Based on the sequential algorithm, the number of studies that had to be omitted from the meta-analysis to drop I2 below the intended threshold (50%) was two for cohort studies addressing the association between smoking and suicide ideation (I2 = 10.5%; OR 2.41; 95% CI: 2.02, 2.80) and four for case-control/cross-sectional ones (I2 = 41.9%; OR 1.653; 95% CI: 1.48, 1.83). Two studies were omitted to drop I2 below the intended threshold for case-control/cross-sectional studies addressing the association between smoking and suicide plan (I2 = 22.3%; OR 2.43; 95% CI: 1.99, 2.86). Finally, the number of studies, which were omitted from the meta-analysis to drop I2 below the desired threshold, was one for cohort studies addressing the association between smoking and suicide attempt (I2 = 0.0%; OR 3.22; 95% CI: 2.99, 3.45) and five for case-control/cross-sectional ones (I2 = 38.9%; OR 2.048; 95% CI: 1.80, 2.29).
