Alcohol Consumption and Risk of Cancer: a Systematic Literature Review

Cancer is a chronic disease that has become, worldwide, an important public health problem. By the year 2030 20.3 million incident cases and 13.2 million deaths from the disease are expected (Ferlay et al., 2010). Currently in Brazil, cancer ranks high as a cause of death, being second only to diseases of the cardiovascular system (Datasus, 2013). Most risk factors for developing cancer are considered modifiable, highlighting a sedentary lifestyle, unhealthy diet, smoking and alcoholism. When the individual is exposed to more than one of these factors, the risk of developing the disease increases. Thus, primary prevention measures are needed, focusing mainly on modifiable risk factors (World Cancer Research Foundation and American Institute for Cancer Research, 2007; Bray et al., 2012). Among the risk factors for cancer that can be modified, tobacco and alcohol use are noteworthy. Alcohol is one of the few psychoactive drugs that are encouraged and accepted by society. Its consumption is increasing worldwide, especially in developing countries. According to Cancer Research UK (2012), there is evidence that, when compared to individuals who do not consume alcohol and do not use tobacco, those who drink and smoke have 50 times more chance of developing some form of cancer. In the UK, alcohol is responsible, every year, for 4% of cancers. Deaths that are related to the consumption of alcoholic beverages amount to 1,804,000 per year or 3.2% of all deaths in the world. In addition, when ingested in excess, alcohol may also be responsible


Introduction
Cancer is a chronic disease that has become, worldwide, an important public health problem. By the year 2030 20.3 million incident cases and 13.2 million deaths from the disease are expected (Ferlay et al., 2010). Currently in Brazil, cancer ranks high as a cause of death, being second only to diseases of the cardiovascular system (Datasus, 2013).
Most risk factors for developing cancer are considered modifiable, highlighting a sedentary lifestyle, unhealthy diet, smoking and alcoholism. When the individual is exposed to more than one of these factors, the risk of developing the disease increases. Thus, primary prevention measures are needed, focusing mainly on modifiable risk factors (World Cancer Research Foundation and American Institute for Cancer Research, 2007;Bray et al., 2012).
Among the risk factors for cancer that can be modified, tobacco and alcohol use are noteworthy. Alcohol is one of the few psychoactive drugs that are encouraged and accepted by society. Its consumption is increasing worldwide, especially in developing countries. According to Cancer Research UK (2012), there is evidence that, when compared to individuals who do not consume alcohol and do not use tobacco, those who drink and smoke have 50 times more chance of developing some form of cancer. In the UK, alcohol is responsible, every year, for 4% of cancers. Deaths that are related to the consumption of alcoholic beverages amount to 1,804,000 per year or 3.2% of all deaths in the world. In addition, when ingested in excess, alcohol may also be responsible of publication, in order to analyze the most recent studies, which generated a result of 32 papers. The selection of cancer types was based on the most frequent topographies in the country presented by Cancer incidence estimates in Brazil (2013), as follows: oral cavity, esophagus, stomach, colon and rectum, larynx, trachea, bronchus and lung, female breast, cervical, endometrial, ovarian cancers, prostate, bladder, central nervous system (CNS), leukemia, non-Hodgkin lymphoma and skin malignancies. Other cancers with lower incidence such as pharynx, pancreas and thyroid were later included in the selection for the important causal relationship with alcohol.
Publications were selected based on their titles and abstracts. We used the following criteria for inclusion: studies that addressed the relationship between alcohol consumption and the risk of cancer using meta-analysis as a study type, published in English. We chose to exclude publications that dealt exclusively with biochemical, molecular and genetic aspects, as well as duplicate articles.
From the reading of the 25 selected articles, we created a table with the following: title, year, author, type of cancer, the inclusion period of the study, number of articles, number of cases, risk measure -including the confidence interval (CI) -and conclusion. For the presentation of the association found in the studies, adjusted analysis was considered, and in its absence, crude analysis. An adaptation of the criteria proposed by Rosenthal (1996), was used to synthesize the magnitude of risk according to the odds ratios (OR) or relative risk (RR); results were presented graphically as: no risk (OR or RR<1): -; low risk (OR or RR=1-1.5): +; moderate risk (OR or RR=1.5-2.5): ++; high risk (OR or RR=2.5-4): +++; and very high risk (OR and RR>4): ++++.

Results
Results were presented considering the main aspects of the articles included in this review, in accordance with the location of the cancer, as follows: digestive organs (stomach, esophagus, pancreas, colon and rectum), respiratory (lung, larynx, oral cavity and pharynx), breast and female genital organs (breast, endometrium and ovary), other cancers (bladder, CNS, thyroid and prostate).

Alcohol consumption and the risk of cancer in organs of the digestive system
Three meta-analysis were found ( Table 1) that addressed the ingestion of alcoholic beverages and the risk of esophageal cancer Lubin et al., 2012). A meta-analysis (Lubin et al., 2012) evaluated the association with esophageal cancer, stratified into the following types: adenocarcinoma of the esophagus, gastroesophageal junction adenocarcinoma and squamous cell carcinoma of the esophagus. After checking the effect of tobacco, increased risk with doseresponse for squamous cells carcinoma of the esophagus from the consumption of 3 doses/day (RR 2.15, 95%CI 1.3-3.6) was found. A study including 20 reports of cancer of the esophagus and gastric cardia, after control of confounding by tobacco use, found no significant association (RR 0.94, 95%CI 0.83-1.06) . Assessing the risk of squamous cell cancer of the esophagus due to the light consumption of alcoholic beverages (up to 1dose/day), without adjusting for tobacco, according to the gender and geographical area, an increased risk among men (RR 1.46, 95%CI 1.19-1.80) and in the Asian population (RR 1.49, 95%CI 1.12-1.98) was observed .
Looking at alcohol consumption and the risk of gastric cancer, a meta-analysis examined the relationship between moderate and excessive intake, anatomical location and dose response from 59 studies (Table 1). In the crude analysis there was an increased risk of cancer among individuals who consumed alcohol in any amount (RR 1.07, 95%CI 1.01-1.13), with this risk becoming higher among those whose consumption was done in excess (>50g/day) (RR 1.14, 95%CI 1.08-1.21) and non-Asians who consumed ≥4 doses/day (RR 1.39, 95%CI 1.14-1.69). When adjusted for smoking, consumption of alcoholic beverages was statistically associated with the risk of gastric cancer (RR 1.12, 95%CI 1.01-1.24) (Tramacere et al., 2012).

. Search and Selection of Publications
Excluded Studies (n=18): 11 were not meta-analysis; 1 focused on the risk after quitting 1 we did not have access to the study 1 focused on another risk factor 4 made reference mainly to biochemical, genetic and molecular aspects.

Search on PubMed:
Cancer, alcohol and meta-analysis  DOI:http://dx.doi.org/10.7314/APJCP.2013.14.9.4965 Alcohol Consumption and the Risk of Cancer: a Systematic Literature Review 1.37) and women (RR 1.08, 95%CI 1.03-1.13) (Ferdiko et al., 2011). Another publication identified a statistically significant association between alcohol consumption and the risk of colon cancer (RR 1.50, 95%CI 1.25-1.79) and rectum cancer (RR 1.63, 95%CI 1.35-1.97), and when analyzing the dose-response it was observed that, for an increase of 100g of alcohol consumed per week among men, there is an 18% increase in the risk of colon cancer (RR 1.18, 95%CI 1.13-1.24) and a 19% increase in the risk of rectal cancer (RR 1.19, 95%CI 1.12-1.26). This effect was not observed among women (Moskal et al., 2007). A meta-analysis studying colorectal cancer found no statistically significant association between this cancer and a mild intake of alcoholic beverages, regardless of stratification for gender and where the study was conducted .
The association between alcohol consumption and cancer of the pancreas was studied by a meta-analysis defined as heavy alcohol intake of 3 or more doses per day (Table 1). An increased risk of 23% was found after controlling tobacco effect (RR 1.23, 95%CI 1.12-1.35) .

Alcohol consumption and the risk of cancer of the respiratory organs
Two meta-analyses were identified ( Table 2) that evaluated alcohol consumption as being a risk of lung cancer (Uehara and Kiyohara, 2010; Bagnardi et al., 2011). A meta-analysis examining individuals who had never used tobacco found association and dose-response between alcohol consumption and the risk of this cancer (RR 1.21, 95%CI 0.95-1.55) . In another study, analyzing the risk adjusted for tobacco, the statistical association between alcohol consumption and lung cancer was also not found to be significant (RR 1.00, 95%CI 0.73-1.26) (Uehara and Kiyohara, 2010).
The risk between alcohol consumption and cancer of the larynx was evaluated in two meta-analyses (Islami et al., 2010;Bagnardi et al., 2012) (Table 2). A metaanalysis that adjusted for the effects of smoking, gender and age, reported a 50% higher risk of laryngeal cancer among those who consumed 1-4 doses/day (RR 1.50, 95%CI 1.23-1.83) and 2.46 times higher among those with a consumption of more than 4 doses/day (RR 2.46, 95%CI 1.88-3.22) (Islami et al., 2010). A meta-analysis examining the risk of laryngeal cancer in relation to the consumption of alcohol (up to 1 dose/day) found no statistically significant association (PR 0.90, 95%CI 0.73-1.10), and the same occurred for geographical area and gender .
Three meta-analyses (Table 2) analyzing the relationship between alcohol consumption and the risk of cancer of the oral cavity and pharynx were identified . It was observed that, without adjustment for smoking, alcohol intake, at any intensity, increases the risk  . In another meta-analysis, when controlling for the effect of smoking, gender and age, authors found no statistically significant association between the consumption of alcoholic beverages (>1dose/day) and   . In unadjusted analyses for tobacco, a meta-analysis reported an increased risk of cancer of the oral cavity and pharynx in individuals who consume up to 1dose/day (RR 1.17, 95%CI 1.06-1.29) in men (RR 1.20, 95%CI 1.06-1.36) and in studies conducted in North America (RR 1.15, 95%CI 1.01-1.30) and Asia (RR=1.34 95% CI 1, 06-1.68) (Bagnardi et al., 2010).

Alcohol consumption and risk of breast and female genital organs cancer
Four publications were found (Table 3) evaluating the association between alcohol consumption and the risk of breast cancer (Hamajima et al., 2002;Key et al., 2006;Bagnardi et al., 2012;Seitz et al., 2012). A meta-analysis classified the articles in scores according to the degree of control for confounding. For studies judged high quality, excess risk related to alcohol drinking was 22% (RR 1.22, 95%CI 1.09-1.37). When the dose-response effect was measured an increased risk of 10% for each 10 g of ethanol/ day (RR 1.10, 95%CI 1.05-1.15) was observed. The risk remained regardless of the type of beverage consumed (beer, wine or spirits) (Key et al., 2006). Another study, after adjustment for smoking, showed a dose-response effect with a 7% increase in the risk of developing breast cancer for each 10g/day of alcohol consumed (p<0.001) (Hamajima et al., 2002). A meta-analysis that adjusted for age, family history, parity, menopause, use of oral contraceptives and hormone replacement therapy, showed an increase of 3% in the risk (RR 1.03, 95%CI 1.00-1.07) (Seitz et al., 2012). Not adjusting for the risk of tobacco consumption, a publication reported a risk increase of 5% for the consumption of 1 dose/day of alcohol (RR 1.05, 95%CI 1.02-1.08) and 6% in studies conducted in North America (RR 1.06, 95%CI 1.02-1.09%) .
The risk of endometrial cancer among drinkers was analyzed by three authors (Friberg et al., 2010;Sun et al., 2011). One meta-analysis did not reported statistically significant association between endometrial cancer and alcohol consumption in smokers (RR 0.88, 95%CI 0.64-1.21) and among nonsmokers (RR 0.86, 95%CI 0.62-1.20) (Sun et al., 2011). After adjustment for smoking, oral contraceptive use and body mass index, another meta-analysis did not observe a statistically significant association between the consumption of alcoholic beverages and the risk of endometrial cancer (RR 1.33, 95%CI 0.92-1.91) (Friberg et al., 2010). A similar result was reported by another unadjusted analysis, independent of the intensity of alcohol consumption .
We identified one meta-analysis assessing alcohol consumption and ovarian cancer, in which no statistically significant association was found (RR 1.00, 95%CI 0.95-1.05), regardless of the intensity of daily consumption .

Alcohol consumption and the risk of bladder, prostrate, central nervous system and thyroid cancer
There were two publications (Table 4) concerning the association between alcohol consumption and the risk of developing bladder cancer (Mao et al., 2010;Pelucchi et al., 2011). After adjustment for smoking, one metaanalysis found no risk among smokers (RR 1.80, 95%CI 0.54-5.99) or nonsmokers (RR 1.19 95%CI 0.85-1.53) Thyroid cancer Mack et al., 2003Mack et al., b 1980Mack et al., -1997  *RR=relative risk; CI=confidence interval; NR=not reported; a Adjusted for tobacco; b Unadjusted; c Adjusted for age, sex, race, education, histology of cancer (papillary and follicular), tobacco and cohort (Mao et al., 2010). Another meta-analysis, after adjustment for smoking, did not observe an increased risk of cancer when consumption was <3 doses/day (RR 0.98, 95%CI 0.89-1.07) and >3 doses/day (RR 0.97, 95%CI 0.72-1.31) . Prostate cancer was the subject of one meta-analysis (Table 4), which examined the risk according to different designs used in the studies. Considering all the designs, authors found a 16% increase in the risk of developing prostate cancer per dose/day consumed (RR 1.16, 95%CI 1.06-1.26). When analyzing by type of study, case-control studies remained statistically significant, with an increased risk of 24% per dose/day (95%CI 1.14-1.34). These results were not adjusted for tobacco consumption (Filmore et al., 2009). CNS tumors were evaluated by one meta-analysis (Table 4) without control for tobacco. Authors observed a statistically significant association in the use of distillates (RR 1.20, 95%CI 1.01-1.42) and in male consumers (RR 1.65, 95%CI 1.27-2.13) (Galeone et al., 2012).
Two papers (Table 4) addressing the relationship between alcohol intake and the risk of developing thyroid cancer were included in this review (Mack et al., 2003;Kitahara et al., 2012). In the pooled analysis, no association was found in the crude analysis (RR 0.80, 95%CI 0.70-0.90), regardless of the type of alcoholic beverage or the amount consumed (Mack et al., 2003). A pooled analysis adjusting for age, sex, race, education, histology of cancer (papillary or follicular), tobacco use and cohort, found a negative association between alcohol consumption and the risk of developing thyroid cancer (RR 0.88,. A reduction in risk for those who had never smoked (RR 0.81, 95%CI 0.67-0.97) and for former smokers was also observed (RR 0.89, 95%CI 0.82-0.98) (Kitahara et al., 2012).
The synthesis of the magnitude of the risk of cancer caused by alcohol consumption is shown in Table 5.

Discussion
To the best of our knowledge, this is the first systematic review of meta-analysis on the effect of alcohol consumption and the risk of developing cancer. Although several methodological limitations can be attributed to the primary studies included in the meta-analysis, this review enables the current scientific evidence about the real risks posed by the consumption of alcohol in cancer development to be summarized.
Alcohol, alone, can not be considered a causative agent of cancer. However, studies with animals show alcohol as a factor that influences the risk of developing the disease in users who consume it excessively. In addition, alcohol can act as a co-carcinogen agent, i.e. the alcoholic beverage associated with other risk factors may cause promotion or acceleration of the disease (Doll et al., 1999). For the American Cancer Society (2012) ethanol, a chemical found in alcoholic beverages, is responsible for the onset of cancer. The amount of drinking by the individual is more important than the type of beverage consumed. Acetaldehyde is the most toxic metabolite of ethanol because it has mutagenic and carcinogenic properties (Seitz and Stickel, 2007). Experiments showed an increase in the appearance of tumors when ethanol was injected on the oral or esophageal mucosa of animals (Seitz et al., 1998). Alcohol acts on the mucosa as a solvent, causing the penetration of carcinogens. Thus the molecular composition of cells is modified, which explains the increased risk of cancer in individuals who combine alcohol and tobacco. The increased risk of cancers associated with overweight and obesity may be indirectly related to alcohol, since it contributes to weight gain. Chronic alcoholism associated with nutritional deficiency increases the risk of cancer (Poschl and Seitz, 2004;Winstanley et al., 2011).
According to Bagnardi et al. (2011) the consumption of alcohol increases the risk of cancer of the oral cavity, pharynx, larynx, esophagus, stomach, colon, rectum, liver, breast and ovaries. It also highlights the existence of a minimal risk related to lung cancer and prostate cancers. The authors emphasize the need for further studies regarding these and other types of cancer, due to the large differences observed in the analyzed publications. A significant difference in risk on females compared to males is also noted. Other authors (Inoue et al., 2007) found no relationship between alcohol consumption and increased risk for the development of cancer in women. These studies also report that the risk becomes greater in individuals who consume alcohol and tobacco simultaneously (Inoue et al., 2007;Bagnardi et al., 2011). However, even though alcohol and tobacco are factors that act synergistically, ensuring the increased risk, alcohol is considered an independent risk factor for many types of cancer (Pelucchi et al., 2006).
There is a definite relationship between alcohol consumption and cancers of the mouth, pharynx, esophagus, liver, colon, rectum and breast. In relation to other types of cancer, there is a suspicion of an association, with the exception of bladder, endometrial and prostate cancer, of which there is no evidence (Boffetta and Hashibe, 2006).
In the meta-analysis that did not control the effect of tobacco on the study of the association between  alcohol consumption and the risk of developing cancer, an increased risk for developing colorectal cancer and prostate cancer, among men, and in both sexes for cancers of the oral cavity and pharynx was evidenced. For these cancers, it is necessary to perform further studies to monitor the effect of tobacco as a possible confounding variable in the observed association. As tobacco is considered an important risk factor for cancer and having as an assumption that smoking is associated with alcohol consumption, because most individuals who smoke also consume alcoholic beverages, adjustment for tobacco as a confounding factor is needed thus avoiding incorrect conclusions about the actual effects of alcohol.
The meta-analysis identified in this literature review suggests that moderate to heavy alcohol consumption after controlling for tobacco increases the risk of cancer of the esophagus, stomach, pancreas, larynx and breast. An inversely related risk in the development of thyroid cancer among ex-smokers and nonsmokers was also observed. However, it would be incorrect to state the protective effect of alcohol, and studies to verify the biological plausibility of this effect are needed.
This study found no association between alcohol consumption as a risk factor for lung, bladder, central nervous system, endometrium and ovary cancers.
This systematic review has limitations inherent to the articles included in the meta-analysis identified for this study. The inclusion of publications with significant associations and exclusion of studies with negative results may be a limitation. Furthermore, since the selection of the types of cancer to be included in this study was based on the most frequent topographies in Brazil (INCA, 2011), cancers of the liver, nasal cavity and paranasal sinuses, to which alcohol was defined as a carcinogen by the IARC (2013), were not included in this review, which may be a limitation. Another limitation is the region where the studies were conducted, which included mainly populations originating from North America, Europe and Asia. In addition, the associations may reflect cultural aspects, customs and lifestyles of each geographic region. Furthermore, no study included data from Latin America, pointing to the need to explore this topic in comprehensive regional studies.
In conclusion, after controlling for the effect of tobacco, this systematic review suggests that moderate to heavy consumption of alcohol increases the risk of developing cancers of the esophagus, stomach, larynx, pancreas and breast. In the meta-analysis that did not control the effect of smoking as a confounding variable in the association between alcohol consumption and the risk of developing cancer, there was evidenced an increased risk of prostate cancer among men, and for both sexes an increased risk of cancers of the oral cavity and pharynx, colorectal and central nervous system. It was observed that alcohol may be an inversely related risk for the development of thyroid cancer among ex-smokers and nonsmokers. Additionally, this meta-analysis review found no association between alcohol consumption and the development of lung, bladder, endometrium and ovary cancers.