Research paper 347

Coffee, tea, and melanoma risk among postmenopausal women Haotian Wu, Katherine W. Reeves, Jing Qian and Susan R. Sturgeon Laboratory research suggests that components in coffee and tea may have anticarcinogenic effects. Some epidemiologic studies have reported that women who consume coffee and tea have a lower risk for melanoma. We assessed coffee, tea, and melanoma risk prospectively in the Women’s Health Initiative – Observational Study cohort of 66 484 postmenopausal women, followed for an average of 7.7 years. Coffee and tea intakes were measured through self-administered questionnaires at baseline and at year 3 of follow-up. Self-reported incident melanomas were adjudicated using medical records. Cox proportional hazard models were used to estimate risk, adjusting for covariates, with person-time accumulation until melanoma diagnosis (n = 398), death, loss to follow-up, or through 2005. Daily coffee [hazard ratio (HR) = 0.87, 95% confidence interval (CI) 0.68–1.12] and tea (HR = 1.03, 95% CI 0.81–1.31) intakes were not significantly associated with melanoma risk compared with nondaily intake of each beverage. No significant trends were observed between melanoma risk and increasing intakes of coffee (P for trend = 0.38) or tea (P for trend = 0.22). Women who reported daily coffee intake at both baseline and year 3 had a significantly decreased

risk compared with women who reported nondaily intake at both time points (HR = 0.68, 95% CI 0.48–0.97). Consistent daily tea intake was not associated with decreased melanoma risk. Overall, there is no strong evidence that increasing coffee or tea consumption can lead to a lower melanoma risk. We observed a decrease in melanoma risk among long-term coffee drinkers, but the lack of consistency in the results by dose and type cautioned against overinterpretation of the results. European Journal of Cancer Prevention 24:347–352 Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

Introduction

for incident melanoma within the Women’s Health Initiative – Observational Study (WHI-OS), a large prospective cohort with a large number of confirmed melanoma cases.

Laboratory and animal studies have suggested that coffee and tea components, including caffeine, may influence melanoma development and metastasis (Huang et al., 1997; Lou et al., 1999; Katiyar et al., 2007). A few prospective studies have addressed the potential role of coffee, tea, or caffeine in melanoma risk (Stensvold and Jacobsen, 1994; Zheng et al., 1996; Veierød et al., 1997; Song et al., 2012). Two studies from Norway reported that intake of five or more cups of coffee per day was associated with a lower melanoma risk in Norwegian women, but not men (Stensvold and Jacobsen, 1994; Veierød et al., 1997). However, each study included no more than 61 incident melanoma cases among women (Stensvold and Jacobsen, 1994; Veierød et al., 1997). The only prospective epidemiologic study examining melanoma risk among tea drinkers showed no altered risk among tea drinkers compared with nondrinkers, on the basis of 76 incident cases (Zheng et al., 1996). A large prospective analysis based on the Nurses’ Health Study did not find significantly increased melanoma risk among high total caffeine, high caffeinated coffee, or high decaffeinated coffee consumers (Song et al., 2012). Given the limited and inconsistent literature, we examined the associations between coffee and tea intakes and the risk

European Journal of Cancer Prevention 2015, 24:347–352 Keywords: coffee, melanoma, tea, Women’s Health Initiative Division of Biostatistics and Epidemiology, University of Massachusetts Amherst, Amherst, Massachusetts, USA Correspondence to Susan R. Sturgeon, Division of Biostatistics and Epidemiology, Arnold House 407, University of Massachusetts Amherst, 715 North Pleasant Street, Amherst, MA 01003-9304, USA Tel: + 1 413 577 1364; fax: + 1 413 545 1645; e-mail: [email protected] Received 16 May 2014 Accepted 5 September 2014

Methods Study population

The WHI-OS is a large multicenter cohort study of 93 676 postmenopausal women aged between 50 and 79 years, who were recruited from 40 centers in the USA between 1993 and 1998 ([No authors listed], 1998). We accessed publicly available data with follow-up through to 12 September 2005. Given the striking difference in incidence rates among racial groups, we excluded nonwhite participants (n = 15 660). We also excluded women who reported previous cancer(s) (except nonmelanoma skin cancer; n = 11 183) or who had missing exposure or outcome information (n = 349). Our primary analyses included 66 484 participants, with a total of 513 650 person years (average 7.73 years). Information on age, race, education, height, weight, waist-to-hip ratio, region of residence, alcohol intake, exercise, hormone use, smoking history, and medication

0959-8278 Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

DOI: 10.1097/CEJ.0000000000000093

348

European Journal of Cancer Prevention 2015, Vol 24 No 4

use was collected at baseline. Data on alcohol intake, waist-to-hip ratio, and smoking history were updated after 3 years. Sun exposure information was only collected 4 years after entry. Participants self-reported sunscreen use, skin reaction to sun exposure, and average daily sun exposure in the summer and other seasons during their childhood, teens, and 30s. Exposure and outcome assessments

Coffee and tea intake were assessed at baseline and 3 years after entry. At baseline, participants were asked ‘Do you usually drink coffee each day?’ Participants who answered ‘yes’ indicated the number of cups of caffeinated and noncaffeinated coffee consumed per day: none, 1 cup/day, 2–3 cups/day, 4–5 cups/day, or 6 or more cups/day. Tea consumption was assessed in an identical manner. Three years after baseline, participants were asked ‘During the past three months, how often did you drink these beverages?’ and were given a list including ‘regular instant coffee’, ‘regular espresso or latte’, ‘other regular drip coffee’, ‘decaf coffee’, and ‘regular tea’. Participants were asked to indicate consumption, in cups, in categories: never or less than once per month, 1–3/month, 1/week, 2–4/week, 5–6/week, 1/day, 2–3/day, 4–5/day, or 6 + /day. For comparison with baseline data, daily intake was calculated using the midpoint value of each range, summed across all relevant fields. The response ‘never or less than once per month’ was assigned a daily value of 0, and the ‘6 + /day’ category a value of 6. Herbal and decaf teas were not included because the baseline questionnaire did not assess either. To investigate the effects of consistent intake between baseline and year 3 on melanoma risk, consumption of less than one cup/day at both baseline and year 3 was classified as ‘consistent nonintake’, higher intake at year 3 compared with baseline was classified as ‘increased intake’, lower intake at year 3 compared with baseline was classified as ‘decreased intake’, and the same level of intake between year 3 and baseline was classified as ‘consistent intake’. The ‘consistent intake’ group was further categorized as follows: ‘ < 1 cup/day’, ‘1 cup/day’, ‘2–3 cups/day’, and ‘ ≥ 4 cups/day’. Outcomes were adjudicated cases of cutaneous melanoma, identified through annual questionnaires ([No authors listed], 1998). A total of 363 cases were included after excluding 35 cases diagnosed in the first year. Statistical analysis

Cox proportional hazard models were used to estimate hazard ratios and 95% confidence intervals. Trend was tested by modeling each intake variable as an ordinal variable and testing its significance with a log-likelihood test. Potential confounders included age, education, income, height, region of residence, BMI, waist–hip ratio, alcohol intake, recreational physical activity, hormone

use, smoking history, total NSAID use, and aspirin use. Sun-related variables were collected after 4 years; thus, data were not available for individuals who died or were lost to follow-up within the first 4 years. Follow-up time was accrued from enrollment to the date of diagnosis, death, or the last date of follow-up, whichever came first. We built parsimonious models using a purposeful selection of covariates (Hosmer and Lemeshow, 1989). Covariates with P-values less than 0.25 in a univariate model were included in the initial multivariable model, and the log-likelihood test was used to remove variables with P-values greater than 0.05 from the multivariable model. To address potential confounding by sun-related variables, hazard ratio estimates were compared between models adjusted for all variables except sun-related variables and models that also included sun-related variables. This comparison was made using the full period of follow-up and again using a restricted follow-up from year 4 onward. For the subanalysis on intake consistency and risk for melanoma, follow-up started on the day of completion of the year 3 follow-up questionnaire. All statistical analyses were carried out using SAS 9.3 (SAS Institute Inc., Cary, North Carolina, USA) or Stata 12 (StataCorp, College Station, Texas, USA).

Results Table 1 shows the baseline characteristics of participants by coffee intake. Compared with women in the lowest level of daily total coffee (< 1 cup/day) intake, women in the highest level of daily total coffee intake (≥4 cups/day) were more likely to drink alcohol, take aspirin, smoke, and spend time outside during summer in their teens and 30s, but were less likely to have BMI greater than 30 kg/m2. Compared with women in the lowest category of tea intake (< 1 cup/day), women in the highest category of tea intake (≥2 cups/day) were slightly less likely to earn $100 000 or more, be alcohol drinkers, and live in the west, but were slightly more likely to develop burns when exposed to sunlight. No differences were observed for smoking and sun exposures across tea intake groups. In multivariable-adjusted analyses, women who had a total coffee intake of at least 1 cup/day had a nonsignificant 13% lower risk for melanoma compared with those who drank up to 1 cup/day of coffee (Table 2). There was no trend of decreasing risk with increasing numbers of cups of total coffee consumed per day. In a separate analysis, no change in risk was observed among those who reported 5 + cups/day of coffee (data not shown). Daily intake of at least one cup of caffeinated or decaffeinated coffee was not related to risk for melanoma. There were also no consistent trends of melanoma risk with increasing numbers of daily cups of coffee consumed for either regular or decaffeinated coffee. Melanoma risk was similar between women who drank at least 1 cup/day of regular tea compared with those who

Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

Coffee, tea, melanoma risk in postmenopausal women Wu et al. 349

Table 1

Distribution of characteristics according to daily coffee intake < 1 cup/day [n (%)]

Age 50–54 55–59 60–69 70–79 Education High school or less Some college Bachelors Graduate degree Income < 20 000 20 000–49 999 50 000–99 999 > 100 000 BMI < 25 25–29.99 30 + Waist–hip ratio < 0.75 0.75–0.85 > 0.85 Height < 158 (Q1) 158–162.1 (Q2) 162.2–166.3 (Q3) > 166.3 (Q4) Alcohol Nondrinkers < 3/week 3–7/week > 7/week Exercise/week None < 60 min 60–150 min > 150 min Region Northeast South Midwest West Hormone use Never Ever Smoke Never Past Current Any NSAID use No Yes Aspirin No Yes Sunscreen use No Yes Skin reaction No burn/tan Burn/tan Burn/tan slightly Burn/no tan Summer/day during childhood ≤ 120 min > 120 min Summer/day during teens ≤ 120 min > 120 min Summer/day during thirties ≤ 120 min > 120 min Other season/day during childhood ≤ 120 min

1 cup/day [n (%)]

2–3 cups/day [n (%)]

4 + cups/day [n (%)]

2684 3463 7369 3989

(15.33) (19.78) (42.10) (22.79)

961 1587 4073 2654

(10.36) (17.11) (43.91) (28.61)

2679 4025 10 225 5579

(11.90) (17.88) (45.43) (24.79)

2063 3026 7340 3295

(13.12) (19.24) (46.68) (20.96)

3065 6083 4493 3721

(17.65) (35.04) (25.88) (21.43)

1908 3360 2308 1630

(20.73) (36.50) (25.07) (17.71)

4631 8297 5370 4062

(20.71) (37.11) (24.02) (18.17)

3115 5719 3752 3033

(19.94) (36.62) (24.02) (19.42)

2181 6891 5124 2045

(13.43) (42.43) (31.55) (12.59)

1108 3869 2605 1007

(12.90) (45.05) (30.33) (11.72)

2756 9226 6436 2549

(13.14) (44.00) (30.70) (12.16)

1824 6438 4693 1722

(12.43) (43.86) (31.98) (11.73)

7660 (44.24) 5388 (31.12) 4267 (24.64)

4007 (43.71) 3115 (33.98) 2045 (22.31)

9447 (42.40) 7843 (35.20) 4991 (22.40)

6601 (42.40) 5558 (35.70) 3408 (21.89)

4455 (25.60) 8724 (50.13) 4224 (24.27)

2162 (23.45) 4690 (50.86) 2369 (25.69)

5561 (24.84) 11 429 (51.05) 5397 (24.11)

4292 (27.46) 7981 (51.06) 3359 (21.49)

4473 4137 4381 4377

(25.75) (23.82) (25.22) (25.20)

2654 2251 2244 2039

(28.89) (24.50) (24.42) (22.19)

5812 5578 5555 5396

(26.01) (24.97) (24.86) (24.15)

3708 3835 3967 4104

(23.75) (24.56) (25.41) (26.28)

8700 5891 1457 1437

(49.76) (33.69) (8.33) (8.22)

3522 3355 1152 1236

(38.01) (36.21) (12.43) (13.34)

6906 8252 3418 3911

(30.71) (36.70) (15.20) (17.39)

5017 5760 2336 2594

(31.94) (36.67) (14.87) (16.51)

2501 2188 3919 8823

(14.35) (12.55) (22.48) (50.62)

1356 1201 2169 4512

(14.68) (13.00) (23.48) (48.84)

3051 2746 5331 11 280

(13.62) (12.25) (23.79) (50.34)

2137 2031 3657 7840

(13.64) (12.97) (23.35) (50.05)

3968 3953 4191 5393

(22.67) (22.58) (23.94) (30.81)

2576 2414 1775 2510

(27.77) (26.03) (19.14) (27.06)

5927 5527 5005 6049

(26.33) (24.56) (22.24) (26.87)

3870 3401 4312 4141

(24.61) (21.63) (27.42) (26.34)

4499 (26.13) 12 720 (73.87)

2431 (26.61) 6704 (73.39)

6041 (27.28) 16 101 (72.72)

4553 (29.49) 10 887 (70.51)

10 540 (60.83) 6280 (36.24) 507 (2.93)

5011 (54.75) 3833 (41.88) 308 (3.37)

10 256 (46.17) 10 637 (47.88) 1321 (5.95)

6245 (40.22) 7681 (49.47) 1600 (10.31)

10 240 (58.50) 7265 (41.50)

5268 (56.80) 4007 (43.20)

12 610 (56.03) 9897 (43.97)

8678 (55.19) 7046 (44.81)

13 806 (78.87) 3699 (21.13)

7198 (77.61) 2077 (22.39)

17 373 (77.19) 5134 (22.81)

12 071 (76.77) 3653 (23.23)

7388 (47.29) 8236 (52.71)

3890 (47.04) 4380 (52.96)

9127 (45.44) 10 958 (54.56)

6644 (47.24) 7419 (52.76)

5431 3906 4226 1920

3065 2046 2192 902

(35.08) (25.23) (27.29) (12.40)

(37.36) (24.94) (26.72) (10.99)

7188 5179 5425 2184

(35.98) (25.93) (27.16) (10.93)

5128 3594 3777 1565

(36.46) (25.55) (26.86) (11.13)

4503 (28.53) 11 279 (71.47)

2456 (29.39) 5901 (70.61)

5721 (28.22) 14 554 (71.78)

4010 (28.20) 10 208 (71.80)

6502 (41.24) 9264 (58.76)

3364 (40.32) 4980 (59.68)

7852 (38.76) 12 408 (61.24)

5497 (38.72) 8700 (61.28)

10 931 (69.29) 4845 (30.71)

5707 (68.34) 2644 (31.66)

13 681 (67.44) 6604 (32.56)

9566 (67.32) 4644 (32.68)

9966 (63.42)

5203 (62.60)

12 800 (63.35)

8934 (63.22)

Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

350 European Journal of Cancer Prevention 2015, Vol 24 No 4

Table 1 (continued) < 1 cup/day [n (%)] > 120 min Other season/day during teens ≤ 120 min > 120 min Other season/day during thirties ≤ 120 min > 120 min

Table 2

1 cup/day [n (%)]

2–3 cups/day [n (%)]

4 + cups/day [n (%)]

5748 (36.58)

3109 (37.40)

7404 (36.65)

5197 (36.78)

11 053 (70.33) 4662 (29.67)

5747 (69.12) 2568 (30.88)

14 077 (69.72) 6113 (30.28)

9859 (69.73) 4280 (30.27)

13 144 (83.48) 2601 (16.52)

6882 (82.68) 1442 (17.32)

16 743 (82.85) 3465 (17.15)

11 733 (82.91) 2419 (17.09)

Adjusted hazards ratios for coffee and tea intake and risk for melanoma Multivariable adjusteda

Age adjusted

Total coffee Nondaily Daily Cups < 1/day 1 + /day 2–3/day 4 + /day P-trend Regular coffee Nondaily Daily Cups < 1/day 1/day 2–3/day 4 + /day P-trend Decaf coffee Nondaily Daily Cups < 1/day 1/day 2–3/day 4 + /day P-trend Tea Nondaily Daily Cups < 1/day 1/day 2 + /day P-trend

Sunlight adjustedb

Cases

Person years

HR (95% CI)

Cases

Person years

HR (95% CI)

Cases

Person years

HR (95% CI)

97 266

136 846 376 523

1 1.00 (0.79–1.26)

90 233

123 313 340 249

1 0.87 (0.68–1.12)

84 207

108 680 300 750

1 0.82 (0.63–1.07)

95 51 121 86

134 955 70 871 173 975 122 737

1 1.03 (0.73–1.44) 0.99 (0.76–1.30) 0.99 (0.74–1.33) 0.93

88 44 107 76

121 678 63 521 157 465 111 243

1 0.89 (0.62–1.28) 0.86 (0.64–1.15) 0.88 (0.64–1.21) 0.38

82 41 95 68

107 317 55 971 139 004 98 985

1 0.89 (0.61–1.31) 0.81 (0.60–1.10) 0.84 (0.61–1.17) 0.22

151 212

214 256 297 487

1 1.01 (0.82–1.25)

134 189

192 888 269 253

1 0.97 (0.77–1.21)

124 167

170 558 237 673

1 0.93 (0.74–1.18)

151 57 117 38

214 256 85 966 156 568 54 954

1 0.94 (0.69–1.28) 1.06 (0.83–1.35) 0.98 (0.69–1.40) 0.81

134 50 106 33

192 888 77 627 141 887 49 740

1 0.86 (0.62–1.19) 1.01 (0.78–1.32) 1.01 (0.69–1.49) 0.88

124 46 92 29

170 558 68 296 125 646 43 731

1 0.85 (0.60–1.19) 0.96 (0.73–1.26) 1.01 (0.67–1.53) 0.91

230 122

333 588 167 472

1 1.06 (0.85–1.32)

210 104

301 434 151 183

1 0.92 (0.72–1.16)

191 94

265 936 134 298

1 0.87 (0.68–1.11)

230 52 61 9

333 588 76 832 72 175 18 465

1 0.98 (0.73–1.33) 1.23 (0.92–1.63) 0.71 (0.36–1.37) 0.68

210 43 53 8

301 434 69 209 65 425 16 548

1 0.82 (0.59–1.14) 1.07 (0.79–1.45) 0.69 (0.34–1.39) 0.62

191 39 47 8

265 936 61 428 58 079 14 791

1 0.77 (0.55–1.09) 1.00 (0.73–1.38) 0.73 (0.36–1.49) 0.44

259 101

375 840 134 044

1 1.09 (0.87–1.38)

233 88

339 406 121 002

1 1.03 (0.81–1.32)

210 80

300 226 106 591

1 1.03 (0.79–1.34)

259 37 61

375 840 63 372 67 591

1 0.85 (0.60–1.20) 1.31 (0.99–173) 0.15

233 32 54

339 406 59 340 60 923

1 0.77 (0.53–1.11) 1.30 (0.97–1.75) 0.26

210 27 51

300 226 50 617 53 570

1 0.71 (0.47–1.06) 1.36 (1.00–1.84) 0.22

CI, confidence interval; HR, hazard ratio. Adjusted for age, height, waist–hip ratio, education, income, alcohol, smoking, region of residence, aspirin, and history of nonmelanoma skin cancer. b Adjusted for previous covariates plus skin reaction to sun, sunscreen use, and summer sunlight exposure in the 30s. a

drank up to 1 cup/day. There was no significant trend with increasing numbers of cups of tea. Compared with the same reference group, those who drank 1 cup/day of tea had a nonsignificant 23% decrease in risk, whereas those who drank at least 2 cups/day had a nonsignificant 30% increase in risk. We assessed sunlight-related variables as potential confounders by applying these variables to the multivariable model and comparing the hazard ratio estimates before and after sunlight adjustment. In both analyses, in which personyear accumulation started at either baseline (Table 2) or year

4 (data not shown), adjusting for sunlight-related variables did not appreciably alter the hazard ratio estimates. We also examined the risk for melanoma by consistency of intake (Table 3). Compared with women with consistent nondaily intake of coffee (< 1 cup/day at baseline and year 3), those who increased intake or reported consistent intake had decreased risks of melanoma, whereas those who decreased intake reported a similar risk. We further examined risk according to the intensity of intake among consistent users and found that all intensities of intake were associated with decreased risk,

Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

Coffee, tea, melanoma risk in postmenopausal women Wu et al. 351

Table 3 Adjusted hazards ratios for intake consistency and risk for melanoma Multivariable adjusteda

Total coffee Consistent abstainers Increased intake Decreased intake Consistent drinkers Cups 1/day 2–3/day ≥ 4/day P-trend Regular coffee Consistent abstainers Increased intake Decreased intake Consistent drinkers Cups 1/day 2–3/day ≥ 4/day P-trend Decaf coffee Consistent abstainers Increased intake Decreased intake Consistent drinkers Cups 1/day 2–3/day ≥ 4/day P-trend Tea Consistent abstainers Increased intake Decreased intake Consistent drinkers Cups 1/day ≥ 2/day P-trend

Cases

Person years

HR (95% CI)

58 17 66 70

59 290 25 776 67 207 95 283

1 0.63 (0.36–1.08) 0.92 (0.65–1.32) 0.68 (0.48–0.97)

6 43 21

14 480 52 450 28 354

0.39 (0.17–0.92) 0.73 (0.48–1.10) 0.72 (0.43–1.20) 0.15

88 16 45 67

98 023 25 774 56 564 73 312

1 0.69 (0.40–1.17) 0.85 (0.59–1.22) 0.95 (0.69–1.31)

11 46 10

15 432 44 368 13 513

0.72 (0.38–1.35) 1.05 (0.73–1.52) 0.87 (0.44–1.69) 0.92

130 12 48 15

154 292 18 077 46 435 26 747

1 0.76 (0.42–1.38) 1.18 (0.85–1.65) 0.62 (0.36–1.05)

4 10 1

8266 14 894 3587

0.55 (0.20–1.49) 0.74 (0.39–1.41) – 0.11

155 6 21 33

178 529 122 287 38 975 23 496

1 0.56 (0.25–1.28) 0.62 (0.39–0.98) 1.54 (1.05–2.24)

10 23

7778 15 718

1.30 (0.68–2.47) 1.65 (1.06–2.56) 0.02

CI, confidence interval; HR, hazard ratio. Model adjusted for age, height, waist–hip ratio, education, income, alcohol, smoking, region of residence, aspirin, and previous history of nonmelanoma skin cancer.

a

but no significant trend was observed. Compared with women with consistent nondaily intake of tea, women who changed their habits were found to have a decreased risk, whereas those who reported consistent intake had a significantly increased risk. Risk increased with increasing numbers of cups of tea among consistent tea users; however, these estimates were based on small numbers.

Discussion In our primary analyses, we observed no statistically significant associations between self-reported intake of coffee and melanoma risk. There was some evidence to suggest that long-term consistent intake of total coffee may reduce the risk for melanoma, as those who consistently drank at least 1 cup/day of coffee showed a significant decrease in risk compared with those who consistently abstained from coffee. However, there was no dose–response relationship observed between the

frequency of coffee intake and melanoma risk among those with consistent intake of coffee. There are notable differences between the results of our study and the results of prior studies. Stensvold and Jacobsen (1994) and Veierød et al. (1997) observed that high coffee intake significantly reduced melanoma risk in Norwegian women by 60% or higher. Regional differences in the amount of coffee consumed in Norway and in the USA may have contributed to the disparity. Chemical composition of coffee varies based on the choice of beans and preparation methods (Mills et al., 2013). Case–control studies have also reported an inverse association between coffee consumption and melanoma risk, although results were not always statistically significant (Østerlind, 1990; Naldi et al., 2004; Fortes et al., 2008, 2013). Case–control studies are prone to recall bias, in which cases report their exposure to coffee and tea differently in comparison with controls. In addition, all these studies were conducted in different populations, which may not be comparable with our population (Østerlind, 1990; Naldi et al., 2004; Fortes et al., 2008, 2013). Song et al. (2012) reported that higher quintiles of caffeine intake are nonsignificantly associated with an increased risk for melanoma (P for trend =0.09), whereas neither caffeinated nor decaffeinated coffee intake was associated with melanoma risk. In our analyses, we observed nonsignificant inverse associations of both caffeinated and decaffeinated coffee with melanoma risk. Thus, our results do not support a role of caffeine in reducing melanoma risk. However, we did not assess the amount of caffeine consumed, thus precluding a direct comparison with the results from the Nurses’ Health Study. A prospective study of postmenopausal women in Iowa by Zheng et al. (1996) reported no change in melanoma risk with varying levels of tea consumption. The small number or cases in this cohort conferred low power to detect a potential association. Findings on tea consumption from case–control studies have been mixed (Østerlind, 1990; Naldi et al., 2004; Fortes et al., 2008, 2013). Our findings on consistent tea drinkers are suggestive of a positive relationship between tea intake and melanoma risk, but our main analyses showed no association between tea consumption and melanoma risk and the hypothesized biologic mechanism does not support a positive association between tea intake and melanoma risk. There are several limitations to our study. Coffee and tea intake data were not assessed identically at baseline and year 3. Our lowest exposure group was defined as less than 1 cup/day, which included minimally exposed women and biased our results toward the null. Many potential covariates were adjusted for, except for information such as family history and history of melanocytic

Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

352 European Journal of Cancer Prevention 2015, Vol 24 No 4

nevi. Sun exposure data were available, but they were collected at year 4 of follow-up and may have been biased by recall in melanoma patients diagnosed before year 4. However, a previous study within the WHI-OS reported no significant differences in self-reported sun exposures between women who were diagnosed with melanoma before and those diagnosed after sun data collection (Gamba et al., 2013). Furthermore, adjustment for the sun exposure variables at either baseline or after time of data collection had negligible, if any, effect on hazard ratio estimates. Finally, as we were only able to assess postmenopausal women, studying the effect of coffee intake at different time points in life may be important. In summary, we did not find evidence that coffee or tea consumption substantially influences melanoma risk in a large well-conducted prospective study of US women.

Acknowledgements This manuscript was prepared using the Women’s Health Initiative (WHI) – Observational Study (OS) research material obtained from the Biologic Specimen and Data Repository Information Coordinating Center of the National Heart, Lung, and Blood Institute (NHLBI). Conflicts of interest

There are no conflicts of interest.

References [No authors listed] (1998). Design of the Women’s Health Initiative clinical trial and observational study. The Women’s Health Initiative Study Group. Control Clin Trials 19:61–109.

Fortes C, Mastroeni S, Melchi F, Pilla MA, Antonelli G, Camaioni D, et al. (2008). A protective effect of the Mediterranean diet for cutaneous melanoma. Int J Epidemiol 37:1018–1029. Fortes C, Mastroeni S, Boffetta P, Antonelli G, Pilla MA, Bottà G, et al. (2013). The protective effect of coffee consumption on cutaneous melanoma risk and the role of GSTM1 and GSTT1 polymorphisms. Cancer Causes Control 24:1779–1787. Gamba CA, Swetter SM, Stefanick ML, Kubo J, Desai M, Spaunhurst KM, et al. (2013). Aspirin is associated with lower melanoma risk among postmenopausal caucasian women: The women’s health initiative. Cancer 119:1562–1569. Hosmer DW, Lemeshow S (1989). Applied logistic regression. New York: Wiley. Huang MT, Xie JG, Wang ZY, Ho CT, Lou YR, Wang CX, et al. (1997). Effects of tea, decaffeinated tea, and caffeine on UVB light-induced complete carcinogenesis in SKH-1 mice: demonstration of caffeine as a biologically important constituent of tea. Cancer Res 57:2623–2629. Katiyar S, Elmets CA, Katiyar SK (2007). Green tea and skin cancer: photoimmunology, angiogenesis and DNA repair. J Nutr Biochem 18:287–296. Lou YR, Lu YP, Xie JG, Huang MT, Conney AH (1999). Effects of oral administration of tea, decaffeinated tea, and caffeine on the formation and growth of tumors in high-risk SKH-1 mice previously treated with ultraviolet B light. Nutr Cancer 33:146–153. Mills CE, Oruna-Concha MJ, Mottram DS, Gibson GR, Spencer JP (2013). The effect of processing on chlorogenic acid content of commercially available coffee. Food Chem 141:3335–3340. Naldi L, Gallus S, Tavani A, Imberti GL, La Vecchia C. Oncology Study Group of the Italian Group for Epidemiologic Research in Dermatology (2004). Risk of melanoma and vitamin A, coffee and alcohol: a case–control study from Italy. Eur J Cancer Prev 13:503–508. Østerlind A (1990). Malignant melanoma in Denmark: occurence and risk factors. Glostrup: [Drucker]: Larsen. Song F, Qureshi AA, Han J (2012). Increased caffeine intake is associated with reduced risk of basal cell carcinoma of the skin. Cancer Res 72:3282–3289. Stensvold I, Jacobsen BK (1994). Coffee and cancer: a prospective study of 43 000 Norwegian men and women. Cancer Causes Control 5:401–408. Veierød MB, Thelle DS, Laake P (1997). Diet and risk of cutaneous malignant melanoma: a prospective study of 50 757 Norwegian men and women. Int J Cancer 71:600–604. Zheng W, Doyle TJ, Kushi LH, Sellers TA, Hong CP, Folsom AR (1996). Tea consumption and cancer incidence in a prospective cohort study of postmenopausal women. Am J Epidemiol 144:175–182.

Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

Coffee, tea, and melanoma risk among postmenopausal women.

Laboratory research suggests that components in coffee and tea may have anticarcinogenic effects. Some epidemiologic studies have reported that women ...
78KB Sizes 2 Downloads 8 Views