Changes in Cigarette Consumption With Reduced Nicotine Content Cigarettes Among Smokers With Psychiatric Conditions or Socioeconomic Disadvantage: 3 Randomized Clinical Trials

Abstract

Importance: This study is part of a programmatic effort evaluating the effects of reducing nicotine content of cigarettes to minimally addictive levels.

Objective: To examine whether very low-nicotine-content (VLNC) cigarettes decrease smoking rates and dependence severity among smokers with psychiatric disorders or socioeconomic disadvantage.

Design, setting, and participants: These 3 randomized clinical trials were performed at the University of Vermont, Brown University, and Johns Hopkins University between October 2016 and September 2019. Participants received 12 weeks of exposure to study cigarettes with nicotine content ranging from levels representative of commercial cigarettes (15.8 mg nicotine/g tobacco) to less than a hypothesized addiction threshold (2.4 mg/g and 0.4 mg/g). Daily smokers from 3 at-risk populations participated: individuals with affective disorders, exemplifying smokers with mental illness; individuals with opioid use disorder, exemplifying smokers with substance use disorders; and women with high school educations or less, exemplifying smokers with socioeconomic disadvantage. Data were analyzed from September 2019 to July 2020.

Interventions: Random assignment to 1 of 3 study cigarettes provided weekly at no cost for 12 weeks.

Main outcomes and measures: The primary outcome was between-group differences in mean total cigarettes smoked daily (CPD) during week 12; secondary outcomes included CPD for study and nonstudy cigarettes and dependence severity across weeks analyzed using analysis of covariance, random coefficients growth modeling, or repeated measures analysis of variance.

Results: A total of 775 participants were included (mean [SD] age, 35.59 [11.05] years; 551 [71.10%] women [owing to 1 population being exclusively women]); participants smoked a mean (SD) of 17.79 (9.18) CPD at study intake. A total of 286 participants were randomized to 0.4 mg/g, 235 participants were randomized to 2.4 mg/g, and 254 participants were randomized to 15.8 mg/g. Participants randomized to VLNC cigarettes had decreased mean [SEM] total CPD during week 12 across populations (Cohen d = 0.61; P < .001). At week 12, mean (SEM) CPD decreased to 17.96 (0.98) CPD in the 0.4 mg/g group and to 19.53 (1.07) CPD in the 2.4 mg/g group, both of which were significantly different from the 15.8 mg/g group (25.08 [1.08] CPD at week 12) but not each other (0.4 mg/g adjusted mean difference: -7.54 [95%CI, -9.51 to -5.57]; 2.4 mg/g adjusted mean difference: -5.34 [95% CI, 7.41 to -3.26]). Several secondary outcomes differed across populations randomized to VLNCs, including mean total CPD across weeks, with linear trends lower in participants receiving 0.4 mg/g (-0.28 [95%CI, -0.39 to -0.18]; P < .001) and 2.4 mg/g (-0.13 [95%CI, -0.25 to -0.01]; P < .001) doses compared with those receiving the 15.8 mg/g dose (0.30 [95% CI, 0.19 to 0.41]). Fagerström Test of Nicotine Dependence mean total scores were significantly lower in participants who received VLNCs (Cohen d = 0.12; P < .001), with those who received the 0.4 mg/g dose (mean [SD] score, 3.99 [0.06]; P < .001 vs 15.8 mg/g) or 2.4 mg/g dose (mean [SD] score, 4.07 [0.06]; P = .01 vs 15.8 mg/g) differing from those who received the 15.8 mg/g dose (mean [SD] score, 4.31 [0.06]) but not from each other.

Conclusions and relevance: These findings demonstrate that decreasing the nicotine content of cigarettes to very low levels reduced smoking rate and nicotine-dependence severity in these high-risk populations, effects that may facilitate successful cessation.

Trial registration: ClinicalTrials.gov Identifiers: NCT02232737, NCT02250664, NCT02250534.

Conflict of interest statement

Conflict of Interest Disclosures: Dr. Higgins reported receiving grants from the National Institutes of Health (NIH) outside the submitted work. Dr Tidey reported receiving grants from National Institute of General Medical Sciences outside the submitted work. Dr Heil reported grants from the NIH during the conduct of the study. Dr Gaalema reported receiving grants from the NIH during the conduct of the study. Dr Hughes reported receiving personal fees from Swedish Match outside the submitted work. Dr Villanti reported receiving grants from the NIH during the conduct of the study. Dr Miller reported employment by the Food and Drug Administration (FDA), Center for Tobacco Products (CTP) outside the submitted work. Dr MacLeod reported receiving grants from the NIH during the conduct of the study. No other disclosures were reported.

Figures

Figure 1.. Participant Enrollment, Randomization, and Retention Flowchart

OAT indicates opioid assisted treatment; PI, principal investigator.

Figure 2.. Number of Total and Study Cigarettes Smoked per Day According to Nicotine Content

Data points are arithmetic means collapsed across participants and populations; error bars indicate SEM. Data not sharing a superscript letter differed significantly in post hoc testing and in slope of linear trends.

Figure 3.. Changes in Nicotine Dependence Severity

A, Data points are arithmetic means collapsed across participants, time, and populations; error bars indicate SEM. Data points not sharing a superscript letter differed significantly by dose in post hoc testing. B, Data points are arithmetic means collapsed across participants and populations at each assessment; error bars indicate SEM. Doses not sharing a superscript letter differed in slope of linear trends.

Figure 4.. Biomarkers of Exposure as a Function of Dose, Time, and Population

A, Data points are arithmetic means across participants and populations and error bars indicate SEM. Doses not sharing a superscript letter differ in linear trends across the 12-week study period. B, Data points are geometric means across participants and time at each dose; error bars represent SEM. Data points not sharing a superscript letter differed significantly in post-hoc testing. C, Data points are geometric means across participants and time at each dose; error bars indicate SEM. Data points not sharing a superscript letter differed significantly in post hoc testing.