Table 1.
Characteristics of studies investigating longitudinal associations between metabolites and dementia risk.
| References | Study population (country) | Study design |
N (% men), mean age (yrs), follow-up time (yrs) |
Biospecimen | Baseline cognition status of participants | Outcome (N); ascertainment |
|---|---|---|---|---|---|---|
| Mielke et al., 2010 | WHAS II (USA) | Cohort, population-based | 100 (0), 74.0, 9.0 | Non-fasting serum | MMSE score ≥24 and none was impairment | Cognitive impairment on psychomotor speed (TMT-A, 24), executive function (TMT-B, 34), verbal immediate memory (HVLT-immediate recall, 27) and delayed (HVLT-delayed recall, 23) memory; first performance at or below the tenth percentile of each age- and education-matched cognitive test |
| Oresic et al., 2011 | PredictAD project (Finland) | Patient cohort (internal cross-validation) | 226 (37.6), 71.0, 2.3 (mean) | Fasting serum in most participants |
MCI | Converters from MCI to AD (52); MCI, AD, and dementia were diagnosed using the criteria of proposed by MCADRC, NINCDS-ADRDA, and DSM-IV, repectively |
| Mapstone et al., 2014 | Rochester/Orange County Aging Study (USA) | Case-control, longitudinal, population-based (internal cross-validation) | 147 (37.4), 80.2, 2.1 (mean) | Fasting plasma | Non-impaired memory | Converters from non-impaired memory to aMCI/AD (74); aMCI and AD were classified as met the criteria of amnestic subtype of MCI, and NINCDS-ADRDA, repectively |
| Mousavi et al., 2014 | Betula study (Sweden) | Case-control, longitudinal, population-based | 93 (32.0), 65.6, 5.0 | Non-fasting serum | Cognitively normal | Dementia (31); dementia, AD, and vascular dementia were diagnosed using the DSM-IV, NINCDS-ADRDA, Gorelick's criteria, repectively |
| Graham et al., 2015 | Belfast City Hospital memory clinic patients (UK) | Case-control, longitudinal (internal cross-validation) | 72 (45.8), 77.9, 2.0 | Fasting plasma | MCI | Converters from MCI to AD (19); MCI and AD were classified as met the criteria of an international working group on MCI, and NINCDS-ADRDA, respectively |
| Casanova et al., 2016 | BLSA (USA) and AGES-RS (Iceland) | Case-control, longitudinal, population-based (internal cross-validation) | BLSA: 192 (51.0), 77.2, 4.3 (mean); AGES-RS:200 (45.5), 78.2, 5.2 (mean) |
Fasting serum | Cognitively normal | AD (93 in BLSA and 100 in AGES-RS); BLSA: DSM-III-R and the NINCDS-ADRDA criteria, for dementia and AD, respectively; AGES-RS: consensus made by a panel that includes a geriatrician, neurologist, neuropsychologist, and neuroradiologist based on a 3-step procedure, including cognitve test (MMSE, DSST), neurologic examination, and medical history and social, cognitive, and daily functioning relevant to the diagnosis |
| Simpson et al., 2016 | BLSA-NI (USA) | Cohort, population-based | 107 (61.0), 72.9, 7.0 (median) | Fasting plasma | Non-dementia | Cognitive decline; CVLT, TMT-A, TMT-B, MMSE, BVRT, CRT, and CLF were used for verbal memory, processing speed, executive function, global cognitive function, visual memory, visuo-spatial ability, and language fluencies, repectively |
| Abdullah et al., 2017 | ADAPT (USA) | Case-control, longitudinal | 195 (50.3), 78.0, 3.0 | Non-fasting serum | Cognitively normal | MCI (15) and AD (8); Petersen criteria and NINCDS-ADRDA were used to diagnose MCI and AD, repectively |
| Bressler et al., 2017 | ARIC-NS (USA) | Cohort, population-based | 6-year cognitive change: 1,035 (33.0), 55.0, 6.0; Incident dementia: 1,534 (36.4), 53.4, 17.1 (median) |
Fasting serum | NR | 6-year cognitive change, difference between the cognitve test scores obtained at baseline and those at follow-up, including DWRT, DSST, and WFT; Dementia (141), hospital records |
| Chouraki et al., 2017 | Framingham offspring Cohort (USA) | Cohort, population-based | 2,067 (47.6), 55.9, 15.8 (mean) | Fasting plasma | Dementia-free | Dementia (93, including 68 AD); DSM-IV, and NINCDS-ADRDA for dementia and AD, respectively |
| Li et al., 2017 | ARIC-NS (USA) | Cohort, population-based | 221 (33.5), 71.3, 7.3 (median) | Fasting plasma | Cognitively normal | MCI (77), dementia (18) and cognitive score change; criterias of NIA-AA and DSM-V were used to classify MCI and dementia, repectively; cognitive test score changes in MMSE, DWRT, DSS, WFT, and a composite global score based on the above tests. |
| Toledo et al., 2017 | Discovery: ADNI-1 (USA) Validation: ERF (the Netherland) RS (the Netherland) |
Discovery: Patient cohort Validation: Cohort, population-based (external validation) |
ADNI: 734 (42.4), 75.1, 3 (median) ERF: 905 (43.7), 48.0, NR RS: 2752 (41.8), 74.2, 9.7 (median) |
Fasting serum | ADNI: 199 cognitivly normal, 358 MCI, and 175 AD ERF and RS: cognitivly normal |
Converters from MCI to AD (NR); MCI and AD were diagnosed using criteria for aMCI and NINDS-ADRDA, respectively Cognitive change: ADAS-Cog13 in ADNI, general cognitive ability (g-factor) in ERS and RS based cognitive tests. |
| Dorninger et al., 2018 | VITA (Austria) | Case-control, longitudinal, population-based | 174 (36.7), 75.8, 7.5 | Fasting plasma | Non-dementia | AD (22); DSM-IV and NINCDS-ADRDA were used to classify dementia and AD, respectively |
| Tynkkynen et al., 2018 | Discovery: Finrisk 97 (Finland) DILGOM (Finland) WH II (UK) EGCUT (Estonia) Validation: Health 2000 (Finland) FHS (USA) RS (the Netherland) ERF (the Netherland) |
Cohort, population-based (external validation) | A total of 22623; Finrisk 97: 4580 (50.4), 55.1, 10.0 (median) DILGOM: 3399 (47.0), 57.5, 7.9 (median) WH II: 4612 (73.5), 55.9, 17.0 (median) EGCUT: 2570 (41.1), 59.6, 7.5 (median) Health 2000: 1166 (42.9), 56.7, 10.0 (median) FHS: 2171 (47.4), 56.7, 17.9 (median) RS: 2759 (42.4), 75.0, 9.4 (median) ERF: 1366 (42.5), 49.2, 11.6 (median) |
Fasting serum in six cohorts (except for Finrisk 97 fast for 4 h); Fasting plasma in FHS | Free of dementia | Dementia (995) and AD (745); based on continuous follow-up health records in the Finrisk97, ERF, WH II, EGCUT, Health 2000, and DILGOM; RS, based on the general practitioner records and cognitive screening met for DSM-III-R; FHS, based on cognitive test met for DSM IV (dementia) and NINCDS-ADRDA (AD). |
| van der Lee et al., 2018 | Cognition analysis: Discovery: RS (the Netherland) ERF (the Netherland) Replication: WH II (UK) NTR (the Netherland) FHS (USA) SHIP (Germany) Dementia analysis: RS (the Netherland) ERF (the Netherland) WH II (UK) VUMC ADC (the Netherland) AgeCoDe (Germany) EGCUT (Estonia) Finrisk 97 (Finland) DILGOM (Finland) |
Cohort, population-based (external validation) | Cognition analysis: Discovery: 5188 in total (2683 (43.1), 48.9 in ERF, 2505 (41.8), 74.2 in RS); Replication: 7,025 in total (4235 (73.8), 55.8 in WH II, 338 (37.6), 40.7 in NTR, 944 (43.6), 50.1 in SHIP, 1508 (54.9), 55.7 in FHS); Dementia analysis: 27,000 in total, 1532 in ERF, 11.3 (mean); 2010 in ERF 7.6 (mean); 4612 in WH II, 16.7 (mean); 2356 in FHS, 15.7 (mean); 1303 (54.9), 64.1 in VUMC ADC, NR; 7517 (45.3), 48.8 in Finrisk 97, 9.7 (mean); 4788 (44.2), 52.3 in DILGOM, 7.7 (mean); 2572 (41.1), 59.1 in EGCUT, 7.0 (mean); 310 (31.6), 84.1, in AgeCoDe, 4.5 (mean) |
Fasting plasma (except for AUMC ADC and Finrisk 97) | Non-dementia for incident dementia analysis |
Dementia (1990) and AD (1356); General cognitive ability score weres selcet as the first PCA component based on cognitive tests, including at least three cognitive domains. Dementia and AD: based on continuous follow-up health records in the ERF, WHII, EGCUT, Finrisk 97, and DILGOM; RS, based on the general practitioner records and cognitive screening met for DSM-III-R; FHS, based on cognitive test met for DSM IV (dementia) and NINCDS-ADRDA (AD); AgeCoDe, based on health records and cognitive tese met for DSM IV. |
| Varma et al., 2018 | BLSA (USA) ADNI (USA) |
BLSA: Cohort, population-based ADNI: Patient cohort, |
BLSA: 207 (48.3), 78.7, 4.3 (mean) ADNI: 767 (57.4), 75.2, 3.0 (mean) |
Fasting serum | Cognitively normal in BLSA; MCI in ADNI |
Incident AD (92) and cognitive performance change in BLSA, converters from MCI to AD in ADNI (185); DSM-III-R and NINCDS-ADRDA was used to indentify dementia and AD, respectively; MCI was diagnosed using Petersen criteria. Cognitive performence was based on neuropsychological tests: CVLT for memory, TMT-A and Digit Forward for attention, TMT-B and Digit backward for executive function, Letter and Semantic Fluency for language, and CDT and CRT for visuo-spatial ability. |
ADAPT, Alzheimer's Disease Anti-inflammatory Prevention Trial; ADAS-Cog13, Alzheimer's Disease Assessment Scale–Cognition, lower levels indicate better cognition; ADNI, Alzheimer's Disease Neuroimaging Initiative; AgeCoDe, German Study on Aging, Cognition, and Dementia; AGES-RS, Age, Gene/Environment Susceptibility-Reykjavik Study; ARIC-NS, Atherosclerosis Risk in Communities-Neurocognitive Study; BLSA-(NI), Baltimore Longitudinal Study of Aging-(neuroimaging substudy); BVRT, Benton Visual Retention Test; CDT, Clock Drawing Test; CLF, Category and Letter Fluency Test; CRT, Card Rotation Test; CVLT, California Verbal Learning Test; DILGOM, Dietary, Lifestyle and Genetic determinants of Obesity and Metabolic Syndrome Study; DSM, Diagnostic and Statistical Manual of Mental Disorders; DSST, Digit Symbol Substitution Test; DWRT, Delayed Word Recall Test; EGCUT, Estonian Biobank (Estonian Genome Center, University of Tartu); ERF, Erasmus Rucphen Family study; FHS, Framingham Heart Study; HVLT, Hopkins Verbal Learning Test; MCADRC, Mayo Clinic Alzheimer's Disease Research Center; MCI, mild cognitive impairment; MMSE, Mini-Mental State Examination; NINCDS-ADRDA, National Institute of Neurologic and Communicative Disorders and Stroke, and Alzheimer's Disease and Related Disorders Association criteria; NTR, Netherlands Twin Registry; NR, not reported; RS, Rotterdam Study; SHIP-Trend, Study of Health in Pomerania–Trend; TMT, Trail Making Test; VUMC ADC, VUMC Amsterdam Dementia Cohort; WFT, Word Fluency Test; WH II, Whitehall II study; WHAS II, Women's Health and Aging Study (WHAS) II; yrs, years.