Pooled Analysis of Circulating Carotenoids and Breast Cancer Risk

Plan to submit an R01 in June 2013 for funding to cover plasma assays for carotenoids in cases and controls or recalibration, data collection and harmonization, data analyses, and manuscript preparation.

Studies of dietary intake, of fruits and vegetables overall and carotenoids specifically, have had mixed results, although a modest inverse association has been suggested in several studies. Several studies of dietary patterns that include higher fruit and vegetable intake observed lower risk of estrogen receptor negative (ER-), but not ER+, breast tumors. Most recently, in the Pooling Project, stronger inverse associations were observed with higher intake of α-carotene, β-carotene, and lutein+zeaxanthin (Zhang XH et al, 2012).

Plasma carotenoids likely provide a better measure of internal exposure, given error in diet assessments, influences of cooking and storage on carotene content, geographic and seasonal variation of foods, and individual differences in absorption. Several prospective studies have been conducted examining the relation of plasma or serum carotenoids with breast cancer risk. While many observed RRs<1, the particular carotenoid associated with risk, as well as the
subgroups in which stronger associations were observed, differed across studies.

Recently, we undertook a pooled analysis of 8 existing cohort studies of plasma carotenoids and breast cancer risk, with a total of 3,055 cases and 3,956 controls (Eliassen et al, JNCI December 2012). To account for laboratory differences and examine population differences across studies, we recalibrated participant carotenoid levels to a common standard by re-assaying 20 plasma or serum samples from each cohort together at the same laboratory. We conducted an aggregated analysis using recalibrated data and quantile cut points established across the pooled study populations.

Statistically significant inverse associations with breast cancer were observed for α-carotene (top vs. bottom quintile RR (95% CI)=0.87 (0.71-1.05), ptrend= 0.04), β-carotene (0.83 (0.70-0.98), p-trend=0.02), lutein+zeaxanthin (0.84 (0.70-1.01), p trend=0.05), lycopene (0.78 (0.62-0.99), p-trend=0.02) and total carotenoids (0.81 (0.68-0.96), p-trend=0.01). β-cryptoxanthin was not statistically significantly associated with risk. For several carotenoids, associations appeared stronger for ER-compared with ER+ tumors (e.g., β-carotene top vs. bottom quintile RR (95% CI)=0.52 (0.36-0.77), p-trend=0.001 for ER-; 0.83
(0.66-1.04), p-trend=0.06 for ER+; p-heterogeneity=0.01).

Several of the associations varied by BMI and smoking status. Interactions with BMI were fairly consistent across carotenoids, but somewhat unexpected with stronger inverse associations with breast cancer risk observed among leaner (BMI<25) women and suggestively or statistically significantly positive associations among obese women (BMI≥30). For example, top vs. bottom quartile RR (95% CI) for total carotenoids=0.65 (0.51-0.83), p-trend=0.001 for BMI <25; 1.74 (0.97-3.13), p-trend=0.04 for BMI ≥30; p-heterogeneity=0.01. Statistically significant interactions also were observed with current smoking for lutein +zeaxanthin and total carotenoids, with stronger inverse associations among current smokers (e.g., total carotenoids top vs. bottom quintile RR (95% CI) nonsmokers= 0.93 (0.77-1.11), p-trend=0.38; current smokers=0.47 (0.30-0.73), p trend= 0.002 (p-heterogeneity=0.01)).

Thus, our results to date are rather intriguing. The inverse associations we observed among ER- tumors highlight carotenoids as one of the first modifiable risk factors for this poor prognosis tumor type. Expanding this analysis to the Cohort Consortium, with far more cases, will allow us to further explore this finding, to determine if the association is with ER- tumors only, or if it applies to both receptor types. In addition, our observed interaction with BMI was not expected and deserves further exploration, as do interactions with other oxidative stress related exposures, including alcohol and smoking.

Our goal is to expand upon our recent pooled analysis of circulating carotenoids and breast cancer risk (~3000 cases from nested case-control studies). We observed significant inverse associations between several carotenoids and overall as well as ER- breast cancer, with significant or suggestive heterogeneity by ER status. With more cases and improved power, this potential heterogeneity could be examined in more detail, as well as interactions with lifestyle factors.

1. Higher levels of plasma carotenoids, specifically α-carotene, β-carotene, lutein/zeaxanthin, lycopene, and total carotenoids, are associated with reduced breast cancer risk. The association is apparent among both estrogen receptor positive (ER+) and ER- tumors, but is stronger among ER- tumors.

2. Lifestyle factors that contribute to oxidative stress, including smoking and alcohol consumption, modify the association of carotenoids with breast cancer risk.

Include nested case-control studies with previously measured plasma carotenoids, as well as those that have not assayed carotenoids yet. 1:1 case:control matching (all available controls if plasma carotenoids already measured in more than 1 control per case). Recalibrate existing data, similar to our published pooled analyses and the vitamin D pooling project of breast/colorectal cancer, and incorporate newly assayed samples. Conduct an aggregated analysis of pooled data, with particular focus on ER status and interactions with lifestyle factors.

The potential for carotenoids to affect ER- tumors appears promising, but no individual cohort can look at this. In our prior analysis with 3000 cases, we still only had ~400 ER- cases and we need more power to tease apart the associations with ER- and ER+ tumors. The interaction we observed with BMI needs further exploration, as well as examining the interaction by ER status, given the provocative findings in our first set of pooled analyses.

100

breast cancer, preferably with hormone receptor status; HER2 status if available

plasma/serum carotenoids data are NOT required; assays will be conducted for cohorts that do not already have them measured.

Reproductive variables, other breast cancer risk factors, and lifestyle factors that contribute to oxidative stress: preferably alcohol intake, smoking status, body mass index. Any factors used to match cases and controls (both blood collection-related and others). If a study does not yet have matched cases and controls, I suggest matching on: date, time, and fasting status at blood collection; age at blood collection; menopausal status; and race/ethnicity.

Yes

Bloods stored at ≤-80 degrees C. Plasma or serum is ok. Prior freeze/thaw is ok.