ABSTRACT
Understanding the true burden of tobacco smoking on adverse pregnancy outcomes is critical in generating appropriate interventions to improve outcomes. Self-reporting of human behaviour that is associated with stigma is associated with underreporting in general and may bias the impact of smoking in studies; however, self-reporting is frequently the most practical method of gleaning this information. The objective of this study was to evaluate concordance between self-reported smoking and concentrations of plasma cotinine, a biomarker of smoking, among participants enrolled in two related HIV cohorts. A total of 100 pregnant women (76 living with HIV [LWH] and 24 negative controls) in their third trimester, and 100 men and non-pregnant women (43 LWH and 57 negative controls) were included. Among all participants, 43 pregnant women (49% LWH and 25% negative controls) and 50 men and non-pregnant women (58% LWH and 44% negative controls) were self-reported smokers. The odds of discordance between self-reported smoking and cotinine levels were not significantly different between self-reported smokers and non-smokers, nor between pregnant women and others, but were significantly increased, regardless of self-reported status, among people LWH compared to negative controls. The overall concordance between plasma cotinine and self-reported data among all participants was 94% with a sensitivity and specificity of 90% and 96%, respectively. Taken together, these data demonstrate that participant surveying in a non-judgemental context can lead to accurate and robust self-report smoking data among both persons LWH and not, including in the context of pregnancy.