Estimating cotinine associations and a saliva cotinine level to identify active cigarette smoking in alaska native pregnant women.

Studies indicate nicotine metabolism varies by race and can change during pregnancy. Given high rates of tobacco use and limited studies among Alaska Native (AN) women, we estimated associations of saliva cotinine levels with cigarette use and second-hand smoke (SHS) exposure and estimated a saliva cotinine cutoff to distinguish smoking from non-smoking pregnant AN women. Using questionnaire data and saliva cotinine, we utilized multi-variable linear regression (n = 370) to estimate cotinine associations with tobacco use, SHS exposure, demographic, and pregnancy-related factors. Additionally, we estimated an optimal saliva cotinine cutoff for indication of active cigarette use in AN pregnant women using receiver operating characteristic (ROC) curve analysis (n = 377). Saliva cotinine significantly decreased with maternal age and significantly increased with cigarettes smoked per day, SHS exposure, and number of previous full term pregnancies. Using self-reported cigarette use in the past 7 days as indication of active smoking, the area under the ROC curve was 0.975 (95 % CI: 0.960-0.990). The point closest to 100 % specificity and sensitivity occurred with a cotinine concentration of 1.07 ng/mL, which corresponded to sensitivity of 94 % and specificity of 94 %. We recommend using a saliva cotinine cutoff of 1 ng/mL to distinguish active smoking in pregnant AN women. This cutoff is lower than used in other studies with pregnant women, most likely due to high prevalence of light or intermittent smoking in the AN population. Continued study of cotinine levels in diverse populations is needed.