Table 2.

Results of the latent variable growth curve model

		MMSE (n= 113)^a		CDR-SoB (n= 118)^b

		Est	P-value	Est	P-value
Intercept	Time since dementia diagnosis	−1.54	<0.001	1.30	<0.001
	Age	0.01	0.918	0.77	0.358
	Sex^c	−0.12	0.908	−0.01	0.946
	Medium education^d	−1.71	0.090	0.93	0.244
	High education^d	0.31	0.827	1.39	0.214
	CHD^e	1.73	0.051	−1.45	0.036

Slope	Time since dementia diagnosis	−0.23	0.385	0.05	0.797
	Age	−0.07	0.383	0.03	0.581
	Sex^c	−1.02	0.131	−0.46	0.366
	Medium education^d	−0.04	0.951	0.30	0.503
	High education^d	−0.35	0.687	0.72	0.283
	CHD^e	−1.03	0.039	1.01	0.008

↵a The fit for the MMSE model was good (χ² [53] = 67.795, P = 0.08, root mean square error of approximation [RMSEA] = 0.050, comparative fit index [CFI] = 0.958).
↵b The fit for the CDR-SoB model was good (χ² [53] = 69.14, P = 0.07, RMSEA = 0.051, CFI = 0.964).
↵c Female sex was used as reference.
↵d Low educational level was used as reference.
↵e Absence of CHD was used as reference. CDR-SoB = Clinical Dementia Rating Sum of Boxes. CHD = coronary heart disease. Est = estimate of the effect on intercept and slope of cognition. MMSE = Mini Mental State Examination.