000-0002-7777-1845) (Professor in Public Health Interventions) was a co-investigator. He co-wrote the original grant application, co-designed the trial and assisted with interpretation of the trial findings and drafting of the report.

Tim Coleman (https://orcid.org/0000-0002-7303-4805) (Professor of Primary Care) was a co-investigator. He co-wrote the original grant application, co-designed the trial and assisted with interpretation of the trial findings and drafting of the report. Professor Coleman is an NIHR Senior Investigator.

Sarah Lewis (https://orcid.org/0000-0001-5308-6619) (Professor of Medical Statistics) was a co-investigator. She co-wrote the original grant application, co-designed the trial, co-wrote the statistical analysis plan and provided statistical oversight and assisted with interpretation of the trial findings and drafting of the report.

Christopher Griffiths (https://orcid.org/0000-0001-7935-8694) (Professor of Primary Care) was the chief investigator (from June 2019 onwards). He oversaw the management of the trial and assisted with interpretation of the trial findings and drafting of the report.

Robert Walton (https://orcid.org/0000-0001-7700-1907) (Clinical Professor of Primary Health Care) was the chief investigator (from the start of the study to June 2019). He oversaw the management of the trial and assisted with interpretation of the trial findings and drafting of the report.

Rachel Whitemore (https://orcid.org/0000-0002-7599-8640) (Senior Trial Manager) was a site co-ordinator for the English sites. She identified recruiting sites, conducted site staff training and monitoring and assisted with the drafting of the report.

Miranda Clark (https://orcid.org/0000-0002-6179-046X) (Senior Trial Manager) was a site co-ordinator for the English sites. She identified recruiting sites, conducted site staff training and monitoring and assisted with the drafting of the report.

Michael Ussher (https://orcid.org/0000-0002-0995-7955) (Professor of Behavioural Medicine) was a co-investigator. He co-wrote the original grant application, co-designed the trial and assisted with interpretation of the trial findings and drafting of the report.

Lesley Sinclair (https://orcid.org/0000-0002-2210-8181) (Research Fellow/Trial Manager) was a site co-ordinator for the Scottish sites. She identified recruiting sites, conducted site staff training and monitoring and assisted with the drafting of the report.

Emily Seager (https://orcid.org/0000-0001-6672-5555) (Research Health Psychologist) was a research assistant who delivered the interventions and contributed to data collection.

Sue Cooper (https://orcid.org/0000-0002-1994-6395) (Principal Research Fellow and NIHR Programme Manager) was a co-investigator. She co-wrote the original grant application, co-designed the trial and assisted with interpretation of the trial findings and drafting of the report.

Linda Bauld (https://orcid.org/0000-0001-7411-4260) (Bruce and John Usher Professor of Public Health) was a co-investigator. She co-wrote the original grant application, co-designed the trial and assisted with interpretation of the trial findings and drafting of the report.

Felix Naughton (https://orcid.org/0000-0001-9790-2796) (Associate Professor in Health Psychology) was a co-investigator. He co-wrote the original grant application, co-designed the trial and assisted with interpretation of the trial findings and drafting of the report.

Peter Sasieni (https://orcid.org/0000-0003-1509-8744) (Academic Director of King’s CTU and Professor of Cancer Prevention) was a co-investigator and lead of the CTU providing trial oversight. He co-wrote the original grant application, co-designed the trial and assisted with interpretation of the trial findings and drafting of the report.

Isaac Manyonda (https://orcid.org/0000-0003-3080-2357) (Consultant in Obstetrics and Gynaecology) was a co-investigator. He co-wrote the original grant application, co-designed the trial and assisted with interpretation of the trial findings and drafting of the report.

Peter Hajek (https://orcid.org/0000-0001-9160-4296) (Professor of Clinical Psychology) was the PI. He led the original grant application, co-designed the trial and led the drafting of the report.

Ethics statement

The study was approved by the National Research Ethics Service Committee London – South East (ref: 17/LO/0962) and the Medicines and Healthcare Regulatory Agency via the CTIMP (Clinical Trial of an Investigational Medicinal Product) Notification Scheme. The trial was conducted in compliance with the Medicines for Human Use (Clinical Trials) Regulations 2004 (SI 2004/1031), Research Governance Framework, Good Clinical Practice guidelines, and the World Medical Association Declaration of Helsinki (1996).

Data-sharing statement

All data requests should be submitted to the corresponding author for consideration. Access to anonymised data may be granted following review.

Disclaimers

This report presents independent research funded by the National Institute for Health and Care Research (NIHR). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, the HTA programme or the Department of Health and Social Care. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, the HTA programme or the Department of Health and Social Care.