Table 1.
Odds ratios (OR) for associations of patient characteristics with breast cancer-specific death among 349 ER-positive breast cancer patients treated with tamoxifen in the Kaiser Permanente Northwest Health Plan (Portland, Oregon)
| Variable | Case patients (n = 97) | Control patients (n = 252) | Univariate OR (95% CI)† | P‡ | Multivariable OR (95% CI) †,§ | P‡ |
|---|---|---|---|---|---|---|
| No. (%)* | No. (%)* | |||||
| Age at diagnosis, y | ||||||
| ≤50 | 29 (29.9) | 73 (29.0) | —|| | |||
| 51–60 | 22 (22.7) | 66 (26.2) | — | |||
| 61–70 | 28 (28.9) | 77 (30.6) | — | |||
| >70 | 18 (18.6) | 36 (14.3) | — | |||
| SEER summary stage at diagnosis | ||||||
| Localized | 41 (42.3) | 112 (44.4) | —|| | |||
| Regional spread | 56 (57.7) | 140 (55.6) | — | |||
| Year of diagnosis | ||||||
| 1990–1993 | 14 (14.4) | 42 (16.7) | —|| | |||
| 1994–1998 | 38 (39.2) | 93 (36.9) | — | |||
| 1999–2002 | 31 (32.0) | 84 (33.3) | — | |||
| 2003–2008 | 14 (14.4) | 33 (13.1) | — | |||
| Progesterone receptor | ||||||
| Negative | 27 (27.8) | 42 (16.8) | Ref. | .03 | ||
| Positive | 70 (72.2) | 208 (83.2) | 0.54 (0.31 to 0.93) | |||
| Missing | 0 | 2 | ||||
| Tumor size, mm | ||||||
| ≤11 | 16 (17.4) | 67 (26.9) | Ref. | .02 | ||
| >11 to 16.5 | 10 (10.9) | 58 (23.3) | 0.77 (0.33 to 1.84) | |||
| >16.5 to 25 | 31 (33.7) | 68 (27.3) | 2.03 (1.00 to 4.14) | |||
| >25 | 35 (38.0) | 56 (22.5) | 2.29 (1.10 to 4.77) | |||
| Missing | 5 | 3 | ||||
| Tumor differentiation | ||||||
| Poor | 34 (37.8) | 52 (21.7) | Ref. | .002 | Ref. | .001 |
| Moderate | 55 (61.1) | 121 (50.4) | 0.77 (0.43 to 1.37) | 0.65 (0.32 to 1.29) | ||
| Well | 1 (1.1) | 67 (27.9) | 0.03 (<0.01 to 0.19) | 0.01 (<0.01 to 0.14) | ||
| Missing | 7 | 12 | ||||
| Chemotherapy | ||||||
| No | 41 (42.3) | 121 (48.2) | Ref. | .27 | ||
| Yes | 56 (57.7) | 130 (51.8) | 1.44 (0.75 to 2.76) | |||
| Missing | 0 | 1 | ||||
| Radiotherapy | ||||||
| No | 37 (38.1) | 88 (34.9) | Ref. | .80 | ||
| Yes | 60 (61.9) | 164 (65.1) | 0.94 (0.56 to 1.56) | |||
| Duration of tamoxifen use, mo | ||||||
| ≤52¶ | 64 (66.0) | 81 (32.1) | Ref. | <.001 | Ref. | <.001 |
| 53 to 61 | 17 (17.5) | 88 (34.9) | 0.18 (0.09 to 0.37) | 0.11 (0.04 to 0.28) | ||
| >61 | 16 (16.5) | 83 (32.9) | 0.15 (0.07 to 0.32) | 0.09 (0.03 to 0.24) | ||
| Per 12 months | 0.59 (0.49 to 0.72) | <.001 | 0.56 (0.44 to 0.71) | <.001 | ||
| Bilateral oophorectomy before follow-up mammogram | ||||||
| No | 96 (99.0) | 248 (98.4) | Ref. | |||
| Yes | 1 (1.0) | 4 (1.6) | —# | |||
| Hormone therapy use at baseline | ||||||
| Nonuser | 47 (48.5) | 104 (41.3) | Ref. | .26 | ||
| Former user | 17 (17.5) | 37 (14.7) | 0.95 (0.45 to 2.00) | |||
| Current user | 33 (34.0) | 111 (44.1) | 0.63 (0.35 to 1.13) | |||
| Antidepressant use at baseline | ||||||
| Nonuser | 88 (90.7) | 207 (82.1) | Ref. | .11 | ||
| Former user | 5 (5.2) | 16 (6.4) | 0.76 (0.26 to 2.23) | |||
| Current user | 4 (4.1) | 29 (11.5) | 0.32 (0.11 to 0.95) | |||
| Smoking status at baseline | ||||||
| Never | 44 (45.4) | 155 (61.8) | Ref. | .004 | Ref. | .01 |
| Ever | 53 (54.6) | 96 (38.3) | 2.07 (1.26 to 3.40) | 2.37 (1.23 to 4.56) | ||
| Missing | 0 | 1 | ||||
| Race | ||||||
| White | 96 (99.0) | 244 (97.2) | Ref. | |||
| Nonwhite | 1 (1.0) | 7 (2.8) | —# | |||
| Missing | 0 | 1 | ||||
| BMI at baseline, kg/m2** | ||||||
| <25 | 22 (26.2) | 76 (33.6) | Ref. | .40 | ||
| 25–29 | 29 (34.5) | 76 (33.6) | 1.37 (0.71 to 2.62) | |||
| 30–34 | 17 (20.2) | 44 (19.5) | 1.63 (0.80 to 3.31) | |||
| ≥35 | 16 (19.1) | 30 (13.3) | 1.86 (0.84 to 4.12) | |||
| Missing | 13 | 26 | ||||
| Per 5kg/m2 increase | 1.12 (0.92 to 1.35) | .25 | ||||
| Change in BMI (follow-up minus baseline), kg/m2** | ||||||
| <-2.5 | 6 (8.3) | 13 (6.7) | 1.10 (0.36 to 3.31) | .90 | ||
| <-1 to -2.5 | 11 (15.3) | 25 (13.0) | 1.08 (0.51 to 2.30) | |||
| -1 to 1 | 39 (54.2) | 104 (53.9) | Ref. | |||
| >1 to 2.5 | 10 (13.9) | 40 (20.7) | 0.86 (0.37 to 1.96) | |||
| >2.5 | 6 (8.3) | 11 (5.7) | 1.55 (0.55 to 4.39) | |||
| Missing | 25 | 59 | ||||
| Per 1kg/m2 change | 1.00 (0.90 to 1.12) | .95 | ||||
| Baseline percent density | ||||||
| Q1¶: ≤15 | 29 (29.9) | 51 (20.2) | Ref. | .35 | ||
| Q2: >15 to 23 | 19 (19.6) | 50 (19.8) | 0.77 (0.38 to 1.56) | |||
| Q3: >23 to 31 | 16 (16.5) | 51 (20.2) | 0.53 (0.25 to 1.11) | |||
| Q4: >31 to 43 | 19 (19.6) | 47 (18.7) | 0.66 (0.33 to 1.33) | |||
| Q5: >43 | 14 (14.4) | 53 (21.0) | 0.49 (0.22 to 1.07) | |||
| Per 10% increase | 0.87 (0.75 to 1.01) | .07 | ||||
| Baseline dense area, cm2 | ||||||
| Q1¶: ≤21 | 29 (29.9) | 50 (19.8) | Ref. | .24 | ||
| Q2: >21 to 30 | 17 (17.5) | 51 (20.2) | 0.49 (0.24 to 1.04) | |||
| Q3: >30 to 42 | 18 (18.6) | 50 (19.8) | 0.60 (0.30 to 1.20) | |||
| Q4: >42 to 57 | 19 (19.6) | 51 (20.2) | 0.55 (0.27 to 1.16) | |||
| Q5: >57 | 14 (14.4) | 50 (19.8) | 0.45 (0.20 to 1.00) | |||
| Per 10cm2 increase | 0.93 (0.83 to 1.03) | .17 | ||||
| Months from baseline mammogram to follow-up mammogram | ||||||
| ≤12 | 40 (41.2) | 115 (45.6) | Ref. | .91 | ||
| >12 to 18 | 13 (13.4) | 34 (13.5) | 1.06 (0.51 to 2.17) | |||
| >18 to 24 | 21 (21.7) | 45 (17.9) | 1.27 (0.66 to 2.44) | |||
| >24 | 23 (23.7) | 58 (23.0) | 1.10 (0.58 to 2.08) | |||
| Months from tamoxifen initiation to follow-up mammogram | ||||||
| ≤6 | 11 (11.3) | 19 (7.5) | 1.75 (0.68 to 4.49) | .41 | ||
| 7 to 9 | 27 (27.8) | 62 (24.6) | 1.45 (0.74 to 2.81) | |||
| 10 to 12 | 26 (26.8) | 88 (34.9) | Ref. | |||
| 13 to 15 | 7 (7.2) | 25 (9.9) | 0.88 (0.33 to 2.30) | |||
| 16 to 18 | 18 (18.6) | 49 (19.4) | 1.20 (0.56 to 2.55) | |||
| ≥19 | 8 (8.3) | 9 (3.6) | 2.70 (0.93 to 7.85) | |||
* Column percentages. BMI = body mass index; CI = confidence interval; ER = estrogen receptor; OR = odds ratio.
† Odds ratio and 95% confidence interval, estimated using conditional logistic regression.
‡ Two-sided P heterogeneity, based on the Wald test.
§ Adjusted for variables retained in this column. All other variables were considered for inclusion using stepwise regression, but were not retained in the model because of lack of statistical significance.
|| Not estimated for matching factors.
¶ Percentiles based on the distribution among control patients.
# Not estimated because of small numbers.
** Odds ratio and confidence interval estimated with multiply imputed data.