Table 1. Comparison of Cognitive Performances Between Tobacco Ever Users and Nonusersa.
| NIH Toolbox Cognition Battery subscale | Study period | |||||||
|---|---|---|---|---|---|---|---|---|
| Wave 1 | Wave 2 | |||||||
| Nonuse, weighted mean (SE) (n = 11 613) | Tobacco ever use, weighted mean (SE) (n = 116) | Adjusted b (SE)b | Adjusted P valuec | Nonuse, weighted mean (SE) (n = 10 117) | Tobacco ever use, weighted mean (SE) (n = 86) | Adjusted b (SE)b | Adjusted P valuec | |
| Dimensional Charge Card Sort | 92.3 (0.3) | 91.5 (1) | −0.4 (1.5) | .77 | NA | NA | NA | NA |
| Flanker Inhibitory Control and Attention | 93.8 (0.3) | 94 (0.7) | 0.2 (1.0) | .83 | 99.9 (0.3) | 100 (1.4) | 0.2 (1.9) | .91 |
| List Sorting Working Memory | 96.1 (0.5) | 93.9 (1) | −2.1 (1.4) | .16 | NA | NA | NA | NA |
| Oral Reading Recognition | 90.6 (0.2) | 88.7 (0.6) | −1.5 (0.6) | .03 | 94.7 (0.3) | 93.1 (0.5) | −2.1 (0.5) | <.001d |
| Pattern Comparison Process Speed | 88 (0.5) | 84.8 (1.2) | −3.5 (1.6) | .04 | 103.3 (0.5) | 101 (1.7) | −2.0 (2.3) | .40 |
| Picture Sequence Memory | 102.6 (0.4) | 100.1 (0.8) | −2.7 (1.2) | .04 | 108.5 (0.4) | 104.7 (2) | −5.4 (1.8) | .007d |
| Picture Vocabulary Tests | 84 (0.5) | 82 (0.8) | −2.9 (0.6) | <.001d | 88.5 (0.5) | 87 (1.1) | −3.0 (0.7) | <.001d |
| Crystalized cognition composite score | 86 (0.3) | 83.8 (0.5) | −2.4 (0.5) | <.001d | 90.5 (0.4) | 89.3 (0.8) | −2.7 (0.8) | .005d |
| Fluid cognition composite score | 91.2 (0.5) | 88.7 (0.7) | −2.4 (1.3) | .08 | NA | NA | NA | NA |
| Total cognition composite score | 85.8 (0.5) | 83 (0.6) | −2.9 (0.5) | .01d | NA | NA | NA | NA |
Abbreviations: NA, not available; NIH, National Institutes of Health.
Multivariate regression analyses were performed where the dependent variables were cognitive performance scores. Sampling weights and site clustering were incorporated in the survey regression analytical procedures for statistical inference at the population level. The independent variable was early age at initiation of tobacco use (ever use) measured at wave 1.
The analysis was adjusted by covariates, including age, sex, race and ethnicity, pubertal stage, substance ever use, parental monitoring, school environment, and study site. A sensitivity analysis was performed to further account for intracranial volume (eTable 4 in the Supplement). Regression coefficients measured the association of early-age initiation of tobacco use as changes in cognitive performance scores between tobacco ever users vs controls (no use).
False discovery rate (FDR) correction was performed to prevent inflation of type I errors.
Indicates FDR of less than 0.05.