Dose-response association between salivary cotinine levels and Mycobacterium tuberculosis infection.

SETTING: Tijuana, Mexico.
OBJECTIVE: To describe the association between salivary cotinine levels and interferon-gamma (IFN-γ) release assay results.
DESIGN: We conducted a cross-sectional study among injection drug users. Salivary cotinine levels were measured using NicAlert, a semi-quantitative dipstick assay. QuantiFERON©-TB Gold In-Tube (QFT-GIT) was used to determine Mycobacterium tuberculosis infection.
RESULTS: Among 234 participants, the prevalence of QFT-GIT positivity for NicAlert cotinine categories 0 (non-smoking), 1 (second-hand smoke exposure or low-level smoking) and 26 (regular smoking) were respectively 42.1%, 46.4% and 65.2% (Ptrend 0.012). We found increasing trends in QFT-GIT positivity (Ptrend 0.003) and IFN-γ concentrations (Spearman's r 0.200, P 0.002) across cotinine levels 0 to 6. In multivariable log-binomial regression models adjusted for education, cotinine levels were not associated with QFT-GIT positivity when included as smoking categories (1 and 26 vs. 0), but were independently associated with QFT-GIT positivity when included as an ordinal variable (prevalence ratio 1.09 per 1 cotinine level, 95%CI 1.021.16).
CONCLUSION: Our findings suggest that a dose-response relationship exists between tobacco smoke exposure and M. tuberculosis infection. Longitudinal studies that use biochemical measures for smoking status are needed to confirm our findings.