Multivariate Cox proportional hazards analysis of genetic polymorphisms in relation to overall survival following breast cancer diagnosis, by ER status*.



DOI:10.1371/journal.pgen.1000054.t004, Dimensions: 603832,




  1. (1) Spanish National Cancer Research Centre, grid.7719.8
  2. (2) Hannover Medical School, grid.10423.34
  3. (3) Helsinki University Central Hospital, grid.15485.3d
  4. (4) German Cancer Research Center, grid.7497.d
  5. (5) University Medical Center Hamburg-Eppendorf, grid.13648.38
  6. (6) Erasmus University Medical Center, grid.5645.2
  7. (7) International Agency For Research On Cancer, grid.17703.32


*Analyses by ER status included data from 12 studies with information on vital status and ER status (CGPS, CNIO-BCS, HABCS, HEBCS, KBCP, kConFab, LUMCBCS, MCCS, PBCS, SASBCS, SBCS, SEARCH).**Per-allele hazard ratios (HR) and observed P values are adjusted for study. Allele changes are (common>rare based on frequencies in European populations): G>A for rs2981582; G>A for rs3803662; T>G for rs889312; A>G for rs13281615 and A>G for rs381798.***Permutation adjusted P values.****Per-allele hazard ratios (HR) and observed P values are adjusted for study, age at diagnosis (continuous), ER status and grade (continuous). Analyses limited to 11 studies with ER and grade information (CGPS, CNIO-BCS, HABCS, HEBCS, KBCP, LUMCBCS, MCCS, PBCS, SASBCS, SBCS, SEARCH).The P values for the interaction between ER status and genotype adjusted for study, grade and age at diagnosis are: 0.60, 0.15, 0.29, 0.45, 0.38 for rs2981582, rs3803662, rs889312, rs13281615, rs381798, respectively.