Table 4.

Random effects summary ORs for childhood leukemia in relation to prenatal maternal occupational exposure: subgroup analyses.

Exposure (no. of risk estimates)	Summary OR (95% CI)	Heterogeneity Q-value
Total quality score < median (12)	2.25 (1.49–3.42)	13.3, p = 0.27
Total quality score ≥ median (4)	1.86 (1.11–3.14)	5.0, p = 0.17
External validity score < median (7)	1.99 (1.12–3.51)	10.7, p = 0.10
External validity score ≥ median (9)	2.18 (1.43–3.31)	9.0, p = 0.34
Exposure measurement score < median (7)	1.44 (0.83–2.51)	5.6, p = 0.47
Exposure measurement score ≥ median (9)	2.45 (1.68–3.58)	11.9, p = 0.16
Confounding score < median (7)	1.71 (0.99–2.96)	8.9, p = 0.18
Confounding score ≥ median (9)	2.38 (1.56–3.62)	10.3, p = 0.24
Well-defineda exposure window (15)	2.06 (1.47–2.90)	19.2, p = 0.16
Unspecified acute leukemia (7)	1.59 (1.02–2.47)	11.0, p = 0.09
ALL (5)	2.64 (1.40–5.00)	1.8, p = 0.77
AML (4)	2.64 (1.48–4.71)	2.8, p = 0.42
Exposed on farm (9)	2.44 (1.53–3.89)	8.7, p = 0.37
Mixed or unknown exposure place (7)	1.81 (1.17–2.81)	9.2, p = 0.16
Pre-1990 (5)	2.12 (1.05–4.25)	4.0, p = 0.40
≥ 1990 (11)	2.10 (1.44–3.08)	15.5, p = 0.11
Population-based case–control studies (15)	2.10 (1.50–2.94)	19.6, p = 0.14
Insecticides (6)	2.72 (1.47–5.04)	6.2, p = 0.29
Herbicides (2)	3.62 (1.28–10.3)	0.8, p = 0.37

^aOr reasonably inferable.