Table 3. Sensitivity and heterogeneity analysis of pooled relative risks of VTE for smokers.

	Ever Versus Never Smoker				Current Versus Never Smoker				Former Versus Never Smoker
	n Studies	RR (95% CI)	I2 (95% CI)	p-Valuea	n Studies	RR (95% CI)	I2 (95% CI)	p-Valuea	n Studies	RR (95% CI)	I2 (95% CI)	p-Valuea
Statistical model
Random effects	32	1.17 (1.09–1.25)	64.53 (48.37–75.63)	<0.001	15	1.23 (1.14–1.33)	64.89 (31.17–79.73)	<0.001	14	1.10 (1.03–1.17)	53.52 (14.82–74.64)	0.009
Fixed effects	32	1.19 (1.15–1.22)			15	1.29 (1.24–1.34)			14	1.09 (1.06–1.12)
Analysis of all studies with
Adjusted risk estimateb	22	1.16 (1.09–1.23)	58.27 (33.09–73.97)	0.01	14	1.25 (1.17–1.35)	59.04 (26.04–77.29)	0.003	14	1.10 (1.03–1.17)	53.52 (14.82–74.64)	0.009
Large cohortc	11	1.17 (1.11–1.23)	34.68 (0.00–67.91)	0.12	9	1.32 (1.26–1.38)	10.61 (0.00–68.53)	0.35	9	1.07 (1.04–1.11)	0.00 (0.00–64.80)	0.52
Analysis of all studies except
Two largest studiesd	30	1.17 (1.07–1.27)	65.99 (50.08–76.83)	<0.001	13	1.19 (1.07–1.33)	64.94 (36.72–80.58)	0.001	12	1.10 (1.01–1.22)	57.90 (20.10–77.82)	0.006
One outlier studye	31	1.17 (1.09–1.25)	64.21 (47.54–75.58)	<0.001	12	1.23 (1.13–1.33)	66.18 (40.56–80.76)	<0.001	12	1.08 (1.05–1.12)	0.00 (0.00–56.59)	0.51

^a p-Value for I².

^bStudies reporting estimates that adjusted for at least one confounding factor.

^cLarge prospective cohort studies with sample size over 15,000.

^dStudies with population over 1 million by Hippisley-Cox J and Sweetland S [26],[29].

^eStudies with largest RR by Dreyer NA for ever smokers, by Hansson PO for current smokers and by Zhu J for former smokers [20],[30],[43].