A receiver operated curve-based evaluation of change in sensitivity and specificity of cotinine urinalysis for detecting active tobacco use

Abstract

Background: Tobacco use has been associated with various carcinomas including lung, esophagus, larynx, mouth, throat, kidney, bladder, pancreas, stomach, and cervix. Biomarkers such as concentration of cotinine in the blood, urine, or saliva have been used as objective measures to distinguish nonusers and users of tobacco products. A change in the cut-off value of urinary cotinine to detect active tobacco use is associated with a change in sensitivity and sensitivity of detection.

Aim: The current study aimed at assessing the impact of using different cut-off thresholds of urinary cotinine on sensitivity and specificity of detection of smoking and smokeless tobacco product use among psychiatric patients.

Settings and design: All the male subjects attending the psychiatry out-patient department of the tertiary care multispecialty teaching hospital constituted the sample frame for the current study in a cross-sectionally.

Materials and methods: Quantitative urinary cotinine assay was done by using ELISA kits of Calbiotech. Inc., USA. We used the receiver operating characteristic (ROC) curve to assess the sensitivity and specificity of various cut-off values of urinary cotinine to identify active smokers and users of smokeless tobacco products.

Results: ROC analysis of urinary cotinine levels in detection of self-reported smoking provided the area under curve (AUC) of 0.434. Similarly, the ROC analysis of urinary cotinine levels in detection of self-reported smoking revealed AUC of 0.44. The highest sensitivity and specificity of 100% for smoking were detected at the urinary cut-off value greater than or equal to 2.47 ng/ml.

Conclusions: The choice of cut-off value of urinary cotinine used to distinguish nonusers form active users of tobacco products impacts the sensitivity as well as specificity of detection.