Smoking during pregnancy is three times more prevalent among Indigenous Australian women than non-Indigenous Australian women (52% v 16%).1 There is a similar increased risk of indoor exposure to tobacco smoke for Indigenous children aged 0–14 years compared with non-Indigenous children (28% v 9%).1
Tobacco use during pregnancy is associated with low birthweight and adverse perinatal health outcomes.2,3 In children, exposure to second-hand smoke (SHS) from tobacco contributes to lower respiratory tract illness, otitis media and chronic middle ear effusion,4,5 and is associated with an increased risk of childhood asthma.6,7
For many women, pregnancy acts as a significant prompt to quit smoking.8,9 Around one in five pregnant smokers quit by the time of their first antenatal visit.10,11 Those who experience social disadvantage, and those who are heavy smokers, less educated, or exposed to SHS in the home are less likely to quit.12,13
Pregnancy may be less of a stimulus to quit smoking for Indigenous Australian women.14 The extent to which smoking is regarded as normal in some Indigenous communities, and high rates of multiple life stressors, teenage pregnancy and single motherhood act to reinforce smoking behaviour in this population.15-18 However, findings from one qualitative study suggested that Indigenous mothers are attentive to actions that protect newborns and young infants from exposure to SHS (eg, establishing smoke-free homes).15
Study participants were a cohort of pregnant Indigenous women from two urban and four remote communities in the NT (68% from Darwin) who were enrolled in a randomised controlled trial of a maternal vaccine between 30 June 2006 and 4 May 2010. The randomised controlled trial has been described elsewhere.19 Eligible participants for the vaccine trial, and consequently for this study, were Indigenous women aged 17–39 years who:
were having a singleton uncomplicated pregnancy (eg, not complicated by significant pre-eclampsia with hypertension, proteinuria and oedema);
lived in one of the two urban and four remote participating communities;
had not received the 23-valent pneumococcal polysaccharide vaccine within the previous 3 years;
intended to live in the study area during the follow-up period; and
were not HIV positive, and had no history of severe allergy, uncontrolled asthma or splenectomy.
Analyses were performed using Stata, version 11 (StataCorp, College Station, Tex, USA). Depending on their distribution, numerical data are described using mean and SD or median and interquartile range (IQR). The median number of smokers in the household was compared between women who were smokers and non-smokers during pregnancy using a Mann–Whitney U test. We used a random-effects logistic model to examine changes in smoking status among women over time, incorporating age, education and community (urban v remote) as covariates.
By 1 June 2010, 162 of the 215 women recruited into the vaccine trial had completed their last study visit at 7 months postpartum (Box 1). Demographic characteristics of the 215 participants are shown and compared with those of reference populations in Box 2.
Overall, 99 of the 215 women (46%) reported smoking during pregnancy, with a median of seven cigarettes per day (IQR, 3–15). Most smokers (70/99; 71%) indicated that they smoked 10 or fewer cigarettes per day. Those who smoked during pregnancy lived with a median of two household smokers, whereas non-smokers lived with a median of one household smoker (P < 0.001).
Of the 66 women who were non-smokers at the antenatal visit, 16 (24%; 95% CI, 14%–35%) reported smoking at 2 months postpartum, and 23 (35%; 95% CI, 23%–47%) were smokers by the time their baby reached 7 months of age (Box 3). Of the 55 women who were smokers at the antenatal visit, 10 (18%; 95% CI, 7.6%–29%) had quit at 1 month postpartum but only two (3.6%; 95% CI, − 1.5%–8.7%) were not smoking by the time of the 7-month visit.
Two random-effects logistic regression models that allowed for dependence between observations within subjects, adjusted for age, education and community location (urban v remote), were also performed. In one model, the missing data were excluded while in the other model, missing data were replaced by the last-known smoking status. Both models confirmed the trends seen in Box 3 (results available from the authors). Mothers with a higher level of education were less likely to smoke (P = 0.003).
Antepartum, 78% of participants (166/214) had smokers living in their households (median of two smokers at home). When limited to 146 households with exposure to cigarette smokers over time (ie, participants who completed a 7-month visit and had at least one smoker in the household at any time point), 69% (101/146) reported being smoke-free antepartum.
Household smoking trends postpartum: The prevalence of exposure to smoking in participants’ households was: 31% (95% CI, 23%–39%; 45/146 households) at the antenatal visit; 12% (95% CI, 6.9%–20%; 14/114 households) at the 1-month visit; and 16% (95% CI, 10%–23%; 23/145 households) at the 7-month visit. The denominator differential over these visits reflects missing data at the study visit time points (Box 1).
Box 4 contrasts the trend in the proportion of women smokers with the proportion of households which allowed family members to smoke inside during the postpartum period.
In this report, we describe important trends in maternal smoking among Indigenous women and exposure of infants to SHS in the postnatal period to 7 months. Among Indigenous women who did smoke during pregnancy, most were light smokers but tended to live with a higher number of other household smokers than did women who did not smoke during pregnancy. Our findings are consistent with those of other researchers who found that, while Indigenous women who smoke during pregnancy have low to medium nicotine dependency,16,23 the number of smokers in the household is significantly associated with smoking among pregnant Indigenous women.24
Our results also highlight that the early months of an infant’s life are a period in which ex-smoker mothers and perhaps some non-smoker mothers will restart or start smoking. We cannot determine whether the increased number of smokers postpartum represents the addition of new smokers or relapsing smokers, but we expect that most were relapsing smokers who quit before or during their pregnancy. In other settings, 65%–80% of smokers who were abstinent during pregnancy started smoking again before their baby was a year old.11 Relapse postpartum is predicted by similar determinants as smoking in pregnancy; in particular, poor education and being exposed to social cues to smoke in the home,10,11 both common scenarios in Australian Indigenous populations. In our sample, mothers with a higher level of education were less likely to smoke postpartum.
Additionally, our findings suggest that, while most Indigenous households include smokers,21 most pregnant Indigenous women and new Indigenous mothers report smoke-free homes, reflecting recent estimates from the NT.21 It appears that the birth of a child does spur some households to change their smoking behaviours, as the proportion of households that allowed family members to smoke inside decreased notably after the birth. This supports qualitative research among pregnant Indigenous women in Western Australia, where there was greater awareness and more actions relating to the effects of smoking on a baby than there was in relation to smoking in pregnancy.15
Our study was a non-random sample, so the extent to which our results can be generalised is uncertain. Participants were generally more educated than the NT Indigenous female population,21 which suggests that the true prevalence of smoking during pregnancy and of household exposure to SHS may well be higher in the NT Indigenous community. We also relied on self-report of smoking status, and this has been shown to underestimate numbers of pregnant smokers.25 Our data on smoking behaviours of mothers and household members were limited — we did not collect data on pregnant women’s smoking history or on previous quit attempts. Finally, our study was powered to test for other primary outcomes, and not the analyses reported here, so there were wide confidence intervals around some of our estimates.
Despite these limitations, we believe the findings support the need for more attention to smoking in pregnancy and postpartum relapse in this population. At a health care service level, a “continuum of care” model of smoking cessation and relapse prevention is required, that starts before conception or in early pregnancy and continues through to the postnatal period. Health care providers need to prioritise smoking-cessation support to Indigenous women around the time of conception, especially as interventions can reduce smoking in pregnancy by 6%, and reduce low birthweight (risk ratio, 0.83; 95% CI, 0.73–0.95).8 In the Indigenous context, interventions should focus on counselling and support strategies to address smoking dependence, barriers to quit, and relapse triggers.23 Ideally, such strategies would involve families and households, and would address sources of stress within the immediate environment, as well as provide alternative coping strategies and ongoing support to better manage personal circumstances.26 More rigorous research is required to evaluate intensive, theoretically sound interventions to prevent relapse among recent quitters in the postnatal period, and to protect the newborn and any other children living in the household against the harms of SHS.27
1 Flowchart showing recruitment and involvement of study participants
* Limited to participants who lived in a household with at least one smoker at any time point.
Provenance: Not commissioned; externally peer reviewed.
Abstract
Objective: To describe the trends in maternal smoking and smoking in the household for a cohort of Indigenous women followed from late pregnancy to 7 months postpartum.
Design and setting: Prospective cohort study embedded within a randomised controlled trial (RCT) performed in the Northern Territory involving participants recruited between 30 June 2006 and 4 May 2010.
Participants: 215 Indigenous women aged 17–39 years who had been recruited into the RCT, 162 of whom had completed their last study visit at 7 months postpartum by 1 June 2010.
Main outcome measures: Smoking status of women, and smoking within their households, in their third trimester, and at 1 month, 2 months and 7 months postpartum.
Results: There were complete data on women’s smoking status for 121 participants. Among these, the self-reported smoking rate was 45% (95% CI, 36%–55%) during pregnancy, increasing to 63% (95% CI, 54%–71%) at 7 months postpartum. Of the 66 women who were non-smokers at the antenatal visit, 23 (35%; 95% CI, 23%–47%) were smoking by the time their baby reached 7 months of age. Thirty-one per cent (95% CI, 23%–39%) of households included people who smoked inside during the antepartum period, whereas 16% (95% CI, 10%–23%) included people who smoked inside at 7 months postpartum.
Conclusions: While an apparent reduction in indoor exposure to tobacco smoke during the postpartum period is encouraging, this is offset by an increase in the proportion of antenatal non-smokers who subsequently reported smoking after the birth of their child. More health care service delivery and research attention needs to be directed to smoking during pregnancy and to postpartum relapse in this population.