Table 2. Sensitivity Analysis on the Association Between Quartile of RAR Parameters and Risk of Parkinson Diseasea.
| RAR parameter | Multivariable-adjusted OR (95% CI)b | |
|---|---|---|
| Multivariable plus nighttime sleep adjustedc | 2-y Time lag | |
| Amplitude, counts/min | ||
| Q4: ≥4209.0 | 1 [Reference] | 1 [Reference] |
| Q3: 3552.1 to <4209.0 | 1.29 (0.57-2.94) | 1.11 (0.51-2.41) |
| Q2: 2929.3 to <3552.1 | 1.42 (0.63-3.20) | 1.31 (0.61-2.80) |
| Q1: <2929.3 | 3.56 (1.68-7.56)d | 2.40 (1.15-5.00)d |
| Per SD decrease (–1098.0) | 1.86 (1.34-2.58)d | 1.54 (1.12-2.12)d |
| P value for trend | <.001 | .01 |
| Mesor, counts/min | ||
| Q4: ≥2433.4 | 1 [Reference] | 1 [Reference] |
| Q3: 2143.5 to <2433.4 | 1.19 (0.52-2.72) | 1.17 (0.53-2.58) |
| Q2: 1859.6 to <2143.5 | 1.32 (0.58-3.01) | 1.10 (0.49-2.45) |
| Q1: <1859.6 | 3.24 (1.52-6.92)d | 2.76 (1.35-5.67)d |
| Per SD decrease (–510.5) | 1.68 (1.21-2.32)d | 1.50 (1.11-2.05)d |
| P value for trend | .001 | .005 |
| Robustness, pseudo-F statistic | ||
| Q4: ≥1318.1 | 1 [Reference] | 1 [Reference] |
| Q3: 970.2 to <1318.1 | 1.53 (0.63-3.71) | 1.19 (0.51-2.74) |
| Q2: 703.1 to <970.2 | 3.29 (1.50-7.22)d | 2.36 (1.13-4.96)d |
| Q1: <703.1 | 3.34 (1.45-7.67)d | 2.33 (1.07-5.07)d |
| Per SD decrease (–512.4) | 1.71 (1.23-2.38)d | 1.42 (1.04-1.93)d |
| P value for trend | <.001 | .009 |
| Acrophase, time of day | ||
| Advanced (<1:04 pm) | 0.84 (0.41-1.76) | 0.65 (0.29-1.46) |
| 1:04 to 3:29 pm | 1 [Reference] | 1 [Reference] |
| Delayed (>3:29 pm) | 0.98 (0.47-2.07) | 0.82 (0.37-1.78) |
Abbreviations: OR, odds ratio; Q, quartile; RAR, rest-activity rhythm.
Using likelihood ratio tests comparing the models containing multiple confounders and traditional indices of sleep disturbances with those models further including the continuous RAR parameters showed a statistically significant improvement in model fit for amplitude (P = .002), mesor (P = .009), and robustness (P = .004).
Adjusted for age, clinic site, race, educational level, number of depressive symptoms, body mass index, physical activity, benzodiazepine use, alcohol and caffeine intake, smoking, cognitive function, and history of hypertension, stroke, coronary heart disease, and type 2 diabetes.
Adjusted for sleep efficiency, total sleep time, apnea-hypopnea index, and periodic limb movement index.
P < .05.