| Effects . | Hazard Difference (95% CI) . | Hazard Ratio (95% CI) . |
|---|---|---|
| Total effect | 55 (35, 74) | 1.32 (1.11, 1.56) |
| Indirect effect by LM | 26 (14, 37) | 1.10 (1.05, 1.16) |
| Indirect effect by LM + Horvath DNAmAA | 25 (12, 39) | 1.10 (1.04, 1.16) |
| Indirect effect by Horvath DNAmAA but not LM | −0.8 (−13.6, 11.9) | 1.00 (0.93, 1.07) |
| Indirect effect by LM + Hannum DNAmAA | 27 (15, 40) | 1.11 (1.05, 1.17) |
| Indirect effect by Hannum DNAmAA but not LM | 1.5 (−11.2, 14.3) | 1.01 (0.94, 1.08) |
| Indirect effect by LM + DNAmPhenoAA | 31 (17, 44) | 1.13 (1.07, 1.19) |
| Indirect effect by DNAmPhenoAA but not LM | 4.7 (−8.3, 17.6) | 1.02 (0.95, 1.10) |
| Indirect effect by LM + DNAmGrimAA | 35 (20, 50) | 1.15 (1.09, 1.21) |
| Indirect effect by DNAmGrimAA but not LM | 8.8 (−4.2, 21.7) | 1.04 (0.97, 1.12) |
| Effects . | Hazard Difference (95% CI) . | Hazard Ratio (95% CI) . |
|---|---|---|
| Total effect | 55 (35, 74) | 1.32 (1.11, 1.56) |
| Indirect effect by LM | 26 (14, 37) | 1.10 (1.05, 1.16) |
| Indirect effect by LM + Horvath DNAmAA | 25 (12, 39) | 1.10 (1.04, 1.16) |
| Indirect effect by Horvath DNAmAA but not LM | −0.8 (−13.6, 11.9) | 1.00 (0.93, 1.07) |
| Indirect effect by LM + Hannum DNAmAA | 27 (15, 40) | 1.11 (1.05, 1.17) |
| Indirect effect by Hannum DNAmAA but not LM | 1.5 (−11.2, 14.3) | 1.01 (0.94, 1.08) |
| Indirect effect by LM + DNAmPhenoAA | 31 (17, 44) | 1.13 (1.07, 1.19) |
| Indirect effect by DNAmPhenoAA but not LM | 4.7 (−8.3, 17.6) | 1.02 (0.95, 1.10) |
| Indirect effect by LM + DNAmGrimAA | 35 (20, 50) | 1.15 (1.09, 1.21) |
| Indirect effect by DNAmGrimAA but not LM | 8.8 (−4.2, 21.7) | 1.04 (0.97, 1.12) |
Absolute (hazard difference per 10 000 person-years) and relative (hazard ratio) size of total effect of low (vs high) educational attainment on all-cause mortality, and for indirect effects by unhealthy lifestyle and morbidities (LM) and epigenetic aging biomarkers (Horvath DNAmAA, Hannum DNAmAA, DNAmPhenoAA, and DNAmGrimAA). All effects are pooled estimates of a single cohort’s effects through an inverse variance-weighted meta-analysis model. The total number of participants/deaths across cohorts was 6 156/1 586.
| Effects . | Hazard Difference (95% CI) . | Hazard Ratio (95% CI) . |
|---|---|---|
| Total effect | 55 (35, 74) | 1.32 (1.11, 1.56) |
| Indirect effect by LM | 26 (14, 37) | 1.10 (1.05, 1.16) |
| Indirect effect by LM + Horvath DNAmAA | 25 (12, 39) | 1.10 (1.04, 1.16) |
| Indirect effect by Horvath DNAmAA but not LM | −0.8 (−13.6, 11.9) | 1.00 (0.93, 1.07) |
| Indirect effect by LM + Hannum DNAmAA | 27 (15, 40) | 1.11 (1.05, 1.17) |
| Indirect effect by Hannum DNAmAA but not LM | 1.5 (−11.2, 14.3) | 1.01 (0.94, 1.08) |
| Indirect effect by LM + DNAmPhenoAA | 31 (17, 44) | 1.13 (1.07, 1.19) |
| Indirect effect by DNAmPhenoAA but not LM | 4.7 (−8.3, 17.6) | 1.02 (0.95, 1.10) |
| Indirect effect by LM + DNAmGrimAA | 35 (20, 50) | 1.15 (1.09, 1.21) |
| Indirect effect by DNAmGrimAA but not LM | 8.8 (−4.2, 21.7) | 1.04 (0.97, 1.12) |
| Effects . | Hazard Difference (95% CI) . | Hazard Ratio (95% CI) . |
|---|---|---|
| Total effect | 55 (35, 74) | 1.32 (1.11, 1.56) |
| Indirect effect by LM | 26 (14, 37) | 1.10 (1.05, 1.16) |
| Indirect effect by LM + Horvath DNAmAA | 25 (12, 39) | 1.10 (1.04, 1.16) |
| Indirect effect by Horvath DNAmAA but not LM | −0.8 (−13.6, 11.9) | 1.00 (0.93, 1.07) |
| Indirect effect by LM + Hannum DNAmAA | 27 (15, 40) | 1.11 (1.05, 1.17) |
| Indirect effect by Hannum DNAmAA but not LM | 1.5 (−11.2, 14.3) | 1.01 (0.94, 1.08) |
| Indirect effect by LM + DNAmPhenoAA | 31 (17, 44) | 1.13 (1.07, 1.19) |
| Indirect effect by DNAmPhenoAA but not LM | 4.7 (−8.3, 17.6) | 1.02 (0.95, 1.10) |
| Indirect effect by LM + DNAmGrimAA | 35 (20, 50) | 1.15 (1.09, 1.21) |
| Indirect effect by DNAmGrimAA but not LM | 8.8 (−4.2, 21.7) | 1.04 (0.97, 1.12) |
Absolute (hazard difference per 10 000 person-years) and relative (hazard ratio) size of total effect of low (vs high) educational attainment on all-cause mortality, and for indirect effects by unhealthy lifestyle and morbidities (LM) and epigenetic aging biomarkers (Horvath DNAmAA, Hannum DNAmAA, DNAmPhenoAA, and DNAmGrimAA). All effects are pooled estimates of a single cohort’s effects through an inverse variance-weighted meta-analysis model. The total number of participants/deaths across cohorts was 6 156/1 586.
This PDF is available to Subscribers Only
View Article Abstract & Purchase OptionsFor full access to this pdf, sign in to an existing account, or purchase an annual subscription.
