Background
First author, year | Study design | Follow-up years | Sample size (n) | Age (years) | Female (%) | Results |
---|---|---|---|---|---|---|
Adiponectin | ||||||
van Himbergen, 2012 [6] | Prospective cohort study | 13 | 840 | 72.6 | 64.4 | Positive association: Higher plasma adiponectin levels were associated with a higher risk of incident all-cause dementia and AD only in females, not in males. |
Teixeira, 2013 [8] | Cross-sectional/prospective cohort study | 2.5 | 157 | 71.4 | 73.6 | (Cross-sectional study) Negative association: Serum adiponectin levels were lower in MCI and AD as compared to normal cognition. (Prospective study) No association between adiponectin levels and incident MCI or AD. |
Kitagawa, 2016 [9] | Prospective cohort study | 6.9 | 466 | 67.8 | 43 | No association between serum adiponectin levels and incident all-cause dementia, AD, and vascular dementia. |
Gilbert, 2018 [10] | Prospective cohort study | 1.2 | 205 | 80.6 | 65.4 | No association between baseline serum adiponectin levels and the course of cognitive decline. |
van Andel, 2021 [7] | Prospective cohort study | 3 | 898 | 69.8 | 53.9 | Positive association: Higher serum adiponectin levels were associated with more decline in general cognitive function and information processing speed only in females. |
Kim, 2021 [11] | Prospective cohort study | 5.3 | 345 | 76.6 | 34.5 | No association between baseline serum adiponectin levels and the course of cognitive decline in MCI. |
Ban, 2007 [12] | Cross-sectional study | 0 | 60 | 79.3 | 45 | No association: Serum adiponectin levels were not significantly different between normal cognition and vascular dementia group. |
Une, 2011 [13] | Cross-sectional study | 0 | 73 | 74.7 | 60.3 | Positive association: Plasma adiponectin levels were higher in MCI and AD as compared to normal cognition. |
Letra, 2019 [18] | Cross-sectional study | 0 | 124 | 73.9 | 68.5 | Positive association: Plasma adiponectin levels were higher in AD as compared to MCI. |
Leptin | ||||||
Lieb, 2009 [5] | Prospective cohort study | 8.3 | 785 | 79 | 62 | Negative association: Higher plasma leptin levels were associated with a lower risk of incident all-cause dementia and AD. |
Holden, 2009 [14] | Prospective cohort study | 5 | 2870 | 73.7 | 51 | Negative association: Higher serum leptin levels were associated with less cognitive decline. |
Gustafson, 2012 [15] | Prospective cohort study | 32 | 1384 | 46.9 | 100 | No association between leptin levels and incident all-cause dementia. |
Zeki Al Hazzouri, 2013 [16] | Prospective cohort study | 4 | 579 | 82.6 | 100 | Negative association: Higher leptin levels were associated with a lower risk of incident all-cause dementia or MCI only in those with BMI < 25 kg/m2. |
Oania, 2015 [17] | Prospective cohort study | 3 | 352 | 74.7 | 36.4 | No association between leptin levels and incident all-cause dementia. |
Gilbert, 2018 [10] | Prospective cohort study | 1.2 | 205 | 80.6 | 65.4 | No association between baseline serum leptin levels and cognitive decline. |
van Andel, 2021 [7] | Prospective cohort study | 3 | 898 | 69.8 | 53.9 | No association between baseline serum leptin levels and cognitive decline. |
Ban, 2007 [12] | Cross-sectional study | 0 | 60 | 79.3 | 45 | No association: Serum leptin levels were not significantly different between normal cognition and vascular dementia group. |
Letra, 2019 [18] | Cross-sectional study | 0 | 124 | 73.9 | 68.5 | No association: Plasma leptin levels were not significantly different between MCI and AD. |
Methods
Participants and data acquisition
Cognitive function assessment
Structural brain MRI analysis
Biomarker measurements
Other clinical variables affecting the levels of adipokines or the risk of AD
Statistical analysis
Results
Baseline characteristics of participants
Aβ (−) (n = 31) | Aβ (+) (n = 125) | p-value | |
---|---|---|---|
Age (years) | 75.3 (8.59) | 74.3 (7.31) | 0.542 |
Sex | |||
Male | 21 (67.7%) | 84 (67.2%) | 1 |
Female | 10 (32.3%) | 41 (32.8%) | |
Education (years) | 15.6 (3.32) | 16.1 (2.85) | 0.461 |
History of hypertension | |||
No | 12 (38.7%) | 50 (40.0%) | 1 |
Yes | 19 (61.3%) | 75 (60.0%) | |
History of diabetes mellitus | |||
No | 26 (83.9%) | 110 (88.0%) | 0.752 |
Yes | 5 (16.1%) | 15 (12.0%) | |
PPARγ agonists usea | |||
No | 31 (100%) | 124 (99.2%) | 1 |
Yes | 0 (0%) | 1 (0.8%) | |
Ever smoker | |||
No | 15 (48.4%) | 75 (60.0%) | 0.333 |
Yes | 16 (51.6%) | 50 (40.0%) | |
History of alcohol abuse | |||
No | 28 (90.3%) | 119 (95.2%) | 0.540 |
Yes | 3 (9.7%) | 6 (4.8%) | |
BMI (kg/m2) | 25.1 (2.50) | 24.6 (2.51) | 0.321 |
Plasma insulin (uIU/mL) | 3.12 (3.37) | 2.32 (2.79) | 0.227 |
eGFR (mL/min/1.73 m2) | 69.4 (16.3) | 69.8 (15.0) | 0.907 |
Plasma adiponectin (μg/mL) | 6.33 (3.70) | 6.77 (3.95) | 0.564 |
Plasma leptin (ng/mL) | 10.5 (9.79) | 9.61 (9.29) | 0.632 |
CSF Aβ (pg/mL)b | 1530 (181) | 625 (182) | < 0.001 |
Number of APOE ε4 allele | |||
0 | 26 (83.9%) | 47 (37.6%) | < 0.001 |
1 | 5 (16.1%) | 61 (48.8%) | |
2 | 0 (0%) | 17 (13.6%) | |
ADAS-Cog | 9.34 (3.93) | 12.4 (4.73) | < 0.001 |
Cortical thickness in PHC, right (mm) | 2.46 (0.340) | 2.34 (0.317) | 0.097 |
Cortical thickness in PHC, left (mm) | 2.49 (0.334) | 2.37 (0.367) | 0.089 |
Cortical thickness in ERC, right (mm) | 3.31 (0.468) | 3.11 (0.502) | 0.041 |
Cortical thickness in ERC, left (mm) | 3.04 (0.554) | 2.98 (0.510) | 0.606 |
Follow-up period (months) | 56.7 (41.7) | 53.7 (37.6) | 0.713 |
Number of repeated measurements | |||
ADAS-Cog | 7.06 (3.48) | 6.44 (2.78) | 0.359 |
MRI | 5.26 (1.75) | 5.34 (2.04) | 0.814 |
Relationship between adipokines and clinical characteristics at baseline
Predictor | Outcome—adiponectin | Outcome—leptin | ||||
---|---|---|---|---|---|---|
Beta | SE | p-value | Beta | SE | p-value | |
Intercept | 1.480 | 0.491 | 0.003 | − 0.952 | 0.587 | 0.107 |
Aβ (+) | − 0.028 | 0.050 | 0.580 | 0.059 | 0.060 | 0.333 |
ADAS-Cog | 0.007 | 0.005 | 0.158 | 0.006 | 0.006 | 0.310 |
PHC, right | 0.115 | 0.081 | 0.160 | − 0.014 | 0.097 | 0.884 |
PHC, left | − 0.263 | 0.076 | < 0.001 | − 0.007 | 0.090 | 0.934 |
ERC, right | 0.003 | 0.051 | 0.958 | − 0.062 | 0.061 | 0.311 |
ERC, left | 0.069 | 0.051 | 0.176 | 0.016 | 0.061 | 0.796 |
Age | 0.000 | 0.003 | 0.935 | − 0.001 | 0.003 | 0.806 |
Sex (female) | 0.125 | 0.043 | 0.004 | 0.481 | 0.051 | < 0.001 |
Education | − 0.003 | 0.007 | 0.707 | 0.000 | 0.008 | 0.966 |
APOE ε4 | 0.009 | 0.029 | 0.762 | − 0.038 | 0.035 | 0.286 |
BMI | − 0.025 | 0.008 | 0.001 | 0.074 | 0.009 | < 0.001 |
Hypertension | − 0.004 | 0.037 | 0.920 | 0.088 | 0.045 | 0.050 |
Diabetes mellitus | − 0.037 | 0.057 | 0.518 | − 0.110 | 0.069 | 0.110 |
PPARγ agonist | 0.306 | 0.232 | 0.190 | − 0.127 | 0.278 | 0.647 |
Smoking | 0.050 | 0.040 | 0.219 | 0.046 | 0.048 | 0.337 |
Alcohol | − 0.015 | 0.083 | 0.858 | − 0.047 | 0.100 | 0.637 |
Insulin | − 0.132 | 0.060 | 0.030 | 0.332 | 0.072 | < 0.001 |
eGFR | 0.000 | 0.001 | 0.832 | − 0.002 | 0.002 | 0.123 |
The impacts of baseline adipokines on changes in cognition and cortical thickness across Aβ conditions
Adiponectin
Outcome | Adiponectin × time interaction | |||||
---|---|---|---|---|---|---|
Aβ (+) | Aβ (−) | |||||
Beta | SE | p-value | Beta | SE | p-value | |
ADAS-Cog | 0.224 | 0.093 | 0.018 | − 0.018 | 0.054 | 0.744 |
PHC, right | − 0.004 | 0.002 | 0.012 | − 0.002 | 0.002 | 0.529 |
PHC, left | − 0.004 | 0.002 | 0.025 | − 0.002 | 0.003 | 0.411 |
ERC, right | − 0.007 | 0.003 | 0.024 | 0.010 | 0.005 | 0.042 |
ERC, left | − 0.005 | 0.003 | 0.122 | 0.002 | 0.004 | 0.690 |
BMI change | − 0.004 | 0.016 | 0.805 | 0.054 | 0.042 | 0.216 |
Leptin
Outcome | Leptin × time interaction | |||||
---|---|---|---|---|---|---|
Aβ (+) | Aβ (−) | |||||
Beta | SE | p-value | Beta | SE | p-value | |
ADAS-Cog | 0.091 | 0.059 | 0.128 | 0.000 | 0.030 | 0.988 |
PHC, right | 0.001 | 0.001 | 0.485 | 0.001 | 0.001 | 0.190 |
PHC, left | 0.001 | 0.001 | 0.428 | 0.000 | 0.001 | 0.774 |
ERC, right | − 0.001 | 0.002 | 0.723 | 0.000 | 0.002 | 0.998 |
ERC, left | 0.000 | 0.002 | 0.833 | 0.000 | 0.002 | 0.890 |