Table 2.

Linear mixed model regression analysis for HD motor progression.

Parameter	Outcome: motor score				Outcome: random slope
	Coef.	SE	T statistic	P value	Coef.	SE	T statistic	P value
Female	3.375	0.474	7.127	1.16 × 10−12	0.016	0.051	0.316	0.8
Age	2.217	0.045	48.878	<5.0 × 10−300	0.013	0.005	2.749	5.99 × 10−3
CAG length	2.743	0.089	30.757	5.27 × 10−191	0.071	0.010	7.449	1.09 × 10−13
Age at motor onset	−1.542	0.049	−31.687	1.24 × 10−201	−0.010	0.005	−2.006	4.49 × 10−2
Education	−1.877	0.193	−9.742	3.09 × 10−22	−0.083	0.021	−4.037	5.49 × 10−5
Visit	3.516	0.078	45.019	<5.0 × 10−300

The table presents the coefficient estimates from two linear regression models: (1) a linear mixed model analysis of 14,850 longitudinal motor scores (dependent variable) across 5204 manifest HD cases from Enroll-HD data 1 and (2) a linear regression model analysis of the resulting random slope estimates (dependent variable) of the same 5204 cases. The linear mixed-effects model included two random-effects (a random intercept term and a random slope term with respect to visit) and several fixed-effect terms (sex, age at baseline, CAG-repeat length, age at motor onset, educational attainment). The empirical Bayes estimate of the random slope was used as a measure of HD motor progression for each of the N = 5204 manifest HD cases. We adjusted the random slope estimate for all fixed effects in our downstream EWAS analysis. The columns report the covariate name, regression coefficient, standard error, Student’s T statistic, and unadjusted two-sided Wald test p value.