Physical activity as an aid to smoking cessation during pregnancy: Two feasibility studies

In the UK, 17% of women admit to smoking throughout pregnancy [1], but this is likely to be an underestimate since these data were collected retrospectively and many pregnant women that smoke deny it [2]. Smoking during pregnancy is harmful to the fetus, mother, and the child after birth. It is a major preventable contributor to miscarriage and stillbirth, being associated with 4000 deaths annually, as well as causing low birth weight with its attended risks [3, 4]. The risks to the mother, in addition to the long term risks reported for smokers in general, include placental abruption [5, 6]. The newborns of smokers have a higher perinatal mortality and are at an increased risk of neo-natal mortality, sudden infant death syndrome, learning difficulties in childhood, problem behaviour, asthma and childhood obesity [7–11].

In the UK approximately half of pregnant smokers stop for at least part of their pregnancy [1], but, of these, two thirds are likely re-start post-natally [12]. Women who smoke throughout pregnancy are more likely to come from economically disadvantaged backgrounds [13]. The most effective smoking cessation therapy in non-pregnant smokers combines behavioural support with nicotine replacement therapy (NRT), bupropion, nortriptyline, or varenicline [14–16]. Of these medications, only NRT has been licensed for use during pregnancy. A recent trial provides evidence that the addition of NRT to cognitive-behavioural therapy for pregnant smokers increases rates of smoking cessation [17]. However, there are concerns about NRT harming the foetus [18] and many women are reluctant to use NRT at all during pregnancy [19]. Behavioural support increases spontaneous smoking cessation rates in pregnancy by up to 7%, such that up to 15% of women who receive this abstain from smoking until the end of their pregnancy [20]. It would therefore seem that treatments that increase the effectiveness of behavioural support deserve concerted research efforts.

Most attempts to stop smoking, if unaided, end within a week [21] at least in part because of the strong urges to smoke experienced by those attempting to stop. Medications that are effective at smoking cessation reduce urges to smoke. A systematic review of short-term studies concluded that moderate intensity physical activity (equivalent to a brisk walk) reduced cravings in abstinent smokers [22]. When undertaken for at least 110 minutes per week during an eight week intervention, this activity has been associated with smoking cessation in non-pregnant women [23]. While there is no definitive evidence that physical activity improves abstinence rates, moderate physical activity is an attractive potential smoking cessation intervention in pregnancy because it is recommended [24] and might prove popular as it could prevent excessive weight gain [25]. There are no reported studies that have investigated physical activity for smoking cessation in pregnancy. Following recommendations for developing complex interventions [26], we conducted two uncontrolled, exploratory phase II studies investigating this topic. Our aim was to assess the feasibility of implementing a physical activity intervention for smoking cessation during pregnancy in terms of (i) recruitment, (ii) attendance, (iii) physical activity adherence and (iv) acceptability.

This paper reports the results of two feasibility studies, referred to as Study 1 and Study 2. The studies were conducted sequentially, and Study 2 was a modified version of Study 1, based on the experience gained from Study 1.

Details of recruitment for the separate studies are given below:

These two studies demonstrate, for the first time, that it is feasible to recruit pregnant smokers to exercise-based smoking cessation programmes. Our recruits were generally positive about both programmes, although the hospital-based scheme was viewed more positively than the programme delivered in women's homes. One quarter of the women recruited to either of the programmes quit smoking to the end of pregnancy, a rate which is similar to that shown for NRT in non-pregnant smokers [16]. However, this may be an optimistic estimate of the expected smoking cessation rates because we verified abstinence by expired CO rather than by a more reliable measure such as cotinine [50]. Moreover, the study did not include a control condition and quit rates might have been higher than usual had a control condition been used because these women were interested in increasing physical activity and therefore may have been more health conscious and more motivated to stop smoking than usual.

The proportion of women in the present study meeting government recommendations for physical activity (22%) was lower than the 36% reported in a recent survey of pregnant smokers [51]. This suggests that the current findings apply to reasonably sedentary pregnant smokers. Consistent with previous surveys of pregnant smokers, walking was the most popular form of activity [38, 39]. The overall percentage of women recorded as smokers at their first antenatal booking visit in this study (9%) is markedly lower than national data which shows that 17% of women continue to smoke throughout pregnancy [1]. This may be partly due to the large proportion of South Asian women in the hospital's antenatal population (national smoking rates have been reported to be as low as 4% among Asian women) [52]. There was a reduction of around 5% in the number of smokers recorded in the PAS in 2006 compared with 2000. This is consistent with national surveys in the UK showing that the proportion of women smoking throughout pregnancy fell from 19% in 2000 to 17% in 2005 [1]. However, these reductions may partly reflect poor rates of enquiry or reduced disclosure as a result of the increased stigma associated with smoking [2, 53]. Rates of disclosure might have been increased by presenting a multiple-choice question about smoking status [54]. The average level of cigarette consumption is consistent with national data [12].

Of those recorded as smokers at their first antenatal booking visit only 12% of the women were recruited to the two studies. In absolute terms this recruitment rate is low and it suggests that the intervention is unlikely to have a major impact on public health. However, the proposed intervention is still important since it is necessary to assess whether interventions for pregnant smokers are effective among those who accept them. We had not anticipated any higher recruitment rates, as according to UK statistics [1] half of all women stop smoking in pregnancy and nearly all of these do so prior to or early in pregnancy without any contact with antenatal care services and without formal support. We were therefore recruiting from the 'recalcitrant' half of women who are left, probably many of whom are reluctant to stop smoking and to receive help with stop smoking by any means. Thus, we are encouraged that our recruitment rate compares favourably with rates for dedicated stop smoking in pregnancy services of 5% for pregnant smokers in a national report [55], and 10% reported by the smoking cessation service in the same hospital from which we recruited. Consistent with a previous survey [51], our rate of recruitment suggests that the vast majority of pregnant women, including those from low-income groups, who are interested in receiving help with stopping smoking would be willing to join a study offering a physical activity intervention. Our findings also suggest that women would volunteer irrespective of whether they were allocated to the control group or the physical activity group. It noteworthy, that of the women recorded as smokers only around 60% could be contacted to be invited to join the study. Evidently, more effective methods are needed for making contact with these women. For example, the SNAP trial [56] proposes asking all pregnant women to complete a form at their ultrasound scan in order to gauge their interest in being recruited for a smoking cessation trial. Many pregnant smokers who report that they have quit smoking are likely to relapse during their pregnancy; therefore recruitment rates might have also been increased by contacting those who said that they had recently quit smoking [57, 58]. Additionally, several women were reluctant to join the study as crèche facilities were not available and future studies may need to offer these.

It is concerning that a substantial proportion of women agreed to join the study but failed to attend their first appointment. This may be partly due to the women agreeing to participate over the telephone, having had little time to consider the commitment. Future studies might reduce the number of early drop-outs by offering the women an introduction session prior to joining the study [29]. Our feasibility studies suggest that successful recruitment requires a multi-faceted approach. Telephoning women identified via the hospital database was effective. Advertising through posters and flyers did not appear to boost recruitment rates, nevertheless the latter approaches are easy to implement and forewarn women that they may be telephoned about the study. All the midwives were willing to refer pregnant smokers to the study and several women were recruited in this way, but it is not clear whether these women would have been recruited anyway via telephone contact. Presentations at nuchal scan clinics were relatively labour intensive and did not yield any recruits. This might have been due to the public nature of such presentations, with the smokers being unwilling to identify themselves in such public fora. In summary, our findings suggest that contacting the women by telephone or through their midwife are likely to be the most effective recruitment methods, from which we would expect to recruit around 50 smokers per year from a hospital with 5000 deliveries.

Many women relapse following childbirth [32] and future studies will need to assess whether a physical activity intervention can reduce this postpartum relapse. Studies might also wish to assess levels of smoking reduction, although many confounding variables will apply, and the studies will require careful design. A 50% or more smoking reduction is associated with increased infant birth weight [59] and physical activity may aid smoking reduction [60]. Additionally, it would be interesting to examine the impact of physical activity on peri-natal outcomes such as antenatal complications, duration of labour and birth weight, but the numbers required would be extremely large and such studies may not be feasible.

The women in our studies reported a reduction in the desire to smoke immediately after a bout of supervised exercise, both prior to quitting, and on the quit day following overnight abstinence. This is consistent with findings in non-pregnant smokers [21], and suggests a plausible mechanism through which physical activity might aid smoking cessation. If this can be shown to be a consistent finding, then controlled studies will be needed to assess how long this benefit is sustained after a bout of exercise. The women also reported mood enhancement after a bout of exercise and again this is encouraging, since enjoyment of exercise is likely to increase adherence [61]. This also suggests another avenue of research, since the mood enhancement suggests that the release of neurotransmitters (e.g. endorphins) may be implicated, and biochemical studies could address this issue, and suggest alternative avenues of therapy.

Reports of increases in levels of physical activity tended to be greater in response to the more extensive intervention used in Study 2, than for the Study 1 intervention, and this was especially the case at the final follow-up. Controlled studies with larger samples are required to compare the impact of different intensities of physical activity intervention on activity levels among pregnant smokers. These studies need to include reliable measures for verfying smoking abstinence, such as cotinine. More objective measures of physical activity, such as accelerometers, also need to be employed. Additionally, in order to reduce the possibility of participants providing socially desirable responses future studies should aim to have separate staff deliver the intervention and collect outcomes data. The interview findings suggest that, in general, the women found the physical intervention acceptable and reported many benefits from the intervention. At the outset, the majority of the women reported that they found the idea of regular exercise appealing and that they thought exercising is something that they should do, but only a minority of the women said they saw themselves as the kind of person who exercises regularly. This suggests that the intervention may need to shape the identities of these women towards being someone who is physically active. Future studies may also benefit by tailoring the intervention to accommodate pregnancy related barriers such as fatigue and feeling self-conscious and vulnerable, and these barriers need to be assessed more systematically.

Our findings suggest that a physical activity intervention is feasible and acceptable as an aid to smoking cessation during pregnancy. A definitive large randomised controlled trial is now required and is under way. If the intervention were shown to be successful for pregnant smokers it would provide a blue print for designing physical activity interventions for other populations of smokers.