Abstract
It has been hypothesized that consumption of citrus, a group of foods particularly rich in a class of photoactive compounds known as furocoumarins, may increase the risk of malignant melanoma. However, this hypothesis has not been rigorously studied in a general sample of US men and women. This study examined the relationship between citrus intake and melanoma risk in participants of the NIH-AARP Diet and Health Study. Among 388,467 adults, 3,894 melanoma cases were identified during a median follow-up of 15.5 years. After adjustment for relevant potential confounders, total citrus consumption was not significantly associated with melanoma risk in this cohort. Among those with higher estimated exposure to ultraviolet radiation, and among those aged 60+ years at baseline, there were significant trends toward increased melanoma risk associated with whole citrus fruit consumption (p-trends = 0.01 and 0.02, respectively), but the hazard ratios of the top consumers (2+ cups per week) versus non-consumers were nonsignificant. Further research is needed to explore associations of citrus with melanoma risk among older adults and those with high sun exposure.
Keywords: diet, cancer, melanoma, NIH-AARP Diet and Health Study, citrus, sun exposure
Introduction:
Skin cancer is the most commonly diagnosed cancer in the US, and the incidence of melanoma, the most serious form of skin cancer, has continued to rise over the past decades (1). An estimated 96,480 new melanoma cases and 7,230 deaths are expected to occur in the US in 2019 (2). While exposure to ultraviolet radiation (UVR) is a well-established risk factor for melanoma, the patterns of UVR exposure that best predict risk are not fully understood and may differ between individuals (3). Furthermore, numerous host traits and environmental factors affect the skin’s response to UVR exposure (4, 5), and recent research has indicated the importance of certain lifestyle factors including dietary exposures in determining melanoma risk (6, 7).
A 2015 study that pooled data from the Nurses’ Health Study (NHS) and the Health Professionals Follow-Up Study (HPFS) showed that greater consumption of citrus products was associated with greater risk of incident melanoma (8). Similarly, we recently found that citrus juice intake was positively associated with melanoma risk among participants of the Women’s Health Initiative (WHI) who spent at least 30 minutes outdoors daily during the summer (9). It has been hypothesized that a class of photoactive compounds known as furocoumarins may be responsible for this observed relationship between citrus and melanoma (10, 11). Furocoumarins are produced by several edible plant species and are especially rich in certain citrus products (12). We recently found that citrus products contribute approximately 93% of total dietary furocoumarins among US adults when considering seven major furocoumarin compounds (13). The phototoxic activities of furocoumarins have been well documented (14). If photoactivated by UVR, particularly within the UVA spectrum, furocoumarins may form monoadducts (15) or interstrand crosslinks with DNA (16). Animal studies have demonstrated that DNA crosslinking by furocoumarins can cause inhibition of DNA synthesis, mutation, recombination, chromosome aberrations, and tumor induction (17–20).
While citrus consumption has been shown to be positively associated with melanoma risk among cohorts of US women (9) and health professionals (8), the relationship has not been rigorously tested in a prospective analysis of a general sample of US adults. Therefore, the aim of this study was to examine the relationship between citrus intake and melanoma risk in a sample of the general US population of men and women.
Methods:
Study Population
The National Institutes of Health (NIH) and the AARP formed the NIH-AARP Diet and Health Study in 1995–1996. Questionnaires were mailed to 3.5 million members of the AARP residing in one of six US states (California, Florida, Pennsylvania, New Jersey, North Carolina, and Louisiana) or two metropolitan areas (Atlanta, Georgia and Detroit, Michigan) (21). Questionnaires asked about demographic and health-related behaviors along with dietary intake. Baseline data were available for 544,735 participants aged 50–71 years at baseline, of whom approximately 60% were male. For this study, we excluded those with a self-reported history of cancer (n = 48,536). Importantly, participants were not asked about history of any type of skin cancer, so the analytic cohort likely includes participants with and without histories of melanoma and nonmelanoma skin cancers. We additionally excluded proxy-responders (n = 70,626), those with a self-identified race/ethnicity other than white non-Hispanic (n = 46,766) due to their much lower risk of melanoma (5), those with extreme energy intakes (defined as greater than two times the interquartile range below the 25th percentile or above the 75th percentile) (n = 10,225), and those with zero years of follow-up (n = 52), yielding an analytical cohort of 388,467 participants.
Follow-Up and Case Ascertainment
Participants were followed from baseline until the date of first melanoma or other cancer diagnosis, death, the end of study follow-up (December 31, 2011), or when the participant moved out of the registry area, whichever came first. Incident cases of cutaneous malignant melanoma were identified through probabilistic record linkage with state cancer registries. Cutaneous malignant melanoma was defined according to the third edition of the International Classification of Disease for Oncology.
Assessment of Citrus Intake
Citrus consumption during the year prior to baseline was assessed through questionnaire items asking about consumption of “orange juice or grapefruit juice” (10 frequency options ranging from never to 6+ times per day, with 3 portion options ranging from less than ¾ cup to more than 1 cup), “oranges, tangerines, tangelos” (10 frequency options ranging from never to 2+ times per day, with 3 portion options ranging from less than 1 orange to more than 1 orange), and “grapefruit” (10 frequency options ranging from never to 2+ times per day, with 3 portion options ranging from less than ½ grapefruit to more than ½ grapefruit).
Assessment of Covariates
Baseline questionnaires were used to gather information on important covariates. History of cigarette smoking was defined as never, former, or current. Body mass index (BMI) was calculated from self-reported height and weight. Education level was grouped into the following categories: ≤11 years, high school graduate, some college or other post-high school training, and college graduate. Average daily alcohol intake over year preceding baseline assessment was estimated from consumption of beer, wine, and liquor. Physical activity during the year preceding baseline assessment was defined as the frequency of activity lasting 20 minutes or longer and that caused increases in breathing or heart rate or sweating. Family history of cancer (excluding skin cancers, which were not ascertained) was also gathered from baseline questionnaires. Erythemal UVR exposure was assessed in this cohort as previously described (22). Briefly, NASA Total Ozone Mapping Spectrometer (TOMS) estimates for noon-time ground-level erythemal UVR measured during the month of July between 1978–1993 and 1996–2005 were averaged. Erythemal UVR exposures were assigned to participants by deterministic linkage of the census tract centroid of baseline residence to the closes point on the TOMS grid.
Statistical Analysis
All statistical analyses were carried out using SAS software, version 9.4 (SAS Institute, Cary, NC). Baseline characteristics of the analytic sample were summarized by level of citrus intake. The association between citrus intake and incident malignant melanoma was assessed using Cox proportional hazards regression models to estimate hazard ratios (HRs) and 95% confidence intervals (CIs). To examine the specificity and consistency of the findings, the HRs of melanoma in association with consumption of non-citrus fruits were also computed. Follow-up time was used as the time metric in proportional hazards modeling. Base multivariable models were adjusted for age and sex, and additional potential confounders were included in full models: cigarette smoking, BMI, education, alcohol intake, physical activity, family history of cancer, and July erythemal UVR exposure. We tested for linear trends in HRs across categories of citrus consumption by conducting linear regression modeling using the median value of the intake level plotted against the corresponding HRs.
To examine the robustness of the data, we conducted additional analyses stratified by UVR exposure level (below and above the median) given that the hypothesized link between citrus and melanoma involves furocoumarin-UVR interaction (11), and by age (<60 years at baseline and ≥ 60 years) because the median melanoma diagnosis age in the US falls within the lower age range in this cohort, between about 55 and 60 years of age (23).
Results:
The mean age of participants at baseline was 62 years and the median follow-up time was 15.5 years. A total of 3,894 cases of malignant melanoma were identified during follow-up. The majority of the eligible cohort was male (59.2%) and 72.9% had at least some college education. The majority were either former or current cigarette smokers (61.5%), and nearly half (46.4%) engaged in physical activity at least three times per week. The median total intake of citrus products was 3.57 cup equivalents per week. At baseline, age and BMI were not associated with citrus intake (Table 1), but citrus consumption differed by several factors including education, physical activity, and smoking history.
Table 1.
Baseline characteristics of NIH-AARP Diet and Health Study participants by level of total citrus consumption
| Total Citrus | None | 0 < cup/wk < 1 | 1 ≤ cups/wk < 3.5 | 0.5 ≤ cups/day < 1 | 1+ cups/day | p-value* |
|---|---|---|---|---|---|---|
| N | 9,610 | 85,589 | 93,920 | 108,950 | 90,398 | |
| Median Total Citrus Intake (cups/wk) | 0 | 0.49 | 2.03 | 5.18 | 10.43 | |
| Age at entry (mean, years) | 61.82 | 60.93 | 61.07 | 61.99 | 62.12 | 0.22 |
| BMI (mean, kg/m2) | 26.52 | 27.13 | 27.26 | 26.86 | 26.87 | 0.96 |
| Sex | <0.001 | |||||
| Male (n, %) | 5,065 (52.71%) | 48,195 (56.31%) | 55,015 (58.58%) | 64,320 (59.04%) | 57,387 (63.48%) | |
| Female (n, %) | 4,545 (47.29%) | 37,394 (43.69%) | 38,905 (41.42%) | 44,630 (40.96%) | 33,011 (36.52%) | |
| Cigarette Smoking | <0.001 | |||||
| Never Smoker (n, %) | 2,660 (27.68%) | 25,620 (29.93%) | 32,561 (34.67%) | 40,566 (37.23%) | 35,615 (39.40%) | |
| Former Smoker (n, %) | 4,559 (47.44%) | 42,460 (49.61%) | 47,130 (50.18%) | 54,888 (50.38%) | 44,659 (49.40%) | |
| Current Smoker (n, %) | 2,035 (21.18%) | 14,765 (17.25%) | 11,214 (11.94%) | 10,045 (9.22%) | 7,148 (7.91%) | |
| Education | <0.001 | |||||
| ≤ 11 years (n, %) | 901 (9.38%) | 5,693 (6.65%) | 4,694 (5.00%) | 5,001 (4.59%) | 4,351 (4.81%) | |
| High School Graduate (n, %) | 2,377 (24.73%) | 19,290 (22.54%) | 17,837 (18.99%) | 20,327 (18.66%) | 15,919 (17.61%) | |
| Some College or other Post-HS training (n, %) | 3,322 (34.57%) | 30,467 (35.61%) | 32,367 (34.46%) | 34,712 (31.86%) | 28,147 (31.14%) | |
| College Graduate (n, %) | 2,724 (28.35%) | 28,050 (32.77%) | 36,945 (39.34%) | 46,667 (42.83%) | 39,911 (44.15%) | |
| Physical Activity1 | <0.001 | |||||
| Never/Rarely (n, %) | 2,663 (27.71%) | 19,428 (22.70%) | 16,124 (17.17%) | 17,157 (15.75%) | 12,460 (13.78%) | |
| 1–3/mo (n, %) | 1,240 (12.90%) | 13,952 (16.30%) | 13,834 (14.73%) | 13,978 (12.83%) | 10,224 (11.31%) | |
| 1–2/wk (n, %) | 1,772 (18.44%) | 18,307 (21.39%) | 21,904 (23.32%) | 24,185 (22.20%) | 18,632 (20.61%) | |
| 3–4/wk (n, %) | 2,055 (21.38%) | 19,805 (23.14%) | 25,460 (27.11%) | 31,291 (28.72%) | 26,761 (29.60%) | |
| 5+/wk (n, %) | 1,794 (18.67%) | 13,563 (15.85%) | 16,025 (17.06%) | 21,669 (19.89%) | 21,769 (24.08%) | |
| Alcohol (g/day)2 | <0.001 | |||||
| None (n, %) | 3,986 (41.48%) | 23,377 (27.31%) | 20,547 (21.88%) | 22,785 (20.91%) | 19,311 (21.36%) | |
| >0 and <5 g/day (n, %) | 3,059 (31.83%) | 33,842 (39.54%) | 38,389 (40.87%) | 43,254 (39.70%) | 35,249 (38.99%) | |
| ≥5 and <20 g/day (n, %) | 1,280 (13.32%) | 15,413 (18.01%) | 20,924 (22.28%) | 25,722 (23.61%) | 21,244 (23.50%) | |
| ≥20 and <30 g/day (n, %) | 318 (3.31%) | 3,824 (4.47%) | 4,780 (5.09%) | 6,297 (5.78%) | 5,343 (5.91%) | |
| ≥30 g/day (n, %) | 967 (10.06%) | 9,133 (10.67%) | 9,280 (9.88%) | 10,892 (10.00%) | 9,251 (10.23%) | |
| July erythemal UVR (J/m2)3 | ||||||
| ≤186.3 (n, %) | 2,158 (22.46%) | 19,577 (22.87%) | 23,027 (24.52%) | 30,881 (28.34%) | 25,728 (28.46%) | <0.001 |
| >186.3–237.0 (n, %) | 2,162 (22.50%) | 22,196 (25.93%) | 23,129 (24.63%) | 26,643 (24.45%) | 20,565 (22.75%) | |
| >237.0–253.7 (n, %) | 2,337 (24.32%) | 20,670 (24.15%) | 22,739 (24.21%) | 23,494 (21.56%) | 18,853 (20.86%) | |
| >253.7 (n, %) | 2,942 (30.61%) | 23,036 (26.91%) | 24,901 (26.51%) | 27,810 (25.53%) | 25,134 (27.80%) | |
| Family history of cancer4 | <0.001 | |||||
| No (n, %) | 4,500 (46.83%) | 39,089 (45.67%) | 42,853 (45.63%) | 49,708 (45.62%) | 41,958 (46.41%) | |
| Yes (n, %) | 4,636 (48.24%) | 42,744 (49.94%) | 46,778 (49.81%) | 54,416 (49.95%) | 44,266 (48.97%) |
Physical activity defined as activities lasting 20 or more minutes that caused increases in breathing or heart rate or worked up a sweat
Alcohol from alcoholic drinks including beer, wine, and liquor
July erythemal UVR exposure was categorized as quartiles
Family cancer history based on participants’ reports of cancer among first-degree family members
for categorical variables, p-value based on Cochran-Mantel-Haenszel Statistics; for continuous variables, p-value represents trend for linear relationship assessed through linear regression modeling assuming citrus consumption was equal to the median value among participants in each category
Total citrus intake was not associated with melanoma risk in this cohort (Table 2). In the age- and sex-adjusted model, higher consumers of whole citrus fruits had increased melanoma risk. The HR (and 95% CI) for those consuming 0.25 – 0.75 cups per week versus non-consumers was 1.16 (1.01, 1.33). HRs for those consuming 0.75 – 2 cups of citrus fruits per week and 2+ cups per week were 1.17 (1.01, 1.35) and 1.19 (1.03, 1.37), respectively (p-trend = 0.22). These associations were attenuated and not significant after further adjustment for other potential confounders. The highest consumers of grapefruit (3+ cups per week) also had somewhat increased risk of melanoma in comparison to non-consumers, with an age- and sex-adjusted HR of 1.21 (1.06, 1.39). However, there were no significant linear trends in HRs among categories of grapefruit consumption, and the positive associations were somewhat attenuated in the multivariable-adjusted model.
Table 2.
Association of citrus product consumption with malignant melanoma in NIH-AARP Diet and Healthy Study participants
| Model Estimates | p-trend1 | |||||
|---|---|---|---|---|---|---|
| Total Citrus | None | 0 < cup/wk < 1 | 1 ≤ cups/wk < 3.5 | 0.5 ≤ cups/day < 1 | 1+ cups/day | |
| No. Cases/No. noncases | 73 / 9,537 | 693 / 84,896 | 973 / 92,947 | 1,190 / 107,760 | 965 / 89,433 | - |
| Age- and Sex-Adjusted HR (95% CI) | 1 (ref) | 1.02 (0.80, 1.30) | 1.27 (1.00, 1.60) | 1.29 (1.02, 1.64) | 1.24 (0.97, 1.57) | 0.24 |
| Multivariable-Adjusted HR (95% CI)* | 1 (ref) | 0.98 (0.77, 1.25) | 1.15 (0.91, 1.46) | 1.16 (0.92, 1.47) | 1.09 (0.86, 1.39) | 0.41 |
| Citrus Juices | None | 0 ≥ cups/wk < 0.25 | 0.25 ≥ cups/wk < 0.5 | 0.5 ≥ cups/wk < 1 | 1+ cups/wk | |
| No. Cases/No. noncases | 278 / 34,329 | 1,538 / 161,631 | 533 / 47,880 | 1,177 / 14,645 | 368 / 36,088 | - |
| Age- and Sex-Adjusted HR (95% CI) | 1 (ref) | 1.11 (0.98, 1.26) | 1.30 (1.12, 1.50) | 1.23 (1.08, 1.41) | 1.11 (0.95, 1.30) | 0.78 |
| Multivariable-Adjusted HR (95% CI)* | 1 (ref) | 1.06 (0.93, 1.20) | 1.20 (1.04, 1.39) | 1.14 (1.00, 1.30) | 1.02 (0.87, 1.19) | 0.90 |
| Citrus Fruits 2 | None | 0 ≥ cups/wk < 0.25 | 0.25 ≥ cups/wk < 0.75 | 0.75 ≥ cups/wk < 2 | 2+ cups/wk | |
| No. Cases/No. noncases | 244 / 28,404 | 843 / 88,853 | 991 / 95,032 | 781 / 74,439 | 1,035 / 97,845 | - |
| Age- and Sex-Adjusted HR (95% CI) | 1 (ref) | 1.08 (0.94, 1.25) | 1.16 (1.01, 1.33) | 1.17 (1.01, 1.35) | 1.19 (1.03, 1.37) | 0.22 |
| Multivariable-Adjusted HR (95% CI)* | 1 (ref) | 1.05 (0.91, 1.21) | 1.07 (0.93, 1.23) | 1.05 (0.91, 1.22) | 1.07 (0.93, 1.23) | 0.36 |
| Grapefruit | None | 0 ≥ cups/wk < 1 | 1 ≥ cups/wk < 2 | 2 ≥ cups/wk < 3 | 3+ cups/wk | |
| No. Cases/No. noncases | 880 / 101,609 | 2,420 / 230,674 | 107 / 9,568 | 203 / 16,672 | 284 / 26,050 | - |
| Age- and Sex-Adjusted HR (95% CI) | 1 (ref) | 1.19 (1.10, 1.28) | 1.25 (1.03, 1.53) | 1.35 (1.16, 1.58) | 1.21 (1.06, 1.39) | 0.40 |
| Multivariable-Adjusted HR (95% CI)* | 1 (ref) | 1.10 (1.02, 1.19) | 1.15 (0.94, 1.41) | 1.21 (1.04, 1.41) | 1.10 (0.96, 1.26) | 0.50 |
| Oranges/Tangerines/Tangelos | None | 0 ≥ cups/wk < 0.25 | 0.25 ≥ cups/wk < 0.75 | 0.75 ≥ cups/wk < 2 | 2+ cups/wk | |
| No. Cases/No. noncases | 422 / 46,140 | 1,301 / 130,143 | 786 / 76,068 | 706 / 66,069 | 679 / 66,153 | - |
| Age- and Sex-Adjusted HR (95% CI) | 1 (ref) | 1.06 (0.95, 1.19) | 1.08 (0.96, 1.22) | 1.13 (1.00, 1.27) | 1.12 (0.99, 1.26) | 0.18 |
| Multivariable-Adjusted HR (95% CI)* | 1 (ref) | 1.03 (0.92, 1.15) | 1.01 (0.90, 1.14) | 1.04 (0.92, 1.18) | 1.03 (0.91, 1.17) | 0.38 |
| Non-Citrus Fruits | < 1 cup/wk | 1 ≥ cups/wk < 7 | 1 ≥ cups/day < 1.5 | 1.5 ≥ cups/day < 2 | 2+ cups/day | |
| No. Cases/No. noncases | 261 / 28,128 | 1,876 / 187,543 | 743 / 72,551 | 484 / 42,800 | 530 / 53,551 | - |
| Age- and Sex-Adjusted HR (95% CI) | 1 (ref) | 1.04 (0.92, 1.19) | 1.06 (0.92, 1.23) | 1.18 (1.01, 1.37) | 1.02 (0.88, 1.18) | 0.65 |
| Multivariable-Adjusted HR (95% CI)* | 1 (ref) | 0.95 (0.83, 1.08) | 0.94 (0.81, 1.08) | 1.03 (0.89, 1.20) | 0.90 (0.77, 1.04) | 0.48 |
Multivariable models adjusted for age (continuous), sex (male/female), cigarette smoking (never, former, current), BMI category (underweight, normal weight, overweight, obese), education (≤11 years, high school graduate, some college or other post-HS training, college graduate), average daily alcohol intake (continuous), physical activity (engaged in physical activity never or rarely, 1–3 times/month, 1–2 times/week, 3–4 times/week, or 5 or more times/week), family history of cancer (yes/no), and July erythemal exposure (quartiles)
p-value represents trend for linear relationship assessed through linear regression modeling assuming citrus consumption was equal to the median value of participants in the category
Citrus fruits calculated as the sum of reported intakes of grapefruits and oranges/tangerines/tangelos
Because grapefruit and total citrus fruit consumption had the most apparent associations with melanoma risk, total citrus fruits and each individual whole citrus fruit were further examined by subgroup of UVR exposure (Table 3) and age (Table 4). Among those in the upper half of UVR exposure, those consuming the most total citrus fruits (2+ cups per week) had somewhat elevated melanoma risk compared to non-consumers in age- and sex-adjusted modeling [HR: 1.25 (1.03, 1.51)]. In the multivariable adjusted model, there was a significant linear trend towards increasing risk of melanoma among higher citrus fruit consumers with higher UVR exposure (p-trend = 0.01), but the HRs for each consumption category were nonsignificant. Neither grapefruit consumption nor orange consumption were associated with increased risk of melanoma within either subgroup of UVR exposure.
Table 3.
Association of citrus consumption with malignant melanoma among subgroups of UVR1 in NIH-AARP Diet and Health Study participants
| Model Estimates | p-trend2 | |||||
|---|---|---|---|---|---|---|
| Citrus Fruits 3 | None | 0 < cups/wk < 0.25 | 0.25 ≤ cups/wk < 0.75 | 0.75 ≤ cups/wk < 2 | 2+ cups/wk | |
| UVR ≤ 237 J/m 2 | ||||||
| No. Cases/No. noncases | 115 / 13,740 | 409 / 46,537 | 510 / 49,338 | 359 / 37,771 | 444 / 46,843 | - |
| Age- and Sex-Adjusted HR (95% CI) | 1 (ref) | 1.02 (0.83, 1.26) | 1.18 (0.96, 1.44) | 1.09 (0.89, 1.35) | 1.11 (0.90, 1.36) | 0.56 |
| Multivariable-Adjusted HR (95% CI)* | 1 (ref) | 0.99 (0.81, 1.22) | 1.08 (0.88, 1.32) | 0.99 (0.80, 1.22) | 1.00 (0.82, 1.23) | 0.80 |
| UVR > 237 J/m 2 | ||||||
| No. Cases/No. noncases | 129 / 14,621 | 433 / 42,208 | 480 / 45,568 | 421 / 36,592 | 590 / 50,874 | - |
| Age- and Sex-Adjusted HR (95% CI) | 1 (ref) | 1.15 (0.94, 1.40) | 1.14 (0.94, 1.39) | 1.24 (1.02, 1.51) | 1.25 (1.03, 1.51) | 0.19 |
| Multivariable-Adjusted HR (95% CI)* | 1 (ref) | 0.99 (0.83, 1.18) | 1.01 (0.85, 1.20) | 1.00 (0.84, 1.19) | 1.07 (0.90, 1.26) | 0.01 |
| Grapefruit | None | 0 ≥ cups/wk < 1 | 1 ≥ cups/wk < 2 | 2 ≥ cups/wk < 3 | 3+ cups/wk | |
| UVR ≤ 237 J/m 2 | ||||||
| No. Cases/No. noncases | 405 / 50,960 | 1,179 / 119,193 | 43 / 4,443 | 96 / 7,979 | 114 / 11,654 | - |
| Age- and Sex-Adjusted HR (95% CI) | 1 (ref) | 1.23 (1.10, 1.38) | 1.20 (0.88, 1.64) | 1.49 (1.19, 1.87) | 1.21 (0.99, 1.50) | 0.56 |
| Multivariable-Adjusted HR (95% CI)* | 1 (ref) | 1.15 (1.03, 1.29) | 1.12 (0.82, 1.54) | 1.35 (1.08, 1.69) | 1.13 (0.91, 1.39) | 0.64 |
| UVR > 237 J/m 2 | ||||||
| No. Cases/No. noncases | 474 / 50,526 | 1,239 / 111,194 | 64 / 5,112 | 107 / 8,673 | 169 / 14,358 | - |
| Age- and Sex-Adjusted HR (95% CI) | 1 (ref) | 1.16 (1.41, 1.29) | 1.27 (0.98, 1.65) | 1.24 (1.00, 1.53) | 1.18 (0.99, 1.41) | 0.39 |
| Multivariable-Adjusted HR (95% CI)* | 1 (ref) | 1.09 (0.98, 1.22) | 1.20 (0.93, 1.56) | 1.14 (0.93, 1.41) | 1.11 (0.93, 1.32) | 0.46 |
| Oranges/Tangerines/Tangelos | None | 0 ≥ cups/wk < 0.25 | 0.25 ≥ cups/wk < 0.75 | 0.75 ≥ cups/wk < 2 | 2+ cups/wk | |
| UVR ≤ 237 J/m 2 | ||||||
| No. Cases/No. noncases | 204 / 22,724 | 652 / 67,807 | 373 / 38,802 | 321 / 32,866 | 287 / 32,030 | - |
| Age- and Sex-Adjusted HR (95% CI) | 1 (ref) | 1.04 (0.89, 1.21) | 1.03 (0.87, 1.22) | 1.06 (0.89, 1.26) | 1.01 (0.84, 1.20) | 0.80 |
| Multivariable-Adjusted HR (95% CI)* | 1 (ref) | 1.02 (0.87, 1.19) | 0.98 (0.83, 1.16) | 1.01 (0.84, 1.20) | 0.96 (0.80, 1.15) | 0.14 |
| UVR > 237 J/m 2 | ||||||
| No. Cases/No. noncases | 218 / 23,352 | 647 / 62,155 | 411 / 37,185 | 385 / 33,130 | 392 / 34,041 | - |
| Age- and Sex-Adjusted HR (95% CI) | 1 (ref) | 1.09 (0.93, 1.27) | 1.13 (0.96, 1.33) | 1.20 (1.01, 1.41) | 1.21 (1.03, 1.43) | 0.10 |
| Multivariable-Adjusted HR (95% CI)* | 1 (ref) | 1.06 (0.91, 1.24) | 1.07 (0.91, 1.27) | 1.13 (0.95, 1.33) | 1.15 (0.97, 1.36) | 0.06 |
Multivariable models adjusted for age (continuous), sex (male/female), cigarette smoking (never, former, current), BMI category (underweight, normal weight, overweight, obese), education (≤11 years, high school graduate, some college or other post-HS training, college graduate), average daily alcohol intake (continuous), physical activity (engaged in physical activity never or rarely, 1–3 times/month, 1–2 times/week, 3–4 times/week, or 5 or more times/week), and family history of cancer (yes/no)
UVR is based on July erythemal UVR exposure level; high versus low UVR is based on categorization as above or below median UVR in the cohort
p-value represents trend for linear relationship assessed through linear regression modeling assuming citrus consumption was equal to the median value of participants in the category
Citrus fruits calculated as the sum of reported intakes of grapefruits and oranges/tangerines/tangelos
Table 4.
Association of citrus consumption with malignant melanoma among subgroups of age in NIH-AARP Diet and Healthy Study participants
| Model Estimates | p-trend1 | |||||
|---|---|---|---|---|---|---|
| Citrus Fruits 2 | None | 0 < cups/wk < 0.25 | 0.25 ≤ cups/wk < 0.75 | 0.75 ≤ cups/wk < 2 | 2+ cups/wk | |
| <60 YEARS at BASELINE | ||||||
| No. Cases/No. noncases | 75 / 10,496 | 312 / 36,010 | 340 / 36,101 | 251 / 26,425 | 241 / 29,556 | - |
| Age- and Sex-Adjusted HR (95% CI) | 1 (ref) | 1.20 (0.93, 1.55) | 1.29 (1.00, 1.65) | 1.31 (1.01, 1.69) | 1.15 (0.89, 1.49) | 0.31 |
| Multivariable-Adjusted HR (95% CI)* | 1 (ref) | 1.17 (0.91, 1.50) | 1.17 (0.91, 1.50) | 1.17 (0.90, 1.51) | 1.03 (0.79, 1.33) | 0.54 |
| 60+ YEARS at BASELINE | ||||||
| No. Cases/No. noncases | 169 / 17,908 | 531 / 52,843 | 651 / 58,931 | 530 / 48,014 | 794 / 68,289 | - |
| Age- and Sex-Adjusted HR (95% CI) | 1 (ref) | 1.03 (0.86, 1.22) | 1.10 (0.93, 1.30) | 1.10 (0.93, 1.31) | 1.19 (1.01, 1.40) | 0.04 |
| Multivariable-Adjusted HR (95% CI)* | 1 (ref) | 1.00 (0.84, 1.18) | 1.02 (0.86, 1.21) | 1.00 (0.84, 1.19) | 1.07 (0.90, 1.26) | 0.02 |
| Grapefruit | None | 0 ≥ cups/wk < 1 | 1 ≥ cups/wk < 2 | 2 ≥ cups/wk < 3 | 3+ cups/wk | |
| <60 YEARS at BASELINE | ||||||
| No. Cases/No. noncases | 297 / 39,025 | 795 / 85,204 | 37 / 3,739 | 38 / 3,914 | 52 / 6,706 | - |
| Age- and Sex-Adjusted HR (95% CI) | 1 (ref) | 1.22 (1.07, 1.30) | 1.29 (0.91, 1.81) | 1.31 (0.94, 1.84) | 1.03 (0.77, 1.38) | 0.87 |
| Multivariable-Adjusted HR (95% CI)* | 1 (ref) | 1.15 (1.01, 1.32) | 1.22 (0.86, 1.71) | 1.20 (0.85, 1.68) | 0.97 (0.72, 1.30) | 0.66 |
| 60+ YEARS at BASELINE | ||||||
| No. Cases/No. noncases | 583 / 62,584 | 1,625 / 145,470 | 70 / 5,829 | 165 / 12,758 | 232 / 19,344 | - |
| Age- and Sex-Adjusted HR (95% CI) | 1 (ref) | 1.17 (1.06, 1.29) | 1.24 (0.96, 1.58) | 1.36 (1.14, 1.62) | 1.26 (1.08, 1.47) | 0.25 |
| Multivariable-Adjusted HR (95% CI)* | 1 (ref) | 1.11 (1.00, 1.22) | 1.15 (0.90, 1.48) | 1.24 (1.04, 1.48) | 1.17 (1.00, 1.36) | 0.26 |
| Oranges/Tangerines/Tangelos | None | 0 ≥ cups/wk < 0.25 | 0.25 ≥ cups/wk < 0.75 | 0.75 ≥ cups/wk < 2 | 2+ cups/wk | |
| <60 YEARS at BASELINE | ||||||
| No. Cases/No. noncases | 139 / 16,243 | 444 / 50,064 | 242 / 28,140 | 220 / 23,066 | 174 / 21,075 | - |
| Age- and Sex-Adjusted HR (95% CI) | 1 (ref) | 1.02 (0.85, 1.24) | 0.98 (0.80, 1.21) | 1.10 (0.89, 1.36) | 0.98 (0.79, 1.23) | 0.91 |
| Multivariable-Adjusted HR (95% CI)* | 1 (ref) | 1.00 (0.83, 1.21) | 0.93 (0.76, 1.15) | 1.04 (0.84, 1.28) | 0.93 (0.75, 1.17) | 0.49 |
| 60+ YEARS at BASELINE | ||||||
| No. Cases/No. noncases | 283 / 29,897 | 857 / 80,079 | 544 / 47,928 | 486 / 43,003 | 505 / 45,078 | - |
| Age- and Sex-Adjusted HR (95% CI) | 1 (ref) | 1.08 (0.94, 1.23) | 1.13 (0.98, 1.30) | 1.14 (0.99, 1.32) | 1.17 (1.01, 1.36) | 0.12 |
| Multivariable-Adjusted HR (95% CI)* | 1 (ref) | 1.06 (0.93, 1.21) | 1.08 (0.93, 1.25) | 1.08 (0.93, 1.25) | 1.11 (0.96, 1.29) | 0.10 |
Multivariable models adjusted for age (continuous), sex (male/female), cigarette smoking (never, former, current), BMI category (underweight, normal weight, overweight, obese), education (≤11 years, high school graduate, some college or other post-HS training, college graduate), average daily alcohol intake (continuous), physical activity (engaged in physical activity never or rarely, 1–3 times/month, 1–2 times/week, 3–4 times/week, or 5 or more times/week), family history of cancer (yes/no), and July erythemal exposure (quartiles)
p-value represents trend for linear relationship assessed through linear regression modeling assuming citrus consumption was equal to the median value of participants in the category
Citrus fruits calculated as the sum of reported intakes of grapefruits and oranges/tangerines/tangelos
Among those who were aged 60+ years at baseline, there was a significant linear trend in risk associated with total citrus fruit intake in the age- and sex-adjusted (p-trend = 0.04) and multivariable-adjusted models (p-trend = 0.02) (Table 4). Additionally, the top citrus fruit consumers (2+ cups per week) had greater malignant melanoma risk versus non-consumers in the age- and sex-adjusted analysis, but not in the multivariable analysis. No other significant associations were observed between consumption of any citrus fruits and melanoma risk either in the older or younger subgroups.
Discussion:
In this study, neither total citrus nor any particular type of citrus product was associated with melanoma risk among the entire cohort. However, among those with higher estimated UVR exposure, and among older participants (60+ years at baseline), there were significant trends toward increased melanoma risk associated with whole citrus fruit consumption. Both of these subgroup trends must be interpreted cautiously because the HRs were not statistically significant.
The relationship between citrus intake and melanoma risk in humans has previously been examined in the HPFS and NHS (8), as well as in the WHI (9). In the combined analysis of HPFS and NHS participants, total citrus, orange juice, and grapefruit were all associated with increased melanoma risk (8). In the WHI, we found that citrus juice consumption was associated with increased melanoma risk among women who reported spending at least 30 minutes outdoors per day in the summer (9). Therefore, while data gathered from health professionals in the HPFS and NHS suggested strong relationships between citrus and melanoma, these findings have not been fully replicated in subsequent studies using data from the WHI and the NIH-AARP Diet and Health Study.
The current study has important limitations that could explain why clear relationships with melanoma risk were not observed. First, our estimate of UVR exposure is based on participants’ geographic location of residence and may not capture true UVR exposure due to lifestyle characteristics and sun-related behaviors. Additionally, we did not have information regarding participants’ history of any type of skin cancer or their skin phototype, which are important risk factors for melanoma (24). Therefore, our analysis may be confounded due to our inability to control for these factors.
Importantly, if a relationship between citrus and melanoma exists, and if furocoumarin exposure mechanistically explains this relationship, investigating exposure to a limited selection of citrus products may not be the optimal approach to examine this relationship. Previous in vitro studies have shown wide variability in the phototoxic and photocarcinogenic potential of various furocoumarins (25, 26). Our previous work has shown that there is wide variability in the furocoumarin compositions of different citrus products, and that several non-citrus foods are key contributors of certain furocoumarin compounds (12). Therefore, intake of the citrus products assessed in the NIH-AARP Study FFQ may not be a strong indicator of the total phototoxic or photocarcinogenic potential of the dietary furocoumarins consumed.
Additionally, while an FFQ is a useful method for assessing habitual dietary intake (27), if citrus furocoumarins can increase melanoma risk by interacting with cellular components in skin when photoactivated, it may be necessary to examine individual episodes of furocoumarin consumption rather than habitual intake. The concentrations of furocoumarins in skin could be influenced by the dosages ingested in a single meal, and tissue concentrations will rise and fall over the hours following ingestion (28–30). Therefore, it may also be critical to examine an individual’s UVR exposure patterns during this post-ingestion period if an interaction between furocoumarins and UVR in the skin can influence skin cancer risk.
In conclusion, this study suggested a positive relationship between citrus fruit consumption and melanoma risk among older individuals and those with higher UVR exposures. However, to confirm and further clarify these findings, future studies with more detailed dietary data and more information on potential confounding risk factors will be needed. Future animal and in vitro studies may also provide critical mechanistic insights.
Acknowledgements:
We are indebted to the participants in the NIH-AARP Diet and Health Study for their outstanding cooperation. We also thank Sigurd Hermansen and Kerry Grace Morrissey from Westat for study outcomes ascertainment and management and Leslie Carroll at Information Management Services for data support and analysis.
Financial Support:
This study was supported by the University of Connecticut USDA Hatch-Multistate Competitive Capacity Grant Program (CONS01012, PI: Dr. Ock K. Chun). In addition, this research was supported [in part] by the Intramural Research Program of the NIH, National Cancer Institute. Cancer incidence data from the Atlanta metropolitan area were collected by the Georgia Center for Cancer Statistics, Department of Epidemiology, Rollins School of Public Health, Emory University, Atlanta, Georgia. Cancer incidence data from California were collected by the California Cancer Registry, California Department of Public Health’s Cancer Surveillance and Research Branch, Sacramento, California. Cancer incidence data from the Detroit metropolitan area were collected by the Michigan Cancer Surveillance Program, Community Health Administration, Lansing, Michigan. The Florida cancer incidence data used in this report were collected by the Florida Cancer Data System (Miami, Florida) under contract with the Florida Department of Health, Tallahassee, Florida. The views expressed herein are solely those of the authors and do not necessarily reflect those of the FCDC or FDOH. Cancer incidence data from Louisiana were collected by the Louisiana Tumor Registry, Louisiana State University Health Sciences Center School of Public Health, New Orleans, Louisiana. Cancer incidence data from New Jersey were collected by the New Jersey State Cancer Registry, The Rutgers Cancer Institute of New Jersey, New Brunswick, New Jersey. Cancer incidence data from North Carolina were collected by the North Carolina Central Cancer Registry, Raleigh, North Carolina. Cancer incidence data from Pennsylvania were supplied by the Division of Health Statistics and Research, Pennsylvania Department of Health, Harrisburg, Pennsylvania. The Pennsylvania Department of Health specifically disclaims responsibility for any analyses, interpretations or conclusions. Cancer incidence data from Arizona were collected by the Arizona Cancer Registry, Division of Public Health Services, Arizona Department of Health Services, Phoenix, Arizona. Cancer incidence data from Texas were collected by the Texas Cancer Registry, Cancer Epidemiology and Surveillance Branch, Texas Department of State Health Services, Austin, Texas. Cancer incidence data from Nevada were collected by the Nevada Central Cancer Registry, Division of Public and Behavioral Health, State of Nevada Department of Health and Human Services, Carson City, Nevada.
Footnotes
Disclosure Statement: Melissa M. Melough, Junichi Sakaki, Linda M. Liao, Rashmi Sinha, Eunyoung Cho, and Ock K. Chun report no conflicts of interest.
References
- 1.Clarke CA, McKinley M, Hurley S, Haile RW, Glaser SL, Keegan THM and Swetter SM: Continued Increase in Melanoma Incidence across all Socioeconomic Status Groups in California, 1998–2012. J Invest Dermatol 137: 2282–2290, 2017. [DOI] [PubMed] [Google Scholar]
- 2.Siegel RL, Miller KD and Jemal A: Cancer statistics, 2019. CA Cancer J Clin 69: 7–34, 2019. [DOI] [PubMed] [Google Scholar]
- 3.Sample A and He YY: Mechanisms and prevention of UV-induced melanoma. Photodermatol Photoimmunol Photomed: 1–12, 2017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Elwood MJ and Jopson J: Melanoma and sun exposure: An overview of published studies. Int J Cancer 73: 198–203, 1997. [DOI] [PubMed] [Google Scholar]
- 5.Rastrelli M, Tropea S, Rossi CR and Alaibac M: Melanoma: Epidemiology, Risk Factors, Pathogenesis, Diagnosis and Classification. Vivo - Int J Exp Clin Pathophysiol Drug Res 28: 1005–1012, 2014. [PubMed] [Google Scholar]
- 6.Loftfield E, Freedman ND, Graubard BI, Hollenbeck AR, Shebl FM, Mayne ST and Sinha R: Coffee drinking and cutaneous melanoma risk in the NIH-AARP diet and health study. J Natl Cancer Inst 107: 1–9, 2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.de Waure C, Quaranta G, Gualano MR, Cadeddu C, Jovic-Vranes A, Djikanovic B, La Torre G and Ricciardi W: Systematic review of studies investigating the association between dietary habits and cutaneous malignant melanoma. Public Health 129: 1099–1113, 2015. [DOI] [PubMed] [Google Scholar]
- 8.Wu S, Han J, Feskanich D, Cho E, Stampfer MJ, Willett WCand Qureshi A a.: Citrus Consumption and Risk of Cutaneous Malignant Melanoma. J Clin Oncol 33: 2500–2508, 2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Melough MM, Wu S, Li W, Eaton C, Snetselaar L, Wallace R, Qureshi AA, Cho E and Chun OK: Citrus Consumption and Risk of Cutaneous Malignant Melanoma in the Women ‘ s Health Initiative Citrus Consumption and Risk of Cutaneous Malignant Melanoma in the Women ‘ s Health Initiative. Nutr Cancer 23: 1–8, 2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Sayre RM and Dowdy JC: The increase in melanoma: Are dietary furocoumarins responsible? Med Hypotheses 70: 855–859, 2008. [DOI] [PubMed] [Google Scholar]
- 11.Melough MM and Chun OK: Dietary furocoumarins and skin cancer: A review of current biological evidence. Food Chem Toxicol 122: 163–171, 2018. [DOI] [PubMed] [Google Scholar]
- 12.Melough MM, Lee SG, Cho E, Kim K, Provatas AA, Perkins C, Park MK, Qureshi A and Chun OK: Identification and Quantitation of Furocoumarins in Popularly Consumed Foods in the U.S. Using QuEChERS Extraction Coupled with UPLC-MS/MS Analysis. J Agric Food Chem 65: 5049–5055, 2017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Melough MM, Kim K, Cho E and Chun OK: Relationship between Furocoumarin Intake and Melanoma History among US Adults in the National Health and Nutrition Examination Survey 2003–2012. Nutr Cancer: 1–9, 2019. [DOI] [PubMed] [Google Scholar]
- 14.Melough MM, Cho E and Chun OK: Furocoumarins: A review of biochemical activities, dietary sources and intake, and potential health risks. Food Chem Toxicol 113: 99–107, 2018. [DOI] [PubMed] [Google Scholar]
- 15.Scott BR, Pathak MA and Mohn GR: Molecular and genetic basis of furocoumarin reactions. Mutat Res Genet Toxicol 39: 29–74, 1976. [DOI] [PubMed] [Google Scholar]
- 16.Cao H, Hearst JE, Corash L and Wang Y: LC-MS/MS for the detection of DNA interstrand cross-links formed by 8-methoxypsoralen and UVA irradiation in human cells. Anal Chem 80: 2932–2938, 2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Zajdela F and Bisagni E: 5-Methoxypsoralen, the Melanogenic Additive in Sun-Tan Preparations, Is Tumorigenic in Mice Exposed To 365 nm U.V. Radiation. Carcinogenesis 2: 121–127, 1981. [DOI] [PubMed] [Google Scholar]
- 18.Lambertini L, Surin K, Ton T vu T, Clayton N, Dunnick JK, Kim Y, Hong HHL, Devereux TR and Sills RC: Analysis of p53 Tumor Suppressor Gene, H-ras Protooncogene and Proliferating Cell Nuclear Antigen (PCNA) in Squamous Cell Carcinomas of HRA/Skh Mice Following Exposure to 8-Methoxypsoralen (8-MOP) and UVA Radiation (PUVA Therapy). Toxicol Pathol 33: 292–299, 2005. [DOI] [PubMed] [Google Scholar]
- 19.Kornhauser A, Wamer WG and Giles AL: Psoralen Phototoxicity: Correlation with Serum and Epidermal 8-Methoxypsoralen and 5- Methoxypsoralen in the Guinea Pig. Science (80-) 217: 733–735, 1982. [DOI] [PubMed] [Google Scholar]
- 20.Forbes PD, Davies RE and Urbach F: Long-Term Toxicity of Oral 8-Methoxypsoralen Plus Ultraviolet Radiation in Mice. J Toxicol - Cutan Ocul Toxicol 9: 237–250, 1991. [Google Scholar]
- 21.Schatzkin A, Subar AF, Thompson FE, Harlan LC, Tangrea J, Hollenbeck AR, Hurwitz PE, Coyle L, Schussler N, Michaud DS, Freedman LS, Brown CC, Midthune D and Kipnis V: Design and serendipity in establishing a large cohort with wide dietary intake distributions: The National Institutes of Health-American Association of retired persons diet and health study. Am J Epidemiol 154: 1119–1125, 2001. [DOI] [PubMed] [Google Scholar]
- 22.Lin S-W, Wheeler DC, Park Y, Cahoon EK, Hollenbeck AR, Freedman DM and Abnet CC: Prospective study of ultraviolet radiation exposure and risk of cancer in the US. Int J Cancer 131: E1015–23, 2012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Garbe C and Leiter U: Melanoma epidemiology and trends. Clin Dermatol 27: 3–9, 2009. [DOI] [PubMed] [Google Scholar]
- 24.Azoury SC and Lange JR: Epidemiology, Risk Factors, Prevention, and Early Detection of Melanoma. Surg Clin North Am 94: 945–962, 2014. [DOI] [PubMed] [Google Scholar]
- 25.Lohr C, Raquet N and Schrenk D: Application of the concept of relative photomutagenic potencies to selected furocoumarins in V79 cells. Toxicol Vitr 24: 558–566, 2010. [DOI] [PubMed] [Google Scholar]
- 26.Raquet N and Schrenk D: Application of the equivalency factor concept to the phototoxicity and -genotoxicity of furocoumarin mixtures. Food Chem Toxicol 68: 257–266, 2014. [DOI] [PubMed] [Google Scholar]
- 27.Dao MC, Subar AF, Warthon-Medina M, Cade J, Burrows T, Golley RK, Forouhi NG, Pearce M and Holmes BA: Dietary assessment toolkits: an overview. Public Health Nutr 22: 404–418, 2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Melough MM, Vance TM, Lee SG, Provatas AA, Perkins C, Qureshi A, Cho E and Chun OK: Furocoumarin Kinetics in Plasma and Urine of Healthy Adults Following Consumption of Grapefruit (Citrus paradisi Macf.) and Grapefruit Juice. J Agric Food Chem 65: 3006–3012, 2017. [DOI] [PubMed] [Google Scholar]
- 29.Tegeder I, Bräutigam L, Podda M, Meier S, Kaufmann R, Geisslinger G and Grundmann-Kollmann M: Time course of 8-methoxypsoralen concentrations in skin and plasma after topical (bath and cream) and oral administration of 8-methoxypsoralen. Clin Pharmacol Ther 71: 153–161, 2002. [DOI] [PubMed] [Google Scholar]
- 30.Wulf HC and Andreasen MP: Concentration of 3H-8-methoxypsoralen and its metabolites in the rat lens and eye after a single oral administration. Investig Ophthalmol Vis Sci 22: 32–36, 1982. [PubMed] [Google Scholar]