Figures
Abstract
Background
Several epidemiological studies have determined the associations between coffee intake level and skin cancer risk; however, the results were not yet conclusive. Herein, we conducted a systematic review and meta-analysis of the cohort and case-control studies for the association between coffee intake level and malignant melanoma (MM) risk.
Methods
Studies were identified through searching the PubMed and MEDLINE databases (to November, 2015). Study-specific risk estimates were pooled under the random-effects model.
Results
Two case-control studies (846 MM patients and 843 controls) and five cohort studies (including 844,246 participants and 5,737 MM cases) were identified. For caffeinated coffee, the pooled relative risk (RR) of MM was 0.81 [95% confidential interval (95% CI) = 0.68–0.97; P-value for Q-test = 0.003; I2 = 63.5%] for those with highest versus lowest quantity of intake. In the dose-response analysis, the RR of MM was 0.955 (95% CI = 0.912–0.999) for per 1 cup/day increment of caffeinated coffee consumption and linearity dose-response association was found (P-value for nonlinearity = 0.326). Strikingly, no significant association was found between the decaffeinated coffee intake level and MM risk (pooled RR = 0.92, 95% CI = 0.81–1.05; P-value for Q-test = 0.967; I2 = 0%; highest versus lowest quantity of intake).
Citation: Liu J, Shen B, Shi M, Cai J (2016) Higher Caffeinated Coffee Intake Is Associated with Reduced Malignant Melanoma Risk: A Meta-Analysis Study. PLoS ONE 11(1): e0147056. https://doi.org/10.1371/journal.pone.0147056
Editor: Erica Villa, University of Modena & Reggio Emilia, ITALY
Received: September 23, 2015; Accepted: December 28, 2015; Published: January 27, 2016
Copyright: © 2016 Liu et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: This work was supported by the China Post-doctoral Science Foundation (NO: 2013M541563) and the peak of six personnel Foundation in Jiangsu Province. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Abbreviations: 95% CI, 95% confidential interval; MM, malignant melanoma; OR, odds ratio; RR, relative risk; UV, ultra-violet
Introduction
Malignant melanoma (MM) is a lethal type of skin cancer, and its incidence has been increasing worldwide for the past several years [1]. Many factors have been identified that may influence the susceptibility of people to MM including ultraviolet (UV)-light exposure, skin phototype, eye and hair color, tanning capacity, presence of freckles and/or precancerous lesions, larger number of common acquired nevi, presence of atypical nevi, and genetic factors [2]. Solar ultraviolet was recognized as one of the most notable environmental factors for MM risk, although many lifestyle and environmental factors influence MM risk through interaction with UV exposure. Several dietary or nutritional factors, including the amount of ingestion of fish, vegetables and fruits, were suggested to have preventative effects against melanoma and a few studies have reported protective roles of beta-carotene and vitamins A, C, D and E on melanoma. However, these identified susceptibility factors only account for a small proportion of melanoma risk [3]. Caffeine, an ingredient found at high concentrations in coffee, has been reported to inhibit UV-induced skin carcinogenesis. Oral administration of caffeine could enhance UV-induced cell apoptosis [4], promote the elimination of DNA damaged cells [4], reduce the inflammation of epidermal cells [5] and inhibit the aberrant DNA methylation of the cells [6]. Moreover, caffeine also inhibits the metastasis of melanoma tumor cells [7] and enhances the radiosensitivity of tumor cells [8]. Therefore, higher coffee intake may have preventative effects against cutaneous melanoma. To date, several epidemiological studies have determined associations between caffeinated coffee intake level and the risk for MM and found that caffeinated coffee may have protective effects against melanoma [9–11]. However, the results for other studies found no such associations [12–14]. In contrast, several epidemiological studies have also evaluated the associations between the decaffeinated coffee and melanoma risk, but none of them found a statistically significant association [9,12,13,15–18]. As the results of the epidemiological studies are inconsistent, we herein systematically reviewed the epidemiological studies that have assessed the associations between the caffeinated or decaffeinated coffee intake level and the risks for MM, and further assessed their associations utilizing meta-analysis methods.
Methods and Materials
Eligibility study identification
To identify the eligibility studies that have determined the association between the coffee intake level and the risk for MM, two authors (Liu and Shen) independently searched the PubMED and MEDLINE databases with the terms “melanoma” or”skin cancer” in combination with “coffee” or “caffeine”. The electronic search was performed in November, 2015 and the eligibility studies should be online published online in peer-reviewed journals. The titles, abstracts and the full manuscripts were independently reviewed and assessed by the two authors, and the third author (Cai) joined the assessment of the manuscript if there was any discrepancy. In addition, we further checked the references of the retrieved studies and the published reviews or comments to identify any missing study in the literature search. In brief, the working flow chart of the eligible literature identified was presented as Fig 1.
Inclusion and exclusion criteria
Epidemiological studies that were to be included in the current meta-analysis studies should meet the following criteria: 1) was a population- or hospital-based, case-control study or a prospective or retrospective cohort study; 2) provided the coffee intake level category and the corresponding relative risk (RR) or the odds ratio (OR) estimates and their confidential intervals (95% CIs) for MM; 3) not overlapped participants with other included studies. If overlapping populations were identified between studies, the most complete one was included in the meta-analysis; 4) be reported in the English language. Here, we excluded conference abstracts, reviews, ecological studies, cellular studies and animal studies. Studies that concerned the non-melanoma skin cancer were also excluded.
Quality assessment of the included studies
We assessed the quality of the evidence, for each study, on the association between caffeinated or decaffeinated coffee intake level and melanoma risk using the Newcastle-Ottawa quality scale [19]. For case-control or cohort studies, a total of nine points were designated based on the characteristics of the studies including: (1) the selection of comparison groups, total score: 4; (2) the comparability of the groups, total score: 2; (3) the quality of the measurement of exposure and outcome, total score: 3. The total score of the adapted studies ranged between 5 to 7, and we deemed those with a score of 6 or lower as poorer quality studies (S1 Table).
Data extraction
The following data was extracted from each study: the last name of the first author, the publication year, the study design (hospital or population based case-control studies or the cohort studies), the country where the study was performed, the sample size (the number of participants for cohort studies and the number of patients and controls for case-control studies), the study period for the cohort studies (the year for the participants recruitment and the last following-up year), the types of coffee (caffeinated or decaffeinated), the method for the exposure assessment, the intake category for the coffee intake level and the corresponding relative risk (RR) or odds ratio (OR) estimates with their 95% confidential intervals (95% CIs), and the adjustments in the statistical analyses. If there were subgroup studies, each subgroup was recognized as individual studies. The meta-analysis studies were performed following the MOOSE (S1 File) and PRISMA guidelines (S2 File).
Statistical analysis
The inverse variance weighting method was applied to calculate the summary estimates and the 95% CIs. The risk estimates for the highest, in comparison with the lowest, category for each study were extracted or calculated, and the squared inverse variance for the logarithm estimate was considered as the appropriate weight for each study. The ORs from case-control studies were assumed as the estimates of the RRs for prospective studies. The DerSimonian and Laird random-effects model that considers both within- and between-study heterogeneity were used in the pooling analyses [20]. The Cochran's Q-test in together with the I2 statistic were used to assess the heterogeneity between the studies. Significant heterogeneity was considered when P < 0.05 for Q-test or the I2 was larger than 25%. To identify any individual study that may significantly affect the pooled estimates, the sensitivity analyses were performed through calculating the pooled risk estimates after excluding each single study from the meta-analysis repeatedly. The publication bias for the studies was assessed with the Begg’s funnel plots and further determined with the Egger’s linear regression test [21,22]. To investigate the influence of the basic characteristics of the individual studies on the heterogeneity between the results of multiple studies, the meta-regression analyses were performed. We also applied the generalized least square estimated trend (GLST) analysis methods proposed by Orsini et al. [23] to evaluate the dose-response relationship between coffee intake and MM risk in the prospective cohort studies. The restricted cubic spline methods with three knots at 10%, 50% and 90% for the distribution of daily caffeinated coffee intake were used to evaluate the linear relationship between coffee intake and MM risk. All statistical analyses were performed with R (www.r-project.org), Stata (version 11.0, StataCorp) and Review Manager (version 5.2.10) software. Two-sided P < 0.05 was considered to be statistically significant.
Results
Literature search and study characteristics
In the literature search, we have identified seven articles, reported from two case-control studies [10,16] and five cohort studies that have determined the association between the coffee intake level and the MM risk [9,11–14]. The basic characteristics of the included studies are presented in Table 1. The two case-control studies, both of which were conducted in Italy, have recruited a total of 846 MM patients and 843 controls. Five prospective cohort reports with eight subgroups involved a total of 844,246 participants and 5,737 cases. Of them, three were performed in the United States and one in Norway (Table 1). According to the Newcastle-Ottawa quality scale guidelines, three studies were designated as being conducted with poorer quality and three studies were conducted with better quality (S1 Table).
Caffeinated coffee and MM risk
These seven articles included 10 individual subgroups have determined the association between the caffeinated coffee intake level and the risk for MM. The pooled RR was 0.81 (95% CI = 0.68–0.97) for those participants with the highest category of caffeinated coffee intake versus lowest category (Fig 2). Significant heterogeneity between the studies was noticed (Q = 24.64, df = 9, P = 0.003; I2 = 63.5%). However, no individual study significantly affected the pooled estimate as suggested by the sensitivity analyses. The funnel plot suggested there was no potential publication bias for the included studies (S1 Fig), which was validated with the Egger’s test (P = 0.897).
The square represents the study-specific RR and the size of the squares reflects the statistical weight in the meta-analysis; horizontal lines represent the 95% confidential intervals (95% CIs); the diamond indicates the overall RR with its 95% CI under the random-effects model.
The meta-regression analyses suggested that none of the study-level characteristics (quality score, study design, gender of the participants, median or average follow-up time, study region, the lower boundary level of the highest and lowest categorized coffee intake level) significantly influence the results of multiple studies (S2 Table). In the stratification studies (Table 2), we found the association between the caffeinated coffee intake level and the MM risk was more prominent in the cohort studies (pooled RR = 0.84, 95% CI = 0.71–0.99) than in the case-control studies (pooled RR = 0.76, 95% CI = 0.36–1.62). The protective effects for the caffeinated coffee against melanoma were statistically significant in women (pooled RR = 0.76, 95% CI = 0.61–0.95) but not in men (pooled RR = 1.11, 95% CI = 0.91–1.36).
To evaluate the dose-response relationship between the caffeinated coffee intake and risk for MM, five articles with eight prospective cohort studies were included in the generalized least squares trend estimation analysis [9,11–14]. We found a linear dose-response association between the caffeinated coffee intake level and MM risk (P for nonlinearity = 0.326; Fig 3). In the dose-response analysis, a 4.5% reduced MM risk (pooled RR = 0.955, 95% CI = 0.912–0.999) was noticed for per 1 cup/day increment of caffeinated coffee intake compared to non-drinkers (Fig 3).
Decaffeinated coffee and MM risk
Four studies (one case-control study and three cohort studies) with six subgroups have evaluated the association between decaffeinated coffee intake level and MM risk [9,12,13,16] (Table 2). The pooled RR was 0.92 (95% CI = 0.81–1.05) for those with highest versus lowest quantity of intake, suggested that decaffeinated coffee may not influence the MM risk (Fig 4). The Q-test (Q = 0.94, df = 5, P = 0.967) and the I2 (0%) statistic suggested that no significant heterogeneity between the studies was evident. None of the study-level characteristics significantly affected the results of the multiple studies as suggested by the meta-regression analyses (S3 Table). The sensitivity analyses also found no individual study that significantly affected the pooled estimate. No publication bias was noticed as suggested by the funnel plot (S2 Fig) and the Egger’s test (P = 0.109). We also found no significant association between the decaffeinated coffee intake level and the MM risk in any stratification studies by study design, participants’ gender, quality score or study region (Table 2).
The square represents the study-specific RR and the size of the squares reflects the statistical weight in the meta-analysis; horizontal lines represent the 95% confidential intervals (95% CIs); the diamond indicates the overall RR with its 95% CI under the random-effects model.
Discussion
Regular coffee contains numerous bioactive compounds including polyphenols, diterpenes, trigonelline and caffeine etc. Epidemiological studies have reported that coffee may have preventative effects against cardiovascular disease and cancer caused mortality [24], type 2 diabetes [25], stroke [26], prostate cancer [27] and endometrial cancer [28]. In contrast, higher coffee intake was found to be associated with increased risk for lung cancer [29], bladder cancer [30] and gastric cancer [31]. Compounds in coffee have been reported to have anticancer activities against UV-light induced skin carcinogenesis in experimental tests [32–34]. It was supposed that higher intake of coffee, which is rich in caffeine, might lead to decreased risk of MM. Here, we firstly evaluated the association between the coffee intake level and the MM risk through meta-analysis methods and found that participants with more caffeinated coffee intake may have reduced MM risk. In contrast, no significant association was noticed between decaffeinated coffee intake level and MM risk. These data further provided evidence for the chemo-preventive effects of caffeine against the MM risk.
There were several mechanisms for the cancer preventative effects of caffeine against MM. Animal studies have suggested that caffeine inhibits the UVB-induced thymine dimers formation and reduces sunburn lesions in the epidermis of experimental mice [35,36]. Topical application of caffeine enhanced the UVB-induced apoptosis, but it was inactive on non-UVB-treated normal skin in the hairless SKH-1 mice [37]. Oral administration pre-treatment of caffeine also increased the p53-positive cells, p21-positive cells and apoptotic sunburn cells in the epidermis [37]. Meanwhile, oral administration of coffee had similar stimulatory effects on UV-induced cellular apoptosis in the epidermis of SKH-1 mice [34]. In tumor cells, caffeine could lead to selective radiosensitization of p53-deficiency cells through activating the p34cdc2 kinase [38] and the radiosensitizing effects of the caffeine were also related to its inhibitory activities on the catalytic activity of both ATM and the related kinase, ATM and Rad3-related (ATR) [39]. Lu et al. found that administration of caffeine enhances the removal of DNA-damaged cells through inhibiting the ATR-mediated phosphorylation of Chk1 and prematurely increasing the number of cyclin B1-containing cells that undergo lethal mitosis [40]. Abnormal activity in the NF-kB signaling pathway is critical to the transformation of skin cells in the tumorigenesis, and caffeine is reported to have inhibitory activities on UV-mediated NF-kB activation in melanoma cells [41]. Caffeine also inhibited the lung metastasis and reduced the tumor growth of the melanoma cells [7,42]. Therefore, there it is biologically plausible that higher caffeine intake via coffee or tea may reduce the MM risk in certain populations.
Although we did not find a statistically inverse relationship between decaffeinated coffee intake level and the MM risk, we could not exclude the possibility that there were other potential anticancer ingredients in coffee besides the caffeine. The predominant chlorogenic acid in coffee, 5-O-caffeoylquinic acid and its metabolite caffeic acid have been reported to suppress the UVB-induced skin carcinogenesis by inhibiting the cyclooxygenase-2 (COX-2) expression [43] or directly inhibiting the ERK1/2 activities [32]. The cellular studies have revealed that the diterpenes cafestol and kahweol could induce the cellular apoptosis [44] and protest against the oxidative stress and DNA damage [45]. In vivo studies have found that topical administration of diterpenes inhibits the inflammation in epidermal cells of mice [5]. The nicotinic acid and nicotinamide generated from trigonelline in coffee roasting have protective roles against UVB-induced skin carcinogenesis in cellular models [46–48]. Considering the important role of caffeine in the suppression of tumorigenesis, the anticancer activities of decaffeinated coffee could be relatively lower and studies with larger sample size are warranted to reveal the anticancer activities of decaffeinated coffee.
Meta-regression studies suggested that no study-level characteristics including study design, study quality score, median or average following-up time, participant gender, or level of coffee intake significantly influence the heterogeneity between the studies. These factors may account for a small proportion of the heterogeneity between the studies. Other factors including the UV exposure level, the amount of the caffeine in the coffee, caffeine intake from other sources, the preparation method of coffee etc. may also contribute to the heterogeneity between the studies. In the stratification studies, we found the evidence for the protective effects of caffeinated coffee was more prominent in cohort studies than in case-control studies, which may be due to the relative higher quality of the cohort studies (S1 Table) and smaller sample size from case-control studies. Interestingly, we found significant protective effects against melanoma in women, but not men, regarding the use of caffeinated coffee. It has been suggested that men usually received less UV exposure than the women due to their different dressing styles, and the incidence of melanoma is higher for women. Moreover, men who have higher caffeinated coffee intake levels usually ingest higher levels of alcohol and/or with heavier smoking habits, both of which are associated with increased risk for melanoma. Thus, the protective roles of caffeinated coffee were not as apparent as in women for men.
There are several limitations for the current study. First, the ranges for coffee intake differed between the studies, which may lead to the heterogeneity of risk estimates between the groups. The dose-response analysis also found a significantly negative relationship between caffeinated coffee intake level and MM risk in which the RR was 0.96 (95% CI = 0.94–0.98) per 1 cup/day caffeinated coffee intake compared to non-drinkers. The coffee the participants consumed are typologically different both in preparation and in the coffee amount present in the cup, which may contribute to the heterogeneity between the studies. Second, all the studies were performed in the United States or European countries, so whether the associations noticed here could be externalized to other populations are still unknown. Larger prospective studies performed in other countries or populations are warranted. Third, the interactions between caffeinated coffee intake level and UV-light exposure level and their roles in the MM risk are largely unknown as caffeine reduced the UV irradiation induced DNA damage and enhanced UV induced cellular apoptosis. Fourth, in the meta-analysis of published studies, publication bias may lead to a biased false positive association as the small studies with null results may not yet published. Although we found no statistically significant publication bias here, we could not exclude the possibility that such bias existed. Last but not the least, the estimates from the individual studies was adjusted for different cofounding factors, which may also contribute to the heterogeneity between the studies and the interactions between the coffee intake level the cofounding factors in the melanoma risk were also hard to determine.
In conclusion, this meta-analysis found that caffeinated coffee intake was negatively associated with the MM risk, but not so for the intake of decaffeinated coffee. Considering the high prevalence of coffee consumption and the rising incidence rate of melanoma, our study may provide important public health implications in the prevention of MM. However, more studies, especially the intervention studies are warranted to fully elucidate the chemo-preventive effects of caffeinated coffee against MM.
Supporting Information
S1 Fig. Funnel plot of the association between higher caffeinated coffee intake and the melanoma risk compared to those of lower intake level for publication bias identification.
https://doi.org/10.1371/journal.pone.0147056.s001
(TIF)
S2 Fig. Funnel plot of the association between higher decaffeinated coffee intake and the melanoma risk compared to those of lower intake level for publication bias identification.
https://doi.org/10.1371/journal.pone.0147056.s002
(TIF)
S1 File. MOOSE Checklist for the meta-analysis study.
https://doi.org/10.1371/journal.pone.0147056.s003
(DOCX)
S2 File. PRISMA Checklist for the meta-analysis.
https://doi.org/10.1371/journal.pone.0147056.s004
(DOC)
S1 Table. The study quality for each subgroup studies according to the Newcastle-Ottawa quality scale.
https://doi.org/10.1371/journal.pone.0147056.s005
(DOCX)
S2 Table. Univariate meta-regression analysis between the logrithm relative risk of melanoma for the highest vs. lowest quantile of caffeinated coffee intake and the basic characteristics of the study.
https://doi.org/10.1371/journal.pone.0147056.s006
(DOCX)
S3 Table. Univariate meta-regression analysis between the logrithm relative risk of melanoma for the highest vs. lowest quantile of decaffeinated coffee intake and the basic characteristics of the study.
https://doi.org/10.1371/journal.pone.0147056.s007
(DOCX)
Author Contributions
Conceived and designed the experiments: JL JC. Performed the experiments: JL BS MS. Analyzed the data: JL JC. Contributed reagents/materials/analysis tools: JL BS MS. Wrote the paper: JL JC.
References
- 1. Jiang AJ, Rambhatla PV, Eide MJ (2015) Socioeconomic and lifestyle factors and melanoma: a systematic review. Br J Dermatol 172: 885–915. pmid:25354495
- 2. Bakos L, Mastroeni S, Bonamigo RR, Melchi F, Pasquini P, Fortes C (2013) A melanoma risk score in a Brazilian population. An Bras Dermatol 88: 226–232. pmid:23739694
- 3. de Waure C, Quaranta G, Gualano MR, Cadeddu C, Jovic-Vranes A, Djikanovic B, et al. (2015) Systematic review of studies investigating the association between dietary habits and cutaneous malignant melanoma. Public Health 129: 1099–1113. pmid:26212104
- 4. Conney AH, Lu YP, Lou YR, Kawasumi M, Nghiem P (2013) Mechanisms of Caffeine-Induced Inhibition of UVB Carcinogenesis. Front Oncol 3: 144. pmid:23785666
- 5. Wei WC, Lin SY, Chen YJ, Wen CC, Huang CY, Palanisamy A, et al. (2011) Topical application of marine briarane-type diterpenes effectively inhibits 12-O-tetradecanoylphorbol-13-acetate-induced inflammation and dermatitis in murine skin. J Biomed Sci 18: 94. pmid:22189182
- 6. Lee WJ, Zhu BT (2006) Inhibition of DNA methylation by caffeic acid and chlorogenic acid, two common catechol-containing coffee polyphenols. Carcinogenesis 27: 269–277. pmid:16081510
- 7. Gude RP, Menon LG, Rao SG (2001) Effect of Caffeine, a xanthine derivative, in the inhibition of experimental lung metastasis induced by B16F10 melanoma cells. J Exp Clin Cancer Res 20: 287–292. pmid:11484989
- 8. Higuchi K, Mitsuhashi N, Saitoh J, Maebayashi K, Sakurai H, Akimoto T, et al. (2000) Caffeine enhanced radiosensitivity of rat tumor cells with a mutant-type p53 by inducing apoptosis in a p53-independent manner. Cancer Lett 152: 157–162. pmid:10773407
- 9. Wu S, Han J, Song F, Cho E, Gao X, Hunter DJ, et al. (2015) Caffeine Intake, Coffee Consumption, and Risk of Cutaneous Malignant Melanoma. Epidemiology 26: 898–908. pmid:26172864
- 10. 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. pmid:23860951
- 11. Veierod 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. pmid:9178814
- 12. Wu H, Reeves KW, Qian J, Sturgeon SR (2015) Coffee, tea, and melanoma risk among postmenopausal women. Eur J Cancer Prev 24: 347–352. pmid:25325307
- 13. Loftfield E, Freedman ND, Graubard BI, Hollenbeck AR, Shebl FM, Mayne ST, et al. (2015) Coffee drinking and cutaneous melanoma risk in the NIH-AARP diet and health study. J Natl Cancer Inst 107.
- 14. Nilsson LM, Johansson I, Lenner P, Lindahl B, Van Guelpen B (2010) Consumption of filtered and boiled coffee and the risk of incident cancer: a prospective cohort study. Cancer Causes Control 21: 1533–1544. pmid:20512657
- 15. Stensvold I, Jacobsen BK (1994) Coffee and cancer: a prospective study of 43,000 Norwegian men and women. Cancer Causes Control 5: 401–408. pmid:7999961
- 16. 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. pmid:15548944
- 17. 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. pmid:22752299
- 18. Ferrucci LM, Cartmel B, Molinaro AM, Leffell DJ, Bale AE, Mayne ST (2014) Tea, coffee, and caffeine and early-onset basal cell carcinoma in a case-control study. Eur J Cancer Prev 23: 296–302. pmid:24841641
- 19. Stang A (2010) Critical evaluation of the Newcastle-Ottawa scale for the assessment of the quality of nonrandomized studies in meta-analyses. Eur J Epidemiol 25: 603–605. pmid:20652370
- 20. DerSimonian R, Laird N (1986) Meta-analysis in clinical trials. Control Clin Trials 7: 177–188. pmid:3802833
- 21. Begg CB, Mazumdar M (1994) Operating characteristics of a rank correlation test for publication bias. Biometrics 50: 1088–1101. pmid:7786990
- 22. Egger M, Davey Smith G, Schneider M, Minder C (1997) Bias in meta-analysis detected by a simple, graphical test. BMJ 315: 629–634. pmid:9310563
- 23. Orsini N B R, Greenland S (2006) Generalized least squares for trend estimation of summarized dose-response data. Stata Journal: 40–57.
- 24. Crippa A, Discacciati A, Larsson SC, Wolk A, Orsini N (2014) Coffee consumption and mortality from all causes, cardiovascular disease, and cancer: a dose-response meta-analysis. Am J Epidemiol 180: 763–775. pmid:25156996
- 25. Jiang X, Zhang D, Jiang W (2014) Coffee and caffeine intake and incidence of type 2 diabetes mellitus: a meta-analysis of prospective studies. Eur J Nutr 53: 25–38. pmid:24150256
- 26. Larsson SC (2014) Coffee, tea, and cocoa and risk of stroke. Stroke 45: 309–314. pmid:24326448
- 27. Liu H, Hu GH, Wang XC, Huang TB, Xu L, Lai P, et al. (2015) Coffee consumption and prostate cancer risk: a meta-analysis of cohort studies. Nutr Cancer 67: 392–400. pmid:25706900
- 28. Zhou Q, Luo ML, Li H, Li M, Zhou JG (2015) Coffee consumption and risk of endometrial cancer: a dose-response meta-analysis of prospective cohort studies. Sci Rep 5: 13410. pmid:26302813
- 29. Xie Y, Qin J, Nan G, Huang S, Wang Z, Su Y (2015) Coffee consumption and the risk of lung cancer: an updated meta-analysis of epidemiological studies. Eur J Clin Nutr. [Epub ahead of print]
- 30. Wu W, Tong Y, Zhao Q, Yu G, Wei X, Lu Q (2015) Coffee consumption and bladder cancer: a meta-analysis of observational studies. Sci Rep 5: 9051. pmid:25761588
- 31. Shen Z, Liu H, Cao H (2015) Coffee consumption and risk of gastric cancer: an updated meta-analysis. Clin Res Hepatol Gastroenterol 39: 245–253. pmid:25459992
- 32. Yang G, Fu Y, Malakhova M, Kurinov I, Zhu F, Yao K, et al. (2014) Caffeic acid directly targets ERK1/2 to attenuate solar UV-induced skin carcinogenesis. Cancer Prev Res (Phila) 7: 1056–1066.
- 33. Lu YP, Lou YR, Peng QY, Nghiem P, Conney AH (2011) Caffeine decreases phospho-Chk1 (Ser317) and increases mitotic cells with cyclin B1 and caspase 3 in tumors from UVB-treated mice. Cancer Prev Res (Phila) 4: 1118–1125.
- 34. Conney AH, Zhou S, Lee MJ, Xie JG, Yang CS, Lou YR, et al. (2007) Stimulatory effect of oral administration of tea, coffee or caffeine on UVB-induced apoptosis in the epidermis of SKH-1 mice. Toxicol Appl Pharmacol 224: 209–213. pmid:17188726
- 35. Zajdela F, Latarjet R (1978) Inhibition of skin carcinogenesis in vivo by caffeine and other agents. Natl Cancer Inst Monogr: 133–140.
- 36. Bowden GT, Fusenig NE (1980) Caffeine inhibition of postreplication repair of UV-damaged DNA in mouse epidermal cells. Chem Biol Interact 33: 101–113. pmid:7438289
- 37. Lu YP, Lou YR, Li XH, Xie JG, Brash D, Huang MT, et al. (2000) Stimulatory effect of oral administration of green tea or caffeine on ultraviolet light-induced increases in epidermal wild-type p53, p21(WAF1/CIP1), and apoptotic sunburn cells in SKH-1 mice. Cancer Res 60: 4785–4791. pmid:10987287
- 38. Yao SL, Akhtar AJ, McKenna KA, Bedi GC, Sidransky D, Mabry M, et al. (1996) Selective radiosensitization of p53-deficient cells by caffeine-mediated activation of p34cdc2 kinase. Nat Med 2: 1140–1143. pmid:8837615
- 39. Sarkaria JN, Busby EC, Tibbetts RS, Roos P, Taya Y, Karnitz LM, et al. (1999) Inhibition of ATM and ATR kinase activities by the radiosensitizing agent, caffeine. Cancer Res 59: 4375–4382. pmid:10485486
- 40. Lu YP, Lou YR, Peng QY, Xie JG, Nghiem P, Conney AH. (2008) Effect of caffeine on the ATR/Chk1 pathway in the epidermis of UVB-irradiated mice. Cancer Res 68: 2523–2529. pmid:18381462
- 41. Ravi D, Muniyappa H, Das KC (2008) Caffeine inhibits UV-mediated NF-kappaB activation in A2058 melanoma cells: an ATM-PKCdelta-p38 MAPK-dependent mechanism. Mol Cell Biochem 308: 193–200. pmid:17932622
- 42. Lentini A, Kleinman HK, Mattioli P, Autuori-Pezzoli V, Nicolini L, Pietrini A, et al. (1998) Inhibition of melanoma pulmonary metastasis by methylxanthines due to decreased invasion and proliferation. Melanoma Res 8: 131–137. pmid:9610865
- 43. Kang NJ, Lee KW, Shin BJ, Jung SK, Hwang MK, Bode AM, et al. (2009) Caffeic acid, a phenolic phytochemical in coffee, directly inhibits Fyn kinase activity and UVB-induced COX-2 expression. Carcinogenesis 30: 321–330. pmid:19073879
- 44. Lee KA, Chae JI, Shim JH (2012) Natural diterpenes from coffee, cafestol and kahweol induce apoptosis through regulation of specificity protein 1 expression in human malignant pleural mesothelioma. J Biomed Sci 19: 60. pmid:22734486
- 45. Lee KJ, Jeong HG (2007) Protective effects of kahweol and cafestol against hydrogen peroxide-induced oxidative stress and DNA damage. Toxicol Lett 173: 80–87. pmid:17689207
- 46. Lang R, Yagar EF, Eggers R, Hofmann T (2008) Quantitative investigation of trigonelline, nicotinic acid, and nicotinamide in foods, urine, and plasma by means of LC-MS/MS and stable isotope dilution analysis. J Agric Food Chem 56: 11114–11121. pmid:19007232
- 47. Thompson BC, Halliday GM, Damian DL (2015) Nicotinamide enhances repair of arsenic and ultraviolet radiation-induced DNA damage in HaCaT keratinocytes and ex vivo human skin. PLoS One 10: e0117491. pmid:25658450
- 48. Thompson BC, Surjana D, Halliday GM, Damian DL (2014) Nicotinamide enhances repair of ultraviolet radiation-induced DNA damage in primary melanocytes. Exp Dermatol 23: 509–511. pmid:24798949