Which measures of cigarette dependence are predictors of smoking cessation during pregnancy? Analysis of data from a randomized controlled trial

Abstract

AimsTo examine the ability of different common measures of cigarette dependence to predict smoking cessation during pregnancy.DesignSecondary analysis of data from a parallel‐group randomized controlled trial of physical activity for smoking cessation. The outcomes were biochemically validated smoking abstinence at 4 weeks post‐quit and end‐of‐pregnancy.SettingWomen identified as smokers in antenatal clinics in 13 hospital trusts predominantly in southern England, who were recruited to a smoking cessation trial.ParticipantsOf 789 pregnant smokers recruited, 784 were included in the analysis.MeasurementsUsing random‐effect logistic regression models, we analysed the effects of baseline measures of cigarette dependence, including numbers of cigarettes smoked daily, Fagerström Test of Cigarette Dependence (FTCD) score, the two FTCD subscales of Heaviness of Smoking Index (HSI) and non‐Heaviness of Smoking Index (non‐HSI), expired carbon monoxide (CO) level and urges to smoke (strength and frequency) on smoking cessation. Associations were adjusted for significant socio‐demographic/health behaviour predictors and trial variables, and area under the receiver operating characteristic (ROC) curve was used to determine the predictive ability of the model for each measure of dependence.FindingsAll the dependence variables predicted abstinence at 4 weeks and end‐of‐pregnancy. At 4 weeks, the adjusted odds ratio (OR) (95% confidence interval) for a unit standard deviation increase in FTCD was 0.59 (0.47–0.74), expired CO = 0.54 (0.41–0.71), number of cigarettes smoked per day 0.65 (0.51–0.84) and frequency of urges to smoke 0.79 (0.63–0.98); at end‐of‐pregnancy they were: 0.60 (0.45–0.81), 0.55 (0.37–0.80), 0.70 (0.49–0.98) and 0.69 (0.51–0.94), respectively. HSI and non‐HSI exhibited similar results to the full FTCD.ConclusionsFour common measures of dependence, including number of cigarettes smoked per day, scores for Fagerström Test of Cigarette Dependence and frequency of urges and level of expired CO, all predicted smoking abstinence in the short term during pregnancy and at end‐of‐pregnancy with very similar predictive validity.

Predictors of smoking cessation during pregnancy: a systematic review and meta‐analysis

“…Of the 40 observational studies, 19 were cross‐sectional and 21 were cohort studies . Fifteen studies used secondary data from RCTs .…”

“…All the studies recruited participants who were at least 16 years of age and the mean age of the women ranged from 16 to 50 years. Of the 17 studies (31.5%) that reported quit attempts, all or most of the participants made a quit attempt in 15 studies , while two other studies reported quit attempts for only 17% and 30% of the participants.…”

“…Sixteen studies, three observational and 13 RCTs , reported predictors of biochemically validated smoking abstinence in pregnancy, using salivary or urine cotinine and/or expired carbon monoxide (CO) or urine thiocyanate assays. Three of the observational studies and two of the RCTs reported predictors of spontaneous quitting during pregnancy without biochemical validation.…”

“…Ten of the observational studies and five of the RCTs had small sample sizes (range = 35–226) and (range = 81–289) , respectively, which were likely to detect only large effect sizes. Of the 13 trials with biochemical validation, nine were powered sufficiently (range = 316–957) to detect small–moderate effect sizes for predictors of abstinence .…”

Large multi-centre pilot randomized controlled trial testing a low-cost, tailored, self-help smoking cessation text message intervention for pregnant smokers (MiQuit)

“…Firth (penalized) logistic regression models 29 were then used to estimate ORs with 95% profile CIs 30 to compare smoking outcomes between treatment groups, adjusting for factors used to stratify the randomization via their inclusion as fixed covariates in each model (trial site, gestation at randomization). Three additional models for all seven smoking outcomes were carried out, each adjusting for one of three baseline variables associated commonly with smoking in pregnancy (heaviness of smoking, partner's smoking status and education) 31, 32, with likelihood ratio tests assessing whether these improved model prediction. Where convergence of a model could not be achieved due to low event rates within small centre sites, these centres were merged to overcome the issue.…”