The association between self-reported varenicline adherence and varenicline blood levels in a sample of cancer patients receiving treatment for tobacco dependence

INTRODUCTION: The degree to which smokers quit successfully with varenicline is strongly associated with their adherence to the medication regimen. Thus, measuring varenicline adherence to identify smokers needing additional intervention is a priority. Few studies, however, have examined the validity of self-reported varenicline adherence, using a biological assessment of adherence as a reference. No study has examined this issue among cancer patients trying to quit smoking, who may show unique patterns of adherence given their medical comorbidity. METHODS: This study used data from 76 cancer patients who received varenicline and provided self-reported varenicline adherence data (pill count) and a blood sample to determine varenicline metabolites 4 weeks after initiating varenicline. RESULTS: Receiver operating characteristic (ROC) curve analyses of plasma varenicline levels showed that 4 ng/ml was the optimal cut-point for differentiating adherence with significant (p's < 0.04) area under the curve values, ranging from 0.73–0.80 for 3-day, 7-day, and 4-week self-reported pill count; specificity values ranged from 0.63–0.78 and sensitivity values ranged from 0.82–0.94. Using this cut-point, adherence was high (88%). However, plasma varenicline levels were weakly correlated with 3-day and 4-week pill count and total pill count (12 weeks) was not correlated with plasma varenicline levels. Patients with head and neck cancer, gastrointestinal cancer, and more advanced disease showed lower varenicline adherence and lower plasma varenicline. CONCLUSIONS: Using the 4 ng/ml cut-point, this study suggests validity of short-term self-reported varenicline adherence among cancer patients undergoing tobacco dependence treatment in contrast to studies in the general population, which supported 12-week pill count.