Table 3.
Comparison of expected and observed cases of invasive breast cancer, and hazard ratios for observed cases, according to Gail model predicted 5-year risk for all women by age group, and for group level risk quintiles (Q1 to Q5) and, within Q5, the upper two deciles of risk (D9 and D10)
| Age (years) | Quantile group | Predicted 5-year risk (range) |
No. of women | Observed (O) breast cancers | Expected (E) breast cancers | Person-years (PY) | O per 10,000 PY | E/0 (95% CI)a | HR (95% CI), p value (Q3 referent) |
HR (95% CI), p value (Q1 referent)b |
|---|---|---|---|---|---|---|---|---|---|---|
| 50–69 | Q1 | 0.6–1.1% | 8041 | 72 | 65 | 34,078 | 21 | 0.90 (0.71–1.15) | 0.59 (0.44–0.79), p < 0.0001 | Referent |
| Q2 | 1.1–1.4% | 8096 | 77 | 86 | 34,160 | 23 | 1.12 (0.89–1.42) | 0.63 (0.47–0.84), p = 0.001 | 1.07 (0.78–1.48), p = 0.679 | |
| Q3 | 1.4–1.7% | 8124 | 122 | 105 | 34,264 | 36 | 0.86 (0.72–1.04) | Referent | 1.70 (1.27–2.27), p < 0.0001 | |
| Q4 | 1.7–2.3% | 7902 | 132 | 131 | 33,254 | 40 | 0.99 (0.84–1.19) | 1.11 (0.87–1.43), p = 0.388 | 1.89 (1.42–2.52), p < 0.0001 | |
| Q5 | 2.3–22.0% | 7995 | 161 | 225 | 33,690 | 48 | 1.40 (1.20–1.64) | 1.34 (1.06–1.70), p = 0.014 | 2.28 (1.73–3.02), p < 0.0001 | |
| D9 | 2.3–3.0% | 3980 | 78 | 88 | 16,790 | 46 | 1.13 (0.91–1.43) | 1.30 (0.98–1.73), p = 0.068 | 2.21 (1.61–3.05), p < 0.0001 | |
| D10 | 3.0–22.0% | 4015 | 83 | 137 | 16,900 | 49 | 1.65 (1.33–2.07) | 1.39 (1.05–1.83), p = 0.022 | 2.35 (1.72–3.23), p < 0.0001 | |
| Total | 0.2–21.7% | 40,158 | 564 | 612 | 169,445 | 33 | 1.09 (1.00–1.18) | |||
| 50–59 | Q1 | 0.6–1.1% | 4046 | 35 | 29 | 17,131 | 20 | 0.83 (0.60–1.19) | 0.83 (0.53–1.30), p = 0.413 | Referent |
| Q2 | 1.1–1.4% | 4054 | 33 | 38 | 17,167 | 19 | 1.14 (0.81–1.66) | 0.78 (0.49–1.23), p = 0.283 | 0.94 (0.58–1.51), p = 0.797 | |
| Q3 | 1.4–1.7% | 4062 | 42 | 46 | 17,117 | 25 | 1.10 (0.81–1.52) | Referent | 1.21 (0.77–1.89), p = 0.413 | |
| Q4 | 1.7–2.3% | 4013 | 67 | 58 | 16,917 | 40 | 0.87 (0.68–1.12) | 1.61 (1.10–2.37), p = 0.015 | 1.95 (1.29–2.93), p = 0.001 | |
| Q5 | 2.3–13.9% | 4041 | 67 | 94 | 17,057 | 39 | 1.40 (1.10–1.80) | 1.60 (1.09–2.35), p = 0.017 | 1.93 (1.28–2.90), p = 0.002 | |
| D9 | 2.1–2.5% | 2028 | 30 | 38 | 8567 | 35 | 1.28 (0.89–1.89) | 1.43 (0.89–2.28), p = 0.138 | 1.72 (1.06–2.80), p = 0.029 | |
| D10 | 2.5–13.9% | 2013 | 37 | 55 | 8490 | 44 | 1.49 (1.08–2.12) | 1.77 (1.14–2.76), p = 0.011 | 2.14 (1.35–3.40), p = 0.001 | |
| Total | 0.2–21.7% | 20,216 | 244 | 264 | 85,388 | 29 | 1.08 (0.96–1.23) | |||
| 60–69 | Q1 | 0.9–1.1% | 4026 | 41 | 39 | 17,022 | 24 | 0.96 (0.71–1.34) | 0.64 (0.43–0.95), p = 0.026 | Referent |
| Q2 | 1.1–1.4% | 3992 | 55 | 48 | 16,833 | 33 | 0.88 (0.68–1.17) | 0.87 (0.61–1.25), p = 0.458 | 1.36 (0.91–2.04), p = 0.134 | |
| Q3 | 1.4–1.7% | 4041 | 64 | 59 | 17,069 | 37 | 0.92 (0.72–1.19) | Referent | 1.56 (1.06–2.31), p = 0.026 | |
| Q4 | 1.7–2.3% | 3946 | 74 | 75 | 16,573 | 45 | 1.02 (0.81–1.30) | 1.19 (0.85–1.67), p = 0.302 | 1.86 (1.27–2.73), p = 0.0001 | |
| Q5 | 2.3–22.0% | 3937 | 86 | 126 | 16,560 | 52 | 1.47 (1.19–1.84) | 1.40 (1.01–1.93), p = 0.044 | 2.18 (1.50–3.16), p < 0.0001 | |
| D9 | 2.8–3.3% | 1953 | 40 | 50 | 8192 | 49 | 1.24 (0.91–1.74) | 1.31 (0.88–1.94), p = 0.182 | 2.04 (1.32–3.16), p = 0.001 | |
| D10 | 3.3–22.0% | 1984 | 46 | 77 | 8368 | 55 | 1.66 (1.25–2.27) | 1.48 (1.01–2.16), p = 0.042 | 2.31 (1.52–3.53), p < 0.0001 | |
| Total | 0.2–21.7% | 19,942 | 320 | 348 | 84,057 | 38 | 1.09 (0.97–1.22) |
CI confidence interval, D, decile, HR hazard ratio, Q quintile
a Chi-squared test across Q1–Q5 (O vs E) were: χ2= 23.0, p = 0.0001 for women 50–69 years old; χ2 = 11.0, p = 0.0262 for women 50–59 years old; and χ2 = 14.4, p = 0.0063 for women 60–69 years old
b Log rank tests for trend across hazard functions Q1–Q5 were χ2 = 52, p < 0.0001 for women 50–69 years old; χ2 = 20, p < 0.0001 for women 50–59 years old; and χ2 = 21, p < 0.0001 for women 60–69 years old