Table 1.
Characteristics of cases and controls in population and non‐population based studies
| Population‐based | Non‐population‐based | |||||||
|---|---|---|---|---|---|---|---|---|
| Characteristic | Controls (no.) | % | Cases (no.) | % | Controls (no.) | % | Cases (no.) | % |
| Age, years | ||||||||
| <40 | 252 | 2.2 | 42 | 2.1 | 590 | 4.7 | 249 | 6.7 |
| 40–49 | 1,247 | 10.9 | 293 | 14.4 | 2,135 | 17.0 | 905 | 24.3 |
| 50–59 | 3,999 | 34.9 | 682 | 33.6 | 4,696 | 37.4 | 1,456 | 39.1 |
| 60–69 | 4,720 | 41.1 | 708 | 34.9 | 3,662 | 29.2 | 882 | 23.7 |
| ≥70 | 1,256 | 10.9 | 303 | 14.9 | 1,464 | 11.7 | 233 | 6.3 |
| Age at menarche, years | ||||||||
| ≤12 | 2,140 | 26.2 | 522 | 27.8 | 3,491 | 40.5 | 1,123 | 42.9 |
| 13 | 1,838 | 22.5 | 431 | 22.9 | 2,263 | 26.3 | 640 | 24.4 |
| 14 | 2,034 | 24.9 | 498 | 26.5 | 1,617 | 18.8 | 471 | 18.0 |
| ≥15 | 2,168 | 26.5 | 427 | 22.7 | 1,245 | 14.4 | 385 | 14.7 |
| Parity | ||||||||
| None | 1,221 | 13.5 | 310 | 15.6 | 1,437 | 16.4 | 396 | 14.3 |
| 1–2 | 5,941 | 65.6 | 1,331 | 66.9 | 4,495 | 51.3 | 1,484 | 53.5 |
| 3–4 | 1,726 | 19.0 | 312 | 15.7 | 2,460 | 28.1 | 798 | 28.8 |
| ≥5 | 175 | 1.9 | 37 | 1.9 | 363 | 4.1 | 94 | 3.4 |
| BMI, kg/m2 | ||||||||
| Among women <50 years | ||||||||
| <25 | 542 | 44.4 | 194 | 51.1 | 672 | 49.7 | 461 | 52.3 |
| 25–30 | 431 | 35.3 | 144 | 37.9 | 387 | 28.6 | 273 | 31.0 |
| >30 | 249 | 20.4 | 42 | 11.1 | 292 | 21.6 | 147 | 16.7 |
| Among women ≥ 50 years | ||||||||
| <25 | 2,775 | 35.6 | 472 | 29.5 | 1,998 | 34.5 | 646 | 37.4 |
| 25–30 | 3,028 | 38.9 | 641 | 40.1 | 2,366 | 40.9 | 679 | 39.3 |
| >30 | 2,005 | 25.7 | 486 | 30.4 | 1,419 | 24.5 | 401 | 23.2 |
| Combined HT Use | ||||||||
| Never | 4,836 | 70.7 | 1,000 | 73.4 | 1,070 | 74.9 | 196 | 75.1 |
| Former | 849 | 12.4 | 117 | 8.6 | 238 | 16.7 | 29 | 11.1 |
| Current | 1,154 | 16.9 | 245 | 18.0 | 120 | 8.4 | 36 | 13.8 |
| Family history | ||||||||
| No | 8,023 | 90.2 | 1,707 | 87.4 | 7,778 | 88.6 | 1,997 | 76.9 |
| Yes | 874 | 9.8 | 247 | 12.6 | 1,004 | 11.4 | 599 | 23.1 |
The study population comprised 11 studies participating in the Breast Cancer Association Consortium (see Supporting Information Table S1 for details of the individual studies) with population (ESTHER, KBCP, MARIE, PBCS, UKBGS) and non‐population (CNIO, kConFab, MCBCS, ORIGO, RBCS, SEARCH) based designs. In a model with study design as the outcome: for controls, the distribution of all the risk factors differed by design (p‐value <0.05); for cases, only menarche and family history were different by design.