During 26 years of follow-up, we ascertained 5,522 incident cases of postmenopausal breast cancer. Of these, 3,314 were ER+ tumors, 826 were ER− tumors, and 1,382 were not classified. At baseline, women with a higher overall and animal low-carbohydrate-diet scores tended to have higher body mass indexes and were more likely to be smokers (). On the other hand, women with higher vegetable low-carbohydrate-diet scores or DASH scores tended to be leaner ( and ). In addition, those with higher DASH scores were less likely to be smokers and were more likely to have higher levels of physical activity. Because of the long follow-up period, we presented dietary intake information in 1994, as that reflects intake at approximately the midpoint of the follow-up period. Higher alcohol intake was observed among women with higher low-carbohydrate-diet scores. However, the difference in the percent of energy from carbohydrates and glycemic load was less pronounced between extreme quintiles of the vegetable low-carbohydrate-diet score than other 2 low-carbohydrate-diet scores. Women with higher DASH scores consumed more energy, with a large portion of the energy intake from carbohydrates, fruits, and vegetables and less energy from animal fat or protein.
Age-Standardized Baseline (1980) Characteristics and 1994 Dietary Intake, by Quintile of Low-Carbohydrate-Diet Score, Nurses’ Health Study, 1980–2006
Age-Standardized Baseline Characteristics and 1994 Dietary Intake, by Quintile of Dietary Approaches to Stop Hypertension Score, Nurses’ Health Study, 1980–2006
After adjustment for potential confounders, neither the low-carbohydrate-diet scores () nor the DASH scores () were associated with overall incidence of breast cancer or ER+ breast cancer. This was true even if we stratified by ER+/progesterone receptor-positive (PR+) tumors and ER+/progesterone receptor-negative (PR−) tumors (data not shown). However, for ER− tumors, an inverse association was observed for both the vegetable low-carbohydrate-diet score (relative risk (RR) comparing the fifth quintile with the first quintile = 0.81, 95% confidence interval (CI): 0.65, 1.01; P trend = 0.03) and the DASH score (RR comparing extreme quintiles = 0.80, 95% CI: 0.64, 1.01; P trend = 0.02). We then further stratified by ER− tumors by progesterone receptor status. Comparing extreme quintiles of the DASH score for ER−/PR− tumors (698 cases), we found that the relative risk was 0.83 (95% CI: 0.65, 1.07; P trend = 0.11) and the corresponding comparison for the vegetable low-carbohydrate-diet score yielded a relative risk of 0.83 (95% CI: 0.65, 1.06; P trend = 0.06). For ER−/PR+ tumors (94 cases), the relative risk comparing extreme quintiles of DASH score was 0.65 (95% CI: 0.31, 1.37; P trend = 0.14), and the same comparison for the vegetable low-carbohydrate-diet score yielded a relative risk of 0.88 (95% CI: 0.44, 1.76; P trend = 0.68).
Relative Risk of Incident Breast Cancer, by Quintile of Low-Carbohydrate-Diet Score, Nurses’ Health Study, 1980–2006
Relative Risk of Incident Breast Cancer, by Quintile of Dietary Approaches to Stop Hypertension Score, Nurses’ Health Study, 1980–2006
Among plant food categories, we observed a significant inverse association between fruit intake and low-protein vegetable intake and ER− tumors when comparing those in the fifth quintile with those in the first (RR = 0.71, 95% CI: 0.55, 0.92; P trend = 0.01 and RR = 0.76, 95% CI: 0.60, 0.95; P trend = 0.048, respectively) (). When these 2 categories were combined, the relative risk of ER− tumors was 0.71 (95% CI: 0.55, 0.90; P trend = 0.005) when comparing the extreme quintiles. When we added nuts and legumes to the combination (i.e., nongrain plant foods), the relative risk did not improve appreciably. A major difference between the DASH score and the vegetable low-carbohydrate-diet score is the dairy component in the DASH score. No association was observed between intakes of all dairy, low-fat dairy, or high-fat dairy and ER+ breast cancer or ER− breast cancer.
Multivariatea Relative Risk of Estrogen Receptor-Negative Breast Cancer, by Quintile of Plant Food Intake, Nurses’ Health Study, 1980–2006
When we included nongrain plant foods into the regression model with the DASH score, the inverse association with ER− tumors was attenuated and no longer statistically significant. However, when we included nongrain plant foods into the regression model with the vegetable low-carbohydrate score, the inverse association for the vegetable score was essentially unchanged (data not shown). Separate analyses focusing on components of the vegetable low-carbohydrate score showed that none of the components was associated with ER− tumors on their own (data not shown).