Table 1.
Selected prospective cohort studies which conducted observational analyses of testosterone as the exposure, and incidence of cognitive impairment or dementia (including dementia due to Alzheimer disease) as the outcome
| Study author and year | Size (N men) | Age (years) | Follow-up (years) | Summary of results |
|---|---|---|---|---|
| Moffat et al. [57] | 407 | 50–91 | 10 | Higher ratio of testosterone to SHBG was associated with better scores on visual and verbal memory, visuospatial functioning, visuomotor scanning, and lower rate of longitudinal decline in memory |
| Moffat et al. [58] | 574 | 32–87 | 19 | 54 men developed dementia due to Alzheimer diseasea. Increased ratio of testosterone to SHBG was associated with decreased risk (hazard ratio 0.74 per 10 nmol/mol increase) |
| Geerlings et al. [59] | 2,974 | 71–93 | 6 | 134 men developed dementia of Alzheimer’s typea. Testosterone was not associated with risk of dementia, higher estradiol concentrations were associated with risk of Alzheimer’s disease (hazard ratio 1.25 per 1 SD increase) |
| Ravaglia et al. [60] | 376 | ≥ 65 | 3.8 | 39 men developed dementia (23 dementia of Alzheimer’s type, 12 vascular dementia)a. Neither cFT nor estradiol concentrations were associated with risk of dementia |
| LeBlanc et al. [61] | 1,022 | ≥ 65 | 4.5 | No association of baseline cFT or calculated free estradiol with change in cognition. Higher SHBG was associated with increased risk of cognitive decline (executive function and motor speed, general cognition) |
| Chu et al. [62] | 155 | ≥ 55 | 1 | 10 men developed dementia.a Higher bioavailable testosterone (measured using ammonium sulphate precipitation) associated lower risk of Alzheimer’s disease at 1 year |
| Carcaillon et al. [63] | 503 | ≥ 65 | 4 | 105 men who developed incident dementiaa, and random sample of 413 men as controls. Non-linear association of dementia with baseline testosterone (hazard ratio lower tertile 2.33, P = 0.026, upper tertile 1.9, P = 0.126, vs middle tertile). Risk of dementia associated with lower bioavailable testosterone was greater in men aged ≥ 80 vs men aged < 80 years |
| Ford et al. [14] | 4,069 | 71–88 | 10.5 | 499 men developed dementiab. Lower baseline testosterone was associated with higher risk of incident dementia (hazard ratio 1.14 per 1 SD decrease), as was lower cFT (hazard ratio 1.18 per 1 SD decrease). Lower estradiol was associated with higher risk of incident dementia (hazard ratio 1.11 per 1 SD decrease) but SHBG was not associated |
| Marriott et al. [13] | 159,411 | 40–69 | 7 | 826 men developed dementia, of which 288 were classified as having Alzheimer’s diseaseb. Lower testosterone concentrations were associated with higher incidence of dementia (overall trend P = 0.001, hazard ratio 1.43 for lowest vs highest quintile), and Alzheimer’s disease (overall trend P = 0.017, hazard ratio 1.80 for lowest vs highest quintile). Lower SHBG was associated with lower incidence of dementia and Alzheimer’s disease (P = < 0.001, hazard ratio 0.66; P = 0.012, hazard ratio 0.53 for lowest vs highest quintile, respectively) |
Multiple criteria were used for ascertainment of dementia outcomes including Diagnostic and Statistical Manual of the American Psychiatric Association (DSM), International Classification of Disease of the World Health Organization (ICD) and the National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA) Alzheimer’s criteria
acognitive assessment undertaken
bdementia diagnosis based on hospital morbidity and other health registry data