Skip to main content
. 2017 Mar 9;109(7):djw302. doi: 10.1093/jnci/djw302

Table 3.

Age-specific hazard ratio estimates for the PRS associations and HR estimates for a PRS x age interaction*

Age category, y BRCA1 carriers
BRCA2 carriers
No. of events HR per unit SD increase in the ER- PRS (95% CI) P No. of events HR per unit SD increase in the overall BC PRS (95% CI) P
Outcome: breast cancer
 18–39 4125 1.63 (1.52 to 1.74) 1731 1.65 (1.44 to 1.88)
 40–49 2557 1.18 (1.13 to 1.23) 4.2 × 10−15 1587 1.22 (1.14 to 1.31) 8.5 × 10−5
 50–59 846 1.14 (1.09 to 1.21) .40 718 1.10 (1.02 to 1.19) .05
 ≥60 269 1.20 (1.11 to 1.29) .33 294 1.12 (1.03 to 1.23) .75
 Interaction HR 0.993 (0.990 to 0.996) 3.3 × 10−6 0.995 (0.991 to 0.999) .01
 Main effect PRS 1.69 (1.50 to 1.91) 1.55 (1.29 to 1.87)
Outcome: ovarian cancer
 18–49 1258 1.55 (1.42 to 1.69) 172 3.05 (2.35 to 3.97)
 50–59 808 1.11 (1.05 to 1.18) 1.1 × 10−9 227 1.52 (1.26 to 1.84) 8.2 × 10−6
 ≥60 396 1.14 (1.06 to 1.21) .67 232 1.21 (1.12 to 1.30) .03
 Interaction HR 0.992 (0.988 to 0.998) .003 0.991 (0.979 to 1.00) .11
 Main effect PRS 1.83 (1.43 to 2.34) 2.48 (1.34 to 4.58)
*

The population-derived polygenic risk score (PRS) for estrogen receptor–negative breast cancer was used for the associations with breast cancer in BRCA1 carriers and the overall breast cancer PRS in BRCA2 carriers. P values relate to the difference in PRS association between each age group from the preceding younger group and to the interaction term. CI = confidence interval; HR = hazard ratio; PRS = polygenic risk score.

P value for a two-sided test using a weighted cohort Cox regression with time to breast or ovarian cancer diagnosis, respectively, as the outcome.

HHS Vulnerability Disclosure