Optimal cut-off point of exhaled carbon monoxide to validate self-reported smoking status in healthy adults

Abstract

Background: There is no current consensus regarding the optimal cut-off point of exhaled carbon monoxide (CO) to distinguish smokers from nonsmokers.

Objectives: To assess the accuracy of an exhaled carbon monoxide cut-off point in order to distinguish actual smokers from nonsmokers among apparently healthy adults.

Methods: We studied 50 current smokers (20 male; 47±12 years; BMI: 26±4 kg/m²), with normal lung function (FEV₁/FVC: 81±6; FEV₁: 84±18%pred) who self-reported their smoking status and habits; and 31 paired non-smokers (11 male; 44±11 years; BMI: 26±4 kg/m²; FEV₁/FVC: 83±6; FEV₁: 102±11%pred). All subjects were submitted to CO assessment (in the group of smokers, after a mean of 10±1.2 hours of cigarette abstinence), using a portable CO monitor (MicroCO®).

Results: Median [interquartile range] levels of CO in the group of smokers and non-smokers were 10 [7-17] and 3 [2-4], respectively. The 6ppm cut-off point suggested by the manufacturer generated a 77% sensitivity and 100% specificity; however a 4.5ppm cut-off point generated the highest combined sensitivity (90%) and specificity (90%). The ROC analysis indicated that the CO monitor provided high diagnostic accuracy to distinguish smokers from nonsmokers [area under the curve = 0.979 p<0,001].

Conclusions: Using a portable CO monitor, a 4.5ppm cut-off point seems more accurate than the cut-off point suggested by the manufacturer in order to distinguish smokers from nonsmokers among apparently healthy adults.