Observational studies of exposure to tobacco and nicotine products: Best practices for maximizing statistical precision and accuracy

Abstract

Non-randomized observational studies can track risk-induction and -reduction associated with real-world use of non-combusted nicotine and tobacco products. The objective of this analysis was to evaluate the precision and accuracy of recent studies and to identify opportunities for further optimizing future study designs. The ROBINS framework for minimizing statistical bias was translated to specific considerations that spanned the selection and quantification of cohorts, exposure, and outcomes. These principles were then considered within the context of a recent comprehensive meta-analysis, representing 107 observational studies, which evaluated the effects of using electronic cigarettes (ECs), combusted cigarettes (CCs) and dual use of both. The meta-analysis had previously reported the relative risk from all-sources, including tobacco use and non-tobacco use. We now report the product use-specific risk associated with displacing CCs with ECs indicated from the primary references, along with observations regarding the precision of characterization of CC and EC exposure in the cited studies.

Keywords: Health sciences; Medicine; Research methodology social sciences.

Graphical abstract

Figure 1

Factors impacting precision and accuracy of observational studies of exposure to electronic cigarettes EC (electronic cigarette); CC (combusted cigarette); OR (odds ratio). ROBINS-E domains aggregate into three categories, spanning characterization of population and sampled cohorts, exposure history, and outcomes and results. Observational studies of EC use may include one or more of the cohorts represented across the three rows: EC users who are never, former, and current (dual) users of CC. The factors here are described specifically for exposure to EC and CC, but can generalize to other tobacco and nicotine products (see also Table S1.1).

Figure 2

EC product characteristics impacting exposure Product characteristics which can impact exposure include: (A) the e-liquid composition, (B) the aerosolization technology, (C) the device platform, and (D) the brand and company of record. Schematic diagram adapted from patent assigned to Shenzhen Smoore Technology, Ltd. This platform and similar variants are incorporated in the NJOY Ace (Altria, Richmond, VA) and VUSE Alto (R.J. Reynolds, Winston-Salem, NC), which are FDA-authorized, later-generation, reusable closed-system, nicotine-salt (1.8%–5.0%) pod products with fixed power settings and higher-capacitance ceramic heating elements which are replaced with each pod (1.8mL–1.9mL e-liquid per pod). See also Document S1.

Figure 3

Comprehensive characterization of CC exposure includes intensity, duration and timing of use CPD (cigarettes per day; smoking intensity); CVD (cardiovascular disease); CC (combusted cigarettes); YO (years old). (A) A comprehensive characterization of CC exposure includes smoking intensity (CPD, or cigarettes per day), duration of regular use, and time quit (if relevant). Top graph illustrates CC smoking intensity (Y axis) vs. time (X axis). Bottom graph illustrates corresponding risk (Y axis) vs. time (X axis). (B) CVD risk (Y axis) vs. CC pack-years (X axis). Redrawn from (Lubin et al., 2016). (C) Excess mortality risk due to CC use (total risk - baseline risk; Y axis) vs. time quit (X axis). Redrawn from (Cho et al., 2024). (D) CVD risk (Y axis) vs. CC use history status (X axis). Redrawn from (Farsalinos et al., 2019).

Figure 4

Lifetime number of EC uses, stratified by EC use category (PATH, Wave 6) PATH Wave 6 unweighted adult EC use data is illustrated. For each cohort, the proportion of the sample which has reported lifetime uses of EC numbering 1 to 10, 11 to 50, 51 to 99, and 100+ times is color-coded. Note: Current use (past 30-day use) sample size does not exactly correspond to the sum of current some-day and current every-day use, and ever-use does not exactly correspond to the sum of former and current use, because of inconsistencies in participant answers across different questions (see also Figure S1.1).

Figure 5

Incremental risk and risk reduction (illustrative example) NS (non-smoker/vaper); CC (combusted cigarette); EC (electronic cigarette). Green arrow represents reduction in risk associated with EC use vs. CC use. In this illustrative hypothetical example, a non–tobacco user (NS) has an adjusted odds ratio (aOR) of 1.0 for a given harm, reflecting the background risk rate. If CC smoking doubles the risk of harm, a subject engaging in CC smoking would have aOR of 2.0. In other words, they would incur a risk of 1.0 from non-tobacco sources, and an additional incremental risk of 1.0 from CC smoking. Likewise, if using EC caused 30% of the incremental risk of CC smoking, use of EC would have aOR of 1.3. This aOR would comprise risk of 1.0 from non-tobacco sources and 0.3 from EC use. In this example, if the subject had used EC instead of smoking CC, their aOR would be 1.3 instead of 2.0. The harm reduction associated with EC would be 70%, due to the incremental risk of 0.3 (i.e., 1.3–1.0) with EC vs. 1.0 for CC (i.e., 2.0–1.0). Note that reduction in all sources of risk is 35% for EC vs. CC users (1.3 vs. 2.0), but harm reduction is 70% (0.3 vs. 1.0 incremental risk due specifically to tobacco product use). See also Equations SI.1-4.

Figure 6

Characterization of risk associated with EC use (case study) aOR (adjusted odds ratio); CC (combusted cigarette); EC (electronic cigarette), Metab. (metabolic disease); Oral Dis. (oral disease). (A) Incremental risk reduction associated with EC vs. CC use across categories of harm. All aOR are cited from (Glantz et al., 2024, Table 1), with the exception of CVD, which reflects re-analysis after exclusions of two references which segmented into harm vs. no-harm cohorts and reported that CC use was not associated with increased risk for CVD.^, (B) Precision of characterization of exposure to CC and EC in CVD and stroke-related references cited by (Glantz et al., 2024). Leftmost column: primary reference cited. Next column: distribution of CC use histories among the EC cohort (light blue = never CC use; medium blue = former CC use; dark blue = current CC use). Rightmost columns: percentage of EC cohort which formerly smoked or currently smoke CC, whether odds adjustments were made for duration or timing of CC use, and whether harm events occurring before the start of EC use were adjusted for. Red X represents adjustments not made. See also Equations SI.1-4, Figures S2.1–S2.4, and Tables S2.1–S2.8 for more detailed information and data analysis.