Table 2. Unweighted absolute risks and number needed to treat (NNT) for events of interest after 10 years of follow-up.
Event of interest | Number ofincluded studies | Absolute risk inextended adjuvanttamoxifen group | Absolute risk inadjuvant tamoxifengroup | Absolute riskdifference | NNT* |
Any recurrence (all patients) | 3 [6], [8], [15] | 12.7% | 14.5% | 1.8% | 57 |
Any recurrence (excluding ER-negative) | 3 [6], [8], [15] | 17.5% | 20.2% | 2.7% | 38 |
Distant recurrence | 1 [8] | 15.0% | 16.2% | 1.3% | 79 |
Death after recurrence | 3 [6], [8], [15] | 7.7% | 8.7% | 1.0% | 101 |
Death without recurrence | 3 [6], [8], [15] | 11.9% | 11.8% | −0.0008% | −1281 |
All cause death | 3 [6], [8], [15] | 15.4% | 16.7% | 1.3% | 79 |
95% CI, 95% confidence intervals. ER, estrogen receptor.
*Negative value suggests more events in extended adjuvant group (i.e. net harm).