Table 1.

Characteristics of the studies on selenium and lung cancer included in the pooled analysis of RRs

AuthorStudy designAssessment methodGenderLocationYears of follow-upNo. of casesRR (95% CI)P for dose-response trend%W
Comstock et al. (11)NCCSerumAllMaryland15–182580.65 (0.41–1.02)0.0812.8
Goodman et al. (19)NCCSerumAllMulticenter, United States5–143561.20 (0.77–1.88)0.4913.3
Kabuto et al. (12)NCCSerumAllJapan11–13770.56 (0.20–1.43)NA2.7
Knekt et al. (26)NCCSerumMaleFinland8–121530.66 (0.37–1.19)0.0017.8
Knekt et al. (13)NCCSerumAllFinland16–19770.41 (0.17–0.94)0.463.6
Ratnasinghe et al. (37)NCCSerumMaleChina4–51081.20 (0.60–2.40)0.525.5
Reid et al. (21)Randomized clinical trialPlasmaAllEastern United States5–13600.74 (0.44–1.24)NA9.9
Garland et al. (39)*NCCToenailFemaleMulticenter, United States3.5474.33 (0.54–34.60)0.170.6
Hartman et al. (40)NCCToenailMaleFinland5–8NA0.20 (0.09–0.44)NA4.2
Hartman et al. (40)NCCToenailMaleFinland5–8NA0.61 (0.27–1.41)NA3.9
van den Brandt et al. (18)CohortToenailAllNetherlands3.33170.50 (0.30–0.81)0.00610.8
Hu et al. (23)Case-controlDiet/FFQAllChina2271.30 (0.70–2.20)0.488.1
Kromhout (24)CohortDiet/FFQMaleNetherlands25630.98 (0.41–2.36)>0.013.5
Zhou et al. (25)Case-controlDiet/FFQFemaleChina2900.76 (0.47–1.15)0.1113.3
  • Note: NA, not available; %W, percentage of the total weight applied to the study in the fixed effects model.

  • * Based on conditional logistic regression with 47 matched pairs. The unadjusted RR reported by the authors was 0.46 (95% CI 0.15–1.44; Ref. 39).

  • Results presented separately by randomization period.

  • CI calculated from data given in the publication.