Skip to main content
. 2009 Feb 12;338:b462. doi: 10.1136/bmj.b462

Table 3.

 Logistic regression models illustrating odds of cognitive impairment (95% confidence intervals) in former and never smokers by salivary cotinine levels

Variable Never smokers (n=2014) Former smokers (n=2795)
Basic adjusted model* Fully adjusted model† Basic adjusted model* Fully adjusted model†
Salivary cotinine fourths (ng/ml):
 Lowest (0.0-0.1) Reference Reference Reference Reference
 Second (0.2-0.3) 1.44 (0.90 to 2.31) 1.42 (0.88 to 2.31) 0.95 (0.63 to 1.45) 0.88 (0.57 to 1.34)
 Third (0.4-0.7) 1.26 (0.77 to 2.08) 1.16 (0.68 to 1.97) 1.26 (0.85 to 1.86) 1.11 (0.74 to 1.67)
 Highest (0.8-13.5) 1.75 (1.10 to 2.78) 1.70 (1.03 to 2.80) 1.62 (1.14 to 2.30) 1.32 (0.92 to 1.91)
P for trend 0.03 0.08 0.005 0.1

Population weights used to adjust for sampling design.

*Adjusted for age, sex, education, and testing interval.

†Adjusted for age, sex, education, testing interval, ethnicity, manual occupation, net wealth, smoking history, obesity, alcohol consumption, physical inactivity, and depressive symptoms.