Table 2.
Estimated tubo-ovarian carcinoma and breast cancer relative risk for RAD51C and RAD51D pathogenic variant carriers
| Cancer and models considered | Age, y | RR (95% CI) | P * | LRT P† | AIC | Best fitting models |
|---|---|---|---|---|---|---|
| RAD51C | ||||||
| Tubo-ovarian carcinoma | ||||||
| Age-constant model | 30–79 | 7.55 (5.60 to 10.19) | 5 × 10-40 | 4335.8 | ||
| Age-specific model for each decade of age | 30–39 | 2.85 (0.46 to 17.70) | 0.04 | 4334.0 | ||
| 40–49 | 5.94 (3.09 to 11.43) | |||||
| 50–59 | 8.55 (5.10 to 14.33) | |||||
| 60–69 | 13.90 (8.45 to 22.88) | |||||
| 70–79 | 2.54 (0.53 to 12.27) | |||||
| Age-specific model, separate parameters for two age groups: 30–50 and 50–80 y | 30–49 | 4.97 (2.75 to 8.97) | 0.048 | 4333.8 | ||
| 50–79 | 9.44 (6.63 to 13.45) | |||||
| Piecewise linear model‡ | 35 | 2.40 | 0.004 | 4328.6 | Yes | |
| 45 | 5.14 | |||||
| 55 | 11.02 | |||||
| 65 | 9.02 | |||||
| 75 | 2.82 | |||||
| Breast cancer | ||||||
| Age-constant model | 20–79 | 1.99 (1.39 to 2.85) | 1.55 × 10-4 | 4346.4 | Yes | |
| Age-specific model, separate parameters for each decade of age | 20–29 | 1.19 (0.09 to 16.12) | 0.37 | 4351.0 | ||
| 30–39 | 3.25 (1.60 to 6.62) | |||||
| 40–49 | 2.50 (1.41 to 4.45) | |||||
| 50–59 | 0.96 (0.34 to 2.71) | |||||
| 60–69 | 1.54 (0.45 to 5.36) | |||||
| 70–79 | 2.57 (0.61 to 10.81) | |||||
| Age-specific model, separate parameters for two age groups: 20–50 and 50–80 y | 20–49 | 2.42 (1.61 to 3.63) | 0.12 | 4346.0 | ||
| 50–79 | 1.36 (0.70 to 2.63) | |||||
| RAD51D | ||||||
| Tubo-ovarian carcinoma | ||||||
| Age-constant model | 30–79 | 7.60 (5.61 to 10.30) | 5 × 10-39 | 4160.0 | ||
| Age-specific model for each decade of age | 30–39 | 3.60 (0.78 to 16.75) | 0.02 | 4155.8 | ||
| 40–49 | 3.19 (1.04 to 9.72) | |||||
| 50–59 | 12.54 (7.62 to 20.63) | — | ||||
| 60–69 | 10.60 (6.10 to 18.41) | — | ||||
| 70–79 | 4.94 (1.34 to 18.26) | — | ||||
| Age-specific model, separate parameters for two age groups: 30–50 and 50–80 y | 30–49 | 3.23 (1.36 to 7.71) | — | 0.002 | 4152.1 | |
| 50–79 | 10.56 (7.48 to 14.91) | — | ||||
| Piecewise linear model§ | 35 | 1.64 | — | 0.002 | 4151.6 | Yes |
| 45 | 4.30 | — | ||||
| 55 | 11.29 | — | ||||
| 65 | 10.16 | — | ||||
| 75 | 5.77 | — | ||||
| Breast cancer | ||||||
| Age-constant model | 20–79 | 1.83 (1.24 to 2.72) | 0.0002 | 4177.9 | Yes | |
| Age-specific model, separate parameters for each decade of age except for 20–39 y age group | 20–39 | 2.25 (1.25 to 4.04) | — | 0.59 | 4183.1 | |
| 40–49 | 1.46 (0.69 to 3.09) | — | ||||
| 50–59 | 1.56 (0.69 to 3.51) | — | ||||
| 60–69 | 1.63 (0.54 to 4.98) | — | ||||
| 70–79 | 4.19 (1.51 to 11.62) | — | ||||
| Age-specific model, separate parameters for two age groups: 20–50 and 50–80 y | 20–49 | 1.84 (1.12 to 3.02) | — | 1.00 | 4179.9 | |
| 50–79 | 1.83 (1.02 to 3.26) | — |
The P values assessing the null hypothesis of RR = 1.00. AIC = Akaike information criterion; CI = confidence interval; RR = relative risk.
Likelihood ratio tests (LRT) comparing each model against the model with a constant RR.
logRR(t) = a + b1(t − 30) if t ∈ [30,60); logRR(t) = a + b1 × 30 + b2(t − 60) if t ∈ [60,80), where a = 0.49 (95% CI = −0.80 to 1.78), b1 = 0.076 (95% CI = 0.023 to 0.13), b2 = −0.12 (95% CI = −0.23 to −0.0036).
logRR(t) = a + b1(t − 30) if t ∈ [30,58); logRR(t) = a + b1 × 28 + b2(t − 58) if t ∈ [58,80), where a = 0.010 (95% CI = −1.49 to 1.51), b1 = 0.097 (95% CI = 0.034 to 0.16), b2 = −0.057 (95% CI = −0.13 to 0.017).