Table 2.

Multivariable Hazard Ratios^a for Post-diagnosis Cruciferous Vegetable Intake in Association with Recurrence and Total Mortality, After Breast Cancer Pooling Project (n=11,390)

	Recurrenceb				Total Mortality
Cruciferous vegetables	Cohort	Events	HR	(95% CI)	Events	HR	(95% CI)
Study-specific quartiles (g/day)c
Q1	2,833	336	1.00	(reference)	454	1.00	(reference)
Q2	2,810	357	1.08	(0.93–1.25)	427	1.01	(0.88–1.15)
Q3	2,905	347	1.01	(0.87–1.18)	419	0.97	(0.84–1.11)
Q4	2,842	381	1.10	(0.95–1.28)	425	0.99	(0.86–1.13)
Ptrend				0.34			0.84
Common Cutpoints (g/day)e
<39	5,447	738	1.00	(reference)	934	1.00	(reference)
39–78	2,806	341	1.00	(0.88–1.15)	378	0.89	(0.79–1.01)
≥78	3,137	342	1.05	(0.89–1.24)	413	1.03	(0.88–1.20)
Ptrend				0.60			0.82

^aHRs were from delayed-entry Cox proportional hazards regression models with study as a stratification variable and adjusted for age at diagnosis, ER/PR status, TNM stage, chemotherapy, surgery, radiotherapy, hormonal therapy, smoking, BMI, exercise, menopausal status, race/ethnicity, and education.

^b566 women were excluded where disease-free follow-up time was ≤ date of exposure/enrollment.

^cStudy-specific quartiles were as follows: <10.7, 10.7–22.4, 22.4–43.5, ≥ 43.5 (LACE); <13.0, 13.0-<23.8, 23.8–44.6, ≥ 44.6 (WHEL); <15.4, 15.4–25.4, 25.4–43.2, ≥ 43.2 (NHS); <61.5, 61.5–93, 93–138.0, ≥ 138.0 (SBCSS).

^dP-values for heterogeneity by study were 0.98 and 0.69 for recurrence and total mortality, respectively (using study-specific quartiles). P-values for heterogeneity by study were 0.89 and 0.91 for recurrence and total mortality, respectively (using common cutpoints).

^e78 grams/day is approximately equivalent to ½ cup/day of broccoli.