Comparison of saliva cotinine and exhaled carbon monoxide concentrations when smoking and after being offered dual nicotine replacement therapy in pregnancy

Abstract Background and Aims Although English Stop Smoking Services routinely offer dual nicotine replacement therapy (NRT) to help pregnant women to quit smoking, little is known about how nicotine and tobacco smoke exposures following this compare with that from smoking. We compared, in pregnant women when smoking and after being offered dual NRT, saliva cotinine and exhaled carbon monoxide (CO) concentrations and numbers of daily cigarettes smoked. Design and Setting Secondary analysis of data from three sequential, observational, mixed‐methods cohort studies conducted as part of the Nicotine Replacement Effectiveness and Delivery in Pregnancy programme. Participants were recruited on‐line or in Nottingham University Hospitals (UK) antenatal clinics between June 2019 and September 2020. Participants Forty pregnant women, who agreed to try stopping smoking. Intervention Participants were offered dual NRT, agreed a smoking quit date and received an intervention to improve adherence to NRT. Measurements Saliva cotinine and exhaled CO concentrations and reported number of cigarettes smoked per day. Findings There were no differences in saliva cotinine concentrations at baseline and day 7 post quit date [n = 20, mean difference = −32.31 ng/ml, 95% confidence interval (CI) = −68.11 to 3.5 ng/ml; P = 0.074, Bayes factor = 0.04]. There were reductions in the reported number of cigarettes smoked per day (n = 26, mean difference = −7 cigarettes, 95% CI = −8.35 to −5.42 cigarettes, P < 0.001) and concurrently in exhaled CO concentrations (n = 17, ratio of geometric means = 0.30 p.p.m., 95% CI = 0.17–0.52 p.p.m.; P < 0.001). Conclusion Pregnant women who smoke and are offered dual nicotine replacement therapy (NRT) appear to show no change in their exposure to cotinine compared with their pre‐NRT exposure levels but they report smoking fewer cigarettes, as validated by reductions in exhaled carbon monoxide concentrations.


INTRODUCTION
Smoking during pregnancy is the largest modifiable risk factor for morbidity and mortality for both the mother and developing fetus [1,2]. It is causally associated with fetal growth restriction and is also associated with other pregnancy complications, including stillbirth, prematurity and miscarriage [3,4]. Parental smoking is a significant risk factor for smoking uptake in young people and smoking is a major risk factor for six of the eight leading causes of global mortality [5,6].
Despite this, pregnant women continue to smoke, with the proportion being 12 and 13% in the United Kingdom and United States, respectively [7,8]. Actual smoking rates are probably higher due to stigma causing an under-reporting of smoking [9].
UK smoking cessation guidelines recommend offering nicotine replacement therapy (NRT) to pregnant women who smoke if behavioural support has been ineffective [10], with guidance globally recommending similar approaches [11][12][13]. NRT is considered safer than smoking because nicotine is provided without the toxins found in cigarettes, such as tar and carbon monoxide (CO) [14]. Routine UK practice involves offering dual NRT as a first-line smoking cessation treatment [15].
NRT is an effective smoking cessation tool outside pregnancy, with dual NRT more effective than mono NRT [16]. NRT may be less effective in pregnancy, but there is some uncertainty surrounding this finding [17]. One hypothesis which may partially explain the apparent reduced efficacy of NRT in pregnancy is that in pregnancy there is acceleration of nicotine metabolism caused by an induction of nicotine metabolizing enzymes [18][19][20]. Consequently, for any given dose, NRT will generate a lower nicotine blood concentration in pregnant women. Such concentrations may be insufficient to alleviate smoking withdrawal symptoms and provide therapeutic benefits. To our knowledge, randomized controlled trials (RCT) investigating the use of NRT as a smoking cessation pharmacotherapy during pregnancy have almost all used a mono NRT regime [17,21]. It may be that the dose of NRT offered in mono NRT treatment regimes is insufficient.
Using dual rather than mono NRT in routine care could help more pregnant women to quit smoking; observational analyses of data from English Stop Smoking Services found being prescribed dual NRT during pregnancy was strongly associated with smoking cessation, whereas there was no association with mono NRT [22]. Women's acceptance and use of NRT can be limited by concerns about fetal harms caused by nicotine [23]. Although such concerns are valid and real to those that hold them, they are very probably unfounded; there is no evidence that nicotine used as NRT instead of smoking causes fetal harm [24]. More information regarding nicotine exposure following dual NRT and after smoking could help to reassure pregnant women about the extent of nicotine exposure when offered dual NRT, but this has so far only been investigated in a cohort of five women [25]. This study therefore aims to compare changes in the indicators of smoking intensity and nicotine exposure in pregnant women when they smoke with those after being offered dual NRT. To achieve this aim, we will compare: (1) cotinine concentrations when smoking with those measured after being offered dual NRT; (2) exhaled CO concentrations when smoking with those measured after being offered dual NRT; and (3) the number of daily cigarettes reported as being smoked, when smoking only and after being offered dual NRT.

Design
This is a secondary analysis of data from three sequential cohort studies in which a prototype intervention intended to improve pregnant women's NRT adherence was piloted on pregnant smokers. Cohorts were part of a larger programme, Nicotine Replacement Effectiveness and Delivery in Pregnancy (N-READY), which aims to test whether the intervention affected NRT adherence and smoking cessation. Ethical approval was given by the Nottingham 1 Research Ethics Committee.
Data were collected at baseline, when participants were smoking and then women were offered dual NRT to start using on an agreed quit date. Data were again collected 7 days post quit date. Figure 1 illustrates the study design and n-numbers at each stage. The preregistered research aims and analysis plan can be found at https://osf.

Participants
Women were eligible if pregnant and with less than 25 weeks' gestation, aged over 16 years, smoked at least 10 cigarettes per day pre-pregnancy and currently smoking, interested in receiving smoking cessation support, willing to try NRT and to set a quit date within 14 days. Women with less than 25 weeks' gestation were enrolled because the original cohort studies piloted study methods, and hence inclusion criteria, to be used in a future RCT to test the efficacy of the intervention being optimized in the cohorts. A gestation limit was specified because an outcome of the RCT will be a measure of prologued smoking cessation between a point earlier in pregnancy and childbirth, and so recruiting women before 25 weeks ensures that the period of smoking cessation monitored would be sufficient to be clinically meaningful. Participants were required to be able to give consent, understand English and own a smartphone, because the study intervention and data collection required participants to open links to websites sent via text. Women already enrolled into a cessation study or using NHS stop smoking support were excluded.

Eligible women were identified at Nottingham University Hospitals
Trust antenatal ultrasound and outpatient clinics. In the third cohort women were also identified through Facebook advertisements, a change necessitated by COVID-19 restrictions on face-to-face recruitment. Recruitment was initially planned to run from 1 June 2019 to 31 March 2020. Due to the first UK lockdown of 2020, collection of exhaled CO samples initially ceased. In the third cohort only, following lockdown, all data and samples were collected remotely, and data collection was completed by September 2020.

Intervention
Participants were offered dual NRT and behavioural support from QuickMist mouth spray). Participants were encouraged to use as much NRT as required to ameliorate symptoms of withdrawal and were advised to continue using NRT during smoking lapses of fewer than 14 days, provided they still aimed to quit smoking. To increase NRT adherence, participants received a prototype intervention integrated into routine NHS support comprising telephone consultations, leaflets, websites and text messages designed to reinforce NRT adherence.

Baseline
Researchers collected data including date of birth, ethnicity, qualifications, gestation and whether women had a partner who smoked.
Researchers asked the number of cigarettes smoked daily, use of ecigarettes, smoking in previous pregnancies and how soon after waking women smoked their first cigarette. In cohort 3, some participants declined to give an exact number of cigarettes smoked daily and instead reported a range (e.g. 8-10 cigarettes).
Researchers obtained saliva samples to measure cotinine and anabasine concentrations and an exhaled CO reading. Cotinine is a F I G U R E 1 Consolidated Standards of Reporting Trials (CONSORT) diagram illustrating the number of participants and data collection at each stage of the study. a Five of 20 women in cohort 3 dropped out after completing the baseline questionnaire and did not receive a stop smoking consultation/offer of NRT; b three of 10 women in cohort 2 could not give a carbon monoxide (CO) sample due to COVID-19 restrictions on faceto-face contact-CO samples were carried out remotely in cohort 3 proxy for nicotine exposure from any source, anabasine is a proxy for nicotine exposure from tobacco smoke and exhaled CO is a proxy for tobacco smoke exposure. In cohorts 1 and 2, after seeking baseline information and samples at a face-to-face consultation, the same researchers immediately delivered study smoking cessation interventions. In cohort 3, questions were asked via telephone and researchers sent out equipment for women to provide saliva and CO samples remotely and arranged appointments with different research team members for remote intervention delivery.

Day 7 post quit date
Using methods particular to each cohort described above we sought saliva and exhaled CO samples. We asked whether participants had A natural log-transformation of exhaled CO concentrations was used to achieve a normal distribution. Paired t-tests were used to assess 'within-participant' differences in saliva cotinine, exhaled CO and number of cigarettes smoked per day at baseline when smoking and day 7 post quit date after being offered dual NRT (alpha 0.05).
For exhaled CO, we present the back-transformed estimates of the baseline and follow-up differences, which is the ratio of the geometric means. Analyses were conducted using Stata version 15.
For the six cohort 3 participants who reported a range instead of an integer for the 'number of cigarettes smoked per day (CPD)', in the main analysis the lower bound of this range was imputed at baseline and the higher bound at follow-up. For example, a participant who reported smoking 8-10 CPD at baseline had this imputed as eight daily cigarettes at baseline, and if the same range was reported at follow-up, 10 was imputed. This was a conservative assumption, but we also report a sensitivity analysis using the upper bound at baseline and the lower bound at follow-up.
Upon analysis completion, we calculated Bayes factors from the differences in saliva cotinine in the whole cohort, saliva cotinine in abstinent women and exhaled CO in non-abstinent women [26]. An expected difference of 141 ng/ml in saliva cotinine concentration for the whole cohort was taken from a study of pregnant women using dual NRT [25]. A meta-analysis comparing nicotine exposure during pregnancy when smoking and when abstinent with NRT provided an expected difference of 75.3 ng/ml in saliva cotinine concentration in abstinent women [27]. An expected difference of 3 parts per million (p.p.m.) in exhaled CO concentration was taken from a study of pregnant women using NRT while currently smoking [28]. A conservative approach for estimation using a half-normal distribution was used, in which the SD is equal to the expected effect size. Figure 1 shows that 50, 189 and 199 (438 total) women were assessed for eligibility to join cohorts 1, 2 and 3, respectively. Forty participants joined the study and provided baseline demographic data, with eight,   Table 3 shows the same comparisons as Table 2    and 48 hours (CO) previously. Hence, with the relatively short followup used, we believe findings are probably valid. Although sample collection was more difficult than anticipated and the study smaller than planned, behavioural changes were marked, so there was still sufficient power to detect these. Additionally, although more heavily nicotine-addicted women failed to return for study treatments, dropout rates were similar to those in routine care to which findings generalize most readily. Finally, this study could be criticized for having researchers collect saliva and exhaled CO samples and then, for Unfortunately, only four paired exhaled CO samples were received when CO monitors were sent to participants for remote sample collection; however, it is probable these measurements were collected reliably, as participants were taught the correct processes [29,30].  [25]. Exhaled CO concentrations were reported as also lower, but no formal statistical comparisons were made and adherence to NRT was not monitored [25]. Our study T A B L E 2 Baseline to day 7 post quit date 'within participant' differences in indicators of smoking intensity. builds upon this work by demonstrating that women who have reported using dual NRT were exposed to less tobacco smoke.

RESULTS
Although, in our study, saliva cotinine concentrations were lower following the use of dual NRT, these differences were neither as marked nor were they statistically significant. This could be because, in the Hegaard study, women were instructed to stop NRT if they smoked at all, whereas we encouraged NRT use during smoking lapses. Our findings are consistent with a secondary analysis of RCT data, which found that in pregnant smokers, compared to when smoking, pregnant women using patch NRT had no greater nicotine exposure but smoked less [28]. Similarly, a systematic review found that, compared to when smoking, pregnant women using mono NRT and abstinent from smoking had reduced saliva cotinine concentrations [27].
Importantly, this study sought women motivated to stop smoking, principally from usual care settings, and study interventions were combined with usual cessation care, such that participants probably found these indistinguishable from routine NHS

CONCLUSION
This study was conducted to address concerns of pregnant women regarding nicotine exposure while using NRT in pregnancy. Compared to when smoking, pregnant women who are offered dual NRT show no increase in their exposure to nicotine, smoke less and exhale less CO, meaning their exposure to cigarette smoke toxins is reduced.
Even within pregnant women who admitted concurrent smoking we found, compared to smoking, an offer of dual NRT resulted in reduced nicotine exposure.