Table 3.
Attributing effects to additive interaction between low educational attainment and high genetic risk on dementia risk
| APOE ε4–related genetic risk | Overall genetic risk (PRS) | ||
|---|---|---|---|
| Main effects | Hazard ratio (95% CI) | Main effects | Hazard ratio (95% CI) |
| High APOE ε4–related genetic risk | 2.69 (2.33–3.12) | High overall genetic susceptibility | 2.37 (2.05–2.75) |
| Low educational attainment | 0.99 (0.85–1.16) | Low educational attainment | 0.98 (0.83–1.15) |
| Joint effect | 3.87 (3.73–4.02) | Joint effect | 3.34 (3.20–3.49) |
| Relative excess risk due to interaction a | Hazard ratio (95% CI) | Relative excess risk due to interaction a | Hazard ratio (95% CI) |
| Relative excess risk due to interaction | 1.19 (0.71–1.67) | Relative excess risk due to interaction | 1.00 (0.59–1.40) |
| P-value | < 0.001 | P-value | < 0.001 |
| Attributable proportion | % (95% CI) | Attributable proportion, % | % (95% CI) |
| High APOE ε4–related genetic risk | 59.0 (46.4–71.5) | High overall genetic susceptibility | 58.7 (45.8–71.6) |
| Low educational attainment | − 0.004 (− 0.05 to 0.05) | Low educational attainment | − 0.01 (− 0.07 to 0.05) |
| Additive interaction | 41.5 (27.6–55.3) | Additive interaction | 42.6 (27.6–57.5) |
Results were adjusted for age, sex, Townsend deprivation index, moderating drinking, current smoking, healthy diet score, obesity, diabetes, hypertension, depression, social isolation, hearing problem, regular physical activity, high cholesterol, cardiovascular disease and traumatic brain injury, the first 10 genetic principal components, relatedness, and genotyping array
aOn the hazard ratio scale, we decomposed the joint excess relative risk (HR11) for both exposures into the excess relative risk for low educational attainment (HR01), high genetic risk (HR10), and relative excess risk due (RERI) to their interaction. Specifically, we have HR11 − 1 = (HR01 − 1) + (HR10 − 1) + RERI