Table 1.
Association of demographic and clinical variables with risk of breast cancer*
| Characteristic | No. of women (n = 2666), No. (%) | No. of women with breast cancer (n = 172)† | No. of person-years of follow-up | HR (95% CI)‡ | Ptrend§ |
| Age at BBD diagnosis, y | <.001 | ||||
| <45 | 696 (26.1) | 29 | 9667 | 1.00‖ | |
| 45–54 | 667 (25.0) | 45 | 9507 | 1.58 (0.99 to 2.52) | |
| ≥55 | 1303 (48.9) | 98 | 16 198 | 2.07 (1.37 to 3.14) | |
| Histology | <.001 | ||||
| Nonproliferative disease | 1556 (58.4) | 64 | 20 859 | 1.00‖ | |
| Proliferative disease without atypia | 954 (35.8) | 84 | 12 524 | 2.09 (1.51 to 2.89) | |
| Atypical Hyperplasia | 156 (5.9) | 24 | 1990 | 3.69 (2.30 to 5.91) | |
| Parity | .23 | ||||
| Nulliparous | 358 (14.6) | 27 | 4579 | 1.00‖ | |
| Parous | 2087 (85.4) | 134 | 27 899 | 0.75 (0.50 to 1.14) | |
| Missing¶ | 221 | 11 | |||
| Body mass index, kg/m2 | .91 | ||||
| 15–22 | 498 (25.2) | 32 | 7396 | 1.00‖ | |
| 23–25 | 506 (25.6) | 32 | 7644 | 0.90 (0.55 to 1.47) | |
| 26–29 | 523 (26.5) | 38 | 7840 | 1.00 (0.62 to 1.61) | |
| 30–66 | 449 (22.7) | 29 | 6635 | 0.92 (0.55 to 1.52) | |
| Missing¶ | 690 | 41 | |||
| Postmenopausal Hormone Use | .44 | ||||
| Never | 947 (44.0) | 54 | 13 214 | 1.00‖ | |
| Ever | 1205 (56.0) | 85 | 17 740 | 1.15 (0.82 to 1.61) | |
| Missing¶ | 514 | 33 | |||
| Menopause status | .61 | ||||
| Premenopausal | 796 (30.7) | 37 | 11 065 | 1.00‖ | |
| Postmenopausal | 1796 (69.3) | 128 | 23 660 | 1.12 (0.69 to 1.81) | |
| Missing¶ | 74 | 7 | |||
| Family history of breast cancer# | .004 | ||||
| None | 1542 (64.3) | 93 | 20 901 | 1.00‖ | |
| Weak | 509 (21.2) | 36 | 7349 | 1.15 (0.78 to 1.70) | |
| Strong | 349 (14.5) | 39 | 4912 | 1.80 (1.24 to 2.61) | |
| Missing¶ | 266 | 4 |
Nested cohort study of women with benign breast disease (BBD) diagnosed at the Mayo Clinic between January 1, 1985, and December 31, 1991. CI = confidence interval; HR = hazard ratio.
After a mean follow-up of 13.3 years, 172 of 2666 women developed breast cancer.
Analysis of breast cancer risk was done using the Cox proportional hazards model. Association of risk with age at BBD diagnosis was unadjusted; all other associations with risk are adjusted for age.
Ptrend values were calculated using two-sided Wald test for trend.
Referent category in the Cox proportional hazards model.
Subjects for whom information for that demographic or clinical variable was not available from questionnaires or medical records.
Strong—at least one first-degree relative with breast cancer before age 50 years, or two or more relatives with breast cancer with at least one being a first-degree relative; weak—any other family of breast cancer; None—no family history of breast cancer.