Table 3.

β coefficients^a of growth curve models for grip strength (GS) over 5-year follow-up by protein intake groups.

Outcome	Effects/variable	Model 1		Model 2		Model 3
		β (SE)	p	β (SE)	p	β (SE)	p
GS (kg)	Intercept	19.29 (0.45)	<0.001	19.50 (0.47)	<0.001	7.0 (1.25)	<0.001
All participants	Protein intake group
	Low protein	−1.62 (0.60)	0.008	−2.03 (0.65)	0.002	−0.27 (0.41)	0.51
	Decline
	Time	−0.76 (0.05)	<0.001	−0.86 (0.08)	<0.001	−0.94 (0.2)	<0.001
	Slopes (rate of decline)
	Protein intake group × Time
	Low protein × Time			0.19 (0.11)	0.08	0.07 (0.11)	0.49
GS (kg)	Intercept	25.30 (0.60)	<0.001	25.35 (0.61)	<0.001	13.73 (3.01)	<0.001
Men	Protein intake group
	Low protein	0.65 (0.90)	0.47	0.54 (0.92)	0.56	0.79 (0.84)	0.35
	Decline
	Time	−1.13 (0.10)	<0.001	−1.18 (0.13)	<0.001	−2.02 (0.36)	<0.001
	Slopes
	Protein intake group × Time
	Low protein × Time		0.61	−0.13 (0.21)	0.54	0.1 (0.20)	0.6
GS (kg)	Intercept	14.54 (0.33)	<0.001	14.6 (0.34)	<0.001	8.8 (1.19)	<0.001
Women	Low protein intake group
	Low protein	−0.93 (0.41)	0.02	−1.05 (0.45)	0.02	−0.83 (0.41)	0.046
	Decline
	Time	−0.53 (0.06)	<0.001	−0.57 (0.08)	<0.001	−0.63 (0.21)	0.003
	Slopes
	Protein intake group × Time
	Low protein × Time			0.07 (0.11)	0.53	0.06 (0.12)	0.59

Model 1 includes a linear trend of time and protein intake group at baseline.

Model 2 includes protein intake and time interaction term.

Model 3 is additionally adjusted for sex (in all participants), anthropometry (height and FFM), health-related factors (number of chronic diseases, self-rated health, cognitive impairment, arthritis in hands), PA, attrition variable, and interaction terms (sex × time, PA × time).

^aParameter estimates β coefficients (SE) of fixed effects with GS longitudinal data. Random effects included both intercept and slopes of GS over 5 years. Time was used as continuous variable. Good protein intake (≥1 g/kg adjusted BW/day) served as a reference group.