Comparing cotinine and NNAL verification of self-reported smoking status among lung cancer screening eligible population from the 2007–2014 National Health and Nutrition Examination Survey (NHANES)

Abstract

Background: Biochemical verification of self-reported smoking status is not common among the population eligible for lung cancer screening (LCS). Methods: We used urinary NNAL (4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol and its glucuronides) and serum cotinine as the gold standard to determine the validity and reliability of self-reported smoking status from the 2007–2014 National Health and Nutrition Examination Survey (NHANES). Results: We found 2.3% (n = 652, equivalent to 5.3 million weighted population) of adults eligible for LCS according to the current United States Preventive Services Task Force guideline. Self-reported current smoking status performed similarly against NNAL and cotinine: sensitivity [89.7% (95%CI: 84.9%–94.5%) vs. 89.5% (95%CI: 84.8%–94.3%)]; specificity [99.7% (95%CI: 99.2%–100.0%) vs. 100% (95%CI:100%–100%)]; positive predictive value (PPV) and negative predictive value (NPV) were 99.8% (95%CI:99.4%–100.0%) versus 100% (95%CI:100%–100%) and 85.3% (95%CI: 79.1%–91.5%) versus 85.1% (95%CI: 79.1%–1.0%), respectively; and Kappa [86.5% (95%CI:80.5%–92.5%) vs. 86.5% (95%CI:80.6%–92.3%)]. Performance measures were better among females than males; worst among the non-Hispanic white and best among other race/ethnicity group. The validity and reliability of self-reported smoking status increased with increasing cutpoint levels of both NNAL and cotinine. Conclusions: Self-reported smoking status among people who are at high risk of lung cancer is reasonably reliable. The difference between using NNAL and cotinine appears to be minimal.