Table 2.

Mendelian randomization analysis of the association between longer LTL score and ependymoma risk in CCRLP, Toronto, and GICC case–control datasets

MR Estimate	Pediatric EPN-PF-A (Toronto)a		Childhood-onsetb (CCRLP)		Adolescent-onsetc (CCRLP)		Adult-onsetd (GICC)
	OR (95% CI)	P-value	OR (95% CI)	P-value	OR (95% CI)	P-value	OR (95% CI)	P-value
IVW	0.51 (0.03, 8.7)	0.64	0.93 (0.15, 5.7)	0.94	91.7 (3.6, 2.3 × 103)	0.0063	2.01 (0.34, 12.0)	0.43
Weighted median	0.61 (0.03, 13.3)	0.76	0.82 (0.08, 7.9)	0.85	368 (4.1, 3.3 × 104)	0.010	2.82 (0.27, 29.0)	0.38
Mode-based	0.67 (0.02, 22.5)	0.82	0.30 (0.02, 4.9)	0.40	690 (0.42, 1.1 × 106)	0.084	4.05 (0.30, 4.01)	0.30
MR Egger	183 (1.0 × 10−6, 3.3 × 1010)	0.59	0.08 (2.8 × 10−4, 23.1)	0.38	793 (0.02, 3.2 × 107)	0.22	294 (1.23, 7.0 × 104)	0.042
(Intercept)	0.61 (0.13, 2.94)	0.45	1.19 (0.81, 1.8)	0.37	0.86 (0.43, 1.7)	0.67	0.70 (0.48, 1.01)	0.059

Adjusted for subject sex and 10 ancestry-informative principal components

Nominally significant P values < 0.05 in bold

^aSample size for Toronto posterior fossa type A pediatric ependymoma case–control study includes 83 cases and 332 controls

^bSample size for CCRLP childhood-onset case–control subset includes 114 cases (age < 12 years) and 696 controls

^cSample size for CCRLP adolescent-onset case–control subset includes 39 cases (age 12–19 years) and 696 controls

^dSample size for Glioma International Case-Control Consortium (GICC) case–control study includes 103 cases (ages 18+) and 3287 controls