TABLE 2.
Association between Center for Epidemiologic Studies Depression Scale (CES-D) score at t1 and CES-D ≥ 17 at t1 and incident mild cognitive impairment (MCI) at t2 for the total sample and stratified by sex.
| Relative risk for incident MCI at t2 | ||||||||
|
| ||||||||
| No MCI at t2 | Incident MCI at t2 | CES-D1 score at t1 | Unadjusted RR (95% CI) | P-value | Adjusted (1) RR (95% CI) | P-value | Adjusted (2) RR (95% CI) | P-value |
| Total1 | ||||||||
| n = 1392 | n = 146 | Increase by one point | 1.052 (1.030–1.074) | <0.001 | 1.054 (1.032–1.077) | <0.001 | 1.053 (1.031–1.076) | <0.001 |
| Men | ||||||||
| n = 651 | n = 71 | Increase by one point | 1.044 (1.010–1.078) | 0.010 | 1.046 (1.013–1.081) | 0.006 | 1.046 (1.012–1.081) | 0.007 |
| Women | ||||||||
| n = 741 | n = 75 | Increase by one point | 1.060 (1.031–1.089) | <0.001 | 1.061 (1.031–1.091) | <0.001 | 1.059 (1.029–1.090) | <0.001 |
|
| ||||||||
| CES-D1 ≥ 17 at t1 | ||||||||
|
| ||||||||
| Total1 | ||||||||
| n = 1301 | n = 126 | No | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | |||
| n = 91 | n = 20 | Yes | 2.04 (1.27–3.27) | 0.003 | 2.24 (1.39–3.61) | 0.001 | 2.22 (1.38–3.58) | 0.001 |
| Men | ||||||||
| n = 616 | n = 66 | No | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | |||
| n = 35 | n = 5 | Yes | 1.29 (0.52–3.21) | 0.58 | N/A2 | N/A2 | ||
| Women | ||||||||
| n = 685 | n = 60 | No | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | |||
| n = 56 | n = 15 | Yes | 2.62 (1.49–4.62) | 0.001 | 2.68 (1.52–4.73) | 0.001 | 2.59 (1.46–4.58) | 0.001 |
MCI, mild cognitive impairment; CES-D, Center for Epidemiologic Studies Depression Scale; RR, relative risk; CI, confidence interval, N/A, not applicable.
Data are from log-linear regression models with a Poisson working likelihood; significant p-values are presented in bold.
(1) Adjusted for age at t1, sex (only for the total sample calculations: male/female) and years of education (≤10 years, 11–13 years, 14–17 years, ≥18 years).
(2) Additionally adjusted for diabetes mellitus at t1 (yes/no), history of coronary heart disease at t1 (yes/no) and history of stroke at t1 (yes/no).
1Missing values for n = 9 participants.
2No calculation due to small sample sizes of incident MCI cases.