Chemoprevention of Colorectal Cancer

Abstract

Although colorectal cancer (CRC) screening has reduced the incidence and mortality from CRC, chemoprevention strategies have potential to further reduce CRC incidence and mortality. Chemoprevention agents might be used for average-risk as well as high-risk groups, and to prevent colorectal cancer recurrence after therapy. CRC chemoprevention agents that have been studied include aspirin, non-aspirin non-steroidal anti-inflammatory drugs, statins, agents that target metabolic pathways, and vitamins and minerals. We review the prospect of chemoprevention of CRC, results from preclinical and human studies, challenges, and future directions.

Colorectal cancer (CRC) is a significant global health burden and is the second and third most commonly diagnosed cancer worldwide in women and men respectively, with more than 140,000 new cases diagnosed in the United States (US) in 2019.^{1, 2} Developing more efficacious treatments and improving screening have led to reductions in CRC incidence and mortality, although the burden remains significant and additional strategies for CRC prevention are needed.^3–8

One strategy to decrease CRC risk that has been extensively studied over the past several decades is chemoprevention. The term chemoprevention was first coined in 1976,⁹ and refers to the use of a synthetic or natural substance to decrease the risk of developing cancer, delay the time of cancer onset, or to reverse the carcinogenesis process (Figure 1). However, finding an effective chemoprevention agent for cancer is not an easy task—only a small number of cancer chemoprevention agents have been approved by the Food and Drug Administration.¹⁰ There are many important caveats to consider when evaluating potential chemoprevention agents. Given the length of time that chemoprevention must be administered to prevent cancer, especially in average risk populations, it is important for these agents to be well tolerated with minimal side effects (Figure 1). The agents must also be available at low cost and easy to administer, with a convenient dosing schedule. These are high baseline requirements that must be met before clinical trials can even begin to evaluate efficacy.

Figure 1.

Complexities of CRC chemoprevention. There are many factors to consider in evaluating CRC chemoprevention agents, including the intrinsic properties of the agent (side effects, cost, ease of administration), mechanisms of chemoprevention, and groups most likely to benefit with the least risk.

Chemoprevention studies for cancer are challenging; they are often long in duration and require a large population and substantial financial investment. For trials of agents that might prevent CRC, a key aspect of study design is to determine whether the endpoint will be precancerous lesions or CRC. Most CRCs develop from adenomas,¹¹ adenomas increase CRC risk,¹² and adenoma removal decreases CRC risk.^{13, 14} Therefore, it is a logical extension that results from studies showing that agents can prevent adenoma formation would lead to the conclusion that they would reduce CRC. Studying the effects of chemoprevention agents on adenoma development is certainly easier and requires less time and fewer patients than studying effects on CRC development. Another potential premalignancy endpoint is aberrant crypt foci (ACF)—precursors of adenomas and CRC initially discovered in mice and validated in studies of humans.^15–17 However ACFs have not been widely used in most chemoprevention trials given their variable rates of detection and uncertainties regarding their relationship to CRC risk.^{18, 19} Another challenge for chemoprevention studies with a CRC endpoint is that participants may modify their CRC risk by undergoing screening—it would be unethical to withhold CRC screening from 1 group of participants, given its clear benefits.^3–7 There are therefore multiple challenges associated with CRC chemoprevention studies. Despite these challenges, numerous chemopreventive agents (Table 1) have been studied. We review findings from these studies.

Table 1.

Agents for Chemoprevention of CRC

Agent	Primary Target	Mechanisms	Highest level of clinical evidence	Significant concerns	References
Aspirin and NSAIDs
Aspirin	COX1 and COX2 (irreversible inhibition)	inhibit prostaglandin synthesis, platelet activation, Wnt signaling to beta catenin, and inflammation	RCT	increased risk of bleeding	21
Non-selective NSAIDs (non-aspirin)	COX1 and COX2 (reversible inhibition)	inhibit prostaglandin synthesis, platelet activation, Wnt signaling to beta catenin, and inflammation	RCT	increased risk of bleeding	71
COX-2 inhibitors	COX2 (reversible inhibition)	inhibits prostaglandin synthesis, platelet activation, Wnt signaling to beta catenin, and inflammation	RCT	increased cardiovascular risk	71
Metabolic agents
Metformin	inhibits mitochondrial complex I to prevent production of mitochondrial ATP	activates AMPK, which inhibits the mTOR pathway and reduces cyclin D1 expression and RB phosphorylation	multiple observational studies, one small RCT	lactic acidosis, gastrointestinal side effects (nausea, vomiting, diarrhea)	97–100
Statins	HMG-CoA reductase (reversible inhibition)	disruption of the mevalonate pathway with downstream effects on membrane integrity, cell signaling, protein synthesis, and cell cycle progression	RCTs with cardiovascular effects as the primary endpoint	myalgias, rhabdomyolysis (rare), elevated transaminases (rarely severe)	120, 121
Long-chain omega-3 polyunsaturated fatty acids	components of phospholipids that form cell membranes	anti-proliferative, apoptotic, and anti-angiogenic properties	RCT	unpleasant smelling breath and sweat, nausea, diarrhea	147–149
Vitamins and minerals
Vitamin A	combines with retinol-binding protein, a plasma-specific transport protein	regulates nuclear receptors that suppress tumor formation, induces apoptosis, and enhances immune function	RCT	toxicity can lead to liver damage, joint pain, alopecia, headaches, vomiting, skin desquamation	210, 211
Vitamin C	cofactor in collagen formation and tissue repair	reduces oxidative stress, enhances immune system	RCT	toxicity can lead to nausea, vomiting, diarrhea, headaches	212
Vitamin E	primarily ends up in cell and organelle membranes	inhibits lipid peroxidation in cell membranes, reduces oxidative stress, inhibits carcinogen production	RCT	excessive levels can lead to increased risk of hemorrhagic stroke	213, 214
Beta-carotene	functions as a provitamin A	conflicting evidence of pro- and anti-oxidant properties	RCT	2 large-scale intervention studies have shown an increase in lung cancer among smokers and individuals exposed to asbestos with beta-carotene supplementation.	215, 216–217
Selenium	trace mineral necessary to make selenium-containing proteins	antioxidant effects are most likely due to the antioxidant activity of proteins that contain selenium as an essential component and not from selenium itself	RCT	toxicity can lead to gastrointestinal distress, alopecia, and nail discoloration	218
Folic acid	coenzyme or co-substrate in single-carbon transfers in the synthesis of nucleic acids and amino acid metabolism	proposed mechanism is through effects on DNA replication, repair, and methylation through the one-carbon metabolic pathway	RCT	high levels of folate supplementation could hide vitamin B12 deficiency or accelerate progression of neoplastic lesions	158, 164, 165
Calcium	incorporated into the skeleton	bile acid-binding capacity, direct effect on calcium-sensing receptors on colonocytes	RCT	hypercalcemia can cause nausea, vomiting, constipation, bone pain, kidney stones, confusion, and palpitations	178, 179
Vitamin D	regulates gene transcription via binding to vitamin D receptors located in cell nuclei	inhibits proliferation and angiogenesis, induces differentiation and apoptosis	RCT	dose-limiting hypercalcemic effects	176, 181, 219
New agents
DFMO	ornithine decarboxylase (irreversible inhibition)	inhibits polyamine synthesis which is important for cell survival	RCT (final results pending)	ototoxicity	194
Erlotinib	EGFR tyrosine kinase inhibitor (reversible inhibition)	inhibits EGFR signaling	RCT	gastrointestinal side effects, rash	199
Curcumin	inhibits reactive-oxygen-generating enzymes, protein kinase C, EGFR	anti-inflammatory activity, induces apoptosis	RCT	gastrointestinal side effects at higher doses	220
Guselkumab	monoclonal antibody against IL23 subunit alpha	inhibits IL23 signaling	RCT underway	immunosuppression	203–206
Hydrogen sulfide- and nitric oxide-releasing NSAIDs	COX1 and COX2 (reversible inhibition)	inhibits prostaglandin synthesis, platelet activation, Wnt signaling to beta catenin and inflammation	preclinical studies	decreased risk of gastrointestinal injury	207

Aspirin

The agent with the strongest data to support its CRC chemoprevention effects is aspirin. Acetylsalicylic acid, officially named aspirin in 1899,²⁰ irreversibly inhibits cyclooxygenase 1 (COX1) and cyclooxygenase 2 (COX2).²¹ Although there is debate over aspirin’s exact mechanism of chemoprevention, it inhibits several CRC-related pathways, including prostaglandin synthesis, platelet activation, Wnt signaling to beta catenin, and inflammation.²¹ Studies of aspirin prevention of death from all cancers are inconclusive, reporting significant²² and non-significant²³ reductions in death. For CRC specifically, the first report of aspirin’s association with decreased CRC risk was a case–control study in 1988, which showed fewer cases of CRC in individuals taking aspirin-containing medications.²⁴ Studies in the 1990s showed that aspirin reduced chemically induced tumor development in rodents and tumor development in genetically engineered mice with polyposis.^25–27 Since then, findings from extensive preclinical and clinical studies on aspirin led the US Preventative Services Task Force (USPSTF) to recommend aspirin for prevention of cardiovascular disease (CVD) and CRC—this was the first time a chemopreventive agent was endorsed for a non-high risk population.^{23, 28, 29} However, this recommendation was only made for adults 50–59 years old, with a 10% or more 10-year risk of CVD, who are expected to live more than 10 years, without an increased risk of bleeding.^{23, 28, 29} For patients 60–69 years old, aspirin use should be individualized. Given the ongoing accumulation of evidence on aspirin, these recommendations might change.²⁸

Although aspirin chemoprevention studies with CRC as an endpoint have produced mixed results, they have mostly found evidence for an association with decreased risk of CRC (Table 2). A case–control study in Denmark found that individuals who continuously filled low-dose aspirin prescriptions for 5 or more years had a decreased risk of CRC (odds ratio, 0.73; 95% CI, 0.54–0.99).³⁰ The Health Professionals Follow-up Study associated aspirin use with decreased risk of CRC in men (relative risk 0.79; 95% CI, 0.69–0.90) after at least 6 years of use, with increasing benefits with higher doses of aspirin.^{31, 32} However, the association between aspirin use and decreased risk of CRC disappeared 4 years after aspirin cessation. The Nurses’ Health Study found similar results, but the association of aspirin use with CRC risk was only significant after 20 years.³³ Combining data from the Health Professionals Follow-up Study and Nurses’ Health Study demonstrated a reduced risk of CRC from aspirin with as little as 0.5–1.5 tablets of aspirin per week.³⁴ Furthermore, among the more than 140,000 participants in the Cancer Prevention Study II Nutrition Cohort, those taking at least 325 mg of aspirin for 5 or more years had decreased CRC risk (relative risk 0.68; 95% CI, 0.52–0.90).³⁵

Table 2.

Randomized Placebo-controlled Trials of Aspirin

Study	Years conducted	Site	Participants	Study information	Results	Conclusions	References
CRC endpoint
British Doctors Aspirin Trial	1978–1984	United Kingdom	5139 Male physicians, born on or after 1900	500 mg daily; no placebo; median treatment 6 years	significant decrease in CRC incidence with aspirin use (hazard ratio, 0.70)	decreases CRC incidence	38
UK-TIA Aspirin Trial	1979–1986	United Kingdom and Ireland	2449 Older than 40 years, recent transient ischemic attack or minor ischemic stroke	300 mg or 1200 mg daily, median treatment 4.4 years	no significant decrease in CRC incidence with aspirin use (hazard ratio, 0.82)	no decrease in CRC incidence	38
Physicians’ Health Study	1982–1988	US	22071 male physicians 40–84 years old	325 mg every other day, study terminated after mean follow-up of 5 years	after early study termination after 5 years due to a significant decrease in cardiovascular events seen in aspirin group, there was no significant reduction in CRC risk with aspirin use (relative risk, 1.15); in post-trial period, after 12 years of additional follow up, aspirin users still had no significantly decreased CRC risk (hazard ratio 1.03)	no decrease in CRC incidence	41, 42
SALT	1984–1990	Sweden	1360 Recent transient ischemic attack	75 mg daily, median treatment 2.7 years	no significant decrease in CRC incidence with aspirin use (odds ratio, 0.71)	no decrease in CRC incidence	39
Thrombosis Prevention Trial	1989–1997	United Kingdom	5085 Men, ages 45–69 years old, at increased risk of vascular events	75 mg daily, median treatment 6.9 years	significant decrease in CRC incidence with aspirin use (odds ratio, 0.61)	decreases CRC incidence	39
Women’s Health Study	1992–2004	US	39876 Women health care professionals, 45 years or older	100 mg every other day	during the trial there was no reduction in CRC risk with aspirin use (relative risk, 0.97); in post-trial period, after 8 additional years of follow up there was a significant decrease in CRC risk in aspirin users (hazard ratio, 0.58), producing overall risk reduction (hazard ratio, 0.80)	decreases CRC incidence	36, 37
CAPP2	1999–2007	multiple	861 Lynch syndrome, 25 years or older	600 mg daily, median follow-up period, 29 months	initial post-trial analysis showed no reduced CRC risk with aspirin use (relative risk, 1.0); longer follow up of participants using aspirin for at least 2 years found reduced risk (hazard ratio, 0.41)	decreases CRC incidence	55, 56
ASPREE	2010–2017	US and Australia	19114 70 years or older (65 years or older for blacks and Hispanics)	100 mg daily, median follow up 4.7 years	participants using aspirin had an increased risk of developing colorectal cancer (hazard ratio, 1.77)	increases CRC incidence	43
Adenoma endpoint
CALGB	1993–2000	US	517 30–80 years old, history of CRC who recently underwent curative resection	325 mg daily, median time until colonoscopy exam is 12.8 months	aspirin reduced risk of adenoma (relative risk 0.65) and a longer time until an adenoma was detected (relative risk, 0.64),	decreases adenoma risk	49
CAPP1	1993–2003	Europe	133 10–21 years old, FAP with no prior colectomy	600 mg twice daily, colonoscopy after 1 year and then annually	aspirin did not significantly reduce polyp count in the rectum and sigmoid (relative risk, 0.77); treatment for aspirin for more than 1 year significantly decreased the largest polyp size from 6mm to 3mm	does not decrease adenoma formation in FAP	56
Aspirin/Folate Polyp Prevention Study	1994–2001	US and Canada	1121 21–80 years old with a history of a prior colonic adenoma	81 mg or 325 mg daily, colonoscopy after approximately 3 years	after at least 1 year, 81 mg associated with lower risk of any adenoma (relative risk, 0.81), and advanced adenoma (relative risk, 0.59),; 325 mg of aspirin was not associated with reduced risk of adenoma	low-dose, but not high-dose aspirin, decreases risk of adenoma recurrence	44
APACC	1996–2001	France	238 (1 year), 185 (4 years) 18–75 years old with a history of a prior colonic adenoma	160 mg or 300 mg daily, colonoscopy after 1 and 4 years	after 1 year, aspirin users had decreased risk of having 3 or more adenomas (relative risk, 0.30) and at least 1 adenoma greater than 5mm (relative risk, 0.44); there was no significant differences in adenoma recurrence after 4 years	decreases adenoma risk after 1 year, but not 4 years	45, 46
UKCAP Trial	1997–2005	United Kingdom and Denmark	853 younger than 75 years old with a history of a prior colonic adenoma	300 mg daily, colonoscopy after approximately 3 years	aspirin use was associated with a significantly decreased risk of a recurrent adenoma (relative risk, 0.79) and advanced adenoma (relative risk, 0.63)	decreases adenoma risk	50
Japan Colorectal Aspirin Polyps Prevention (J-CAPP)	2007–2009	Japan	311 40–60 years old, prior endoscopically removed colonic adenoma or adenocarcinoma	100 mg daily, colonoscopy after 2 years	aspirin use reduced risk of adenoma and CRC (odds ratio, 0.60), and produced lower risk in nonsmokers (odds ratio, 0.37); no reduced risk of adenoma or CRC in smokers taking aspirin	decreases adenoma and CRC risk, especially in nonsmokers	47

Randomized placebo-controlled trials of aspirin have primarily been secondary analyses of trials with primary CVD endpoints. The large Women’s Health Study found no reduction in CRC risk during the 10-year active trial.³⁶ However a significant reduction in aspirin users was noted after extended post-trial follow-up (hazard ratio, 0.80; 95% CI, 0.67–0.97).³⁷ Although neither the UK transient ischemic attack (UK-TIA) aspirin trial nor the Swedish aspirin low dose trial (SALT) found that aspirin reduced risk of CRC, pooling of the data with the British Doctors Aspirin Trial and the Thrombosis Prevention Trial showed aspirin use decreased CRC incidence (hazard ratio, 0.76; 95% CI, 0.60–0.96) and mortality (hazard ratio, 0.65; 95% CI, 0.48–0.88).^{38, 39} Furthermore, the association between aspirin use and decreased risk of CRC was more pronounced with 5 or more years of aspirin therapy and after 10 years following aspirin initiation.³⁸ Additional meta-analyses also showed that aspirin decreases the incidence and mortality from CRC.^{23, 40} Other studies have failed to show an association such as the Physician’s Health Study, which showed no significant decrease in CRC risk with aspirin use after 12 years of follow up.^{41, 42} Additionally, the Aspirin in Reducing Events in the Elderly (ASPREE) trial, surprisingly, found increased all-cause and cancer-related mortality and increased risk of CRC with aspirin use (hazard ratio, 1.77; 95% CI, 1.02–3.06).⁴³ Studies with a colonic adenoma endpoint also show mixed results (Table 2). The Aspirin Folate Polyp Prevention Study showed low-dose, but not high-dose, aspirin decreased risk of adenoma (relative risk, 0.81; 95% CI, 0.69–0.96) and advanced adenoma/carcinoma (relative risk, 0.59; 95% CI, 0.38–0.92).⁴⁴ The Association pour la Prévention par I’Aspirine du Cancer Colorectal (APACC) trial found individuals with prior colonic adenomas taking lysine acetylsalicylate had decreased risk of having 3 or more adenomas (relative risk 0.3; 95% CI, 0.10–0.89) or an adenoma of 5 mm or larger (relative risk 0.44; 95% CI, 0.24–0.82) after 1 year,⁴⁵ but no decrease in risk after 4 years.⁴⁶ In a Japanese cohort with endoscopically resected colonic adenomas or adenocarcinoma, aspirin reduced the risk of developing a colonic adenoma or adenocarcinoma by 40% (95% CI, 0.36–0.98).⁴⁷ Interestingly, subgroup analyses found more prominent risk reduction in current nonsmokers, whereas aspirin paradoxically increased CRC risk in smokers. Smoking also negated the protective effects of aspirin in a US-based population.⁴⁸

In the Colorectal Adenoma Prevention Study from the cooperative trials group Cancer and Leukemia Group B (CALGB), individuals with prior CRC randomly assigned to groups that received aspirin had a lower risk of developing colonic adenomas (relative risk, 0.65; 95% CI, 0.46–0.91) and developed adenomas at later times (relative risk, 0.64; 95% CI, 0.43–0.94).⁴⁹ Furthermore, the United Kingdom Colorectal Adenoma Prevention (UKCAP) study showed that aspirin decreased risk of adenoma (relative risk, 0.79; 95% CI, 0.63–0.99) and advanced adenoma (relative risk, 0.63; 95% CI, 0.43–0.91).⁵⁰ Similar results were observed in other studies^{32, 51} and are supported by a meta-analysis.⁵² Data from the Nurses’ Health Study showed the greatest reduction in adenoma risk for participants who took 14 or more tablets of aspirin per week (relative risk, 0.49; 95% CI, 0.36–0.65).⁵³

Aspirin chemoprevention of CRC has also been studied in high-risk groups. The Colorectal Adenoma/Carcinoma Prevention Programme 1 (CAPP1) study found that aspirin did not decrease colonic polyp burden in familial adenomatous polyposis (FAP), but 1 year or more of aspirin use decreased the largest polyp size.⁵⁴ The CAPP2 study, which has been the only randomized placebo-controlled trial of aspirin with CRC as the primary endpoint, investigated use of 600 mg of aspirin daily in individuals with Lynch syndrome. Although the initial post-intervention analysis showed no significant differences,⁵⁵ later analysis showed that for participants of at least 2 years, aspirin significantly reduced CRC risk (hazard ratio, 0.41; 95% CI, 0.19–0.86).⁵⁶ Aspirin also reduced CRC risk in patients with Lynch syndrome in the Colon Cancer Family Registry.⁵⁷ A meta-analysis showed that, in individuals with a prior history of CRC, aspirin decreased CRC mortality (hazard ratio, 0.76; 95% CI, 0.66–0.88).⁵⁸ Interestingly, among aspirin users, those with colorectal tumors with mutations in PIK3CA had longer survival times than patients whose tumors did not have PIK3CA mutations.⁵⁹

One concern with long-term aspirin use is the side effect of gastrointestinal bleeding. Participants in the ASPREE trial who took 100 mg of aspirin daily had an increased risk of major hemorrhage (hazard ratio, 1.38; 95% CI, 1.18–1.62).⁶⁰ Similarly participants in the Use of Aspirin to Reduce Risk of Initial Vascular Events in Patients at Moderate Risk of Cardiovascular Disease (ARRIVE) trial who took 100 mg of aspirin daily had more bleeding events (0.97% vs 0.46%; hazard ratio, 2.11; 95% CI, 1.36–3.28), but no difference in rate of fatal bleeding.⁶¹ Aspirin users in the Women’s Health Study had a slightly increased risk of gastrointestinal bleeding and peptic ulcer disease, but no difference in deaths related to gastrointestinal bleeding,³⁷ whereas other studies found no significant increase in any serious adverse event, including gastrointestinal bleeding.⁴⁴

The Aspirin for Dukes C and High Risk Dukes B Colorectal Cancers (ASCOLT) study has examined the effects of aspirin in adjuvant treatment for individuals with previously treated CRC.⁶² Similarly, the Dutch trial A Trial of Aspirin on Recurrence and Survival in Colon Cancer Patients (ASPIRIN)⁶³ and the ADD-Aspirin trial⁶⁴ are each investigating whether aspirin can prevent recurrence of CRC in patients with non-metastatic CRC who underwent treatment. The CAPP3 study is exploring lower doses (100 mg and 300 mg) of aspirin for chemoprevention of CRC in patients with Lynch syndrome.⁶⁵ Furthermore, the Aspirin Intervention for the Reduction of Colorectal Cancer Risk (ASPIRED) study of individuals with prior colonic adenomas aims to clarify the mechanism of aspirin chemoprevention through biosample collection.⁶⁶ Given the surprising increased cancer-related mortality observed in the ASPREE trial, the ASPREE-XL study will provide longitudinal follow-up data to clarify these findings. Clarification of the mechanisms of aspirin chemoprevention is also important, because a study demonstrated that inhibition of platelet aggregation alone with clopidogrel reduced CRC risk, similar to aspirin.⁶⁷ Additional important questions include better defining what subgroups would benefit most, because aspirin reduced risk for colorectal tumors that express high levels of COX2,⁶⁸ whereas other studies found that aspirin effects on CRC risk varied based on 2 single nucleotide polymorphisms.⁶⁹ Studies of aspirin use to augment CRC screening found aspirin use to be as effective as screening with the fecal occult blood test or flexible sigmoidoscopy in reducing CRC mortality, although these findings require validation.⁷⁰

Studies of aspirin have indicated its chemopreventive effects for CRC. Aspirin likely leads to a modest reduction in adenoma and CRC risk, although only after prolonged, continuous use and accompanied by an increased risk of bleeding. Currently, aspirin is recommended for CRC chemoprevention in select individuals with increased CVD risk, as well as individuals with Lynch syndrome. However, there are many questions that remain, including the dose and frequency of aspirin needed, the target population, and the magnitude of the risk reduction with aspirin use in individuals undergoing screening.

Non-aspirin NSAIDs

Non-aspirin NSAIDs inhibit COX1 and COX2, except unlike aspirin this inhibition is competitive.⁷¹ In the 1970s it was recognized that colorectal tumors express high levels of prostaglandin E₂ (PGE₂).^{72, 73} PGE₂ synthesis requires COX2, so NA-NSAIDs might prevent development of CRC. Studies in rodent models found NA-NSAIDs to prevent enteric tumorigenesis to some degree.^74–76 In case–control studies in humans, NA-NSAID use was associated with decreased risk of CRC. A case–control study from Denmark found individuals who filled NA-NSAID prescriptions to have a reduced risk of CRC (odds ratio, 0.57; 95% CI, 0.44–0.74), especially those with prescriptions for long-term, high-dose NSAIDs.³⁰ Women in Wisconsin who used NA-NSAIDs at least twice per week for 1 year or more were less likely to receive a diagnosis of CRC (odds ratio, 0.43; 95% CI, 0.20–0.89).⁷⁷ Case–control studies also found that use of a broad range of NSAIDs, including selective and non-selective COX2 inhibitors, reduced risk of CRC⁷⁸ and adenoma⁵¹. A recent meta-analysis of 23 studies demonstrated that NA-NSAIDs decreased risk of CRC (odds ratio, 0.74; 95% CI, 0.67–0.81), and subgroup analysis found the largest protective effects in women, in patients taking higher doses of NSAIDs, in patients of white race, and against distal colorectal tumors.⁷⁹

In patients with FAP, 9 months treatment with the NA-NSAID sulindac significantly reduced colon polyp number and diameter (Table 3).⁸⁰ A similar reduction in polyp number were observed in the residual rectum of patients with FAP after longer term follow up.⁸¹ However, sulindac did not prevent the initial development of adenomas in patients who had not yet developed the colonic polyposis phenotype.⁸² Sulindac also significantly decreased the formation of ACF, and in individuals who had colon polyps removed, this drug significantly decreased risk for subsequent colonic adenomas.⁸³ The chemopreventive effects of other non-selective NA-NSAIDs have been studied in high-risk populations. A study of individuals with Lynch syndrome from the Colon Cancer Family Registry showed that taking ibuprofen for 1 month or more was associated with decreased risk of CRC.⁵⁷ A phase IB study of patients with Lynch syndrome is underway to examine the preventative effects of naproxen.⁸⁴

Table 3.

Randomized Placebo-controlled Trials of Non-aspirin NSAIDs

Study	Years conducted	Site	Participants	Study information	Study results	Conclusions	References
Primary Chemoprevention of Familial Adenomatous Polyposis with Sulindac	1993–2001	US	41 8–25 years old, FAP prior to the start of adenoma formation in the colon	75 mg or 150 mg sulindac twice daily, 4 years of treatment	sulindac did not decrease frequency of participants developing rectosigmoid colonic adenomas or significantly change mean number or size of polyps	sulindac does not prevent development of adenomas in patients with FAP	82
Effects of celecoxib, a COX2 inhibitor, in patients with FAP	1996–1998	US and United Kingdom	77 18–65 years old, FAP with some residual colorectum	100 mg or 400 mg celecoxib twice daily, for 6 months	individuals taking 400 mg of celecoxib twice daily had a 28% decrease in mean number of polyps and a 31% reduction in total polyp diameter; no significant reductions with 100 mg of celecoxib twice daily	high-dose celecoxib decreases adenomas in patients with FAP	90
APC Trial	1999–2002	US, Canada, Australia, United Kingdom	2035 31–88 years old, with multiple prior colonic adenomas or a colonic adenoma 5mm or greater in size	200 mg or 400 mg celecoxib twice daily, colonoscopy after 1 year and 3 years	after 3 years, celecoxib use significantly decreased risk of adenoma detection for low dose (relative risk, 0.67), and high dose (relative risk, 0.55); increase in cardiovascular events at low dose (relative risk, 2.6) and high dose (relative risk, 3.4)	celecoxib decreases adenoma risk but increases cardiovascular complications	87
Adenomatous Polyp Prevention On Vioxx (APPROVE) Trial	2000–2004	29 countries	2587 40 years or older with a prior history of a colonic adenoma	25 mg rofecoxib daily, colonoscopy after 1 year and 3 years	rofecoxib reduced risk of adenoma recurrence (relative risk, 0.76), and advanced adenoma (relative risk, 0.70),; increased thrombotic cardiovascular risk (relative risk, 1.89) and upper gastrointestinal serious events (relative risk, 4.91)	rofecoxib decreases adenoma risk but increases cardiovascular and gastrointestinal complications	92
Prevention of Colorectal Sporadic Adenomatous Polyps Trial	2001–2005	32 countries	1561 30 years or older with a prior history of a colonic adenoma	400 mg celecoxib daily, colonoscopy after 1 year and 3 years	celecoxib decreased risk of finding any adenoma (relative risk, 0.64) and advanced adenoma (relative risk, 0.49) with nonsignificant increase in cardiovascular risk	celecoxib decreases adenoma risk	86

Given the gastrointestinal bleeding risks associated with non-selective NA-NSAIDs,⁸⁵ likely related to COX1 inhibition, selective COX2 inhibitors might be safer chemopreventive agents. The Prevention of Colorectal Sporadic Adenomatous Polyps trial found that celecoxib reduced adenoma detection after 3 years (relative risk, 0.64; 95% CI, 0.56–0.75).⁸⁶ The Adenoma Prevention with Celecoxib (APC) trial also found that celecoxib decreased adenoma detection after 3 and 5 years.^87,88 Interestingly in this cohort, polyps that expressed COX2 or did not express 15-prostaglandin dehydrogenase, which degrades PGE₂, had a reduced risk of adenoma formation.⁸⁹ High-dose celecoxib was also effective in patients with FAP, decreasing colorectal polyp burden by 31% after 6 months.⁹⁰ Rofecoxib also reduced adenoma formation,^{91, 92} but it was withdrawn from the market by the Food and Drug Administration due to increased cardiovascular risk.⁹³ Similar safety concerns were raised in the APC trial, in which celecoxib use was associated with an increased risk of cardiovascular events.⁹⁴ Ultimately, a review by the USPSTF stated that although NA-NSAIDs, including selective COX2 inhibitors, reduced the risk of colonic adenomas and CRC, these agents were associated with significant cardiovascular events and increased risk of gastrointestinal side effects.⁹⁵ Due to these concerns, NA-NSAIDs have not been widely accepted as potential CRC chemopreventive agents except for in high-risk populations such as individuals with FAP.

Metabolic Agents

Metformin

Type 2 diabetes is an independent risk factor for CRC leading to interest in anti-diabetic medications as potential CRC chemoprevention agents.⁹⁶ Metformin is a biguanide compound often used first-line in type 2 diabetes. Proposed anti-neoplastic properties include activation of adenosine monophosphate-activated protein kinase (AMPK), which inhibits the mammalian target of rapamycin (mTOR) pathway to prevent cell proliferation.^97–99 Metformin might also slow tumor growth by inhibiting cyclin D1 expression or Rb phosphorylation.¹⁰⁰ Metformin reduced polyp formation in a genetic mouse model of FAP¹⁰¹ and reduced ACF in mice given a chemical carcinogen.⁹⁹ However, results from epidemiology and clinical studies vary. Epidemiology studies showed a decreased risk of CRC with metformin use,^102–108 no association,^109–111 or increased risk.^{112, 113} Observational studies are subject to multiple time-related biases (immortal time bias, time window bias, and time-lag bias). A study of 47,531 patients with type 2 diabetes that used methods to reduce time-related biases showed decreased risk of CRC with metformin use for 5 or more years in men (hazard ratio, 0.65; 95% CI, 0.45–0.94), but not in women.¹¹⁴ Many epidemiology studies have not accounted for potential cancer-modifying effects of concomitant anti-diabetic medications. Restricting a larger meta-analysis to studies adjusting for concomitant anti-diabetic medication use showed a more modest (11%) reduction in CRC incidence among metformin users (adjusted odds ratio, 0.88; 95% CI, 0.78–0.99).¹¹⁵

Metformin has also been evaluated for its chemopreventive effect on adenomas. A meta-analysis of 10 studies found an inverse association between metformin use and colorectal adenoma risk (odds ratio, 0.76; 95% CI, 0.63–0.92); the protective effect remained after adjusting for confounding variables. Subgroup analyses were performed in 3 groups: those with diabetes, those with a history of CRC or adenomas, and those with no history of adenomas regardless of diabetes status.¹¹⁶ Metformin reduced risk of adenomas in patients with diabetes (odds ratio, 0.75; 95% CI, 0.62–0.91) with a trend toward lower risk of adenomas in metformin users with a history of CRC/adenomas.

The first randomized controlled trial (RCT) to examine the chemoprevention effect of low-dose metformin (250 mg/day) on metachronous colorectal polyps or adenomas showed that over 1 year, the incidence of total polyps (relative risk, 0.67; 95% CI, 0.47–0.97), as well as adenomas alone (relative risk, 0.60; 95% CI, 0.40–0.92), were significantly lower with metformin use.¹¹⁷ Limitations of this study included the small sample size, short length of follow up, and generalizability due to the high-risk population (approximately 70% with a history of advanced adenoma or multiple adenomas). This RCT was included in a recent meta-analysis that found reductions in risk of adenoma (odds ratio, 0.75; 95% CI, 0.59–0.97) and colorectal tumors (odds ratio, 0.87; 95% CI, 0.70–0.87) in patients with type 2 diabetes taking metformin.¹¹⁸ Overall, the data support the use of metformin as a first-line oral agent in diabetic patients at high risk for CRC. For the general population, definitive RCTs are needed to better establish the chemopreventive effect of metformin, which should include populations at average risk and high risk for CRC and adenomas.

Statins

Statins are 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase inhibitors commonly prescribed for their lipid-lowering properties. Approximately 25% of Americans older than 40 years take a statin for CVD prevention.¹¹⁹ Statins competitively inhibit HMG-CoA reductase, the rate-limiting enzyme of the mevalonate pathway;¹²⁰ disruptions of this pathway in neoplastic cells might reduce tumor initiation, growth, or metastasis.¹²¹ Statins inhibited proliferation of human CRC cell lines and promoted apoptosis.^{122, 123} In mice with carcinogen-induced or genetically induced colorectal neoplasia, statins reduced colorectal tumor development alone¹²⁴ or in combination with NSAIDs.^{125, 126}

Similar to studies of metformin, however, results from epidemiologic and clinical studies of statins and colorectal neoplasia have produced inconsistent results. These include observational studies and secondary outcomes from RCTs examining effects of statins on cardiovascular events. A well-publicized observational study from northern Israel found that 5 or more years of self-reported statin use was associated with a 45% reduction in CRC risk (odds ratio, 0.55; 95% CI, 0.40–0.74).¹²⁷ Interestingly, a subsequent study using the same population showed that specific polymorphisms in the HMG-CoA reductase gene modify the protective association between statins and CRC risk.¹²⁸ A study of a large cohort of US veterans also showed a similar reduction in CRC risk with statin use (hazard ratio, 0.65; 95% CI, 0.55–0.78).¹²⁹ In contrast, several meta-analyses of case–control and cohort studies have shown smaller risk reductions (risk estimates 0.86–0.91)^130–132 or no association.^{133, 134} The inconsistent results from observational studies could result from healthier behaviors among statin users compared with non-users, residual confounding from over the counter NSAID use that is difficult to capture with administrative data, different hydrophilicity of specific statins,^{135, 136} differential effects on colon vs rectal cancers,^{127, 137} or issues with selection bias and immortal-time bias.¹³⁸

Analyses of data from RCTs evaluating statins and cardiovascular events have provided opportunities to examine effects of statin use on overall and subtype-specific cancer risk. Multiple meta-analyses have found no association between statin use and CRC risk,^{130, 131, 133} including a meta-analysis of 27 RCTs published by the Cholesterol Treatment Trialists’ Collaboration.¹³⁹ One of the main criticisms of conclusions made from these data is the short duration of follow up (often less than 5 years).¹⁴⁰ However, 2 RCTs with longer follow-up times (the West of Scotland Coronary Prevention Study¹⁴¹ and the Heart Protection Study)¹⁴² found no difference in CRC incidence after 10 and 11 years follow up, respectively. In addition, data on cancer incidence and mortality are not systematically collected in RCTs of cardiovascular effects, leading to possible ascertainment bias, and many of these trials are conducted in high-risk populations with competing risks.¹⁴⁰

A few studies have evaluated the effects of statin use on incidence of colorectal adenomas and yielded mixed results. A study of 2626 veterans with colon adenomas removed at index colonoscopy found a 49% reduction in recurrence of adenomas and a 29% reduction in risk of advanced adenomas in subjects who took statins continuously for 3–5 years.¹⁴³ In contrast, secondary analyses of multiple adenoma chemoprevention trials found no association between statin use and recurrent, multiple, or advanced adenomas.¹⁴⁴ Analyses of data from the APC trial found a 39% increase in risk of adenomas with statin use for 3 or more years.¹⁴⁵ So, we cannot unequivocally confirm the significant association between statin use and decreased risk of CRC or adenoma. Well-designed studies are needed to determine the effects of long-term statin use, necessary dosage, ideal combinations with other chemopreventive agents, and subgroups most likely to benefit from statin therapy.¹⁴⁰

Long-chain omega-3 polyunsaturated fatty acid supplements

Long-chain omega-3 polyunsaturated fatty acids, which reduce inflammation, are predominantly found in dietary sources such as dark fish.¹⁴⁶ Eicosapentanoic acid (EPA) and docosahexanoic acid (DHA) are used in treatment for coronary heart disease and hypertriglyceridemia, with good safety profiles.¹⁴⁷ Studies of rodents and CRC cell lines demonstrated the antineoplastic effects of EPA and DHA, including anti-proliferative, apoptotic, and anti-angiogenic properties.^147–149 A proof of concept RCT of EPA (2 g/day for 6 months) for patients with FAP found significant reductions in number and size of rectal adenomas.¹⁵⁰

Based on this information, data from the Vitamins and Lifestyle study, from approximately 68,000 residents from Washington state 50–76 years old, were examined for an association between EPA and DHA intake (via fish oil supplements) or dark fish consumption and CRC risk.¹⁵¹ Individuals who took fish oil supplements (4 or more days/week) for 3 or more years, compared with non-users, had a 49% lower risk of CRC (95% CI, 0.26–1.00). This was due to decreased risk of tumors in the colon (hazard ratio, 0.37; 95% CI, 0.15–0.91) rather than rectum (hazard ratio, 0.98; 95% CI, 0.35–2.69), and reductions in CRC in men (hazard ratio, 0.22; 95% CI, 0.06–0.90) rather than in women (hazard ratio, 0.85; 95% CI, 0.39–1.80). The multicenter, randomized Systematic Evaluation of Aspirin and Fish Oil (SEAFOOD) Polyp Prevention trial compared the effects of EPA and ASA, alone or in combination, vs placebo in prevention of colorectal adenomas. It found no effect for EPA or ASA on the adenoma detection rate at the 1-year surveillance colonoscopy. However, there was a difference based on polyp histology and location.¹⁵² EPA was significantly associated with decreases in left-sided and conventional adenomas, but not right-sided or serrated lesions, whereas ASA significantly reduced numbers of conventional adenomas, serrated lesions, and right-sided polyps. Despite their excellent safety and tolerability profiles, further studies are needed on the effects of EPA and DHA in CRC and adenoma chemoprevention.

Vitamins and Minerals

Antioxidants

Antioxidants are found in fruits, vegetables, and over-the-counter dietary supplements. High concentrations of free radicals in cells can lead to DNA, protein, and cell membrane damage; antioxidants help to reduce this oxidative stress by neutralizing free radicals (Table 1). Multiple observational studies of the role of antioxidants (vitamins A, C, and E, beta-carotene, and selenium) in CRC prevention have yielded mixed results. A pooled analysis of 13 cohort studies found that intake of vitamins A, C, and E from diet alone were not associated with CRC risk.¹⁵³ However, total intake of vitamins C and E from diet and supplements combined showed a modest decrease in CRC risk (vitamin C relative risk, 0.81; 95% CI, 0.71–0.92 and vitamin E relative risk, 0.78; 95% CI, 0.66–0.92). After adjustment for total folate intake, the relative risk, for vitamins C and E decreased slightly, but remained significant.

Observational studies of vitamins and minerals are particularly prone to recall bias and difficulties in accurate assessment of dietary intake. There is also potential confounding by other healthy behaviors shared among individuals who regularly take dietary supplements and information on CRC screening practices and over-the-counter NSAID use is often lacking. Therefore, RCTs provide the most reliable source for information about antioxidants. Meta-analyses of RCTs examining the effects of antioxidants on CRC and adenoma risks have not produced encouraging results.^{154, 155} A meta-analysis of 12 RCTs concluded that vitamins A, C, and E, selenium, and beta-carotene are not effective chemopreventive agents for colorectal neoplasia in the general population, alone, or in combination with other antioxidants or chemopreventive agents.¹⁵⁵ In fact, 1 study included in this meta-analysis showed an increased risk of adenoma development among users of vitamin E (relative risk, 1.74; 95% CI, 1.09–1.79) or vitamin E plus beta-carotene (relative risk, 1.63; 95% CI, 1.01–2.64).¹⁵⁶ This evidence is consistent with the USPSTF recommendations against the use of beta-carotene and vitamin E for cancer prevention and their conclusions that there is insufficient evidence to recommend multivitamins or other single- or paired-nutrient supplements as chemopreventive agents.¹⁵⁷

Folic Acid

In addition to containing antioxidants, fruits and vegetables are good sources of folate (folic acid), which is part of the 1-carbon metabolic pathway required for DNA synthesis, repair, and methylation.¹⁵⁸ Disruptions of this pathway can contribute to carcinogenesis. Epidemiology studies have associated a low-folate diet with increased risk of colorectal neoplasia.^159–161 The Nurses’ Health Study¹⁶² and the Canadian National Breast Screening Study¹⁶³ found an inverse association between folate intake and CRC risk (relative risk, 0.25; 95% CI, 0.1–0–.51 and relative risk 0.69; 95% CI, 0.52–0.93, respectively). However, rodent studies have found folate deficiency to reduce risk, and folate supplementation to increase risk, of colorectal tumor development.^{158, 164, 165} Folate intake might therefore protect against adenoma formation but promote progression of existing colorectal neoplasias.¹⁶⁶

RCTs examining the effect of folic acid supplementation on recurrence of colorectal adenomas as the primary endpoint have reported conflicting results, ranging from a 56% decrease in adenoma recurrence,¹⁶⁷ to a 67% increase in advanced adenomas,¹⁶⁸ to no significant effect.^{50, 169, 170} A meta-analysis confirmed that there were no significant effects on adenoma recurrence in high-risk (relative risk, 0.93; 95% CI, 0.61–1.41) or average-risk (relative risk 1.13; 95% CI, 0.77–1.64) populations.¹⁷¹ This finding was replicated in a meta-analysis of 8 RCTs,¹⁷² in which subgroup analyses found no effects of ethnicity, sex, or body mass index. So, there is no convincing evidence that folic acid is an effective chemopreventive agent for CRC or adenomas in average-risk or high-risk populations. Studies to evaluate the effects of folate intake should include longer durations of follow up, consider various methods of folate status assessment,¹⁷³ and might need to account for the effects of polymorphisms in genes involved in the folate metabolism pathway.¹⁷⁴

Calcium and Vitamin D

Mouse studies have shown calcium and vitamin D to have potential anti-neoplastic effects in the colon.^175–177 Calcium might prevent colorectal carcinogenesis via its bile acid-binding capacity and/or direct effects on calcium-sensing receptors on colonocytes.^{178, 179} Colonocytes express vitamin D receptors and activation of these receptors inhibits proliferation and angiogenesis, induces differentiation, and promotes apoptosis in epithelial tissues.^{176, 180, 181} Epidemiology studies have found calcium and vitamin D to reduce the risk of colorectal neoplasia by 20%–30%,^182–184 although findings from some of the studies were limited to the distal colon and rectum.^{185, 186} Data from RCTs, however, have been inconsistent. The Calcium Polyp Prevention Study Group trial assigned 930 individuals with prior adenomas to groups given 3 g calcium carbonate or placebo and found a significantly decreased risk of adenomas in the calcium group (relative risk, 0.85; 95% CI, 0.74–0.98).¹⁸⁷ In contrast, in the Women’s Health Study,¹⁸⁸ in which approximately half of participants received 500 mg of calcium carbonate and 200 IU vitamin D₃ twice daily vs placebo for an average of 7 years, no significant difference was observed in CRC incidence between the groups. Inconsistent results from some RCTs have been attributed to short duration of follow up and possible suboptimal doses of calcium and vitamin D.¹⁶⁶

After the success of the Calcium Polyp Prevention Study Group trial,¹⁸⁷ Baron et al¹⁸⁰ examined the effects of calcium and vitamin D on prevention of adenomas in individuals with a history of 1 or more adenoma removal within 120 days prior to enrollment. They addressed the concern of suboptimal doses of calcium and vitamin D in earlier studies by substantially increasing both doses, yet remaining within what was believed to be a safe range for daily intake (2000 IU vitamin D and 2.5 g calcium). Participants were randomly assigned to groups that received vitamin D₃ (1000 IU), calcium carbonate (1200 mg), calcium plus vitamin D₃, or placebo. There was no significant reduction in adenoma risk in any of the groups over a 3–5-year period. One potential explanation for this observation could be that calcium and vitamin D₃ affect later stages of adenoma development. Secondary analyses of data from this trial examined post-treatment occurrence of conventional adenomas¹⁸⁹ and serrated polyps.¹⁹⁰ The relative risk of conventional adenomas did not differ significantly among groups.¹⁸⁹ However, the calcium-alone groups had a significant increase in sessile serrated adenomas and polyps (relative risk, 2.65; 95% CI, 1.43–4.91), as did the groups that received calcium plus vitamin D₃ (relative risk, 3.81; 95% CI, 1.25–11.64).¹⁹⁰ One limitation of the serrated polyp analysis was that the original trial was powered to evaluate conventional adenomas, which are more common. In addition, detection of sessile serrated adenomas and polyps has continued to improve over time.¹⁹⁰ The USPSTF and the Institute of Medicine have called for new trials of higher doses of vitamin D to clarify the risk:benefit ratio of vitamin D supplementation in prevention of cancer and cardiovascular risk.¹⁹¹ The vitamin D and omega-3 trial was designed to evaluate the effects of vitamin D₃ (2000 IU/day), with or without marine omega-3 fatty acids (1000 mg/day), vs placebo on primary prevention of CVD and invasive cancer (any type).¹⁹² The investigators found no significant difference in incidence of CRC among groups. So, there are not enough data to support use of calcium or vitamin D supplements for prevention of CRC or adenomas. Further studies are needed to define the potential role of calcium and vitamin D in chemoprevention of colorectal neoplasia, such as investigations into different effects on tumor or polyp location, polyp histology, and modification of effect by variants in the vitamin D receptor gene.¹⁹³

Agents in Development

Many of the newest chemopreventive agents for colorectal neoplasia have been or are being tested in patients with FAP. The agents with the most promise are difluoromethylornithine (DFMO) and erlotinib. DFMO is an irreversible inhibitor of ornithine decarboxylase, an enzyme required for polyamine synthesis. Adenomatous colon polyps and colorectal tumors have increased levels of ornithine decarboxylase and polyamines compared with the normal mucosa.¹⁹⁴ A proof of principle trial, published in 2008,¹⁹⁵ found DFMO to significantly reduce adenoma recurrence in individuals with prior adenomas. In an international RCT, celecoxib was compared to celecoxib plus DFMO in patients with FAP.¹⁹⁶ A separate RCT evaluated the efficacy and safety of the combination of DFMO and sulindac, compared with each monotherapy, for 2 years in individuals with FAP.¹⁹⁷ Celecoxib in combination with DFMO reduced the mean number of adenomas that developed in patients by 13%, but this value was not significantly different than for celecoxib alone.¹⁹⁶ Final results of the DFMO and sulindac trial are pending. The main concern about DFMO is associated ototoxicity. Erlotinib is a tyrosine kinase inhibitor that blocks epidermal growth factor receptors. Findings from studies of APC mutant zebrafish¹⁹⁸ and human colon cancer cell lines^{198, 199} indicated that APC inactivation and EGFR signaling promote COX2 expression and development of intestinal neoplasia. In a trial of the effects of sulindac (150 mg twice daily) plus erlotinib (75 mg daily) vs placebo on development of duodenal neoplasia in patients with FAP,²⁰⁰ researchers found a 70% reduction (95% CI, 29%–109%) in colorectal polyp burden in trial subjects.²⁰¹

Other agents studied in patients with FAP include curcumin and guselkumab. An RCT found no significant difference in mean number or size of polyps in 44 patients with FAP given 100% pure curcumin (1500 mg twice daily) vs placebo for 12 months.²⁰² Interleukin 23 (IL23) signaling promotes CRC progression^203–205; a trial is underway to investigate the effects of guselkumab, an antibody against IL23, in patients with FAP (clinicalTrials.gov no: ).²⁰⁶ Researchers have also aimed to develop NSAIDs with improved gastrointestinal safety profiles, such as hydrogen sulfide- and nitric oxide-releasing NSAIDs. These compounds have decreased risks for producing gastrointestinal injury and reduce numbers of ACF and adenomas in rodents and human cancer cell lines^207–209 However, these agents are not ready for widespread clinical use.

Future Directions

Studies of CRC chemoprevention have been increasing. Although aspirin and NA-NSAIDs have shown the most promise, recommendations for their use as chemopreventive agents have been limited to individuals with increased risk of CVD or CRC predisposition syndromes, such as Lynch syndrome or FAP. For most people, the ideal CRC chemoprevention agent is elusive. Challenges to CRC chemoprevention include identification of new targetable neoplastic pathways in the colon and ways to use combinations of agents to increase efficacy and minimize toxicity. Given the overall low risk of CRC in average-risk populations, it is important to determine whether more common intermediate endpoints, such as ACF or adenomas, can be used. It is also important to identify subgroups most likely to benefit from chemoprevention agents with the lowest level of risk—possibly based on genetic factors that affect response to therapy or history of polyps and polyp subtype.

CRC chemoprevention studies face challenges such as the need for funding to support long studies that enroll large numbers of patients, and the need to validate findings in different ethnic groups and in different locations. It is important to collect accurate data on risk—many potential chemoprevention agents are available as over the counter medications or supplements and their widespread use can confound study results. It is unlikely that chemoprevention will ever replace CRC screening as the primary method for prevention. Increases in uptake of screening and decreases in CRC incidence and mortality will make it even more difficult to demonstrate the efficacy of chemoprevention strategies in clinical trials. Studies in populations undergoing regular CRC screening will therefore need to demonstrate a larger protective effect to show significant chemoprevention in addition to screening. In conclusion, the ideal chemopreventive agent for CRC is one that is broadly effective, safe, inexpensive, widely available, and easy to administer. Although finding a chemoprevention agent that satisfies these criteria is challenging, the possibility of decreasing CRC risk and reducing its morbidity and mortality make CRC chemoprevention an endeavor worthy of continued pursuit.

Abbreviations:

ACF

aberrant crypt foci

CRC

colorectal cancer

CI

confidence interval

COX

cyclooxygenase

CVD

cardiovascular disease

DHA

docosahexanoic acid

DMFO

difluoromethyl ornithine

EGFR

epidermal growth factor receptor

EPA

eicosapentanoic acid

FAP

familial adenomatous polyposis

NSAID

non-steroidal anti-inflammatory drug

NA-NSAID

non-aspirin non-steroidal anti-inflammatory drug

PGE₂

prostaglandin E2

US

United States

USPSTF

US Preventive Services Task Force