Reliance on self-reporting underestimates pregnancy smoking rates in Scotland, with more than 2400 pregnant smokers estimated to be missed each year
- Correspondence to Van T Tong
Centers for Disease Control and Prevention, Division of Reproductive Health, 4770 Buford Highway NE, MS K-22, Atlanta, GA 30341, USA;
Smoking during pregnancy is a preventable cause of infant morbidity and mortality, yet prenatal smoking is a persistent problem, with an estimated 6–22% of women in high-income countries reporting smoking cigarettes during pregnancy.1 Clinic-based interventions are effective in increasing cessation and improving pregnancy outcomes2; however, delivery of these interventions depends on the ability of providers to assess a woman's smoking status and readiness to quit accurately.
Shipton and colleagues sought to estimate the biochemically validated prevalence of women in need of smoking cessation services. Using a random sample of all 2004 births from the West of Scotland, they compared routinely collected self-reported smoking status of pregnant women at maternity booking (8–12 weeks of gestation) and serum cotinine levels from prenatal screening (15–16 weeks of gestation). They found that maternal self-reported smoking status underestimated smoking prevalence by 25%. The authors' findings are consistent with previously published clinical and observational studies in which maternal non-disclosure of smoking status was high.3 Because women in the current study did not know in advance that their smoking status would be validated with biochemical measures, the non-disclosure rates reported here can probably be generalised to similar clinical settings in which women's self-reported smoking status is not biochemically validated.
The authors suggest that more accurate methods are required to identify pregnant smokers for cessation services. Glasgow has recently implemented routine carbon monoxide breath tests at maternity booking. However, this policy raises a number of important issues. One major difficulty in the application of biochemical validation of smoking status is determining the best cut point to identify active smokers. In the USA, for example, it was recently recommended that the cotinine cut point for active smoking be lowered from 14 to 3 ng/ml in serum, reflecting the reduction in heavy second-hand smoke exposure among non-smokers in the USA.4 Ideal cut points for identifying active smokers may vary among countries and settings. Further complicating matters is the fact that cotinine concentrations are decreased by approximately 40% during pregnancy because of increased metabolism of cotinine.5 Guidelines for cut points of cotinine in pregnant women have not been updated in recent years. Although carbon monoxide breath tests are relatively inexpensive and easy to implement, they have lower sensitivity and specificity for detecting active smokers than cotinine testing using serum, urine or saliva.6 In addition, research suggests that asking about smoking in particular ways can increase disclosure rates in smokers. For example, compared with asking a yes/no question to screen for tobacco use, a multiple-choice question that includes categories for quitting and reducing number of cigarettes smoked yields more accurate self-reports.7 Before biochemical validation of smoking status is implemented in routine prenatal care, local decision-makers should weigh the costs, ease of administration and ability of a particular biomarker accurately to identify smokers accurately.6
Perhaps the most challenging aspect of determining whether to recommend biochemical validation of smoking status during pregnancy is the lack of evidence relating to smoking cessation interventions among women who do not disclose their smoking status to their providers. Smoking cessation trials have necessarily focused on women who admit to smoking and who are willing to participate in a cessation trial. It is unknown whether women who do not admit to smoking will benefit from traditional clinical interventions such as the Five As. Providers may also have difficulty confronting these women about their smoking and referring them to cessation services. More research is required on smoking cessation interventions for this population of women.
Although increasing referrals to cessation services is clearly important, it should be noted that such efforts may have only modest effects on overall prenatal smoking prevalence. It has been estimated that universal implementation in the USA of clinic-based augmented smoking cessation interventions for pregnant women would reduce smoking prevalence by only 1.4 percentage points (from 15.8% to 14.4%), even with 100% disclosure of smoking status.8 This reflects the modest effect of behavioural interventions on quit rates.2 Therefore, tobacco control efforts should also focus on prevention and helping women to quit smoking before they become pregnant, when they have more treatment options, such as pharmacotherapy.9 The WHO recommends that all countries implement a comprehensive tobacco control strategy (eg, increasing tobacco prices, smoke-free policies, health warnings, total bans on tobacco advertising and provision of cessation services) to reduce initiation and increase cessation of tobacco use among the population.10 These strategies will help to reduce prenatal smoking and may have a greater population impact than cessation services alone.
Competing interests The authors report no competing interests. The findings and conclusions are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.