TABLE 1

Pooled estimates of the association between tuberculosis and subsequent lung cancer diagnosis or mortality

Model 1#Model 2Model 3+
Studies nPooled estimate§ (95% CI)Studies nPooled estimate§ (95% CI)Studies nPooled estimate§ (95% CI)
Cohort studies
 Lung cancer diagnosis8ƒ2.96 (2.28–3.83)71.77 (1.41–2.22)51.51 (1.30–1.76)
  AdenocarcinomaNA32.00 (0.93–4.31)32.00 (0.93–4.31)
  Small cell carcinomaNA30.88 (0.34–2.26)30.88 (0.34–2.26)
  Squamous cell carcinomaNA32.01 (1.00–4.03)32.01 (1.00–4.03)
 Lung cancer mortality72.97 (2.14–4.11)21.62 (1.18–2.21)NA
Case–control studies
 Lung cancer diagnosis412.00 (1.65–2.41)231.76 (1.41–2.19)191.74 (1.42–2.13)
  Adenocarcinoma102.27 (1.46–3.52)81.96 (1.20–3.21)61.51 (0.92–2.48)
  Small cell carcinoma42.26 (1.13–4.52)31.50 (0.95–2.39)22.05 (0.42–10.03)
  Squamous cell carcinoma34.51 (2.75–7.39)42.43 (1.34–4.41)32.23 (0.85–5.86)
 Lung cancer mortality12.86 (1.87–4.45)NANA

NA: not applicable (no study reporting this estimate). #: unadjusted estimates of lung cancer and previous tuberculosis; : adjusted for age and any assessment of smoking; +: adjusted for age and quantitatively assessed smoking; §: hazard ratio for cohort studies and odds ratio for case–control studies; ƒ: studies from Taiwan used the same database and duplication of information was possible, so we considered only the cohort with the largest sample size [14].