Adjusted IRRs and 95% CIs for atrial fibrillation according to characteristics of the exposure to the death of a child
| Exposure | No of events | Rate/105 person-year | Age-adjusted IRR (95% CI) | Multivariable IRR (95% CI)* | P value† |
| Unexposed | 72 232 | 121.2 | 1.00 | 1.00 | – |
| All deaths | 2473 | 236.4 | 1.10 (1.07 to 1.15) | 1.12 (1.08 to 1.17) | – |
| Cause of death of the deceased | |||||
| Death due to CVD | 139 | 569.6 | 1.34 (1.13 to 1.58) | 1.42 (1.20 to 1.69) | Reference |
| Death due to other causes | 2334 | 228.5 | 1.09 (1.05 to 1.14) | 1.11 (1.06 to 1.16) | <0.0001 |
| Age of the deceased child at loss (in years) | |||||
| ≤1 | 624 | 122.1 | 1.27 (1.18 to 1.38) | 1.17 (1.08 to 1.27) | 0.0118 |
| 2–12 | 337 | 169.7 | 1.10 (0.98 to 1.22) | 1.07 (0.96 to 1.19) | 0.2041 |
| 13–18 | 193 | 244.0 | 0.94 (0.82 to 1.09) | 0.95 (0.82 to 1.10) | Reference |
| 19–29 | 655 | 370.1 | 1.00 (0.93 to 1.08) | 1.04 (0.96 to 1.13) | 0.2573 |
| >29 | 664 | 826.7 | 1.14 (1.05 to 1.23) | 1.26 (1.17 to 1.36) | 0.0005 |
| No of remaining live children at loss | |||||
| 0 | 375 | 133.2 | 1.25 (1.13 to 1.39) | 1.13 (1.02 to 1.26) | 0.5530 |
| 1–2 | 1578 | 248.2 | 1.09 (1.03 to 1.14) | 1.09 (1.04 to 1.15) | Reference |
| ≥3 | 520 | 403.7 | 1.07 (0.98 to 1.16) | 1.21 (1.11 to 1.32) | 0.0508 |
| No of losses during follow-up‡ | |||||
| 1 | 2353 | 228.6 | 1.10 (1.06 to 1.15) | 1.12 (1.07 to 1.17) | Reference |
| ≥2 | 95 | 240.7 | 1.10 (0.90 to 1.34) | 1.12 (0.91 to 1.38) | 0.9956 |
*Adjusted for sex, age at follow-up, calendar year at follow-up, country of birth, highest educational attainment, history of psychiatric disorders and CVDs.
†We tested the differences in the strength of the associations between the different exposure categories and atrial fibrillation risk using the contrast statement in Poisson regression model in SAS. We set as reference group the group who lost a child due to CVDs in the analysis with cause of death, the group who lost a child aged 13–18 years in the analysis with age of the deceased child at loss, the group with 1–2 children left in the analysis with number of remaining live children at loss and the group who lost one child during follow-up in the analysis with the number of losses during follow-up.
‡The analyses were performed among those who had not lost any children before baseline.
CI, confidence interval; CVD, cardiovascular diseases; IRR, incidence rate ratio.