Table 4– Logistic regression model with the primary exposure variable of well water arsenic exposure, controlling for smoking category, with the outcome variable of interest of the presence or absence of dyspnoea
Independent variablesSmokers plus ex-smokers#Current smokersEx-smokers+
OR (95% CI)p-valueOR (95% CI)p-valueOR (95% CI)p-value
Well arsenic§
 Reference1.001.001.00
 High1.32 (1.15–1.52)<0.0011.33 (1.14–1.54)<0.0011.51 (1.27–1.79)<0.001
Smokingƒ
 Smoker/ex-smoker1.44 (1.16–1.78)0.001
 Current smokers1.38 (1.09–1.75)0.008
 Ex-smokers1.64 (1.22–2.20)<0.001
Age1.03 (1.02–1.04)<0.0011.03 (1.02–1.04)<0.0011.03 (1.02–1.04)<0.001
Education0.96 (0.92–0.98)<0.0010.96 (0.94–0.98)<0.0010.96 (0.94–0.99)0.002
Sex##1.58 (1.28–1.95)<0.0011.56 (1.24–1.96)<0.0011.60 (1.24–2.05)<0.001
Systolic BP1.00 (0.99–1.00)0.4881.00 (0.99–1.00)<0.6921.00 (0.99–1.01)0.648
BMI group¶¶
 11.16 (0.99–1.32)0.0541.13 (0.96–1.32)<0.1391.24 (1.04–1.49)0.018
 21.37 (1.04–1.80)0.0261.42 (1.07–1.88)<0.0161.24 (0.91–1.70)0.174
  • BP: blood pressure; BMI: body mass index. #: included the complete dataset, with smokers and ex-smokers combined versus nonsmokers; : included smokers only (ex-smokers were excluded from analysis); +: included ex-smokers only (current smokers were excluded from analysis); §: the reference well water arsenic concentration was <50 μg·L−1, with high arsenic being ≥50 μg·L−1; ƒ: for all three smoking groups the reference was nonsmokers; ##: the reference for sex was male; ¶¶: BMI categorised as low (<18.5 kg·m−2; group 1) and high (≥25 kg·m−2; group 2).