The study included 3326 subjects. The mean age was 60.99 ± 7.94 ranging from 50 to 94 years old. Demographic characteristics of the participants are reported in Table 1. A Mann-Whitney U test indicated that the age was higher for men (Mdn = 60.5) than for women (Mdn= 58), Z= -10.14, P<0.001. Also, the result of this test indicated that men had a higher education (Mdn=5) than women (Mdn=2), Z=-15.02, P<0.001.
Table 1. Demographic characteristics of the study participants
Variable
|
Male
N %
|
Female
N %
|
Total
N %
|
Age (years)
Mean ± SD
50-54
55-59
60-64
65-69
70-74
≥75
|
62.45 ± 8.23
267 17.2
467 30.0
337 21.7
210 13.5
128 8.2
147 9.4
|
59.72 ± 7.44
545 30.8
546 30.8
315 17.8
165 9.3
110 6.2
89 5.0
|
60.99 ± 7.94
812 24.4
1013 30.5
652 19.6
375 11.3
238 7.2
236 7.1
|
Education (years)
0
1-5
6-12
≥13
|
268 17.3
585 37.7
549 35.4
151 9.7
|
630 35.7
727 41.2
335 19.0
74 4.2
|
898 27.1
1312 39.5
884 26.6
225 6.8
|
We calculated the percentage of persons classified as having a cognitive impairment at different MOCA cutoff values (51.3 % of persons who had a score of >= 24 on the MMSE, scored abnormally on the MOCA less than 23. 82.1 % of persons who had a conservative >= 24 on the MMSE, scored abnormally on the MOCA less than 26.
The chi-square test shows that MoCA significantly (p-value<0.001(classified more persons as cognitively impaired than the MMSE (72% and 90% vs 45.1%, respectively; using a cutoff score of 24 on the MMSE, 23 and 26 on the MoCA) (Table 2).
Table 2. Percentage of patients classified as cognitively impaired
MOCA Cutoff
Scores
|
Abnormal
MOCA
|
Abnormal
MMSE (<24)
|
Proportion of MOCA Impaired
Patients with Normal MMSE
|
<23
|
2394(72%)
|
1500(45.1%)
|
51.3%
936/1826
|
<26
|
2995(90%)
|
45.1%
|
82.1%
1500/1826
|
MOCA: Montreal Cognitive Assessment, MMSE: Mini-Mental State Examination
Cutoff points
Graphic representations of the ROC curves are provided in Figure 1. The optimal cutoff point for maximum accuracy (Youden J index) and the respective values of sensitivity, specificity, PPV, NPV are described in Table 3. The cutoff point of below 25 yielded the highest Youden index for the MMSE in discrimination between MCI and healthy basis of MOCA<23 with an AUC of 0.9 (95% CI: 0.89-0.91) and MOCA<26 with an AUC of 0.87 (95% CI:0.86-0.89). With these cutoff points, MMSE had good sensitivity, specificity, PPV and NPV.
Table 3. The optimal cutoff point for maximum accuracy (Youden J index) and the respective values of sensitivity, specificity, PPV, NPV for evaluation the diagnostic accuracy of MMSE
|
Cutoff
|
Youden index J
|
Sensitivity, %
|
Specificity, %
|
PPV
|
NPV
|
MMSE FOR MOCA<23
|
≤23
|
0.56
|
60.90
|
95.49
|
97.2
|
47.8
|
≤24
|
0.63
|
71.14
|
91.95
|
95.8
|
55.4
|
≤25
|
0.66
|
81.24
|
84.23
|
93.0
|
63.6
|
≤26
|
0.57
|
88.35
|
68.56
|
87.8
|
69.6
|
≤27
|
0.42
|
95.28
|
46.46
|
82.1
|
79.3
|
MMSE FOR MOCA<26
|
≤23
|
0.48
|
49.92
|
98.49
|
99.6
|
15.3
|
≤24
|
0.57
|
59.10
|
97.58
|
99.6
|
20.9
|
≤25
|
0.62
|
69.05
|
92.75
|
98.9
|
24.9
|
≤26
|
0.58
|
78.20
|
80.06
|
97.3
|
28.9
|
≤27
|
0.49
|
88.41
|
60.12
|
95.3
|
36.4
|
MOCA: Montreal Cognitive Assessment, MMSE: Mini-Mental State Examination, PPV: Positive Predictive Value, NPV: Negative Predictive Value
The mean score of MoCA was 18.87(±5.31) and 23.67 (±3.84) for MMSE. The mean (SD) scores according to the age group, and education level are shown in Table4.
Table 4. MoCA and MMSE scores on Age group and Education level
Variable
|
MOCA
|
MMSE
|
Mean ± SD
|
Median (IQR)
|
Mean ± SD
|
Median (IQR)
|
AGE
|
50-54
|
20.38±4.78
|
21(17-24)
|
24.50±3.42
|
25(22-27)
|
55-59
|
19.83±4.98
|
20(16-24)
|
24.34±3.54
|
25(22-27)
|
60-64
|
18.75±5.08
|
19(15-22)
|
23.58±3.63
|
24(21-26)
|
65-69
|
17.74±5.06
|
18(14-22)
|
23.21±3.75
|
23(21-26)
|
70-74
|
16.52±5.18
|
17(13-20)
|
22.24±4.01
|
22(20-26)
|
>75
|
14.00±5.49
|
13(10-18)
|
20.29±4.53
|
20(17-24)
|
EDUCATION
|
0
|
13.71±4.37
|
14(11-17)
|
20.24±3.68
|
20(18-23)
|
1-5
|
19.02±3.93
|
19(16-22)
|
23.87±2.96
|
24(22-26)
|
6-12
|
22.33±3.64
|
23(20-25)
|
25.94±2.65
|
26(24-28)
|
>13
|
25.10±2.33
|
25(24-27)
|
27.28±2.04
|
28(26-29)
|
MOCA: Montreal Cognitive Assessment, MMSE: Mini-Mental State Examination, SD: standard deviation, IQR: interquartile range
The mean and SD values of the MMSE and MOCA total scores for each of the age- education- group are shown in Table 5. The median with IQR values was also reported since the information could be useful for clinical reference.
A two-way ANCOVA was run to examine the effect of age and education level on MoCA and MMSE score, while controlling for a gender effect.
There was a significant interaction between the effects of age and education level on MMSE, while controlling for gender effect, F= 1.92, p = .02, partial η2 = .01. The effect size is negligible. Simple main effects analysis showed that people with higher education level have a significantly higher score in MMSE than people with lower education level at any age level (p < .05). Also, Simple main effects analysis showed that in people with high education (>13 years), there was no significant difference in the scores of MMSE between any age level (p > .05). But, in people with education years less than 12 years, the scores of MMSE significantly decrease with increase age (p < .05).
There was no significant interaction between the effects of age and education level on MoCA, while controlling for gender effect, F= 1.48, p = .14, partial η2 = .01. But, the main effects of education (F= 552.50, p < .001, partial η2 = .34) and age (F= 19.34, p < .001, partial η2 = .03) were statistically significant. Partial eta squared showed that the relative impact of education is more than age. Bonferroni post hoc test showed that people with higher education level have a significantly higher score in MoCA than people with lower education level(p<0.05). Also, Bonferroni post hoc test showed that people with higher age have a significantly lower score in MoCA (p<0.05). In general, the above results show that the effect of education on the MMSE and MoCA score is more important than the age.
Table 5. MoCA and MMSE scores on Age- and Education-stratified groups
Age
|
Education
|
MOCA
|
MMSE
|
Mean ± SD
|
Median (IQR)
|
Mean ± SD
|
Median (IQR)
|
50-54
|
0
|
14.61±4.08
|
15(12-17)
|
20.95±3.62
|
21(19-23)
|
1-5
|
19.76±3.75
|
20(17-23)
|
24.16±2.86
|
24(22-26)
|
6-12
|
22.97±3.40
|
23(21-25)
|
26.09±2.55
|
26(25-28)
|
>13
|
25.42±2.25
|
25(24-27)
|
27.37±2.11
|
28(26-29)
|
55-59
|
0
|
14.97±4.36
|
15(12-18)
|
21.21±3.55
|
21(19-24)
|
1-5
|
19.34±4.09
|
20(17-22)
|
24.09±3.03
|
24(22-27)
|
6-12
|
23.14±3.11
|
23(21-26)
|
26.46±2.28
|
27(25-28)
|
>13
|
25.12±2.31
|
25(24-27)
|
27.45±1.81
|
28(27-29)
|
60-64
|
0
|
13.89±4.02
|
14(11-17)
|
20.41±3.24
|
20(18-23)
|
1-5
|
18.80±3.75
|
19(16-21)
|
23.64±2.85
|
24(22-26)
|
6-12
|
22.12±3.54
|
22(20-25)
|
25.96±2.56
|
26(25-28)
|
>13
|
25.19±2.27
|
26(24-26)
|
26.68±2.54
|
27(26-28)
|
65-69
|
0
|
13.81±4.08
|
14(11-16)
|
20.54±3.33
|
21(18-23)
|
|
1-5
|
17.98±3.60
|
18(15-20)
|
23.65±2.92
|
24(22-26)
|
|
6-12
|
21.52±3.79
|
22(18-24)
|
25.43±2.96
|
26(23-28)
|
|
>13
|
23.90±2.57
|
24(22-25)
|
27.45±1.96
|
27(26-29)
|
70-74
|
0
|
12.95±3.99
|
13(11-16)
|
19.75±3.48
|
20(18-22)
|
|
1-5
|
17.85±3.20
|
18(16-20)
|
23.43±2.80
|
24(21-26)
|
|
6-12
|
20.54±3.98
|
21(20-23)
|
24.85±3.19
|
25(22-27)
|
|
>13
|
25.10±2.38
|
26(24-27)
|
27.70±1.25
|
28(27-28)
|
>75
|
0
|
10.69±4.33
|
11(7-13)
|
17.57±3.90
|
18(15-20)
|
1-5
|
16.73±4.12
|
16(13-19)
|
22.51±3.10
|
22(20-25)
|
6-12
|
18.16±4.14
|
18(16-21)
|
24.05±3.07
|
24(22-26)
|
>13
|
23.80±2.49
|
23(22-24)
|
26.20±1.30
|
26(25-27)
|
MOCA: Montreal Cognitive Assessment, MMSE: Mini-Mental State Examination, SD: standard deviation, IQR: interquartile range