Selection criteria limit generalizability of smoking pharmacotherapy studies differentially across clinical trials and laboratory studies: A systematic review on varenicline

Background The selection criteria used in clinical trials for smoking cessation and in laboratory studies that seek to understand mechanisms responsible for treatment outcomes may limit their generalizability to one another and to the general population. Methods We reviewed studies on varenicline versus placebo and compared eligibility criteria and participant characteristics of clinical trials (N = 23) and laboratory studies (N = 22) across study type and to nationally representative survey data on adult, daily USA smokers (2014 National Health Interview Survey; 2014 National Survey on Drug Use and Health). Results Relative to laboratory studies, clinical trials more commonly reported excluding smokers who were unmotivated to quit and for specific medical conditions (e.g., cardiovascular disease, COPD), although both study types frequently reported excluding for general medical or psychiatric reasons. Laboratory versus clinical samples smoked less, had lower nicotine dependence, were younger, and more homogeneous with respect to smoking level and nicotine dependence. Application of common eligibility criteria to national survey data resulted in considerable elimination of the daily-smoking population for both clinical trials (≥47%) and laboratory studies (≥39%). Relative to the target population, studies in this review recruited participants who smoked considerably more and had a later smoking onset age, and were under-representative of Caucasians. Conclusions Results suggest that selection criteria of varenicline studies limit generalizability in meaningful ways, and differences in criteria across study type may undermine efforts at translational research. Recommendations for improvements in participant selection and reporting standards are discussed.