Table 2.

Logistic Regression Model of the Additive Effects of mtDNA Haplotype and the DXS8090(166)-DXS1068(258) Haplotype in Males Homoplasmic for 11778A→G, 14484T→C, and 14495A→G^[Note]

Variables in the Equation	P	OR (95% CI)
Non mtDNA haplogroup J	.043	2.45 (1.03–5.82)
DXS8090(166)-DXS1068(258)	.001	33.84 (4.38–261.53)
Constant	<.01	.02

Note.— As expected, the frequency of non-3460A→G haplogroup J pedigrees (23%) was greater than the European average (found to be 9%–14% in the European populations studied here). The effect of each genetic variable was approximately the same when modeled independently as when both variables were included in the same regression model. Modeling did not identify a significant direct interaction between mtDNA haplotype and DXS8090(166)-DXS1068(258). χ² was greatest for the additive model (χ²=35.68; P<.0001). Together, these findings indicate that each genetic factor that exerts its effect is independent and additive. Logistic regression showed an increased risk of visual failure in non–haplogroup J pedigrees in this cohort. A formal study of the penetrance of LHON in haplogroup J pedigrees versus non-J pedigrees had not been performed previously.