Introduction: Cigarette preference increases as a function of nicotine content, but preference can be shifted by manipulating cigarette cost. The aim of the present study is to model whether the behavioral-economic metric of unit price (cigarette cost/nicotine content) accounts for cigarette preference shifts and whether preference changes to very low nicotine content cigarettes (VLNCs) are associated with corresponding changes in smoking rate. Methods: 169 daily smokers from populations vulnerable to smoking completed sessions in which choices between smoking normal nicotine content (NNC) (15.8mg/g) and VLNC (0.4mg/g) cigarettes were concurrently available. In Condition 1, choices for both products were available ad-lib at an equal cost of 10 responses/choice. In Condition 2, VLNCs were again available ad-lib at 10 responses/choice, but NNCs were available on a progressive-ratio (PR) schedule wherein response cost (and unit price) increased following each NNC choice (10,160,320…8400 responses/choice). Results were analyzed using ANOVAs and a binomial test (p<.05). Results: Participants preferred NNCs over VLNCs in Condition 1, but shifted preference to VLNCs in Condition 2 (p<.001) immediately before the point in the PR progression where unit price for NNCs exceeded unit price for VLNCs (p<.001). Additionally, this preference shift corresponded with reduced total cigarette consumption compared to Condition 1 levels (p<.001). Conclusions: These results suggest that unit price of nicotine underpins cigarette preference and may provide a metric by which regulators can predict product preference and potentially impact it through policy. These results also demonstrate that VLNCs sustain lower smoking rates than NNCs even under acute laboratory conditions.

Rationale: Nicotine is the addictive component in cigarettes which maintains cigarette smoking that subsequently leads to morbidity and mortality. There are growing regulatory efforts to lower the nicotine content in cigarettes so that they are minimally addictive. Valid methods for assessing the abuse liability of cigarettes are essential to these efforts. While subjective effect measures and hypothetical purchase tasks are appealing because they are far easier to administer, it is unclear whether these methods can be used to evaluate acute relative reinforcing, a critical component of abuse liability. This secondary analysis sought to evaluate the utility of one subjective effects measure, the modified Cigarette Evaluation Questionnaire (mCEQ), and one hypothetical purchase task, the Cigarette Purchase Task (CPT), for predicting acute relative reinforcing efficacy as measured by concurrent choice Self-Administration (SA) Method: Current smokers (N=169) belonging to one of three vulnerable populations (socioeconomically disadvantaged women of childbearing age, opioid-maintained individuals, or individuals with affective disorders) participated in a multi-site, double blind study evaluating research cigarettes with varying levels of nicotine content (0.4, 2.4, 5.2, 15.8 mg/g). In Phase 1 (4 sessions, 1 research cigarette per session) participants completed the mCEQ and CPT following ad-lib smoking of the research cigarette. In Phase II (6 sessions) cigarette preference was assessed using two-dose concurrent choice tests. Difference scores were calculated for each of the five mCEQ subscales and five CPT indices for all six possible dose comparisons evaluated in Phase II. We evaluated the utility of the mCEQ subscale and CPT index difference scores for predicting preference for the higher dose in a given dose comparison using a mixed-model of repeated measures analysis of variance. Finally, we used stepwise regressions to determine which subscales and indices served as independent predictors of concurrent choice SA. Results: Among mCEQ subscales, higher Satisfaction and Enjoyment of Respiratory Tract Sensation were independently predictive of higher dose preference in the choice testing regardless of dose comparison. There was a significant Satisfaction X Vulnerable Population interaction where increases in Satisfaction difference scores corresponded to greater changes in higher dose preference among socioeconomically disadvantaged women of childbearing age compared to other Vulnerable Populations. Among CPT indices, Elasticity was the only independent predictor of choice. However, there was a significant Elasticity X Dose Comparison X Vulnerable Population interaction associated with its predictive utility where the relationship between elasticity and choice differed by dose among opioid-maintained individuals. In a final model, including all subscales and indices, Satisfaction and Enjoyment of Respiratory Tract Sensations remained the only significant predictors of choice. Discussion: Concurrent choice testing, subjective effects and hypothetical purchase tasks capture some common features of abuse liability. Concurrent choice testing and the Satisfaction subscale were the most concordant measures. The observation that CPT indices are not robust predictors of choice in a concurrent arrangement suggests this measure may have greater utility for capturing individual differences as opposed to isolating the acute relative reinforcing effects of nicotine. Nevertheless, all three measures can contribute to efforts to assess the abuse liability of cigarettes varying in nicotine dose and important work aimed at regulating these products to improve human health.