Cancer Prevention

50% of all cancers are preventable.
Specific factors associated with cancer risk include the following:

Tobacco use is responsible for 90 percent of all lung cancer deaths and is tied to multiple other cancers.
The association of dietary fat, fruits, vegetables, and fiber with cancer risk is largely unconfirmed. Red meat consumption may promote colorectal cancer and a high intake of tomatoes probably decreases prostate cancer risk.
Vitamin D may reduce the risk of colorectal cancer. Calcium intake, at a minimum of 700 mg/day, may protect against colorectal cancer but high calcium intake (>2000 mg/day) increases risk for prostate cancer. Folate in diet has been associated with a decreased risk of colon and breast cancer, especially in women who drink alcohol; data on folic acid or multivitamin supplementation are inconsistent.
Alcohol intake, even in moderate quantities, increases the risk for colon, breast, esophageal, and oropharyngeal cancer ( but decreases cardiovascular mortality and all cause mortality.)
Physical activity is inversely related to risk for colon and breast cancer. Excess weight increases the risk of multiple cancers.
Skin cancer is directly related to sun exposure, and melanoma rates are increasing. A history of blistering sunburns is of particular risk for melanoma; cumulative sun exposure has more impact on non-melanoma cancers.
HPV, HCV, HTLV1, HIV, EBV, and H pylori have been linked to human cancers. Exposure prevention, screening, vaccination for HPV, and early treatment for abnormal cervical findings and HIV infection can prevent cancer.
Chemoprevention may be helpful in high-risk patients, but risks and benefits should be weighed carefully. Aspirin and NSAIDs offer protection against adenomatous polyps and colorectal cancer, and long-term use in low doses likely decreases cancer-related mortality risk from other solid tumors.
Tamoxifen decreases incidence of breast cancer in high-risk women but increases the risk for thromboembolic disease and early-stage endometrial cancer.Raloxifene is an alternative but has not been evaluated in premenopausal women. Aromatase inhibitors also decrease the risk of breast cancer but may cause muscle and joint pain and are not recommended for premenopausal women. (See "Selective estrogen receptor modulators and aromatase inhibitors for breast cancer prevention.

Cancer Prevention

It is estimated that 50 percent of cancer is preventable [4].Multiple cancer risk factors have been identified; tobacco use, excess weight, poor diet, and inactivity account for two-thirds of all cancers in the US [5]. In one study, nine modifiable risks were identified as the cause of 35 percent of cancer deaths worldwide: smoking, alcohol use, diet low in fruit and vegetables, excess weight, inactivity, unsafe sex, urban air pollution, use of solid fuels, and contaminated injections in health-care settings [6]. The International Agency for Research on Cancer (IARC) has identified and tabulated over 100 human carcinogens [7].Lifestyle factors have been linked to a variety of malignancies, including the most common in the developed world: lung, colorectal, prostate, and breast cancer [8]. In a longitudinal study, participants who had all four lifestyle factors (never smoking, BMI <30, physical activity >3.5 hours weekly, prudent diet) had approximately one-third the risk of cancer compared to those who had none of these factors [9]. A comprehensive systematic review with a global focus conducted by the World Cancer Research Fund came to similar conclusions regarding dietary, weight, and activity factors [10]. A study of an index, based on recommendations from the World Cancer Research Fund and the American Institute of Cancer Research related to weight management, physical activity, plant and animal food consumption, breastfeeding in women, and alcohol intake, found a significant association of higher scores (better compliance) with risk reduction for total cancer and multiple specific cancers (colorectal, stomach, breast, endometrium, lung, kidney, liver, esophagus) but not for prostate, ovarian, pancreatic, or bladder cancer [11]. Despite a robust knowledge of what factors decrease cancer risk, implementation of cancer prevention lags [12].

This topic reviews the major modifiable cancer risk factors and briefly addresses chemoprevention. Lifestyle issues which are associated with an increased risk of cancer are also risk factors for other diseases, such as stroke, heart disease, and diabetes. (See "Overview of primary prevention of coronary heart disease and stroke" and "Prediction and prevention of type 2 diabetes mellitus".)

TOBACCO USE — Tobacco use is the most preventable cause of cancer, and accounts for 21 percent of worldwide total cancer deaths [2]. Approximately one-half of all smokers die of a tobacco-related disease, and adult smokers lose an average of 13 years of life due to this addiction [13,14]. (See "Cigarette smoking and other risk factors for lung cancer".)

Smoking is responsible for approximately 30 percent of all cancer-related deaths in the US [15]. It is the strongest risk factor for lung cancer, increasing risk 10 to 20-fold [16,17]. Smoking is also implicated as a causative factor for leukemia as well as cancers of the oral cavity, nasal cavity, paranasal sinuses, nasopharynx, larynx, esophagus, pancreas, liver, stomach, cervix, kidney, large bowel, and bladder [18,19]. Some evidence also ties smoking to more aggressive prostate cancers [20,21], while the relationship of smoking and breast cancer risk is controversial. (See relevant topic reviews on the risk factors for all of these cancers).

Tobacco acts on multiple stages of carcinogenesis: it delivers carcinogens directly to tissues, causes irritation and inflammation, and interferes with the body's natural protective barriers [22]. The dangers of tobacco are most commonly associated with cigarette smoking, but also occur with cigar, pipe, smokeless tobacco, and exposure to environmental (second-hand) tobacco smoke. (See "Secondhand smoke exposure: Effects in adults".)

Significant health benefits accompany quitting, even for longtime tobacco users. Smoking cessation leads to reduced risk of most tobacco-related diseases and a decrease in all cause mortality. The health benefits of quitting can be seen at all ages and can be measured almost immediately after cessation [23]. (See "Patterns of tobacco use".)

Tobacco prevention and cessation — It is crucial to both prevent initiation and promote cessation of tobacco use, given the tremendous harm of tobacco dependency. Programs and policies that reduce youth initiation and facilitate smoking cessation must be implemented in both clinical and community settings [24]. Responsibilities for health care providers include advice and counseling, referrals to behavioral therapy and support groups, and prescriptions for nicotine replacement and other medications [25,26]. (See "Smoking cessation counseling strategies in primary care" and "Overview of smoking cessation management in adults".)

EXCESS SUN EXPOSURE — Over 1 million cases of skin cancer, including basal cell and squamous cell carcinoma, are diagnosed each year. The American Cancer Society estimates over 68,000 cases of malignant melanoma in the US in 2010 [27], and the incidence continues to rise. Although most skin cancers are curable, in 2011 there are projected to be 8790 deaths related to melanoma [28]. Radiation from the sun is the primary cause of both melanomatous and non-melanomatous skin cancer. Ultraviolet radiation causes genetic mutations and interferes with the cutaneous immune system, limiting the body's ability to reject abnormal cells. Risk of squamous cell and basal cell cancer appear to correlate with total lifetime sun exposure. Cumulative sun exposure may also increase melanoma risk, but repeated intense exposures leading to blistering burns may be even more dangerous [29].

Ultraviolet exposure from tanning beds has been classified as a human carcinogen, with a 75 percent increase in risk for melanoma in patients who utilized tanning booths before age 35 [30]. (See "Risk factors for the development of melanoma".)

Recommendations for sun protection — Targeted interventions for high-risk characteristics such as fair skin, large number of nevi, or positive family history, fails to identify an adequate proportion of people who develop disease [31]. All individuals should limit the time spent in the sun, especially between the hours of 10 am and 3 pm, wear hats, sunglasses, and other protective clothing, and use sunscreen. Because the majority of lifetime sun exposure usually occurs during childhood and adolescence, protective behaviors early in life will provide the greatest benefit. In addition, the World Health Organization has recommended against tanning bed use by anyone under the age of 18 [32]. (See "Primary prevention of melanoma" and "Sunburn".)

PHYSICAL ACTIVITY — Decreased physical activity appears to increase the risk for cancer [8]. Over 60 percent of US adults are not regularly active, including 25 percent who are almost entirely sedentary [33]. It is estimated that sedentary lifestyle is associated with 5 percent of cancer deaths [34]. For people who do not smoke, exercise is one of the most important modifiable risk factors (along with weight control and dietary choices) [35]. (See "Overview of the benefits and risks of exercise".)

In a Japanese cohort, physical activity was associated with a decreased risk for colon, liver, pancreatic, and stomach cancer [36]. The most compelling data are in colon and breast cancer [37-40]:

One study demonstrated a negative correlation between moderate to strenuous exercise and ER-negative, but not ER-positive, breast cancer [39].
In a meta-analysis of 52 studies, there was a significant 24 percent reduced risk of colon cancer when comparing the most versus the least active individuals across all studies (RR 0.76, 95% CI 0.72 to 0.81) [37]. A subsequent meta-analysis of 21 studies found that the risk of colorectal cancer was reduced by 27 percent, comparing the most and least physically active participants, and the risk reduction was the same for both proximal and distal cancers [41]. Physical activity also decreases risk of colon polyps by 15 percent as reported in a meta-analysis of 20 studies [42].
The association between physical activity and decreased risk for breast and colon cancer has been demonstrated across levels of obesity, suggesting that the protective effect of activity goes beyond its impact on body weight [36,43-45].
Limited evidence suggests that activity offers some protection against endometrial and prostate cancer [37,46,47].

Several mechanisms have been proposed to explain the possible protective effect of physical activity including: reduction in circulating levels of insulin, hormones, and other growth factors; impact on prostaglandin levels; improved immune function, and altered bile acid metabolism [48-50]. Physical activity during certain periods of life, such as adolescence, may offer additional protection against disease, particularly for breast cancer [51,52]. The optimal duration, intensity, and frequency of physical activity that may afford cancer protection is unknown.

EXCESS WEIGHT — Almost 65 percent of US adults are overweight; over 30 percent are considered obese [53]. The obesity epidemic affects people of all ages, socioeconomic levels, geographic regions, and ethnicities, and causes significant medical consequences. (See "Health hazards associated with obesity in adults".)

Treatment for obesity is discussed separately. (See "Overview of therapy for obesity in adults".)

Excess weight and obesity are associated with an increase in the risk of multiple cancers including colorectal, postmenopausal breast, endometrial, renal, and esophageal cancer, with a population attributable risk from 9 percent (postmenopausal breast cancer) to 39 percent (endometrial cancer) [37]. A meta-analysis of prospective epidemiologic studies reported significant direct associations with additional cancers including pancreas, thyroid, non-Hodgkin lymphoma, leukemia, and myeloma [54]. Emerging evidence suggests a role for obesity in aggressive prostate cancer [55].

Obesity has been estimated to cause 20 percent of all cancers [4]. Bariatric surgery has been associated with a 60 percent reduction in cancer mortality (5.5 versus 13.3 per 10,000 person-years) over seven year follow-up [56].

Weight gain is also associated with cancer risk. Men who gained ≥21 kg after age 20 had a 60 percent higher risk of colorectal cancer than men who gained only 1 to 5 kg [57]. In addition, women who lost ≥10 kg after menopause and kept it off saw a 50 percent reduction in breast cancer risk [58].

The mechanisms underlying the effect of obesity and weight change likely vary by tumor site with the current focus on endogenous estrogens in breast and endometrial cancer and on insulin and inflammation in colon cancer [4].

DIET — A variety of dietary factors have been studied in relation to cancer. Overall, dietary fat, fruits, vegetables, and fiber have not consistently been shown to affect cancer risk [59] . Intake of other nutrients, particularly certain micronutrients, may offer a degree of protection against certain malignancies. (See "Healthy diet in adults" and"Colorectal cancer: Epidemiology, risk factors, and protective factors" and "Risk factors for prostate cancer".)

Inconsistencies in results of nutritional (diet or supplement) studies can be attributed to multiple factors [60]. Observational studies are subject to imprecision in diet recall, and confounding factors that influence the risk of cancer and occur disproportionately among individuals exposed and not exposed to the nutrient of interest. Randomized controlled trials may yield inaccurate results because of poor adherence to the dietary intervention, insufficient follow-up time, wrong dose or form of the nutrient, or a study population that is replete in the nutrient studied. Additionally, studies tend to focus on one nutrient in isolation, when whole foods or the full composition of a diet may correlate better with cancer risk than any single component.

Dietary fat — Dietary fat has been extensively studied as a possible factor explaining the variation in international cancer rates. No clear link has been found between total fat intake and colon or breast cancer; the data are somewhat more convincing for prostate cancer [61]. (See "Dietary fat".)

While some reports suggest that a diet low in animal fat and/or cholesterol may be protective against colorectal cancer, epidemiologic studies of colon cancer, controlled for caloric intake, have shown no consistent association between total fat intake and cancer risk. In the largest trial to examine this issue, the prospective Women's Health Initiative Dietary Modification Trial, 48,835 women were randomly assigned to a behavioral modification program to decrease dietary fat or a usual diet control group [62]. The intervention group, compared to controls, reduced the percentage of calories consumed as fat (by 10.7 percent in year one, 8 percent at year six). There was no difference in incidence of colorectal cancer during 8.1 years of follow-up. (See "Colorectal cancer: Epidemiology, risk factors, and protective factors", section on 'Diet'.)
Animal and ecologic (international) studies show a positive correlation between fat consumption and increased breast cancer risk. Results of case-control and cohort studies have been mixed, however. As an example, in the same cohort from the Women's Health Initiative cited above, the incidence of breast cancer in 8.1 years of follow-up was not statistically different between low dietary fat and control groups; there was a trend to breast cancer reduction in the low fat intervention group, especially for women with high baseline fat intake who were adherent to the diet modification [63]. (See "Factors that modify breast cancer risk in women".)
A diet high in animal fat may be an important factor in the development of prostate cancer. In particular, intake of large amounts of alpha-linoleic acid and low amounts of linoleic acid appear to increase risk; this combination is common in red meat and some dairy products [64]. The mechanism of this association may be that serum levels of testosterone are lower in men who decrease their fat intake. (See "Risk factors for prostate cancer", section on 'Animal fat'.).

While total fat does not appear to impact cancer risk, questions remain as to whether particular types of fat (saturated, unsaturated, or trans fats) affect risk differently, and whether fat intake in childhood or adolescence carries a greater risk than intake during adulthood. One consistent finding is that excess calories from any source leads to weight gain and an increase in the risk of multiple cancers [65]. (See 'Excess weight' above.)

Red meat — High intake of red meat including beef, pork, veal, and lamb, is associated with an elevated risk of colorectal cancer in both men and women [66-69]. One large study (n = 148,610) concluded that the risk of colon cancer was increased in those with a long-term high intake of processed meat (relative risk [RR] 1.5, 95% CI 1.04-2.17) and with a high ratio of red meat to poultry and fish (RR 1.53, 95% CI 1.08-2.18) [70].

Red meat consumption has also been associated with an increase in cancer-specific mortality. In one large population study (n = 617,119), an association between cancer-specific mortality and red meat intake, comparing highest and lowest quintiles of consumption (HR 1.2, 95% CI 1.1-1.3), was found in individuals aged 50 to 71 years; specific types of cancer were not reported [71]. Findings were similar in another study of over 121,000 men and women where red meat consumption was also associated with an increased risk of cancer mortality (HR 1.16, 95% CI 1.09-1.23) [72].

The mechanisms for this increased risk have not been determined but several factors have been suggested including heme content in the meat, animal fat, and carcinogens produced when the meat is cooked at high temperatures. It isn’t known if risk varies with different animal raising strategies (eg, grass-fed beef).

Fruits and vegetables — Despite suggestions from case control studies that high intake of fruit and vegetables is associated with a significant reduction in cancer, prospective studies have found less consistent results [73-76]. Data from the EPIC study, a cohort study of nearly 500,000 European men and women followed for nine years, found only a weak association between increased intake of fruits and vegetables with overall risk of cancer (HR 0.97, 95% CI 0.96-0.99) [77].

Many epidemiologic studies [78,79], though not all [80], suggest a weak association between the intake of a diet high in fruits and vegetables and protection from colorectal cancer. A pooled analysis of fourteen cohort studies (n >750,000), including the previous study [80], concluded that eating more than 800 g fruit and vegetables daily, compared to less than 200 g, decreased risk for distal colon cancer (RR 0.74) but not for proximal cancer [81]. A subsequent meta-analysis of 19 cohort studies concluded a weaker protective effect, with most of the risk reduction attributable identified at a far lower threshold (100 g/day) of fruit and vegetable intake [82]. A meta-analysis of the relationship between fiber and colorectal cancer found no significant association between fruit, vegetable, or legume fiber and the incidence of colorectal cancer [83].

Evidence is slightly stronger for a link between prostate cancer and tomato products. Several cohort studies have demonstrated a significant reduction of prostate cancer risk in men with the highest intake of tomatoes and tomato products [84,85]. However, a systematic review performed by the FDA found "very limited evidence" to support an association between tomato consumption and reduced risk of prostate or other (ovarian, gastric, and pancreatic) cancers [86]. Lycopene, a carotenoid found in tomatoes, has been postulated to be responsible for this benefit. A systematic review identified three trials of lycopene and prostate cancer, but only one with incident disease as an endpoint, and found a decrease in risk but no difference in serum lycopene levels between groups, concluding that there was insufficient evidence to determine whether lycopene is related to prostate cancer risk [87]. Other studies have failed to support the hypothesis [86,88]. (See "Risk factors for prostate cancer" and "Chemoprevention strategies in prostate cancer".)

A meta-analysis found that intake of high amounts of soy (20 mg per day of isoflavone) in Asian women was associated with a decreased risk for breast cancer, compared to Asian women consuming lower amounts (5 mg daily) [89]. However, even the lowest intake of soy isoflavones in the Asian population was more than fivefold the "high" intake (0.8 mg per day) of women in Western countries, where studies have not shown a protective effect for soy. In another meta-analysis, Chinese women who were in the highest quintile of soy intake had a decreased risk of lung cancer compared to those in the lowest quintile [90]. Increased flavonoids found in tomatoes, green peppers, berries, and citrus fruits have been associated with a modest decrease in breast cancer risk in Western populations [91].

Dairy — A meta-analysis of 21 studies evaluating the relationship of dairy food intake and ovarian cancer found no evidence of association in case control studies (RR 0.96) but three prospective cohort studies did demonstrate increased risk of ovarian cancer with high intake of dairy foods (RR 1.13, 95% CI 1.05-1.22) [92]. Inaccuracy of retrospective diet reports in the case control studies may affect their reliability [93]. However, a subsequent cohort study found no increase in risk of ovarian cancer with dietary dairy or lactose intake [94]. Thus, the relationship of dairy intake to ovarian cancer is uncertain.

Several studies suggest that intake of low-fat dairy products may protect against breast cancer, mainly in premenopausal women [95-98]. In the largest prospective cohort study of over 88,000 women in the Nurses' Health Study, there was an inverse association between breast cancer risk and the intake of low-fat dairy products, calcium (mainly dairy intake), and vitamin D (mainly non-dairy intake) in premenopausal but not postmenopausal women [95]. In contrast, a large pooled analysis of eight prospective studies mainly comprising postmenopausal women did not find a strong association between dairy intake and breast cancer risk [99].

Calcium intake, as determined by a food frequency questionnaire, was associated with decreased total cancer risk over seven years in women, but not in men, in a large US cohort of 567,000 participants aged 50 to 71 years [100]. A decrease in cancers of the digestive tract (particularly colon cancer) with higher dairy food and calcium intake was noted in both men and women, however (for highest versus lowest quintile, RR 0.84 in men and 0.77 in women).

The association between calcium intake and colorectal cancer risk is discussed below. (See 'Calcium' below.)

Fiber — Fiber intake is associated with a reduction in the risk of heart disease [101,102] and diabetes [103,104], but its effect on cancer risk reduction is less certain. Results have been variable among large epidemiologic studies and meta-analyses, and the degree of protection from dietary fiber, if any, will remain unsettled until prospective intervention studies are done. In the absence of randomized trials, observational data may be confounded by the relationship of fiber intake with other micronutrients and with other choices related to lifestyle and diet. (See "Colorectal cancer: Epidemiology, risk factors, and protective factors", section on 'Fiber'.)

Several large epidemiologic studies have reported a significant inverse associated between fiber intake and colorectal cancer risk. As an example, a large European study involving 519,978 patients found that intake of dietary fiber was inversely related to colon cancer incidence (adjusted RR 0.58, 95% CI 0.41-0.85), comparing the highest to lowest quintiles of fiber intake [105]. However, the results may have been confounded by lack of control for folate intake, and fiber may have served as a proxy for this micronutrient. (See 'Folate and other B vitamins' below.)

Several large cohort studies from the US have not shown an association between the development of colorectal adenomas or cancer risk and either fiber intake in general or cereal fiber in particular [67,106-108]. A pooled analysis of 13 prospective cohort studies (725,628 men and women followed up to 20 years) found the inverse association of fiber with the risk of colorectal cancer was no longer apparent after accounting for other dietary risk factors [107]. Likewise, a meta-analysis of five studies concluded that fiber did not affect the incidence or recurrence of adenomatous polyps (the precursor to colorectal adenocarcinomas) within a two to four year follow-up period [108]. However, a meta-analysis of 16 studies from multiple countries did find a relationship between decreased risk of colorectal cancer and total dietary fiber intake (relative risk [RR] of colon cancer for each 10 g/day intake was 0.90, 95% CI 0.86-0.84) [83]. This study also found that the decreased risk was associated with cereal fiber but was not significant with fruit, vegetable, or legume fiber.

Glycemic load — Insulin and insulin-like growth factors promote cell proliferation, and it is hypothesized that hyperinsulinemia may promote certain cancers [109]. An increased risk for certain cancers has been associated with diabetes (primarily type 2) [110]. Patients with diabetes have a two-fold or greater risk of cancers of the liver, pancreas, and endometrium and a slightly lower but increased risk for cancers of the colon, breast and bladder; the risk of prostate cancer is decreased in patients with diabetes [111].

Glycemic load is a function of a food's glycemic index (a measure of how rapidly and to what extent the blood glucose level rises), carbohydrate content per serving, and frequency of intake. Intake of foods with a high glycemic index has been evaluated for association with risk of cancer. These studies have yielded equivocal results. (See"Dietary carbohydrates".)

A large study involving participants in the National Institutes of Health-AARP Diet and Health Study found no significant association between glycemic index or glycemic load (calculated from food questionnaires) with 15,000 cases of cancer in women or 33,000 cases of cancer in men identified over nine years [112].
One case control study of breast cancer reported an association with high glycemic load [113], but large cohort studies have not confirmed this finding [114-118]. A meta-analysis of prospective cohort studies found no significant association between glycemic load and breast cancer risk, but noted that a diet with a high glycemic index was positively associated with increased breast cancer risk [119].
Glycemic load was not associated with prostate cancer risk in a large cohort study [120].
Studies of colorectal cancer and glycemic load have also shown inconsistent results [64,121-124]. Data from the Women's Health Study, evaluating dietary habits of 174 incident colon cancer patients in a cohort of 38,451 women, showed that dietary glycemic load was significantly associated with an increased risk of colorectal cancer (adjusted RR 2.85, 95% CI 1.4 to 5.8), comparing highest to lowest quintiles of glycemic load [121]. On the other hand, a large study of Canadian women found no increase in risk [123]. Evaluation of data from the Nurse's Health Study and from the Health Professionals Follow Up Study found a small increase in risk among men with high dietary glycemic load (RR 1.32, 95% CI 0.98-1.79) but no association among women [124]. A meta-analysis found no association of glycemic load or glycemic index with colorectal cancer risk [125].

Omega-3 fatty acids and dietary fish — A systematic review of prospective studies evaluating the effect of omega-3 fatty acid consumption on tumor incidence concluded that there is no association between omega-3 fatty acids and cancer risk for 11 different types of cancer [126]. Ten studies evaluated in this review reported significant findings, but individual studies indicated both increased and decreased risk with no consistent pattern. A subsequent randomized trial found an increase in cancer risk for women treated with omega-3 fatty acids, but not for men [127].

While an association has not been found for dietary supplementation with omega-3 fatty acids and cancer incidence, an association was found in a systematic review of 41 observational studies for fish consumption and a decreased incidence of colorectal cancer [128]. In that analysis, including case control and cohort studies, an inverse relationship between fish intake and rectal cancer was demonstrated (OR 0.79, 95% CI 0.65-0.97), while an inverse trend was suggested for colon cancer (OR 0.96, 95% CI 0.81-1.14).

VITAMINS AND MICRONUTRIENTS — Multiple observational and prospective studies of the use of supplemental vitamins and minerals to prevent cancer have been disappointing [129]. A systematic review of 38 studies found that neither vitamin C nor vitamin E supplementation was beneficial for prevention of the cancers evaluated [130]. A 2006 National Institutes of Health (NIH) consensus conference panel concluded that "present evidence is insufficient to recommend either for or against the use of multivitamin supplements by the American public to prevent chronic disease" [131]. A subsequent long-term randomized trial (mean 9.4 years treatment) in 8000 women found no evidence that supplementation with vitamin C, E, or beta-carotene (singly or in combination) decreased cancer incidence or cancer mortality [132]. Additionally, two long term observational studies, one including over 160,000 women with follow-up of approximately eight years [133] and another including over 180,000 multiethnic participants with eleven year follow-up [134], found no association between multivitamin use and risk of cancer. (See "Vitamin supplementation in disease prevention".)

However, the large (n = 14,641) randomized Physicians’ Health Study II found that supplementation with a multivitamin resulted in a small reduction in total cancer that narrowly reached statistical significance (HR 0.92, 95% CI 0.86-0.998) [135]. The study population was male physicians 50 years and older at recruitment, and thus generally well nourished and highly educated. With a mean follow-up of 11.2 years, there was a reduction in total cancer from 18.3 to 17.0 events per 1000 person-years, comparing groups assigned to multivitamins and placebo, respectively. No difference was found in the secondary outcome of decreased incidence of specific cancers, and there was no impact on cancer mortality. For men with a baseline history of cancer, multivitamin use was associated with a reduction in total cancer (HR 0.73, 95% CI 0.56-0.96).

Given the variability of study findings, with only marginal benefit in some and no benefit in others, meta-analysis incorporating the prior studies would be helpful. Pending such analysis, it has not been established that multivitamin and mineral supplements provide added benefit to a balanced, healthful diet for most individuals [136]. The only prospective trial of vitamin supplementation (selenium plus vitamin E and beta-carotene) to show a durable finding of fewer deaths in the treated group was conducted in remote Linxian Province, China, a region where the population consumes a poor diet, and enrollees likely suffered deficiencies of the nutrients tested [137].

Vitamin D — Studies of the relationship between vitamin D intake or serum levels of 25(OH)D and cancer risk have been inconsistent [138]. Studies vary in regard to participants (sex, baseline serum levels), types of cancer evaluated and dose of vitamin D. Overall, it does not appear that vitamin D supplements should be prescribed to decrease cancer risk [139].

Colorectal cancer — Vitamin D may reduce the risk of colon cancer [139-143] but the data do not show a consistent relationship [144,145]. A pooled analysis of 10 cohort studies found a trend towards reduction of colorectal cancer with increasing vitamin D intake, but the protective effect was only statistically significant among those individuals with the highest vitamin D intake [146]. Vitamin D may decrease cancer risk through improved calcium absorption.

Prostate and breast cancer — Although there is a theoretical basis to predict an effect of vitamin D on prostate cancer [147], at least two studies did not demonstrate such a relationship [148,149]. (See "Risk factors for prostate cancer", section on 'Calcium and vitamin D'.)

Vitamin D plus calcium supplementation does not appear to affect risk for breast cancer. In the Women's Health Initiative trial, 36,000 postmenopausal women were randomly assigned to take 1000 mg elemental calcium with 400 IU vitamin D daily, or placebo, for a mean of seven years (with hip fracture rate as the primary outcome) [150]. There was no difference between the two groups in the incidence of invasive breast cancer, a secondary outcome of the trial.

Total cancer — An increment of 25 nmol/L in the serum 25(OH)D level was projected to result in a 17 percent reduction in total cancer risk in men, extrapolated from data from the US Health Professionals Follow-Up Study [151]. This incremental level of serum 25(OH)D is not readily achieved with diet (one glass of milk is predicted to increase the plasma level only by 2 to 3 nmol/L), and would require supplementation with at least 1500 IU vitamin D daily.

A randomized trial comparing four-year supplementation with calcium (1400 to 1500 mg daily), calcium plus vitamin D (1100 IU daily) or placebo in 1179 women ≥55 years found a decreased risk of cancer for both calcium alone and calcium combined with vitamin D [152]. After the first year, there was a large risk reduction in women who received vitamin D (RR 0.23, 95% CI 0.1-0.6). This large magnitude of effect seems difficult to explain in light of inconsistent results from other studies.

Data from nearly 17,000 participants in NHANES III, a large national nutritional survey, found that overall cancer mortality risk was not related to baseline 25(OH)D levels, with a mean study follow-up of 13.4 years [139]. Interestingly, cancer mortality risks were increased for some men with higher baseline 25(OH)D levels (>100 nmol/L compared with <37.5 nmol/L), although there was a suggestion of an inverse relationship for colorectal cancer. For women, however, higher serum levels were associated with a decreased cancer mortality risk for those living in the north and sampled in warmer months.

Calcium — Increased calcium intake has been linked to reduced risk of colorectal cancer but may be associated with an increased risk of prostate cancer. There may be a minimum level of calcium intake, around 700 mg/day, that confers protection against colorectal cancer without significantly increasing prostate cancer risk.

Colorectal cancer — Multiple observational studies have demonstrated that higher calcium intake (either dietary or supplemental) is associated with a reduced risk of colorectal cancer [153,154]. As an example, in a combined cohort from the Health Professionals Follow-up Study and Nurse's Health Study, the risk of distal, but not proximal, colon cancer was reduced in subjects who took up to 1250 mg/day elemental calcium versus <500 mg/day (RR 0.58, 95% CI 0.32-1.05) [155].

Calcium supplementation appears to prevent the recurrence of colorectal adenomas. A meta-analysis of three randomized, placebo-controlled trials of subjects with colorectal adenomas (total n = 1485) concluded that the risk of recurrence was significantly lower in patients randomized to calcium supplementation (RR 0.80, 95% CI 0.68-0.93) [156]. As a result of these data, calcium supplementation has been recommended for the primary or secondary prevention of colonic adenomas by the American College of Gastroenterology [157]. (See "Approach to the patient with colonic polyps".)

Despite these benefits in adenoma prevention trials, whether calcium supplementation reduces the risk of colorectal cancer is unproven.

A protective effect of calcium supplementation could not be shown by the Women's Health Initiative (n = 36,282), which found no significant decrease in incidence or stage of colorectal cancer in the group who had been randomly assigned to receive calcium 500 mg and vitamin D 200 IU twice daily, compared to placebo [158]. The average age of women at the start of randomization was 62 years, and follow-up was seven years. Given the known slow progression rate for colorectal cancer, seven years may be too short an interval to find an effect on cancer incidence, and longer-term follow-up for this study population is planned. However, it has been noted that the baseline mean calcium intake in the Women’s Health Initiative participants (1151 mg/day) was above the threshold for effect indicated by previous studies and thus increasing intake would be expected to show no effect [159].

Prostate cancer — Case-control and prospective studies of calcium and prostate cancer have reported inconsistent results [160]. Three large cohort studies found an increased risk of prostate cancer with different measures of calcium intake [161-163]. Two other prospective studies found no association [164,165]. (See "Risk factors for prostate cancer", section on 'Calcium and vitamin D'.)

The risk of prostate cancer may be increased with high, but not moderate, calcium intake. One study of 3811 incident cases of prostate cancer found that total calcium over 2000 mg/day from both diet and supplementation was linked to a 20 percent increase in prostate cancer risk (RR 1.2, 95% CI 1.0-1.6) [163]. High dietary calcium (≥2000 mg/d) was associated with an even greater increased risk of prostate cancer (RR 1.6, 95% CI 1.1-2.3), but moderate dietary calcium was not. Data from another study showed stronger associations between high calcium intake (≥2000 mg/d) and total, advanced, and metastatic disease (RR 1.71, 2.97, and 4.57, respectively) [162].

It has been suggested that high calcium levels may increase prostate cancer risk by down-regulating the active form of vitamin D, thus interfering with vitamin D's proposed inhibition of tumor growth and metastasis.

Selenium — Animal studies suggest that selenium decreases the risk of a variety of tumors, and some epidemiologic studies have shown an inverse relationship between selenium and cancer [166-168]. One study using NHANES III data from 14,000 adults found an inverse association between selenium levels and cancer mortality at levels of selenium up to 130 ng/ml, but an increase in mortality at levels >150 ng/ml [169].

A placebo-controlled randomized trial of selenium for prevention of nonmelanoma skin cancer showed a significant mortality reduction in cancers of the lung, colon, and prostate [170]. A systematic review of the effects of antioxidant supplements on cancer included four randomized trials of selenium alone or in combination with other supplements and found that selenium reduced the overall risk of cancer in men (RR 0.77, 95% CI 0.64-0.92) but not in women [171]. However, a later meta-analysis of randomized trials of antioxidant therapy included five trials of selenium and found no significant cancer risk reduction (RR 0.62, 0.36-1.08) [172].

Based on preliminary evidence from earlier smaller trials, the role of selenium in decreasing the incidence of prostate cancer was evaluated in the much larger Selenium and Vitamin E Cancer Prevention Trial (SELECT), which included over 35,000 men [173]. The trial was stopped prematurely for futility as neither vitamin E nor selenium protected participants from prostate cancer and there was a nonsignificant trend toward increased risk for diabetes in the selenium group [174]. (See "Chemoprevention strategies in prostate cancer", section on 'Selenium' and "Risk factors for type 2 diabetes mellitus", section on 'Selenium'.)

Vitamin E — Current evidence does not support a role for vitamin E supplementation in the prevention of cancer, and some evidence suggests that vitamin E may be harmful.

In long term follow-up (7 up to 12 years) of the SELECT trial, the risk of prostate cancer was higher in the men who were assigned to take vitamin E alone, compared to placebo (HR 1.17, 99%CI 1.004-1.36) [175].
Vitamin E (600 IU alpha-tocopherol every other day) did not prevent invasive cancer in a 10-year follow-up to the Women's Health Study, evaluating healthy women age 45 years and older (mean age 55 years) [176]. Neither vitamin E nor vitamin C decreased risk for prostate or total cancer in a Physicians Health Study randomized trial of 14,641 men aged 50 and older followed for eight years [177].
In a meta-analysis of six randomized trials, vitamin E supplementation had no effect on cancer incidence or cancer mortality [171].

Folate and other B vitamins — Folate is present in green, leafy vegetables, fruits, cereals and grains, nuts, and meats. Folic acid, a synthetic form included in supplements, has many of the same biologic effects as folate, but is more bioavailable. Folate is important in DNA synthesis, methylation, and repair, as well as in the regulation of gene expression.

The role of folate or folic acid in cancer prevention is uncertain. Folate has been associated with a decreased risk for colon and other cancers, especially in individuals who consume alcohol, in observational studies. However, some randomized trials have suggested the possibility that folic acid may increase risk for cancer.

Support from observational studies for folate as a factor in cancer protection is as follows:

Subjects with lower levels of methylenetetrahydrofolate reductase, an enzyme involved in folate metabolism, have a reduced risk of colon cancer [178], as well as cancers of the esophagus, stomach and pancreas [179].
An inverse relationship was found between folate intake and the risk of developing adenomatous polyps in the combined analysis of data from the Nurses’ Health Study and Health Professionals Follow-Up Study [180].
The Nurses’ Health Study demonstrated a decreased risk of colon cancer (RR 0.25, CI 0.13-0.51) in women who took folic acid-containing multivitamins for at least 15 years, suggesting that folate levels may be particularly important in the early stages of colorectal cancer development [181]. Findings from the Women's Health Initiative data, however, showed that increased dietary folate and vitamin B6 intake lowered colorectal cancer risk, while vitamin supplementation with folic acid and B6 did not [182].
Dietary folate, but not vitamin supplementation, was associated with a reduced risk for pancreatic cancer in a Swedish cohort of 82,000 men and women prospectively followed for seven years [183]. A meta-analysis found that dietary folate was associated with reduced risk for esophageal squamous cell cancer (RR 0.66, 95% CI 0.53-0.83), and pancreatic cancer (RR 0.49, 95% CI 0.35-0.67), comparing highest versus lowest intake categories [179].

However, a meta-analysis of case-control studies and observational studies did not demonstrate an association between low dietary folate intake and breast cancer [184]. The study found evidence of publication bias in previous studies that had suggested an association.

In contrast to biologic and observational evidence supporting a role for folate in cancer prevention, randomized trials of folic acid supplementation have not shown benefit and some trials have raised the possibility of harm.

A meta-analysis based on individual patient data from eight randomized trials of patients with increased cardiovascular risk (n = 37,485) found no significant effect offolic acid supplementation in varying doses (0.8 to 40 mg per day) on cancer incidence (RR 1.05, 95% CI 0.98-1.13) or cancer mortality (1.00, 0.85-1.18) with a median follow-up of five years [185]; findings were similar in a subsequent trial in patients with cardiovascular disease [127].
The largest controlled trial to evaluate folic acid supplementation in patients with colorectal adenomas found no decrease in new adenomas at three and six-year follow-up, but an increase in risk of advanced colon lesions and in noncolorectal cancer in patients who received folic acid, compared to controls [186].
Combined analysis of two trials evaluating vitamin supplementation in patients with ischemic heart disease found an increased risk for cancer incidence (primarily lung cancer) (HR 1.21, 95% CI 1.03-1.41) and cancer mortality (HR 1.38, CI 1.07-1.79) at three year follow-up in participants who received folic acid plus vitamin B12 supplementation for three years [187].
The incidence of prostate cancer was increased in the group of men randomly assigned to receive folic acid (1 mg/day) in a placebo-controlled trial of chemoprevention for colorectal polyps [188].

On the other hand, a study of 99,523 US participants enrolled in the Cancer Prevention Study II Nutrition cohort that examined the association between folate intake (dietary or supplemental) and colorectal cancer risk failed to document an increased risk of colorectal cancer from high folate intake [189]. Additionally, a randomized trial of supplementation for up to nine years with combination B vitamins (folic acid 2.5 mg, vitamin B6 50 mg, and vitamin B12 1 mg) or placebo in women at increased risk for cardiovascular disease found no difference in the incidence of colorectal adenoma [190].

Serum levels of other B vitamins have been associated with reduced cancer risk in observational studies:

A meta-analysis of prospective studies found an inverse association for risk of colorectal cancer with vitamin B6 intake and blood levels of pyridoxal 5'-phosphate, the active form of vitamin B6 [191]. However, it is not known whether supplementation with vitamin B6 will decrease cancer risk.
In a prospective study over eight years, involving more than 500,000 participants, having a serum level above median for B6 and methionine was associated with a lower risk of lung cancer, in both never and current smokers [192]. In this cohort, a lower risk of lung cancer was also seen in smokers (past or current) with higher levels of folate, but not in those who never smoked.

Folate in alcohol users — The interaction between folate and alcohol may be important in cancer prevention (see 'Alcohol' below). Alcohol consumption is known to both interfere with folate availability and increase the risk of colon and breast cancer. In one study, the increase in colon cancer risk associated with alcohol use was not seen in men with the highest folate intake [193]. Similarly, the increased risk of breast cancer associated with alcohol use was most pronounced in women with the lowest folate intake [194,195]. An inverse relationship between folate intake and breast cancer was most evident among women whose alcohol consumption was greater than 15 g (about one drink) per day (figure 1). Increasing levels of plasma folate correlated with decreased risk of breast cancer particularly among women who consumed alcohol [196].

Iron — Observational studies suggest that increased iron stores or dietary iron may be associated with increased risk for cancer [197,198]. A randomized trial conducted to evaluate the benefits of phlebotomy in patients with peripheral artery disease found a significant reduction in cancer incidence at six months (HR 0.65, 95% CI 0.43-0.97) in patients assigned to the phlebotomy group, compared to controls [199]. This finding warrants confirmation.

Other — Other vitamin supplements have been evaluated, with variable findings:

High doses of beta-carotene have been associated with an increased incidence of lung cancer [171,200,201]. Beta-carotene doses used in some of these trials may also be found in multivitamin formulations used to promote visual health. Beta-carotene did not decrease cancer incidence in studies of American women [202] and men [203]. (See "Cigarette smoking and other risk factors for lung cancer", section on 'Beta-carotene supplementation'.)
Supplementation with a combination of beta-carotene, selenium, and zinc decreased the incidence of noncardia stomach cancer, but not other intestinal malignancies, in a population in rural China with baseline deficiencies in micronutrients [204].

ALCOHOL — Excess alcohol consumption increases the risk of multiple cancers (See "Overview of the risks and benefits of alcohol consumption", section on 'Alcohol and specific illnesses'.).

A prospective study of over one million women (average age 56 years) found that, at an average follow-up of 7.5 years, 10 g of alcohol (one drink) per day increased the risk for cancers of the oropharynx, esophagus, larynx, rectum, liver, and breast [205]. The overall risk for cancer was increased 6 percent (95% CI 4 to 7 percent) per consumption of 10/g day alcohol of any type. The increased risk for cancers involving the upper respiratory and digestive tract was seen only in current smokers.
Increased risk for breast cancer with moderate alcohol consumption was found in the Women's Health Study [206] as well as in the Nurse’s Health Study [207].
The European Prospective Investigation into Cancer and Nutrition (EPIC) study in eight European countries found alcohol consumption could account for 10 percent of the attributable risk for any cancer in men, and three percent in women [208]. The risk was greatest for consumption of more than recommended upper limits of alcohol, and was greatest for upper gastrointestinal and hepatocellular cancer.
It has been estimated that worldwide 3.6 percent of cancers are associated with chronic alcohol drinking [209].

Moderate alcohol use has beneficial effects on cardiovascular health, but the increased cancer risk may offset such benefits [210]. (See "Overview of the risks and benefits of alcohol consumption", section on 'Cardiovascular disease'.)

Several mechanisms have been postulated to account for the carcinogenicity of alcohol [209]. Its solvent properties may allow carcinogens to penetrate cell membranes. Alcohol increases estrogen levels and impacts folate metabolism. Alcohol may also act as an irritant, causing increased cell production; as a transporter carrying carcinogens; as an inhibitor of DNA methylation; or as a prometabolite for identified carcinogens such as acetaldehyde [209,211,212].

For unclear reasons, moderate consumption of alcohol has been associated with a decreased risk of renal cell carcinoma. (See "Epidemiology, pathology, and pathogenesis of renal cell carcinoma", section on 'Alcohol'.)

INFECTION — It is estimated that 17 percent of all new cancers worldwide are due to infections [15]. Viruses may increase cancer risk through cellular transformation, disruption of cell cycle control, increased cell turnover rates, and immune suppression [213].

Multiple links between viral agents and cancer have been established:

Human papillomavirus (HPV) with cervical and other anogenital cancers as well as squamous cell cancers of the head and neck [214]. (See "Virology of human papillomavirus infections and the link to cancer" and "Human papillomavirus associated head and neck cancer".)
Hepatitis B (HBV) and C (HCV) with hepatocellular carcinoma [215]. (See "Epidemiology and etiologic associations of hepatocellular carcinoma".)
Human T-cell lymphotropic virus (HTLV-I) with adult T cell leukemia [216]. (See "Treatment of large granular lymphocyte leukemia".)
Human immunodeficiency virus (HIV-I) with Kaposi sarcoma, non-Hodgkin lymphoma [15], and multiple non-AIDS defining malignancies [217]. (See "HIV infection and malignancy: Epidemiology and pathogenesis".)
Human herpes virus 8 (HHV-8) with Kaposi sarcoma and primary effusion lymphoma [218,219]. (See "Virology, epidemiology, and transmission of human herpesvirus 8 infection".)
Epstein-Barr virus (EBV) with Burkitt lymphoma [218]. (See "Pathobiology of Burkitt lymphoma".)

The majority of these viruses are spread through contact with infected blood or body fluids, thus offering opportunities for prevention. Vaccinations for HBV and HPV are particularly promising. (See "Prevention of sexually transmitted infections" and "Hepatitis B virus vaccination" and "Adolescent sexuality", section on 'STIs and HIV'.)

Strategies to prevent transmission through infected blood and blood products must also be implemented. Examples include use of sterile disposable needles for a single patient in healthcare settings, needle exchange programs, regulation of tattooing, continued screening of blood, organ, and semen donors, and the development of artificial blood products.

For some viruses, interventions are available to prevent or delay progression to cancer after infection.

A quadrivalent HPV vaccine is recommended for girls and women aged 9 to 26 years of age. Additionally, cervical cancer screening has dramatically reduced the incidence of cervical cancer where screening is widely available [220]. (See "Recommendations for the use of human papillomavirus vaccines" and "Screening for cervical cancer: Rationale and recommendations".)
Retroviral therapy for HIV infection has greatly altered the course of disease and associated cancers. Highly active antiretroviral therapy (HAART) has been shown to reduce the incidence of AIDS-related lymphoma [221]. (See "Selecting antiretroviral regimens for the treatment naive HIV-infected patient".)
Decreasing the hepatitis B viral load by treatment with interferon or nucleoside/tide analogues in patients with chronic hepatitis B infection was associated with a decreased risk for hepatoma [222,223]. (See "Overview of the management of hepatitis B and case examples".)
Excess alcohol use may play a role in cancer development in patients with chronic HBV and HCV infections and should be avoided. Preliminary data suggest that antiviral therapy may reduce the risk of cancer in patients with chronic HCV infections, with reduction of HCV RNA, but the long-term effect of antiviral therapy on cancer risk is not known [224].

In addition to viral agents, the bacterium Helicobacter pylori (H. pylori) has been associated with gastric cancer [225] and with MALT lymphomas. (See "Association between Helicobacter pylori infection and gastrointestinal malignancy".)

CHEMOPREVENTION — For several cancers, prophylactic medication can reduce cancer risk for high risk individuals. The risk/benefit ratio for chemoprevention must be evaluated for individual cases.

Breast cancer

Tamoxifen — Tamoxifen is a selective estrogen receptor modulator (SERM) with both estrogen agonist and antagonist properties.

Data from four randomized trials and a meta-analysis suggests a statistical decrease in the risk of hormone receptor-positive invasive breast cancer with five years oftamoxifen therapy [226-232]. Neither the individual studies nor the meta-analysis have shown whether a reduction in the incidence of breast cancer will lead to a reduction in overall mortality or breast cancer-related mortality [232]. (See "Selective estrogen receptor modulators and aromatase inhibitors for breast cancer prevention", section on 'Premenopausal women'.)

Tamoxifen is approved in the United States for the prevention of breast cancer in high-risk women. High-risk women, as defined in the trials of tamoxifen for breast cancer prevention, include those with a history of lobular carcinoma in situ (LCIS), a five-year estimated risk for breast cancer of at least 1.66 percent as determined by the Gail model, or family history of the disease [233]. (See "Risk prediction models for breast cancer screening", section on 'Breast Cancer Risk Assessment Tool (Gail model)'.)

The acceptance of tamoxifen as a preventive agent has been limited by its association with serious adverse events, including endometrial cancer and thromboembolic events [234]. The greatest clinical benefits with the least side effects appear to be derived in premenopausal women (who are less likely to have thromboembolic sequelae and uterine cancer), women without a uterus, and women at higher breast cancer risk. Models to assess benefit and risk of tamoxifen in individual women are available. (See"Selective estrogen receptor modulators and aromatase inhibitors for breast cancer prevention".)

Because of the potential for serious side effects, the US Preventive Services Task Force (USPSTF) recommended against routine use of tamoxifen for breast cancer prevention in women of average risk in a 2002 guideline [235]. They recommended that health care providers discuss the risks and benefits of tamoxifen for breast cancer prevention with women at high risk of breast cancer and low risk of adverse effects from the medication. The USPSTF clinical practice guideline for prevention of breast cancer, as well as other USPSTF guidelines, can be accessed through the website for the Agency for Healthcare Research and Quality at www.ahrq.gov/clinic/uspstfix.htm.

Raloxifene — Raloxifene is another SERM that is approved for the prevention of osteoporosis, and for breast cancer prevention in postmenopausal women at high risk for invasive breast cancer. In the STAR breast cancer prevention trial that directly compared both agents, raloxifene was slightly less effective than tamoxifen [236]. However, raloxifene is associated with fewer of the most serious side effects associated with tamoxifen [234,236]. (See "Selective estrogen receptor modulators and aromatase inhibitors for breast cancer prevention", section on 'Raloxifene'.)

The 2002 USPSTF guidelines for breast cancer chemoprevention did not address the use of raloxifene for breast cancer prevention [235]. However, raloxifene is a reasonable alternative to tamoxifen for postmenopausal women with a high risk for breast cancer (a history of LCIS or a five-year estimated risk for breast cancer of at least 1.66 percent as determined by the Gail model [233]) and who still have an intact uterus or need treatment for osteoporosis.

There are no data on the use of raloxifene in premenopausal women, and it is potentially teratogenic. Thus, at present, use of raloxifene should be restricted to postmenopausal women.

Aromatase inhibitors — Aromatase inhibitors (AIs) such as anastrozole, letrozole, and exemestane suppress plasma estrogen levels in postmenopausal women by inhibiting or inactivating aromatase. AIs appear to be associated with a lower risk of life-threatening adverse events than are SERMs.

In the placebo-controlled NCIC CTG MAP.3 trial, exemestane reduced the risk of breast cancer in women at high risk for the disease by approximately 65 percent, with minimal side effects [237]. In this trial, high-risk was defined as age ≥ 60 years, estimated risk of breast cancer ≥ 1.66 percent over five years from the Gail model [233], history of atypical ductal hyperplasia or atypical lobular hyperplasia, lobular carcinoma in situ (LCIS), or ductal carcinoma in situ (DCIS) treated with mastectomy. (See"Selective estrogen receptor modulators and aromatase inhibitors for breast cancer prevention", section on 'Aromatase inhibitors'.)

These data establish AIs as an option for breast cancer prevention in high-risk postmenopausal women. However, questions remain as to long-term effects of an AI on bone loss and cardiovascular risk, since follow-up data for the MAP.3 trial is limited. Furthermore, joint and muscle symptoms associated with AIs may limit patient acceptance of this medication for preventive purposes. Additionally, exemestane is more costly than tamoxifen or raloxifene, although an alternative AI is available as a less expensive generic and has similar activity to exemestane when used for adjuvant therapy. (See "Adjuvant endocrine therapy for hormone receptor-positive breast cancer", section on 'Side effects' and "Adjuvant endocrine therapy for hormone receptor-positive breast cancer", section on 'Aromatase inhibitors'.)

AIs are generally avoided in premenopausal women because of concerns that reduced feedback of estrogen to the hypothalamus and pituitary will increase gonadotropin secretion and stimulate the ovary. Thus, use of an AI for breast cancer prevention should be restricted to postmenopausal women.

Aspirin and other anti-inflammatory drugs — Several theories have been proposed for why aspirin and other NSAIDs are effective in reducing colorectal cancer risk, and possibly effective for other cancers. These medications may cause cell cycle arrest or apoptosis (programmed cell death) of abnormal cells. Reduced risk may also relate to irreversible inhibition of cyclooxygenase-2. Inhibition of this enzyme decreases the synthesis of prostaglandins, which may inhibit tumor growth. Finally, aspirin may influence intracellular signaling through inhibition of phospholipase activity.

Colorectal cancer — Regular use of aspirin and other nonsteroidal anti-inflammatory drugs (NSAIDs) has been shown to decrease the risk of adenomatous polyps and colorectal cancer [238-241]. In high-risk patients with Lynch Syndrome, a randomized controlled trial found 600mg/day of aspirin decreased the risk of colorectal cancer by 60 percent [242]. The optimal dose of aspirin for patients at lower risk, however, has not been established [243]. (See "NSAIDs (including aspirin): Role in prevention of colorectal cancer".)

Accumulated evidence indicates that long-term usage of aspirin is necessary to reduce risk. In addition, it appears that daily dosing (at a minimum of 75 mg) is needed to reduce risk.

Full dose aspirin (≥325 mg) taken daily for a minimum of five years, compared with no aspirin, decreased the incidence of colorectal cancer in the Cancer Prevention Study II Nutrition Cohort (RR 0.68, 95% CI 0.52-0.90) [240].
Low-dose aspirin (100 mg every other day), when compared with placebo, did not decrease the incidence of colorectal cancer in the Women's Health Study of healthy women over age 45 years [244].
The Nurses' Health Study data showed that long-term use (>10 years) was necessary, suggesting action early in the pathway to carcinogenesis [245].
Combining data from randomized trials and observational data, one investigational group estimated that a duration of five years of use of 300 mg or more of aspirin per day is required to achieve prevention with a ten year latency [241].
Another study reviewed data from four randomized trials and found that aspirin use (75 to 300 mg daily) for five years or longer reduced risk for colon cancer (HR 0.76, 95% CI 0.50-0.96) and that there was no incremental benefit for doses greater than 75 mg [246]. A fifth trial reviewed in this study, however found that the risk of fatal colorectal cancer was greater for patients taking 30 mg compared to those taking 283 mg daily.
A subsequent analysis of six randomized trials by the same author found that aspirin taken for at least five years reduced the 20-year risk of death from colorectal cancer (OR 0.55, 95% CI 0.41-0.76) [247].

Recommendations from the US Preventive Services Task Force (USPSTF) in 2007 (preceding publication of many of the studies discussed above) noted that chronic use ofaspirin, at doses suggested to decrease the incidence of colorectal cancer, increases the risk for gastrointestinal bleeding and hemorrhagic stroke, as well as renal failure and hypertension [248]. Existing recommendations from the USPSTF and the American Cancer Society do not recommend aspirin use because of complication concerns [248,249] but do not consider the recent findings for a strong benefit with long-term use.

Other cancers — Data regarding aspirin in the prevention of cancers other than colorectal are less consistent, but seem to suggest that daily (but not alternate day) dosing of aspirin decreases the risk of cancer but with a long latency, requiring long-term follow-up of individual patient data to assess the impact of aspirin on cancer prevention.

The observational Cancer Prevention Study II Nutrition Cohort (n = 146,000), with 18,000 cancers diagnosed during a 10 year follow-up period, found that the overall cancer incidence in men was lower for those who took ≥325 mg aspirin daily for at least five years compared to no aspirin use; cancer incidence in women was lower, but the difference was not statistically significant [240]. In addition to colorectal cancer, long-term aspirin use was associated with a lower incidence of prostate cancer (RR 0.81, 95% CI 0.7-0.94) and a trend to lower incidence in breast cancer (RR 0.83, 0.63-1.10).
Low dose aspirin (100 mg every other day) did not prevent total cancer death, or incidence of breast, colorectal, or lung cancer, when compared with placebo, at 10 year follow-up in the Women's Health Study randomized trial of healthy women over age 45 years [244].
A systematic review and meta-analysis included data from eight randomized trials involving over 25,000 patients participating in cardiovascular outcome trials who were randomly assigned to take either aspirin (in varying doses) or placebo [250]. After five years of use, aspirin decreased the risk of death due to cancer (HR 0.66, 95% CI 0.50-0.87). The effect was seen by five years for cancers of the esophagus, pancreas, brain, and lung, and by ten years for stomach, colorectal, and prostate cancer. The benefit was not related to aspirin dose (75 mg or greater) or sex. Thus, this study suggests that low dose aspirin, with a lower risk of adverse side effects, may be sufficient for a prevention benefit.
Another meta-analysis, which failed to account for duration of use, however, included over 100,000 participants with six-year follow-up who had low to moderate risk of cardiovascular disease and found no significant decrease in cancer mortality for regular aspirin use (OR 0.93, 95% CI 0.84-1.03) with a 30 percent increase in nontrivial bleeding events [251]. When the analysis was restricted to daily aspirin use, there was still no significant decrease in cancer risk, but a further increase in bleeding risk.
In a meta-analysis of six trials of low-dose aspirin for primary prevention (35,535 participants), aspirin reduced the incidence of cancer from three years on (OR 0.76, 95% CI 0.63-0.93) and appeared to be equally effective in women and men [247]. In this same study, aspirin reduced cancer deaths from five years on (OR 0.63, 95% CI 0.47-0.86). The number of cancers in the six trials was too small to determine effects on specific cancers. While the effect of aspirin increased with trial duration, effects on vascular events and extracranial bleeding decreased over time. Overall, aspirin reduced the risk of the composite outcome (major vascular event, cancer, or fatal extracranial bleed, HR 0.88, 95% CI 0.82-0.94).
Two other analyses found that data from case control studies correlated with findings from randomized trials, indicating that the effect of aspirin on cancer prevention is largest on gastrointestinal cancers (esophageal, gastric, biliary, in addition to colorectal) [252]. Additionally, randomized trials suggest that aspirin reduces the risk of cancer with distant metastasis, both at initial presentation and on subsequent follow-up [253].

5-Alpha reductase inhibitors and prostate cancer — The randomized Prostate Cancer Prevention Trial (PCPT) evaluated the use of finasteride, a 5-alpha reductase inhibitor, as a chemopreventive agent in 19,000 men who were at increased risk for prostate cancer [254]. The incidence of prostate cancer was decreased in the finasteride group, compared to men in the placebo group (18.4 versus 24.4 percent) but there was an apparent increase in the absolute number and proportion of high grade tumors with finasteride. (See "Chemoprevention strategies in prostate cancer", section on '5-Alpha reductase inhibitors'.)

An international, double-blind, placebo-controlled chemoprevention trial of dutasteride (REDUCE) in 8336 men found a statistically significant 23 percent decrease in the incidence of prostate cancer [255]. However, the reduction was significant only among lower grade tumors (Gleason scores of 5 or 6) with no significant difference for Gleason scores 7 to 10 [255]. As part of an analysis for the US Food and Drug Administration (FDA), regrading of the specimens using a modified Gleason score subsequently found that there was an absolute increase of 0.5 percent in the incidence of tumors with modified Gleason scores 8 to 10 (RR 2.06, 95% CI 1.13-3.75) [256]. Although detection bias based upon the effect of finasteride on prostate size or PSA sensitivity had been suggested as an explanation for the increased incidence of high grade tumors in the PCPT trial, it would not account for a similar finding in biopsy samples from the REDUCE trial.

The advisory committee to the FDA concluded that finasteride and dutasteride do not have a favorable risk-benefit profile for the chemoprevention of prostate cancer in healthy men, with an estimation that one additional high grade cancer would occur for every three to four lower grade cancers that would be prevented [256].

Metformin — The anti-diabetic drug metformin is associated with reduced cancer incidence in patients with type 2 diabetes.

In an observational study of 4085 type 2 diabetic patients in the United Kingdom who used metformin from 1994 to 2003, investigators found that the risk of cancer was reduced by 40 percent among those who took metformin compared with individually matched diabetic patients who did not take metformin (7.3 versus 11.6 percent, HR 0.46, 95% CI 0.40-0.53) [257].
In a prospective study of 1353 patients with type 2 diabetes in the Netherlands, during 9.6 years of follow-up, cancer mortality was significantly reduced comparing those who used metformin with those who did not (HR 0.43, CI 0.23-0.80) [258].
A systematic review and meta-analysis included 11 independent studies contributing 4042 cases of cancer and 529 deaths in patients with diabetes [259]. Both cancer incidence and cancer mortality were reduced by 30 percent among users of metformin. Significant reduction in the incidence of pancreatic and liver cancer was noted; reductions for breast, colon and pancreas were observed, though were not statistically significant. There was a trend toward a dose-response relationship. Two subsequent meta-analyses in patients with diabetes found an association between metformin use and a significant decrease in colorectal cancer risk [260] and a decrease in risk of colorectal, liver, and lung cancers [261].
A registry-based case control study in the UK reported metformin use decreased pancreatic cancer risk among women by 60 percent but had no significant effect among men [262].

Among postulated mechanisms for such a benefit are the inhibition of cancer cell growth and suppression of HER2 overexpression and inhibition of mTOR [263-265].

These results may reflect increased risk due to use of other regimens for diabetes rather than decreased risk due to use of metformin. Ongoing randomized trials in patients with early stage breast cancer may directly address the benefit in the setting of established cancer therapies [266].

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, “The Basics” and “Beyond the Basics.” The Basics patient education pieces are written in plain language, at the 5th to 6th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10th to 12th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on “patient info” and the keyword(s) of interest.)

Beyond the Basics topics (see "Patient information: Medications for the prevention of breast cancer (Beyond the Basics)")

SUMMARY AND RECOMMENDATIONS — Many cancers are preventable. Basic lifestyle changes can have a tremendous impact on the rates of cancer. The fact that such changes also protect against other chronic diseases (cardiovascular disease, stroke, and diabetes) makes the case for prevention even more compelling.

General lifestyle recommendations include:

Avoid tobacco
Be physically active
Maintain a healthy weight
Eat a diet rich in fruits, vegetables, and whole grains and low in saturated/trans fat
Limit alcohol
Protect against sexually transmitted infections
Avoid excess sun
Get regular screening for breast, cervical, and colorectal cancer

Specific factors associated with cancer risk include the following:

Tobacco use is responsible for 90 percent of all lung cancer deaths and is tied to multiple other cancers. (See 'Tobacco use' above.)
The association of dietary fat, fruits, vegetables, and fiber with cancer risk is largely unconfirmed. Red meat consumption may promote colorectal cancer and a high intake of tomatoes probably decreases prostate cancer risk. (See 'Diet' above.)
Vitamin D may reduce the risk of colorectal cancer. Calcium intake, at a minimum of 700 mg/day, may protect against colorectal cancer but high calcium intake (>2000 mg/day) increases risk for prostate cancer. Folate in diet has been associated with a decreased risk of colon and breast cancer, especially in women who drink alcohol; data on folic acid or multivitamin supplementation are inconsistent. (See 'Vitamins and micronutrients' above.)
Alcohol intake, even in moderate quantities, increases the risk for colon, breast, esophageal, and oropharyngeal cancer. (See 'Other' above.)
Physical activity is inversely related to risk for colon and breast cancer. Excess weight increases the risk of multiple cancers. (See 'Physical activity' above and'Excess weight' above.)
Skin cancer is directly related to sun exposure, and melanoma rates are increasing. A history of blistering sunburns is of particular risk for melanoma; cumulative sun exposure has more impact on non-melanoma cancers. (See 'Excess sun exposure' above.)
HPV, HCV, HTLV1, HIV, EBV, and H pylori have been linked to human cancers. Exposure prevention, screening, vaccination for HPV, and early treatment for abnormal cervical findings and HIV infection can prevent cancer. (See 'Infection' above.)
Chemoprevention may be helpful in high-risk patients, but risks and benefits should be weighed carefully. Aspirin and NSAIDs offer protection against adenomatous polyps and colorectal cancer, and long-term use in low doses likely decreases cancer-related mortality risk from other solid tumors. (See 'Chemoprevention' above.)
Tamoxifen decreases incidence of breast cancer in high-risk women but increases the risk for thromboembolic disease and early-stage endometrial cancer.Raloxifene is an alternative but has not been evaluated in premenopausal women. Aromatase inhibitors also decrease the risk of breast cancer but may cause muscle and joint pain and are not recommended for premenopausal women. (See "Selective estrogen receptor modulators and aromatase inhibitors for breast cancer prevention".)

Use of UpToDate is subject to the Subscription and License Agreement.
REFERENCES