Table 2.
Multivariable-adjusted Cox regression analyses for different events in women aged 49–75 years at DCIS diagnosis
| Method of detection | Total number of events | Person-time, years | HR (95% CI) | P value |
|---|---|---|---|---|
| Ipsilateral invasive breast cancer | ||||
| Non-screening-related | 98 | 18,710 | ref | |
| Screen-detected | 231 | 50,422 | 0.75 (0.59–0.96) | 0.024 |
| Interval | 34 | 6359 | 1.02 (0.68–1.51) | 0.941 |
| Contralateral invasive breast cancer | ||||
| Non-screening-related | 105 | 18,825 | ref | |
| Screen-detected | 243 | 50,467 | 0.86 (0.67–1.10) | 0.224 |
| Interval | 30 | 6387 | 0.83 (0.54–1.26) | 0.382 |
| All-cause mortality | ||||
| Non-screening-related | 398 | 19,361 | ref | |
| Screen-detected | 594 | 51,740 | 0.85 (0.73–0.98) | 0.028 |
| Interval | 68 | 6570 | 0.73 (0.56 − 0.96) | 0.025 |
Analysis was performed with age as the primary time scale and time since DCIS diagnosis (0–5, 5–10, and ≥10 years) as the secondary time scale. We adjusted for period of ductal carcinoma in situ (DCIS) diagnosis, DCIS grade and DCIS treatment (time-varying). HR hazard ratio, CI confidence interval