Table 3.

Linear mixed-effect regressions examining variation in physical performance over 5 years according to nutritional deficits^a.

	Within-group 5-year change from baseline β (95% CI); p-value	P for trend	Between-group difference
			Unadjusted model β (95% CI); p-valuec	Adjusted modelb β (95% CI); p-valuec
Outcome: SPPB score (0–12)
Nutritional deficits		<0.001
No deficit,	−0.43 (−0.86, 0.01); 0.057		Ref.	Ref.
1 deficit,	−0.47 (−0.84, −0.11); 0.011		−0.05 (−0.62, 0.52); 0.871	−0.03 (−0.59, 0.53); 0.921
≥2 deficits)	−0.80 (−1.17, −0.42); <0.001		−0.37 (−0.95, 0.20); 0.411	−0.23 (−0.80, 0.33); 0.843
Outcome: gait speed (m/s)
Nutritional deficits		<0.001
No deficit,	−0.09 (−0.16, −0.02); 0.010		Ref.	Ref.
1 deficit,	−0.05 (−0.11, 0.01); 0.052		0.03 (−0.05, 0.12); 0.879	0.03 (−0.06, 0.11); 0.813
≥2 deficits)	−0.08 (−0.14, −0.03); 0.005		0.01 (−0.08, 0.09); 0.879	−0.01 (−0.10, 0.08); 0.813
Outcome: chair rise time (s)
Nutritional deficits		<0.001
No deficit,	0.77 (−0.08, 1.62); 0.075		Ref.	Ref.
1 deficit,	0.46 (−0.24, 1.17); 0.198		−0.31 (−1.41, 0.80); 0.585	−0.18 (−1.26, 0.90); 0.739
≥2 deficits)	1.13 (0.39, 1.86); 0.003		0.36 (−0.77, 1.48); 0.585	0.40 (−0.71, 1.50); 0.739

CI, confidence interval; Ref, reference group; SPPB, Short Physical Performance Battery.

^aCutoff of nutritional deficits: vitamin D <20 ng/ml, homocysteine >14 μmol/L, omega-3 index ≤ lower quartile (≤4.87%).

^bAdjustments for age, sex, Multidomain Alzheimer Preventive Trial (MAPT) groups, education, body mass index, physical activity status, and time interactions.

^cP-value adjusted for multiple comparisons using the Benjamini-Hochberg procedure.