Antioxidant Micronutrients and the Risk of Renal Cell Carcinoma in the Women’s Health Initiative Cohort

Abstract

BACKGROUND

Renal cell carcinoma (RCC) is the eighth leading cancer among women in incidence and commonly is diagnosed at a more advanced stage. Oxidative stress has been considered to play an important role in the pathogenesis of RCC. Various dietary micronutrients have antioxidant properties, including carotenoids and vitamins C and E; thus, diets rich in these nutrients have been evaluated in relation to RCC prevention. The objective of this study was to explore the correlation between antioxidant micronutrients and the risk of RCC.

METHODS

In total, 96,196 postmenopausal women who enrolled in the Women’s Health Initiative (WHI) between 1993 and 1998 and were followed through July 2013 were included in this analysis. Dietary micronutrient intake was estimated from the baseline WHI food frequency questionnaire, and data on supplement use were collected using an interview-based inventory procedure. RCC cases were ascertained from follow-up surveys and were centrally adjudicated. The risks for RCC associated with intake of α-carotene, β-carotene, β-cryptoxanthin, lutein plus zeaxanthin, lycopene, vitamin C, and vitamin E were analyzed using Cox proportional hazards regression adjusted for confounders.

RESULTS

Two hundred forty women with RCC were identified during follow-up. Lycopene intake was inversely associated with RCC risk (P = .015); compared with the lowest quartile of lycopene intake, the highest quartile of intake was associated with a 39% lower risk of RCC (hazard ratio, 0.61; 95% confidence interval, 0.39–0.97). No other micronutrient was significantly associated with RCC risk.

CONCLUSIONS

The current results suggest that further investigation into the correlation between lycopene intake and the risk of RCC is warranted.

INTRODUCTION

In 2014, it is estimated that there will be 63,920 new cases of kidney cancer in the United States, with 13,860 estimated deaths from the disease and an overall 5-year survival rate of approximately 70%.¹ In the past 3 decades, the incidence of kidney cancer has reached a plateau, with cancers generally diagnosed at earlier stages.² However, kidney cancer remains the sixth most common cause of cancer among men and the eighth most common among women in the United States.¹ Renal cell carcinoma (RCC) arises from the renal parenchyma and constitutes the majority (90%) of all cancers of the kidney.²

Various exposures have been linked to the development of RCC, including tobacco, obesity, hypertension, and lack of physical activity,^3–8 but direct mechanistic explanations have not yet been elucidated. It has been hypothesized that the risk of RCC may be associated with compromised hemodynamics or metabolic disturbances as they relate to factors induced by hypoxia (hypoxia-inducible factors) and von Hippel-Lindau proteins and the presence of oxidative stress.^9,10 Previous studies have suggested that micronutrients consumed through diet or dietary supplementation, including vitamins C and E and carotenoids, are capable of inhibiting oxidative DNA damage, mutagenesis, and tumor growth.^11,12 However, many studies have indicated that there is no significant relation between RCC and antioxidant micronutrient intake,^13–15 whereas others have provided supporting evidence that some micronutrients may have a protective effect.^16–19 Some of the discrepancy in results may be caused in part by sample selection and/or relevant reproductive variables, such as history of hysterectomy and oral contraceptive use in studies that included women.^20,21 Given the inconsistency in the literature, it is important to further probe the potential effect of micronutrients on RCC risk in a larger data set.

The Women’s Health Initiative (WHI) is a large, multisite clinical trial and observational cohort in which comprehensive measures of micronutrient intake were collected at baseline among participants who were followed over an average of 11 to 12 years.^14,20,22 In the current analysis, we used data from the WHI that contained information on several baseline characteristics, including participants’ reproductive history, to assess the correlation between prior intake of micronutrients and subsequent risk of RCC in postmenopausal women.

MATERIALS AND METHODS

Study Population

The WHI includes 3 clinical trials and an observational study in which a total of 161,808 postmenopausal women ages 50 to 79 years were enrolled from 1993 to 1998 in 40 clinical centers throughout the United States with information on follow-up available through July 2013. All participants provided written informed consent, and institutional review board approval was obtained from all respective clinical centers. In the current analysis, we included all participants in the clinical trials and the observational study except for women who enrolled in the Dietary Modification clinical trial (n = 48,835) who had a history of any cancer other than nonmelanoma skin cancers at study entry^7,10,12 or who had implausible energy intakes (<600 or >5000 kcal daily) at baseline (n = 4067). This resulted in an inclusive total of 96,196 women in our analytic cohort.

Dietary Data

Dietary intake was assessed using the WHI semiquantitative food frequency questionnaire, which was administered to study participants at baseline to describe the frequency and serving sizes of 122 specific food items over the previous 3 months.²³ Intake of each micronutrient under analysis was assessed by multiplying the serving portions by the nutrient content information obtained from the University of Minnesota Nutrition Coordinating Center databases.²⁴ To estimate supplement intake, women were instructed to bring bottles of nutritional supplements to their baseline clinic visit, at which time the doses of each nutritional supplement were transcribed and processed through a computerized inventory procedure.²⁵ For the purposes of this study, micronutrients with known antioxidant properties were selected for analyses, including β-carotene (total and dietary), α-carotene (dietary), β-cryptoxanthin (dietary), lutein plus zeaxanthin (dietary; these data were recorded as a group), lycopene (dietary), vitamin C (total and dietary), and vitamin E (α-tocopherol [total and dietary]). Each micronutrient was categorized by quartile (defined by the distribution of intake in women without kidney cancer [noncases] only). Supplement intake quantiles were categorized using the method described by Cui et al,²⁶ with all “zero” responses in the first quartile and all nonzero responses categorized by tertile, because the majority of responses were “zero.”

Case Ascertainment

Diagnoses of RCC were initially based on self-reported questionnaires. Pertinent medical records and pathology reports were used to verify the diagnoses at each clinical center by local adjudicators. Subsequently, a blinded review of these reports was completed at the WHI Clinical Coordinating Center (Seattle, Wash) for centralized adjudication using the National Cancer Institute’s Surveillance, Epidemiology, and End Results coding system.²⁷

Statistical Analysis

The baseline characteristics of women who had a diagnosis of RCC (cases) were compared with those of noncases using chi-square tests for categorical variables and t tests for continuous variables. Annualized rates of RCC were calculated according to the use of each micronutrient separately as well as total micronutrient intake, including dietary and supplemental intake when appropriate.

Tests for proportional hazards assumptions were performed using a Cox model, which included micronutrient use and the interaction of micronutrient use with length of follow-up, and testing for a zero coefficient on the interaction term. A Cox proportional hazards modeling approach was used to compute hazard ratios (HRs) and 95% confidence intervals (CIs) for kidney cancer associated with micronutrient intake categorized by quartiles. Cohort members were censored at the time of the following events: death from a cause other than kidney cancer, diagnosis with cancer other than kidney cancer or nonmelanoma skin cancer, or last follow-up. A univariate, stepwise method was used to evaluate potential base model covariates. Variables that were evaluated for inclusion in the base model were: age group at screening (ages 50–59 years, 60–69 years, or 70–79 years), race/ethnicity (white, African American, Hispanic, Native American/Alaskan Native, Asian/Pacific Islander, or other), education (0–8 years, some high school, high school diploma/General Educational Development test, or college or higher), hypertension (never, untreated hypertensive, or treated hypertensive), age at first birth (never pregnant or pregnant at ages <20 years, 20–29 years, or ≥30 years), number of full term pregnancies (1, 2, 3, or >3), smoking status (never, past, or current), oral contraceptive use, history of hysterectomy, body mass index (BMI) (<25 kg/m², 25–29.9 kg/m², or ≥30 kg/m²), and history of oophorectomy. An α of .20 was used as a cutoff level for stepwise selection. Models examining dietary exposure to the micronutrients that had both dietary and supplemental sources also were adjusted for supplemental intake of the micronutrient; similarly, models examining supplemental sources were also adjusted for dietary intake of that micronutrient. Testing for linear trends in the nutrient quartiles was performed using standard orthogonal contrasts.

A post-hoc sensitivity analysis was performed excluding the women with kidney cancer who were diagnosed within 2 years of baseline for all micronutrients. Similarly, we repeated the analyses with time-varying dietary data for those women who had micronutrient intake data available at 3 years of follow-up. Dietary data were available for 91,138 women (95%), and supplement data were available for 69,897 women (73%) at 3 years of follow-up.

RESULTS

In total, 240 women in the study population were diagnosed with RCC during follow-up. The median follow-up for the entire sample was 4339 days (approximately 12 years), and the median time to kidney cancer diagnosis was 2518 days (approximately 7 years). Over the follow-up period, 6821 women were censored at the time of death, and 14,021 were censored at the time they were diagnosed with cancers other than kidney or nonmelanoma skin cancers.

Information on all baseline characteristics of the study population of 96,196 women was available for at least 97% of participants except for age at first birth (80% available), the number of pregnancies (90% available), and the number of weight cycles (37% available), defined as weight changes >10 pounds. Compared with the characteristics of women in the noncase group at baseline, women in the case group were more likely to have less education, treated hypertension, no use of oral contraceptives, a history of hysterectomy, higher BMI, a history of oophorectomy, and lower physical activity (Table 1). No significant differences were observed between women with RCC compared with unaffected women for age at screening, race/ethnicity, age at first birth, the number of pregnancies, smoking status, or the number of weight cycles.

TABLE 1.

Baseline Characteristics of Renal Cell Carcinoma Cases and Noncases

	No. of Women (%)
Characteristic	Noncases, n = 95,956	RCC Cases, n = 240	P
Age at screening, y
<50–59	30,999 (32)	65 (27)
60–69	42,494 (44)	108 (45)
70 to ≥79	22,463 (23)	67 (28)
Total	95,956	240	.13
Race/ethnicity
White	80,295 (84)	203 (85)
Black or African American	7398 (8)	18 (8)
Hispanic/Latino	3832 (4)	9 (4)
Asian or Pacific Islander	2670 (3)	3 (1)
American Indian or   Alaska Native	405 (<1)	4 (2)
Other	1104 (1)	2 (1)
Total	95,704	239	.16
Missing	252 (<1)	1 (<1)
Education
None to some HS	5103 (5)	24 (10)
HS diploma/GED	16,150 (17)	45 (19)
Vocational, training school,   some college, or associate   degree	35,174 (37)	95 (40)
College degree or more	38,765 (40)	75 (31)
Total	95,192	239	.002
Missing	764 (<1)	1 (<1)
Hypertension
Never	62,607 (67)	129 (56)
Untreated hypertensive	7482 (8)	20 (9)
Treated hypertensive	22,967 (25)	82 (35)
Total	93,056	231	<.001
Missing	2900 (3)	9 (3)
Age at first birth, y
Never had term pregnancy	2494 (3)	1 (1)
<20	11,448 (15)	35 (18)
20–29	56,226 (73)	147 (74)
≥30	7188 (9)	17 (9)
Total	77,356	200	.13
Missing	18,600 (20)	40 (17)
No. of pregnancies
1	6606 (8)	16 (7)
2	18,893 (22)	41 (18)
3	21,311 (25)	53 (24)
>3	39,584 (46)	112 (50)
Total	86,394	222	.51
Missing	9562 (10)	18 (8)
Smoking status
Never smoked	48,094 (51)	128 (54)
Past smoker	39,905 (42)	89 (37)
Current smoker	6625 (7)	21 (9)
Total	94,624	238	.24
Missing	1332 (1)	2 (<1)
Oral contraceptive use
No	56,830 (59)	161 (67)
Yes	39,126 (41)	79 (33)
Total	95,956	240	.013
History of hysterectomy
No	58,524 (61)	117 (49)
Yes	37,356 (39)	123 (51)
Total	95,880	240	<.001
Missing	76 (<1)	0 (0)
BMI, kg/m2
<25	37,329 (39)	71 (30)
25 to <30	32,615 (34)	90 (38)
≥30	25,011 (26)	78 (33)
Total	94,955	239	.007
Missing	1001 (1)	1 (<1)
History of oophorectomy
No	76,827 (82)	177 (77)
Yes	16,952 (18)	54 (23)
Total	93,779	231	.036
Missing	2177 (2)	9 (4)
No. of times weight   went up/down >10 lbs
1–3	15,259 (43)	43 (42)
4–6	11,412 (32)	31 (30)
7–10	5049 (14)	12 (12)
11–15	1543 (4)	6 (6)
>15	2250 (6)	11 (11)
Total	35,513	103	.38
Missing	60,443 (63)	137 (55)
Physical activity:	13.5 ±14.32	11.5 ±12.1	.037
Mean ± SD,   MET h/wk

Abbreviations: GED, General Educational Development test; HS, high school; MET, metabolic equivalent; RCC, renal cell carcinoma; SD, standard deviation.

The risk of RCC was inversely associated with quartiles of lycopene intake (P_trend = .029) (Table 2) when adjusting for intake of the other micronutrients examined. These results remained consistent after additional adjustment for age, clinical trial arm (clinical trial or observational study), race, education, BMI, hypertension, smoking, oral contraceptive use, hysterectomy ever, oophorectomy ever, physical activity, and energy intake (P_trend = .015). The number of pregnancies and weight cycling were excluded from the analysis because of large numbers of missing observations. In the fully adjusted model, compared with the first (lowest) quartile of lycopene intake, the risk of RCC was significantly reduced for the fourth (highest) quartile of intake (HR, 0.61; 95% CI, 0.39–0.97). There was no significant association by quartile of intake for any of the other micronutrients evaluated (total β-carotene, dietary β-carotene, α-carotene, β-cryptoxanthin, lutein plus zeaxanthin, total vitamin C, dietary vitamin C, total vitamin E, or dietary vitamin E). It is noteworthy that, when dietary intake and supplemental intake were examined separately for β-carotene, vitamin C, and vitamin E, no associations were observed. Finally, there was no substantial correlation (Pearson correlation coefficient, ≤0.47) among the different micronutrient variables to suggest the presence of collinearity (Table 3).

TABLE 2.

Renal Cell Carcinoma Hazard Ratios for Baseline Intake of Carotenoid and Vitamins C and E in the Women’s Health Initiative Cohort

Variable	Quartile 1	Quartile 2	Quartile 3	Quartile 4	P for Linear Trend
Total β-carotene intake, mcg	≤2462.1	2462.2–5030.7	5030.8–7875.2	≥7875.3
Noncases: No. (%)	23,924 (25)	23,933 (25)	23,938 (25)	23,943 (25)
RCC cases: No. (%)	70 (29)	62 (26)	57 (24)	51 (21)
Unadjusted HR [95% CI]a	1.00	0.93 [0.64–1.37]	0.90 [0.58–1.39]	0.90 [0.55–1.46]	.70
Multivariable adjusted HR [95% CI]b	1.00	0.96 [0.64–1.43]	0.96 [0.61–1.53]	0.90 [0.53–1.52]	.80
Dietary β-carotene intake, mcg	≤1798.3	1798.4–2789.2	2789.3–4320.1	≥44320.2
Noncases: No. (%)	23,933 (25)	23,935 (25)	23,934 (25)	23,942 (25)
RCC cases: No. (%)	63 (26)	61 (25)	62 (26)	54 (23)
Unadjusted HR [95% CI]a	1.00	0.99 [0.68–1.44]	1.03 [0.70–1.53]	0.91 [0.58–1.43]	.86
Multivariable adjusted HR [95% CI]c	1.00	1.09 [0.73–1.62]	1.12 [0.74–1.71]	1.00 [0.61–1.62]	.94
Supplemental β-carotene intake, mcg	0.0	0.05–4499.4	4500–4500.03	≥4500.03
Noncases: No. (%)	52,008 (54)	10,086 (11)	19,943 (21)	13,913 (14)
RCC cases: No. (%)	138 (57)	21 (9)	52 (22)	29 (12)
Unadjusted HR [95 CI]a	1.00	1.02 [0.56–1.85]	1.38 [0.80–2.37]	1.00 [0.58–1.71]	.44
Multivariable adjusted HR [95 CI]d	1.00	1.02 [0.55–1.88]	1.27 [0.73–2.22]	0.99 [0.57–1.70]	.59
Dietary α-carotene intake, mcg	≤343.1	343.2–572.7	572.8–927.2	≥927.3
Noncases: No. (%)	23,930 (25)	23,942 (25)	23,925 (25)	23,947 (25)
RCC cases: No. (%)	66 (28)	54 (23)	71 (30)	49 (20)
Unadjusted HR [95% CI]a	1.00	0.90 [0.62–1.31]	1.34 [0.92–1.96]	1.07 [0.65–1.78]	.18
Multivariable adjusted HR [95% CI]b	1.00	0.99 [0.66–1.48]	1.42 [0.94–2.14]	1.22 [0.72–2.08]	.12
Dietary β-cryptoxanthin intake, mcg	≤82.4	82.5–140.5	140.6–203.7	≥203.8
Noncases: No. (%)	23,927 (25)	23,939 (25)	23,936 (25)	23,942 (25)
RCC cases: No. (%)	69 (29)	57 (24)	60 (25)	54 (23)
Unadjusted HR [95% CI]a	1.00	0.87 [0.61–1.25]	0.96 [0.63–1.47]	0.94 [0.59–1.51]	.89
Multivariable adjusted HR [95% CI]b	1.00	0.91 [0.62–1.32]	0.95 [0.60–1.50]	0.93 [0.56–1.53]	.92
Dietary lutein + zeaxanthin intake, mcg	≤915.6	915.7–1369.6	1369.7–2183.2	≥2183.3
Noncases: No. (%)	23,924 (25)	23,935 (25)	23,934 (25)	23,951 (25)
RCC cases: No. (%)	72 (30)	61 (25)	62 (26)	45 (19)
Unadjusted HR [95% CI]a	1.00	0.84 [0.59–1.21]	0.89 [0.60–1.33]	0.69 [0.43–1.11]	.36
Multivariable adjusted HR [95% CI]b	1.00	0.82 [0.55–1.22]	0.94 [0.61–1.43]	0.69 [0.41–1.15]	.51
Dietary lycopene intake, mcg	≤2727.6	2727.7–4242.1	4242.3–6427.6	≥6427.7
Noncases: No. (%)	23,927 (25)	23,925 (25)	23,942 (25)	23,950 (25)
RCC cases: No. (%)	69 (29)	71 (30)	54 (23)	46 (19)
Unadjusted HR [95% CI]a	1.00	1.01 [0.71–1.42]	0.77 [0.53–1.12]	0.69 [0.45–1.05]	.029
Multivariable adjusted HR [95% CI]b	1.00	0.98 [0.68–1.42]	0.74 [0.50–1.11]	0.61 [0.39–0.97]	.015
Total vitamin C intake, mg	≤96.3	96.4–168.5	168.6–585.6	≥585.7
Noncases: No. (%)	23,926 (25)	23,935 (25)	23,934 (25)	23,943 (25)
RCC cases: No. (%)	68 (28)	60 (25)	61 (25)	51 (21)
Unadjusted HR [95% CI]a	1.00	0.99 [0.63–1.56]	1.12 [0.67–1.87]	0.96 [0.58–1.61]	.79
Multivariable adjusted HR [95% CI]b	1.00	1.05 [0.65–1.71]	1.22 [0.71–2.12]	1.12 [0.65–1.91]	.42
Dietary vitamin C intake, mg	≤61.2	61.3–96.6	96.7–135.2	≥135.3
Noncases: No. (%)	23,932 (25)	23,934 (25)	23,940 (25)	23,938 (25)
RCC cases: No. (%)	64 (27)	62 (26)	56 (23)	58 (24)
Unadjusted HR [95% CI]a	1.00	0.97 [0.68–1.38]	0.87 [0.59–1.29]	0.91 [0.60–1.38]	.52
Multivariable adjusted HR [95% CI]c	1.00	1.08 [0.74–1.57]	0.92 [0.60–1.40]	1.02 [0.66–1.58]	.67
Supplemental vitamin C intake, mcg	0.0	0.009–89.2	90.0–557.1	≥560.0
Noncases: No. (%)	40,512 (42)	17,907 (19)	17,182 (18)	20,349 (21)
RCC cases: No. (%)	111 (46)	39 (16)	44 (18)	46 (19)
Unadjusted HR [95% CI]a	1.00	0.83 [0.43–1.62]	1.07 [0.62–1.85]	0.98 [0.57–1.68]	.49
Multivariable adjusted HR [95% CI]d	1.00	0.95 [0.48–1.89]	1.22 [0.69–2.14]	1.17 [0.66–2.05]	.24
Total vitamin E intake, IU	≤8.5	8.6–36.0	36.1–405.7	≥405.8
Noncases: No. (%)	23,926 (25)	23,930 (25)	23,937 (25)	23,945 (25)
RCC cases: No. (%)	69 (29)	64 (27)	58 (24)	49 (20)
Unadjusted HR [95% CI]a	1.00	1.01 [0.69–1.48]	0.91 [0.57–1.43]	0.80 [0.50–1.28]	.36
Multivariable adjusted HR [95% CI]b	1.00	1.03 [0.67–1.56]	0.88 [0.53–1.46]	0.81 [0.49–1.33]	.33
Dietary vitamin E intake, IU	≤5.6	5.7–7.6	7.7–10.4	≥10.4
Noncases: No. (%)	23,927 (25)	23,943 (25)	23,945 (25)	23,929 (25)
RCC cases: No. (%)	69 (29)	53 (22)	52 (22)	66 (28)
Unadjusted HR [95% CI]a	1.00	0.76 [0.53–1.10]	0.76 [0.52–1.11]	0.99 [0.68–1.44]	.99
Multivariable adjusted HR [95% CI]c	1.00	0.79 [0.53–1.17]	0.76 [0.49–1.18]	0.99 [0.60–1.62]	.92
Supplemental vitamin E intake, IU	0.0	0.02–30.0	30.1–400.0	≥400.03
Noncases: No. (%)	39,155 (41)	23,117 (24)	15,365 (16)	18,313 (19)
RCC cases: No. (%)	111 (46)	56 (23)	36 (15)	37 (15)
Unadjusted HR [95% CI]a	1.00	0.76 [0.41–1.41]	0.77 [0.47–1.26]	0.63 [0.35–1.12]	.20
Multivariable adjusted HR [95% CI]c	1.00	0.75 [0.39–1.42]	0.78 [0.47–1.30]	0.67 [0.37–1.22]	.32

Abbreviations: CI, confidence interval; HR, hazard ratio; RCC, renal cell carcinoma;

^aThis model includes all of the micronutrients.

^bThe adjusted model for multivariate analysis includes all of the micronutrients, age, clinical trial, race, education, body mass index, hypertension, smoking status, oral contraceptive use, hysterectomy ever, oophorectomy ever, physical activity, and energy intake. The number of pregnancies and weight cycling were excluded from the analysis because of large numbers of missing observations.

^cThis model includes all variables from the multivariable analysis plus supplemental intake of this micronutrient.

^dThis model includes all variables from the multivariable analysis plus dietary intake of this micronutrient.

TABLE 3.

Intake Correlation Between Micronutrients

	Pearson Correlation Coefficient
Micronutrient	β-Carotene, mcg	α-Carotene, mcg	β-Cryptoxanthin, mcg	Lutein Plus Zeaxanthin, mcg	Lycopene, mcg	Vitamin C, mg	Vitamin E, IU
β-Carotene, mcg	1.00	0.38	0.25	0.32	0.23	0.36	0.27
α-Carotene, mcg	0.38	1.00	0.47	0.45	0.39	0.11	0.07
β-Cryptoxanthin, mcg	0.25	0.47	1.00	0.36	0.33	0.14	0.06
Lutein + zeaxanthin, mcg	0.32	0.45	0.36	1.00	0.32	0.08	0.04
Lycopene, mcg	0.23	0.39	0.33	0.32	1.00	0.09	0.05
Vitamin C, mg	0.36	0.11	0.14	0.08	0.09	1.00	0.38
Vitamin E, IU	0.27	0.07	0.06	0.04	0.05	0.38	1.00

When we excluded all individuals who developed cancer (other than nonmelanoma skin cancer) or who died within the first 2 years of follow-up, there were 213 kidney cancer cases and 93,966 noncases available for the sensitivity analysis. The results were similar to those from the main analysis, in that dietary lycopene was the only micronutrient associated with a reduced risk of RCC incidence (fourth quartile: HR, 0.58; 95% CI, 0.36–0.95; P_trend = .01). Likewise, results from the time-varying analysis were similar, suggesting that lycopene was associated with a decreased risk of RCC (fourth quartile: HR, 0.64; 95% CI, 0.42–0.98; P_trend = .05), and no other micronutrients were associated with risk.

DISCUSSION

In our analysis of correlations between micronutrient intake and the risk of RCC in the WHI, only lycopene intake was inversely associated with the risk of RCC. Comparison of intake by quartiles and multivariable analyses did not reveal any relation between the other assessed micronutrients and RCC. It is noteworthy that no carotenoids, including lycopene, can be synthesized by the body; rather, they are obtained exclusively from dietary sources. Approximately 85% of dietary lycopene originates from tomato fruit or tomato-based products, including juice, ketchup, soup, pizza, and pasta sauces. Other sources include watermelon, pink grapefruit, guava, and papaya.²⁸

The molecular mechanisms behind the metabolism and effect of individual carotenoids on cancer risk have not been clearly elucidated. However, there are distinguishing features regarding lycopene versus other carotenoids that may provide clues regarding a potential anticancer effect of lycopene. First, in contrast to provitamin A carotenoids (α-carotene, β-carotene, and β-cryptoxanthin), it suggested that lycopene, along with other carotenoids without provitamin A activity (lutein and zeaxanthin), has a different enzymatic affinity. Thus, the absorption and metabolism of lycopene may not be regulated based on vitamin A levels through intestine-specific homeobox (ISX) transcription factor.²⁹ Second, lycopene is considered to be 1 of the most hydrophobic carotenoids.³⁰ This physicochemical distinction implicates its ability to affect various transcriptional pathways at the molecular level differently from other carotenoids. Finally, it has also been demonstrated that lycopene has the greatest singlet oxygen-quenching activity of any of the carotenoids.³¹ These characteristics have made lycopene a good candidate for research as a potential anticancer agent.

The correlation between lycopene intake and a lower risk of RCC observed in our study has not been observed in other prospective studies^13,20 or case-control analyses.^16,19,32 The Alpha-Tocopherol and Beta-Carotene (ATBC) prospective study demonstrated that vitamin E—but not vitamin A, vitamin C, β-carotene, α-carotene, β-cryptoxanthin, lutein plus zeaxanthin, or lycopene—was directly correlated with increased RCC risk.¹³ Another prospective study using databases from the Nurses’ Health Study (NHS) and the Health Professionals Follow-Up Study (HPFS) demonstrated that RCC risk was inversely correlated with vitamin A, vitamin C, β-carotene, α-carotene, β-cryptoxanthin, and lutein plus zeaxanthin, but not with lycopene or vitamin E.²⁰ Previous case-control studies also reported null associations between lycopene and RCC.^16,19,32

For the other micronutrients, pooled analyses of available prospective cohorts on RCC and micronutrient intake have indicated a reduced risk of RCC with increased β-carotene intake.³³ β-Carotene intake also was inversely associated with RCC in 2 case-control studies,^19,32 but not in others.^16,34 The results from case-control studies also were largely conflicting regarding vitamins C and E.^{16,18,19,32,34} It is noteworthy that results from the Iowa prospective study on postmenopausal women indicated that vitamin C intake increased the risk of RCC, whereas vitamin E intake decreased the risk.²² However, the null association we observed for vitamin C and vitamin E intake correlated with results from the pooled analysis.³³

The discrepancy between our results and those published in the literature, especially regarding lycopene, may be caused in part by different study design methods and sample selection. For example, the ATBC prospective study included only men.¹³ Also, the NHS was based only on women ages 30 to 55 years, and the subset analyses evaluating the association between diet and ther risk of RCC in postmenopausal women were limited to only 86 women with RCC. The analysis of lycopene intake in the NHS also was restricted to a smaller portion of the study, involving only 106 women with RCC, because there was a lack of data from the questionnaires during the earlier part of the study relative to other micronutrients.²⁰ Although previous studies using an Iowa database explored the relation between micronutrients and RCC in postmenopausal women (ie, vitamin E and vitamin C were associated with decreased and increased risk of RCC, respectively), the intake of carotenoids was not addressed.^14,22 The smoking status of a study population may also be a contributing factor. Specifically, the ATBC trial, which was solely based on smokers, demonstrated a null association between lycopene and RCC. Similarly, 1 case-control study reported a significant inverse association between lycopene and RCC in the never-smoker group but not in the ever-smoker group.³⁴ Beyond these observations, however, to our knowledge, no previous studies have explored whether smoking masks or enhances the physiologic effects of micronutrient intake. In our study, a subgroup analysis based on smoking status could not be performed because of the lack of statistical power: only 7% of the sample were current smokers, and only 21 cases occurred among smokers. The discrepancy may also be attributed to differences in dietary intake by the study population. The average intake of lycopene in the ATBC trial ranged from approximately 100 to 1000 µg daily; whereas, in the NHS and the HPFS, the range was roughly between 3000 and 16,000 µg daily, similar to the WHI, in which the intake ranged from 1000 to 10,000 µg daily. Certainly, it is important to note that all databases depend on unique dietary questionnaires and are subject to variation across studies. To our knowledge, ours is the first prospective study to evaluate and demonstrate a significant association between lycopene intake and RCC in postmenopausal women.

The association of lycopene intake has been investigated previously in relation to the risk of other cancers. A pooled analysis of 8 prospective breast cancer studies demonstrated that, along with α-carotene, β-carotene, and lutein plus zeaxanthin, lycopene intake was inversely associated with breast cancer risk.³⁵ A similar relation was observed in the WHI, which demonstrated a slightly lower risk of hormone-sensitive breast cancer when comparing the highest quintile versus the lowest quintile of lycopene intake (HR, 0.85; 95% CI, 0.73–1.00).²⁶ In that analysis, a lower risk of hormone-sensitive breast cancer also was associated with dietary—but not total or supplementary—intake of α-carotene and β-carotene, and no relation with intake of vitamin C or E was observed. In addition, a large meta-analysis of 11 case-control studies and 10 cohort studies demonstrated a reduced risk of prostate cancer associated with tomato, tomato products, or lycopene intake.^36,37

In addition to the epidemiologic evidence, in vitro cell line studies have indicated that lycopene inhibits the proliferation of human breast, endometrial, and lung cancer cells with greater efficacy than α-carotene or β-carotene.³⁸ In fact, various mechanisms for the effect of lycopene and its breakdown derivatives on cancer cells have been proposed, including induction of the electrophile/antioxidant response element (EpRE/ARE) transcription system, inhibition of insulin-like growth factor (IGF) effects by interacting with IGF-binding proteins (IGFBPs), and suppression of the activity of nuclear transcription factor κB (NF-κB).³⁹ Accordingly, in multiple animal models, lycopene also exhibited renoprotective effects against nephrotoxins such as cisplatin and ochratoxin A,^40,41 which have a known association with increased RCC risk.^42,43 Furthermore, clinically, the expression of IGF I receptor is associated with poor survival in patients with RCC,⁴⁴ and NF-κB has widely been considered a potential molecular target for RCC therapeutics based on its role in RCC pathogenesis.^45,46 Taken together, these mechanisms may explain in part the suggested protective effect of lycopene against RCC risk observed in our study.

There are several notable strengths in this study. Our analysis was based on a large, multisite population with a relatively substantial number of RCC cases. It featured comprehensive information on both supplemental and dietary micronutrient intake. Because the study focused on postmenopausal women and obtained data on baseline reproductive characteristics, we were able to control for variables like sex, menopausal status, oral contraceptive use, and prior gynecologic surgeries. However, the following limitations of the study design should be considered. First, our analysis was based on micronutrient intake measured at baseline and not on plasma levels of antioxidants. Our study also did not take into account micronutrient ingestion over time, although it should be noted that the total intake of fruits and vegetables in adult females has not changed substantially in the past 2 decades.^47,48 Moreover, the observation of a protective effect of lycopene remained within our sensitivity time-varying analysis among the subset of women who had 3 years of dietary follow-up data. Second, data regarding RCC subtypes were not sufficient for analysis, which is important to note because the exact pathogenesis may differ based on histologic subtype. Finally, because of the nature of the study, our results merely suggest a preventive role of lycopene on RCC.

Conclusions

Increased lycopene intake among postmenopausal women in the WHI was associated with a lower risk of RCC. Because there are no current guidelines for RCC prevention or screening, further studies on the mechanism of the potential effects of lycopene on the risk of RCC would be valuable.