Lifestyle and Cancer Prevention

The Role of Alcohol in Cancer

In 2007, after a week-long review of the evidence by experts, the WHO International Agency for Research on Cancer (IARC) concluded that cancers of the mouth, pharynx, larynx, esophagus, liver, colorectum, and breast were causally related to alcohol consumption. Bagnardi and colleagues provided the most recent relative risk (RR) estimates for the effects of alcohol consumption on these cancers, and others, based on 572 studies with over 486,000 cancer cases. In this meta-analysis, light (≤12.5 g of alcohol per day), moderate (≤50 g/day), and heavy (>50 g/day) consumption of alcohol all increased the risk of oral cavity–pharyngeal and esophageal squamous cell carcinoma (ESCC) in a dose-dependent fashion, with heavy drinkers having approximately five times the risk of these cancers compared with nondrinkers and occasional drinkers (oral cavity and pharynx: RR, 5.13 [95% CI, 4.31–6.10]; ESCC: RR, 4.95 [95% CI, 3.86–6.34]). These were by far the largest RRs detected in the study. Light drinking was not associated with cancer of the larynx, although moderate drinking increased the risk of this cancer by 44% (RR, 1.44 [95% CI, 1.26–1.66]), and heavy drinking more than doubled the risk (RR, 2.65 [95% CI, 2.19–3.19]). Female breast cancer also exhibited a dose-response relationship with alcohol, with light drinking being associated with a 4% increase in risk (RR, 1.04 [95% CI, 1.01–1.07]), moderate drinking conferring a 23% increase in risk (RR, 1.23 [95% CI, 1.19–1.28]), and heavy drinking resulting in a 61% risk increase (RR, 1.61 [95% CI, 1.33–1.94]). As with laryngeal cancer, light drinking was not associated with CRC in either men or women, but moderate and heavy drinking significantly increased the risk of this cancer for men by 21% (RR, 1.21 [95% CI, 1.11–1.32]) and 53% (RR, 1.53 [95% CI, 1.30–1.80]), respectively. Cancers for which the risk was elevated only for heavy drinkers included cancers of the gallbladder (RR, 2.64 [95% CI, 1.62–4.30]), liver (RR, 2.07 [95% CI, 1.66–2.58]), stomach (RR, 1.21 [95% CI, 1.07–1.36]), pancreas (RR1.19 [95% CI, 1.11–1.28]), and lung (RR, 1.15 [95% CI, 1.02–1.30]). The elevated risk of lung cancer may be due to residual confounding by smoking, as a previous meta-analysis did not show an association between alcohol and lung cancer in never smokers. Other findings from the 2014 meta-analysis include a possible positive association between alcohol use and melanoma and prostate cancer; an inverse linear association with lymphomas; and no associations between alcohol use and ovarian, cervical, endometrial, bladder, and brain cancers. With regard to prostate cancer, a 2016 systematic review and meta-analysis by Zhao and colleagues demonstrated a statistically significant dose-response relationship between level of alcohol consumption and increased risk of prostate cancer, with risks being of similar magnitude to the risks detected in the Bagnardi and colleagues meta-analysis. Nevertheless, the literature is mixed on this topic, and the World Cancer Research Fund (WCRF)/AICR currently states that the evidence is “limited—no conclusion” for alcoholic beverages and prostate cancer.

An important aspect of alcohol use that has been more difficult to study, and was not and could not be examined in the comprehensive meta-analysis by Bagnardi and colleagues, is pattern of use. However, in a 2015 study by Cao and colleagues, regularity of drinking and heavy episodic drinking were not associated with cancer risk, after adjustment for total alcohol intake, in the Nurses' Health Study and the Health Professionals Follow-Up Study. Analyses of beverage type have not produced consistent findings; and in general the most-consumed beverage in each study is the one that demonstrates the greatest risk. The amount of alcohol consumed appears to be more important than the type.

Further supporting the role of alcohol in cancer are studies that demonstrate the reversal of risk seen after cessation of drinking. At least three pooled analyses have suggested significant reductions in risk of head and neck, esophageal, and liver cancers after alcohol cessation, although significant timeframes were required for former drinkers' risk to equal the risk of never-drinkers. In a 2013 meta-analysis of nine case-control studies examining the effect of cessation on laryngeal and pharyngeal cancers, a risk reduction of 2% per year was seen, on average, for quitters. Although it took over 35 years for the risks of laryngeal and pharyngeal cancers among quitters to equal the risks of never-drinkers, 5 years was enough time to see a 15% reduction in the alcohol-related elevated risk of these cancers. In 2011, Rehm and colleagues examined the effect of alcohol cessation on risk of esophageal and head and neck cancers in a pooled analysis of 13 studies. This report identified significantly increased risks for both cancers in the 5- to 10-year period immediately following cessation, with risks then declining but not reaching levels of never-drinkers until more than 20 years after quitting. Another meta-analysis, by Heckley and colleagues, found a significantly increased risk of liver cancer in the 5 to 10 years after cessation, with a 6% to 7% reduction in risk per year thereafter, with an estimated 23 years required for risk equivalence between quitters and never-drinkers. The results of these studies are consistent with the long latency period of most cancers but nevertheless demonstrate the reversibility of the increased risk conferred by alcohol. Despite long timeframes required for risk among quitters to equal that of never-drinkers, these findings suggest that alcohol cessation efforts are likely to result in a public health benefit because significant and often substantial decreases were seen in the risks of these alcohol-associated cancers in the short to medium term.

Finally, some studies have documented protective effects of low-to-moderate alcohol consumption on other health outcomes, notably those related to coronary heart disease (CHD) and myocardial infarction. However, there is some evidence to suggest that the observed protective effect of alcohol on cardiovascular disease seen in many epidemiologic investigations is due to residual or unmeasured confounding and/or misclassification bias. Yet, as reported in a review of findings from the Nurses' Health Study, moderate (0.1–14.9 g/day) alcohol intake remained associated with a significantly reduced risk of CHD in that cohort even after controlling for critical lifestyle risk factors and incorporating detailed information on amount and frequency of consumption. The review also notes that the same level of consumption was associated with an increased risk of breast cancer and bone fractures among the Nurses’ Health Study women. Such a comprehensive assessment of all the potential health effects of alcohol consumption is challenging, and the balance of risks and benefits is likely to differ by population and/or other factors. At a more global level, a 2015 prospective cohort study of nearly 115,000 adults from 12 countries across the human development index concluded that current drinking was not associated with a net health benefit. However, the follow-up time of this study was less than 5 years, and more time is likely needed to observe the full range of risks and benefits associated with low-to-moderate drinking. More research is needed in this area to better understand the balance of risks and benefits of alcohol intake in particular settings.

In sum, according to the latest consensus of evidence, alcohol has been shown to convincingly increase the risk of at least six cancers: (1) oral and pharyngeal cancer; (2) laryngeal cancer; (3) ESCC; (4) CRC (in men only); (5) breast cancer; and (6) liver cancer. The risk is strongest for oral and pharyngeal cancers, laryngeal cancer, and ESCC. A strong, positive dose-response relationship is seen for all of these cancers with the exception of liver cancer, in which only heavy consumption appears to increase risk. There is evidence for probable increased risk for stomach cancer and for CRC in women. There is no safe level of alcohol use for cancer prevention, because even relatively low levels have been shown to increase risk of particular cancers, and the amount of alcohol consumed matters more than the type of beverage consumed. And as with smoking, the increased risk conferred by alcohol use is reversible, with former drinkers assuming the same level of risk as never-drinkers within 20 to 35 years of cessation.

Because alcohol use varies across the globe, the attributable fractions of cancer cases and deaths due to alcohol also vary depending on world region and country. According to recent estimates, alcohol was responsible for 5.5% of cancer cases and 5.8% of cancer deaths globally in 2012. By gender, men's alcohol-attributable fractions were twice those of women's for both overall incidence and mortality. By geographic region, the attributable fractions for incidence and mortality were highest in the Western Pacific region, where alcohol accounted for 7.1% of cancer cases and 7.8% of cancer deaths; but among men in this region, the fractions rose to 9.1% and 9.6%, respectively. This region includes China, Japan, Korea, Australia, New Zealand, and many other smaller Asian countries and islands of the South Pacific. Unsurprisingly, the lowest attributable fractions were found in the Eastern Mediterranean region, which consists largely of Muslim countries. In this region, alcohol was responsible for less than 1% of both cancer cases and deaths. Among cancer sites, ESCC had the largest number of cases (44.7%) and deaths (43.7%) attributable to alcohol, followed by oral cavity and pharynx (36.7% of cases and 34.2% of deaths due to alcohol) and larynx (26.1% of cases and 24.7% of deaths due to alcohol). Less than 10% of colorectal and breast cancer cases and deaths were due to alcohol. Among women, breast cancer was the number one cancer attributable to alcohol, whereas ESCC was number one among men.

Proposed Mechanisms Linking Alcohol With Cancer

The mechanisms behind the carcinogenic effects of alcoholic beverages are not entirely clear, although they are likely to be organ specific. Although ethanol itself is considered carcinogenic, it is its metabolite, acetaldehyde (AA), that is likely responsible for much of the carcinogenic effect of alcohol consumption, at least within the upper and lower gastrointestinal (GI) tracts, where it could come into direct contact with tissues and where its concentration would be high enough to exert an effect. Ethanol is metabolized into AA by the alcohol dehydrogenase (ADH) and cytochrome P450 2E1 enzymes; AA is subsequently degraded to acetate (noncarcinogenic) by the aldehyde dehydrogenase (ALDH) enzyme. Salivary AA increases as ethanol ingestion increases, and increased levels of AA have been documented in the saliva of those with and without cancer after moderate alcohol consumption; and experiments in rats and piglets have demonstrated mutagenic levels of colonic AA after administration of ethanol. AA has been shown to induce DNA damage through the formation of adducts and the inhibition of DNA repair, as well as being shown to alter DNA methylation. The effect of alcohol and AA has been shown to be modulated by genetic factors, mainly polymorphisms in the genes for ADH and ALDH that affect enzymatic activity and the amount of AA within saliva and other tissues, and the risk of alcohol-related cancers. As an example, the ALDH2*2 allele is prevalent among Asians, and those who are homozygous for this allele are unable to metabolize AA and therefore cannot tolerate alcohol. Those who are heterozygous for ALDH2*2 can metabolize only small amounts of AA and generate threefold higher concentrations of AA in their serum and saliva when they drink, compared with those who are homozygous for the normal allele (ALDH2*1). A number of studies carried out in the Japanese population have established the mutant ALDH2*2 allele as a strong risk factor among drinkers for cancers of the upper aerodigestive tract—most particularly, esophageal cancer. The mechanism behind the increased risk of carriers of the mutant ALDH allele may be the formation of DNA adducts, as studies in Japanese drinkers have demonstrated that those with the mutant allele harbor more AA-derived DNA adducts in their peripheral lymphocytes than do drinkers without the allele, in addition to having more sister chromatid exchanges and micronuclei.

Within the liver, it is likely that oxidative stress plays an important role in alcohol-related carcinogenesis, because hepatic levels of AA are relatively low after intake of alcohol. The CYP2E1 enzyme involved in the metabolism of ethanol to AA is capable of producing reactive oxygen species (ROS) that result in the formation of lipid peroxidation products (e.g., 4-hydroxynonenal) that are capable of forming highly mutagenic DNA adducts. One such adduct results in a mutation of TP53 that leads to cells being more resistant to apoptosis and confers a growth advantage. Similar to ADH and ALDH, CYP2E1 contains polymorphisms that have been associated with different levels of enzyme activity. However, results of individual studies examining an association of these variants with liver cancer in various populations have been inconclusive. A 2017 meta-analysis of 16 studies including 1737 liver cancer cases and 2947 controls found no association between two CYP2E1 variants and risk of liver cancer. In addition to CYP2E1-related mechanisms, chronic consumption of alcohol may also result in inflammation-driven oxidative stress, hepatic iron overload, and increased nitric oxide production, all additional sources of ROS.

Within the breast, alcohol's effect on estrogen levels is thought to be an important mechanism of carcinogenesis, in addition to the aforementioned mechanisms of alcohol metabolism, because breast tissue has the ability to metabolize ethanol at low levels. Estrogens have a proliferative effect on breast epithelial cells, and it has been well documented that prolonged exposure to estrogens increases the risk for breast cancer. And alcohol has been shown to increase estrogen and other sex hormone levels in both premenopausal and postmenopausal women. A meta-analysis of eight prospective studies involving postmenopausal women documented significant increases in all sex hormones in women who consumed 20 g of alcohol per day compared with nondrinkers. A randomized trial of 51 postmenopausal women showed that serum estrone sulfate and dehydroepiandrosterone sulfate (DHEAS) were significantly elevated by 8% and 5%, respectively, in those who consumed 15 g of alcohol per day; and in women who used hormone replacement therapy (HRT), the effect of alcohol was even more potent, resulting in a threefold increase in plasma estradiol among those who drank 15.6 g of alcohol per week. Among premenopausal women, one of the earliest studies documented a significant increase in plasma estradiol levels in healthy women after acute administration of alcohol (0.695 g/kg), whereas levels did not change significantly after ingestion of a placebo. Controlled-diet studies in this population have found similar results. Authors of some of these studies have postulated that the increase in circulating sex hormones may be due to an increase in the hepatic redox state that is associated with the catabolism of ethanol. Another possible mechanism for the alcohol-mediated increase in circulating sex hormones is the enhanced conversion of testosterone to estrogen resulting from increased aromatase activity that is seen after chronic alcohol exposure. These effects likely operate through estrogen receptor (ER)–dependent pathways. Epidemiologic studies strongly suggest that alcohol use is more strongly associated with ER+ tumors than with ER− tumors. In addition, experimental studies have shown that ethanol can stimulate proliferation of ER+ breast cancer cells but not ER− ones, and furthermore that ethanol increases transcriptional activation of the ER. However, the ER pathway alone cannot explain the effects of alcohol on ER+ tumors. Other mechanisms that are currently being investigated include alcohol's potential ability to promote the epithelial-mesenchymal transition in breast cancer cells, to activate stromal cells and influence tumor-stroma interactions, and to lead to epigenetic dysregulation by possibly interfering with folate metabolism.

Within the colorectum, changes in methylation and folate metabolism are also believed to be responsible for at least some of the carcinogenic effects of alcohol. Numerous epidemiologic studies have documented a protective effect of folate on risk of CRC and adenomatous polyps. Alcohol is a known inhibitor of folate metabolism, leading to reduced methionine synthesis and subsequently resulting in abnormal DNA methylation, which is frequently observed in colorectal neoplasia. Many studies have documented an interaction between folate and alcohol, wherein those with low folate intake and high alcohol consumption have a much greater risk of CRC than do those exposed to either risk factor alone. Heavy use of alcohol can lead to reduced intake of foods rich in folate and other B vitamins and may interfere with intestinal absorption and/or the metabolism of these key nutrients. The effect of folate and alcohol on risk of CRC may be modified by known polymorphisms in the genes involved in DNA methylation and alcohol and folate metabolism. Perhaps the most extensively studied polymorphisms are those in the MTHFR gene, which is central to the formation of methionine. One polymorphism in particular, C677T, appears to interact with levels of folate and alcohol intake to influence risk of CRC. In addition to alcohol's antifolate effects, other alcohol-related carcinogenic mechanisms are likely at play in the colorectum, including inflammation and the production of ROS and AA.

Other mechanisms that may be responsible for some of the carcinogenic effects of alcohol include the presence of other carcinogens in alcohol in addition to ethanol and AA, such as N -nitrosamines, and the ability of alcohol to enhance the absorption or activity of other carcinogens (e.g., tobacco).

Evidence-Based Interventions for Cancer Prevention Related to Alcohol Use

Published cancer prevention recommendations from the ACS and the AICR/WCRF suggest that if alcohol is consumed, it should be in moderation, which is defined as up to one drink per day for women and two per day for men. Interventions to reduce alcohol consumption are available at both the personal (or clinical) and population level. At the personal level, the US Preventive Services Task Force (USPSTF) recommends “that clinicians screen adults aged 18 years or older for alcohol misuse and provide persons engaged in risky or hazardous drinking with brief behavioral counseling interventions to reduce alcohol misuse.” This is a Grade B recommendation, meaning that there is either high certainty that the net benefit of this intervention is moderate or that there is moderate certainty that the net benefit is moderate to substantial. As of the writing of this chapter, this recommendation was in the process of being updated by the USPSTF. Screening for alcohol can be done in one of three ways: with (1) the Alcohol Use Disorders Identification Test (AUDIT); (2) the abbreviated AUDIT-Consumption (AUDIT-C) screen, or (3) the single-question screen, “How many times in the past year have you had five (for men) or four (for women and all adults older than 65 years) or more drinks in a day?” Each of these screening tools has demonstrated good sensitivity and specificity for detecting alcohol misuse across multiple populations, but the AUDIT screen has been the most widely studied in the primary care setting. After screening, brief interventions that require multiple contacts between provider and patient, with each contact lasting at least 6 to 15 minutes, demonstrate the strongest evidence for effectiveness; single-contact interventions lasting 5 minutes or less have limited effect. The USPSTF documents that these types of interventions delivered in the primary care setting can positively affect unhealthy drinking behaviors in adults, including reducing the amount of alcohol consumed weekly, reducing the proportion of persons who engage in episodes of heavy drinking, and increasing long-term adherence to drinking recommendations. It is important to note that the USPSTF recommendation distinguishes between those who “misuse” alcohol and those who “abuse” alcohol and are dependent on it. There is limited evidence that these types of brief behavioral interventions are ineffective in the setting of alcohol abuse or dependence. In such cases the benefits of specialty treatment are well established, and this type of treatment is recommended.

At the population level, a number of actions can be taken by local, state, and/or federal authorities to reduce the use of alcohol and prevent its associated harms. Many of these actions mirror actions used in tobacco control. Taxation, restriction of availability, and regulation of marketing and advertising are common, effective tools not only for tobacco control, but also for alcohol control. Taxes and pricing policies are perhaps some of the more impactful interventions that can be implemented not only to reduce drinking, but also to reduce alcohol-related mortality. A 2009 meta-analysis of 112 studies providing over 1000 estimates of the tax/price–consumption relationship found that a 10% increase in the price of alcohol resulted in a 5% reduction in drinking; and that price affected drinking of all types of beverages and across all populations of drinkers, from light to heavy users. Authors noted that effects were large compared with what is typically seen for preventive interventions. A subsequent 2010 systematic review of 50 studies containing 340 estimates by the same group documented that tax/price policies could significantly reduce alcohol-related morbidity and mortality, with a 35% reduction in alcohol-related mortality expected from a doubling of alcohol taxes. Because of their relatively low cost to implement and their documented large and significant impacts on drinking and related outcomes, tax and pricing interventions are strongly recommended by both the US Community Preventive Services Task Force ( https://www.thecommunityguide.org/topic/excessive-alcohol-consumption ) and the WHO for population-level alcohol control.

Role of Obesity in Cancer

In obesity, adipose tissue typically accumulates in two different regions: the android region or abdominal area, and the gynoid region, which consists of the hips and thighs. Android obesity differs from gynoid obesity such that it contains two different types of adipose tissue deposits, subcutaneous and visceral. The latter, visceral adipose tissue (VAT), is more metabolically active compared with subcutaneous adipose tissue, is considered the impetus for metabolic dysfunction, and may contribute to increased cancer risk in obese individuals. Therefore, excessive accumulation of VAT is considered to be a major contributor to obesity's role in cancer.

Epidemiologic evidence supports the relationship between VAT and increased cancer risk. In an observational study conducted by Britton and colleagues, adiposity measurements in four different areas (subcutaneous, visceral, periaortic, and pericardial) were compared with incident cancer, cardiovascular disease, and all-cause mortality over a median 5-year follow-up period. In this study, 3086 middle-aged men and women (average age, 50 years ± 10) were included. VAT, but not other adipose areas, was significantly associated with cancer incidence ( n = 141 incident cancer cases; hazard ratio [HR], 1.43 [95% CI, 1.12–1.84]) when the model was adjusted for clinical factors (e.g., age, sex, blood pressure, body mass index). When the sample was stratified by sex, there were no significant differences observed among adipose areas and cancer incidence; however, HRs were higher in men compared with women for VAT. These findings support the relationship between VAT and cancer incidence in middle-aged men and women.

Similarly, Murphy and colleagues compared the relationship of subcutaneous, visceral, intramuscular, and total-body adiposity with incident cancer over a 13-year follow-up period in a study that included 2519 older men and women (average age, approximately 74 years). In considering all incident cancers ( n = 617), a significant association between total adiposity, VAT, and cancer incidence was observed in women (HR per standard deviation [SD] increase, 1.14 [95% CI, 1.01–1.30]; and 1.15 [95% CI, 1.02–1.30], respectively) but not in men when clinical factors were adjusted for. Regarding obesity-related cancers ( n = 224), after adjustment for clinical factors, a significant association between total adiposity but not other adiposity measures, and obesity-related cancer incidence was observed in women (HR per SD increase, 1.23 [95% CI, 1.03–1.46]) and a significant association between VAT and obesity-related cancer incidence was observed in men (HR per SD increase, 1.30 [95% CI, 1.06–1.60]). Overall, evidence from these prospective cohort studies support the role of obesity, and specifically VAT, in cancer.

Proposed Mechanisms Linking Obesity With Cancer

Although the relationship between obesity and cancer is evident from observational data, the physiologic mechanisms explaining this relationship have recently emerged in the literature. Essentially, the excessive accumulation of adipose tissue, characterized by adipocyte hypertrophy and hyperplasia, leads to a state of adipocyte hypoxia and resultant macrophage activity, and altered adipokine profiles; these factors promote an interplay among systemic inflammation, sex hormones, and insulin resistance, all of which elicit conditions that promote tumorigenesis.

According to a review conducted by Teoh and Das, compared with lean counterparts, adipocyte samples from obese humans and animal models revealed a state of hypoxia within the adipose tissue via presence of hypoxia-inducible factor (HIF), lower levels of oxygen, and lactate. This demonstrated state of hypoxia causes macrophage recruitment via signaling pathways such as chemokine ligand 2, interleukin (IL)-1β/C-X-C motif chemokine ligand 12. Macrophages remain in the adipose tissue because of increased expression of netrin-1, a protein that has been proposed to block migration of macrophages and is expressed when fatty acid palmitate is saturated. Of note, fatty acid palmitate is saturated within the obese state secondary to the increase in lipolysis and consequent increase in circulating free fatty acids.

Next, the macrophages transition from an M2 prorepair to an M1 proinflammatory state, activating nuclear factor-κB (NF-κB) and signal transducer activator of transcription 3 (STAT3). NF-κB has been linked with reduction in cellular senescence, facilitation of the epithelial-mesenchymal transition, and an increase in proinflammatory cytokines such as tumor necrosis factor–α (TNF-α) and IL-6. STAT3 increases expression of cyclins and antiapoptotic proteins, resulting in an increase in cellular proliferation and decrease in cellular apoptosis. These cytokines play different roles depending on cancer type. For example, in breast cancer, macrophage activation of NF-κB leads to increased expression of cytokines (i.e., IL-1β, TNF-α, prostaglandin E2 [PGE ₂ ]) that increase transcription of CYP19, a transcription factor of aromatase, and result in increased bioavailability of estrogen. Overall, macrophage activity secondary to hypoxic conditions from obesity promotes tumorigenic conditions.

Along with macrophage activity, adipokine profiles, notably adiponectin and leptin, are altered in the obese state. Normal physiologic levels of adiponectin are linked with enhancement of insulin sensitivity and reduction in angiogenesis. However, in the obese state, adiponectin levels decrease owing to instability of adiponectin mRNA and subsequent reduction in gene expression. In contrast, under normal physiologic conditions, leptin promotes satiety and feeding cessation; yet in the obese state, leptin levels increase. The increase in leptin is associated with increases in cellular proliferation and inhibition of apoptosis in breast, ovarian, and cervical cancers. Furthermore, increased leptin has demonstrated an association with colon cancer in men. Currently proposed mechanisms for these associations vary depending on cancer site. For example, in ovarian, endometrial, and breast cancers, some studies have shown that cell growth is facilitated by leptin binding to ER-α. Overall, there is evidence to support obesity-induced alterations in adipokines as an enabler of tumorigenesis.

Considering that adiponectin is associated with insulin sensitivity, in the obese state a reduction in adiponectin also contributes to a reduction in insulin sensitivity and subsequent insulin resistance. Insulin resistance facilitates the tumor microenvironment via two interrelated mechanisms. First, increased insulin levels result in a reduction of insulin-like growth factor–binding proteins 1 and 2 (IGFBP1, IGFBP2) and subsequent increase in insulin-like growth factor 1 (IGF1), ultimately resulting in increased cellular proliferation and decreased apoptosis via the AKT/mTOR pathway. Second, increased insulin, IGF1, and TNF-α reduce the amount of sex hormone–binding globulin produced, which increases cancer risk because of the resultant increase in levels of sex hormones. Although the mechanisms related to estrogen and cancer risk vary depending on cancer type, according to Teoh and Das, estrogen may cause replication errors during DNA synthesis and increase cellular proliferation; hence the number of mutations accumulates over time, yielding tumor growth. In addition to insulin serving as a contributing factor for increased estrogen and protumorigenic effects, it is important to note that adipose tissue is a major site for estrogen synthesis in postmenopausal women and men; thus, increased adipose tissue leads to increased estrogen production, and consequently a greater ratio of circulating estrogens to androgens may lead to cancer. Taken together, it is evident that excessive accumulation of adipose tissue induces protumorigenic conditions.

Role of Physical Inactivity in Cancer

Current public health guidelines recommend 150 minutes per week of moderate-intensity or 75 minutes per week of vigorous-intensity aerobic activity, and at least 2 days per week of strength training exercise. Compliance with these guidelines may be measured subjectively, through self-report, or objectively, through measurement or estimation of peak oxygen consumption via a maximal or submaximal cardiopulmonary exercise test respectively. Overall, both subjective and objective evidence is available linking regular, structured physical activity with reduced risk of chronic disease and certain cancers such as colon and breast cancer. An inverse relationship has been demonstrated between increased physical activity levels and reduction in cancer risk. Moreover, a retrospective analysis conducted by Lee and colleagues found that individuals who did not engage in regular, structured activity were 1.32 times more likely to develop colon cancer and 1.33 times more likely to develop breast cancer than physically active individuals.

In addition to structured physical activity, when lifestyle activity decreases, it is typically replaced by increased sedentary behavior. A meta-analysis by Schmid and Leitzmann noted that colon cancer risk increased by 8% and endometrial cancer risk by 10% for every 2-hour increase in total sitting time per day. In addition, in a prospective cohort study conducted by Nomura and colleagues, a significant association with sitting time and breast cancer incidence was observed when individuals spent more than 10 hours sitting per day. Taken together, sedentary behavior is linked with increased risk of colon, breast, and endometrial cancer.

Proposed Mechanisms Linking Physical Activity With Reduced Cancer Risk

Evidence supports the impact of physical activity on cancer risk reduction through mechanisms related to obesity, skeletal muscle cytokine production, and immune function. As mentioned previously, obesity-induced insulin resistance promotes tumorigenesis. However, physical activity may counteract this by improving insulin sensitivity through alterations in VAT deposition and training adaptations of the skeletal muscle.

Exercise without weight loss or changes in diet has demonstrated effective reduction in VAT, the primary fat depot connected to increased metabolic dysfunction and promotion of tumorigenesis. The proposed mechanism for VAT reduction with exercise alone is attributed to an increase in β-adrenergic receptor activity in adipocytes, which results in an increase in lipolysis from abdominal fat depots compared with other fat depots during exercise. Moreover, the reduction in VAT results in an increase in adiponectin. As mentioned previously, adiponectin is linked with enhancement of insulin sensitivity and reduction of angiogenesis.

In addition, in the obese state, impaired glucose tolerance at the level of the skeletal muscle exacerbates insulin resistance secondary to the increase in circulating free fatty acids and consequent increase in diacylglycerol instead of glucose for substrate utilization. However, exercise-induced skeletal muscle contraction prompts increased translocation of GLUT4 receptors to the T tubules and plasma membrane of the working muscles, which improves skeletal muscle glucose uptake and whole-body insulin sensitivity. Therefore, regular physical activity, even in the obese state, improves insulin sensitivity and prevents increases in bioavailability of sex hormones, two factors that help inhibit tumorigenesis.

Along with obesity-related mechanisms, there are also mechanisms unique to skeletal muscle activity—notably, release of cytokines from contracting skeletal muscle, commonly referred to as myokines. To date, identified myokines include IL-6, irisin, IL-15, calprotectin, myonectin, oncastatin M (OSM), and secreted protein acidic and rich in cysteine (SPARC). According to a review by Goh and colleagues, during exercise myokines are released from the working skeletal muscles, which increases activity of antiinflammatory cytokines (i.e., IL-1 receptor agonist and IL-10), decreases activity of proinflammatory cytokines, and attenuates migration of M1 macrophages and T cells to areas with high expression of inflammatory cytokines or inflammation (i.e., adipose tissue and/or the tumor microenvironment).

Currently, promising evidence is available regarding the roles of exercise-induced OSM and irisin activity in breast cancer prevention, and SPARC in colon cancer prevention. Within the context of breast cancer prevention, Hojman and colleagues treated MCF-7 cells (malignant, ER+, breast epithelial cells) with mouse serum collected immediately after exercise and demonstrated a 52% reduction in cell proliferation and 54% increase in caspase activity. Furthermore, whereas various myokines were upregulated after exercise, OSM demonstrated the greatest effects on cellular proliferation, caspase activity, and apoptosis. The efficacy of OSM was further supported in a model in which OSM signaling was blocked.

Similarly, Gannon and colleagues assessed the influence of irisin activity specifically on malignant breast epithelial cells. Overall, irisin reduced cellular proliferation in aggressive breast epithelial cells through caspase activity and suppression of NF-κB. Taken together, current evidence supports OSM and irisin as two contributing myokines connected to breast cancer prevention through their impact on caspase activity.

Colon cancer is also linked with myokine activity. Specifically, the myokine SPARC has been linked with reduced risk of colon cancer. Aoi and colleagues assessed SPARC in relation to exercise and colon cancer among a colon-26 carcinoma cell model, two different animal models, and a human subject exercise intervention. In humans, SPARC levels released from working muscles during exercise increased linearly with training, yet resting levels remained unchanged, supporting the role of SPARC as a myokine. In the first animal model, exercised mice with azoxymethane (AOM)-induced colon tumorigenesis demonstrated a reduction in aberrant crypt foci and aberrant crypts compared with AOM nonexercised and control groups. These findings were also observed in the cell line model and lend support to the role of exercise in the reduction of colon tumorigenesis. In addition, the second animal model compared sedentary and exercised SPARC knockout mice to exercised wild-type mice and demonstrated an increase in apoptotic cells via caspase-3 and caspase-8 activity in exercised wild-type mice, with no change in apoptosis observed in either group of the SPARC knockout mice, supporting SPARC's role in attenuation of tumorigenesis. Similarly, in a review conducted by Sanchis-Gomar and colleagues, evidence supports SPARC activation of caspase-3 and caspase-8 to promote cellular apoptosis and prevent aberrant crypt foci formation in colon cancer. Overall, evidence supports SPARC's role as a myokine and as a contributing factor in reducing colon cancer.

Interconnected with myokine activity is the impact of exercise-induced activation of natural killer (NK) cells in relation to physical activity and cancer prevention. In a review by Idorn and colleagues a reduction in tumor incidence of greater than 60% was observed due to exercise-induced activation of NK cells among various mouse models. In humans, NK cells seem to function similarly, such that NK cells recognize and kill cancer cells. During exercise the myokine IL-15 activates NK cells and T cells via transpresentation of IL-15/IL-15 receptor alpha complex. In addition, with exercise, the natural increase in epinephrine released from β-adrenergic receptors facilitates the release and mobilization of NK cells. Moreover, NK cells, compared with other lymphocytes, have the greatest density of β-adrenergic receptors. Taken together, improvements in fat deposition and insulin sensitivity, and the activity of myokines and NK cells, are all viable proposed mechanisms explaining the role of physical activity in cancer prevention.

Dietary Components Linked With Increased Cancer Risk

Dietary Components Linked With Decreased Cancer Risk

Obesity, Physical Inactivity, and Dietary Recommendations and Resources for Cancer Prevention

Recommendations for cancer prevention are shown in Table 22.1 . As evidenced by results from the systematic review conducted by Birks and colleagues, weight loss in overweight or obese individuals is linked with reduction in cancer incidence. The underlying theme in weight loss is energy balance—moving more (increasing energy expenditure) and eating less (reducing energy intake). Even small changes can result in large effects over time.

At initiation of an exercise regimen, it is recommended to start with at least 5 days per week of moderate-intensity aerobic exercise (40%–60% heart rate reserve or V̇ o ₂ reserve) for at least 30 minutes per exercise session, with eventual progression to the goal of at least 60 minutes per day of aerobic exercise at more vigorous intensities (i.e., >60% heart rate reserve or V̇ o ₂ reserve). If an individual is unable to complete a continuous exercise session, it is recommended to break up the exercise time into 10-minute increments. Regarding diet, a caloric reduction of 500 to 1000 kcal/day is typically recommended to elicit a 1- to 2-pound weight loss per week. Dietary changes that are often suggested for weight loss include reducing portion size and avoiding high–energy-dense foods, especially those with low nutrient value (e.g., potato chips, sugar-sweetened beverages).

There are countless resources available to assist individuals with energy balance to promote weight loss and maintenance of a healthy weight. The CDC and healthfinder.gov provide a step-by-step guide to help individuals with weight loss. This guide includes strategies, such as goal setting and self-monitoring, to establish healthier dietary habits in efforts to incorporate the behavior change component into weight loss.

Moreover, the CDC website provides details on the activity guidelines, tips on how to incorporate structured physical activity within one's daily routine and increase lifestyle activity, how to begin an exercise program, how to measure exercise intensity, and other worthwhile messages. These resources, in coordination with worksite and community efforts (e.g., the local gym, YMCA, or community recreation center; religious organizations; home exercise videos; wearable devices such as an accelerometer or pedometer), aid in promoting increases in physical activity. In addition, the AICR's website is also a useful resource, providing various aids to assist in meeting cancer prevention guidelines, including a visual aid to assist with portion control and consumption of a balanced diet.

The Role of Ultraviolet Radiation in Cancer

Exposure to UVR, whether from the sun or from a tanning device, is a major risk factor for all forms of skin cancer (basal cell carcinoma [BCC], squamous cell carcinoma [SCC], and melanoma). Despite being highly preventable, skin cancer is the most common form of cancer in the United States, with 5 billion cases treated annually at a cost of $8.1 billion. Melanoma accounts for a small fraction of these cases, approximately 63,000, but it is the deadliest form of skin cancer, with nearly 9000 deaths each year. Unlike many other common cancers, data suggest that skin cancer incidence is rising, underscoring the need for a focus on evidence-based preventive actions.

The risk of skin cancer varies by skin type, race or ethnicity, the pattern of UVR exposure, and the age at which exposure occurs. In general, lighter-skinned individuals have a higher risk of all skin cancers than do darker-skinned individuals. The Fitzpatrick skin type classification system classifies individuals based on their likelihood of tanning or burning, with six categories (I–VI) ranging from “always burns and never tans” to “never burns, deeply pigmented, least sensitive.” Those who are classified as types I and II have the highest risk of burns, damage from UVR, and skin cancer.

Regarding race, non-Hispanic whites have the highest rates of invasive melanoma incidence (24.7 per 100,000) and mortality (3.4 per 100,000), whereas African-Americans have the lowest rates (1.0 per 100,000 and 0.4 per 100,000, respectively).

The type of skin cancer for which one is at risk also varies according to his or her pattern of UVR exposure. Chronic continuous exposure, such as that experienced by an outdoor worker, is more strongly associated with SCC, whereas intermittent exposure, such as that experienced by someone vacationing at the beach, is more strongly associated with BCC and melanoma. In a meta-analysis of 34 studies, intermittent sun exposure was associated with a significant 61% excess risk of melanoma, whereas chronic exposure was slightly inversely associated with risk, although this was not significant. Numerous studies have documented the increased risk of skin cancer associated with sunburns. Sunburns during childhood are strongly associated with risk of future melanoma, although data suggest that sunburns at any age can significantly increase the risk of melanoma. A meta-analysis of 26 studies found that the risk of melanoma increases with increasing number of sunburns during all life periods (i.e., childhood, adolescence, and adulthood). Consequently, prevention efforts should occur across the lifespan.

Although the sun is the most common source of UVR exposure, tanning beds are also an important source, particularly for non-Hispanic white women aged 18 to 25, of whom nearly one-third report tanning indoors. It is estimated that indoor tanning results in 400,000 cases of skin cancer in the United States each year, and 6000 of these are melanoma. In general, the risk of skin cancer increases with increasing use, with younger and more frequent users having a steeply elevated risk. For “ever” indoor tanning, the excess risks of melanoma, BCC, and SCC are 20%, 29%, and 67%, respectively. When “ever” indoor tanning is restricted to those younger than 35 or to those considered frequent users, the excess risk dramatically increases. Although the exact magnitude of the increased risk conferred by indoor tanning varies quite a bit within the literature (e.g., owing to differences in study settings and populations), studies consistently show significantly elevated risk. Important to note, there is no evidence to suggest that indoor tanning is safer than tanning outdoors or that it protects from the effects of future sun exposure.

Proposed Mechanism Linking Ultraviolet Radiation to Skin Cancer

UVR from the sun consists of three different ultraviolet (UV) wavelengths—UVA (315–400 nm), UVB (280–315 nm) and UVC (100–280 nm). UVC rays do not reach the Earth's surface owing to stratospheric ozone and so are not considered a source of UVR damage to the skin. The remaining UVR reaching the surface of the earth is 95% UVA and 5% UVB. UVB rays have more energy but are not absorbed as deeply into the skin as UVA rays, which have less energy but can penetrate into the dermis. UVB is significantly more carcinogenic than UVA, because it can directly and indirectly damage DNA, whereas UVA is weakly absorbed by DNA and primarily causes DNA damage through indirect photosensitization reactions.

UVB directly damages DNA through the formation of DNA “photoproducts” or dimers formed by the covalent bonding of two adjacent pyrimidine bases as a result of the absorption of UVR photons. A few different types of dimers can form, but the one most frequently induced by UVB is the cyclobutane pyrimidine dimer (CPD). If these are not remedied by the nucleotide excision repair system, they can result in C to T transitions or CC to TT tandem mutations. These types of mutations were once thought to be a signature of UVB damage; however, this has been called into question based on data showing that UVA exposure can also result in CPDs, although it is likely through an indirect mechanism. C to T mutations in the TP53 gene are likely an early event in the development of SCC, whereas they occur later in BCC and melanoma. Genes coding for proteins of the Hedgehog signaling pathway, mainly the tumor-suppressor gene PTCH, are frequently mutated in BCC, and many of these mutations are C to T and CC to TT transitions. In melanoma, the PTEN and CDKN2A genes have also been shown to harbor C to T mutations. However, melanomas from intermittently sun-exposed areas also show a high frequency of T:A to A:T mutations, such as the well-known V600E mutation in BRAF, suggesting another type of UVR-induced mutation that is not fully understood.

UVA rays primarily cause damage to DNA indirectly. Other chromophores, such as cytochromes and porphyrins, initially absorb UVA energy and then transfer it to either DNA (a type I photosensitized reaction), which results in CPD formation, or to molecular oxygen (a type II photosensitized reaction), generating ROS that ultimately damage the DNA. In addition to DNA, oxidative stress also damages other cellular components, which can further contribute to carcinogenesis.

Another mechanism of UVR photocarcinogenesis may relate to the ability of UVA and UVB to induce persistent genomic instability through shortening or loss of telomeres as well as through a “bystander effect,” whereby cells not exposed to UVR are also damaged. This may occur through the generation of increasingly damaged progeny of UVR-irradiated cells.

Important to note, UVR is a known immunosuppressant. The ability of UVR to suppress certain parts of the immune system was first established by Kripke and Fisher in experiments in mice showing that UVR prevented the rejection of highly immunogenic transplanted UVR-induced skin tumors. In addition, UVR is used for its immunosuppressant effects as local therapy for various skin diseases, such as psoriasis. Furthermore, it is well known that organ-transplant recipients who have the immune system suppressed to prevent organ rejection are susceptible to skin cancer. UVB rays are thought to be primarily responsible for the immunosuppressive capability of UVR, but UVA rays may also play a role. UVR can induce both systemic and local immunosuppression, as well as influencing both acquired and innate immunity. UVB rays suppress the immune system by inducing immunosuppressive mediators, by damaging and triggering the premature migration of antigen-presenting cells required to stimulate antigen-specific immune responses, by inducing the generation of suppressor cells, and by inhibiting the activation of effector and memory T cells. Immunosuppression via UVA is less well understood but may relate to its ability to produce ROS and reactive nitrogen species as well as to its ability to penetrate more deeply into skin tissue, reaching many more types of cells than UVB, including T cells, mast cells, and granulocytes.

In sum, both UVA and UVB are able to induce carcinogenic changes to DNA and other cellular components, generally through unique, although at times overlapping, mechanisms. UVR is both a carcinogen and an immunosuppressive agent that has been implicated in all forms of skin cancer.

Evidence-Based Interventions for Cancer Prevention Related to Ultraviolet Radiation Exposure

Reducing exposure to UVR can reduce skin cancer risk. There are many actions that individuals and communities can take to reduce their UVR exposure. Unfortunately, many individuals are not aware of the risks posed by UVR exposure. Data indicate that one-third of US adults have been sunburned in the past year and that most do not take the recommended preventive actions. In addition, indoor tanning rates are high in some groups, and skin cancer rates are increasing. These factors led the Surgeon General to release a report in 2014 that serves as a call to action to prevent skin cancer. This report outlines the preventive actions that can be taken today by various constituencies and at various levels of society ( Table 22.3 ). The report also sets five goals to support skin cancer prevention across the country and discusses several strategies to achieve each goal ( Table 22.4 ).

The use of sunscreens is a recommended action for skin cancer prevention, although there are few studies demonstrating the skin cancer–protective effect of these products. Data from the Australian Nambour Skin Cancer Prevention Trial ( N = 1621) suggest that the daily use of sun protective factor (SPF) 15+ sunscreen can reduce the risk of SCC and melanoma. After 8 years of follow-up, SCC incidence in those randomized to daily sunscreen application was 35% lower (0.65 [95% CI, 0.43–0.98]) than in those randomized to discretionary use (including no use). Ten years after cessation of the trial, overall melanoma risk was reduced by 50% (0.50 [95% CI, 0.24–1.02]) and invasive melanoma risk was reduced by 73% (0.27 [95% CI, 0.08–0.97]) in the daily sunscreen group compared with the discretionary use group. No clear benefit for BCC was observed. Numerous cohort and case-control studies have examined the relationship between sunscreen use and skin cancer, but results are conflicting and most have serious methodological limitations relating to measures of sun exposure, sunscreen use, and adjustment for important confounders. For various reasons, the evidence that sunscreens prevent skin cancer may remain elusive. Nevertheless, studies have shown that sunscreens are capable of preventing or delaying solar keratosis, a premalignant skin cancer lesion, and the efficacy of sunscreens to prevent sunburns, an established risk factor for skin cancer, is undisputed.

There is much confusion among the public regarding the type, amount, and frequency of application when using sunscreens. The American Academy of Dermatology states that sunscreens should include broad-spectrum protection (protects against UVA and UVB rays), should have an SPF of at least 30, and should be water resistant. It further recommends that sunscreen should be applied 15 minutes before going outside and then every 2 hours or after swimming or sweating, and that approximately 1 ounce of sunscreen should be used—enough to fill a shot glass—to cover all exposed skin.

Although there is currently a great deal of research and controversy surrounding the ingredients of sunscreen and their safety, the risk-benefit ratio indicates that sunscreens are appropriate to include as a skin cancer preventive intervention.

β-Carotene

Based on previous findings from observational and epidemiologic studies that diets rich in β-carotene are associated with reduced cancer risk, two large phase III clinical trials were conducted to test the efficacy of β-carotene for the prevention of lung cancer.

Selenium

Budesonide

The effects of inhaled budesonide (800 µg twice daily) compared with placebo for 1 year was the focus of a phase IIb trial that included 202 current and former smokers with computed tomography (CT) screen-detected lung nodules. This study was nested within the large phase III Continuous Observation of Smoking Subjects (COSMOS) lung cancer screening trial, which was composed of 5203 participants. Budesonide was found to be well tolerated but not effective in reducing lung nodule size compared with placebo in the per-person analysis, which was the primary end point (reduction rate: budesonide 2%, placebo 1%). However, budesonide was found to produce significant regression of existing nonsolid and partially solid nodules (lung nodules most likely representing adenocarcinoma precursors). Long-term follow-up of 5 years showed a significant reduction in nonsolid lung nodule size after treatment with budesonide compared with placebo. The mean maximum diameter of nonsolid nodules in the budesonide arm dropped from 5.03 mm at baseline to 2.61 mm after 5 years. Notably, budesonide was not found to reduce the size of solid nodules. In addition, neither the number of new lesions nor the number of lung cancers differed in the patients treated with budesonide or placebo. Despite the lack of effect on lung cancer incidence in this phase II trial (which was not powered to detect a difference in lung cancer incidence), budesonide remains a promising agent for lung cancer prevention.

Nonsteroidal Antiinflammatory Drugs

Iloprost

A phase II study of iloprost was conducted to determine its efficacy in reversing premalignant histologic changes in the bronchial epithelium in current and former smokers. A total of 152 patients at high risk of lung cancer were treated with iloprost (a 50 to 150 mg/day dose escalation for 2 months, followed by 150 mg/day for 4 months) or placebo. Former smokers treated with iloprost versus placebo demonstrated a significant improvement in endobronchial histologic findings (58.6% versus 28.6%, respectively), whereas current smokers did not. These results show that iloprost is a promising drug for lung cancer prevention. A future phase III trial will be required to conclusively demonstrate the effectiveness of this promising intervention.