Aspirin and the prevention of colorectal neoplasia

Abstract

Baron JA, Cole BF, Sandler RS, Haile RW, Ahnen D, Bresalier R, McKeown-Eyssen G, Summers, RW, Rothstein R, Burke CA, Snover DC, Church TR, Allen JI, Beach M, Beck GJ, Bond JH, Byers T, Greenberg ER, Mandel JS, Marcon N, Mott LA, Pearson L, Saibil F, van Stolk RU (Norris Cotton Cancer Center, Dartmouth-Hitchcock Medical Center, Lebanon, New Hampshire; Dartmouth Medical School, Hanover, New Hampshire; University of North Carolina School of Medicine, Chapel Hill; University of Southern California School of Medicine, Los Angeles; Veterans Affairs Medical Center, Denver; University of Colorado School of Medicine, Denver, Colorado; Henry Ford Health Sciences Center, Detroit, Michigan; University of Toronto, Toronto, Ontario, Canada; the Roy J. and Lucille A. Carver College of Medicine, University of Iowa, Iowa City; Cleveland Clinic Foundation, Cleveland, Ohio; Fairview Southdale Hospital, Minneapolis; University of Minnesota, Minneapolis; University of Minnesota School of Medicine, Minneapolis; Minnesota Gastroenterology, Minneapolis; the Veterans Affairs Medical Center, Minneapolis, Minnesota; and Rollins School of Public Health, Emory University, Atlanta, Georgia). A randomized trial of aspirin to prevent colorectal adenomas. N Engl J Med 2003;348:891–899.Sandler RS, Halabi S, Baron JA, Budinger S, Paskett E, Keresztes R, Petrelli N, Pipas JM, Karp DD, Loprinzi CL, Steinbach G, Schilsky R (Departments of Medicine and Epidemiology, University of North Carolina, Chapel Hill; Cancer and Leukemia Group B Statistical Center and Department of Biostatistics and Bioinformatics, Duke University Medical Center, Durham, North Carolina; Departments of Medicine and Community and Family Medicine, Dartmouth Medical School, Hanover, New Hampshire; Wake Forest University School of Medicine, Winston-Salem, North Carolina; Weill Medical College of Cornell University, New York; Roswell Park Cancer Institute, Buffalo, New York; Eastern Cooperative Oncology Group, M.D. Anderson Cancer Center, Houston, Texas; North Central Cancer Treatment Group, Mayo Clinic, Rochester, Minnesota; and Cancer and Leukemia Group B, Central Office of the Chairman, Chicago, Illinois). A randomized trial of aspirin to prevent colorectal adenomas in patients with previous colorectal cancer. N Engl J Med 2003:348:883–890.Colorectal cancer is a major cause of cancer-associated mortality worldwide. Globally, it is the fourth most common cancer in males and third most common in females with mortality paralleling incidence. It is the second most common cause of cancer-associated mortality in the United States. Colorectal cancers arise through complex interactions involving environmental and genetic factors. Chemoprevention involves the use of natural or synthetic agents to reverse, suppress, or prevent the occurrence of cancer. Few prospective randomized trials, however, have shown a significant effect of any given agent in altering the natural history or occurrence of colorectal cancer or its precursor lesion, the adenomatous polyp.Two recently published prospective randomized trials (N Engl J Med 2003;398:891–899, 2003;398:883–890) have now demonstrated that aspirin can reduce the recurrence of colorectal adenomas in at-risk individuals. Long-term use of aspirin and other nonsteroidal anti-inflammatory drugs (NSAIDs) has been associated with a 40%–50% reduction in colorectal cancer-associated morality in clinical case-control and cohort studies (N Engl J Med 1995;333:605–614, Cancer 2002;94:950–963). Likewise, NSAIDs have been shown to be associated with regression of adenomas in short-term prospective randomized trials in high-risk patients with familial adenomatous polyposis (Gastroenterology 1991;101:635–639, Br J Surg 1993;80:1618–1619, N Engl J Med 1993;328:1313–1316, N Engl J Med 2000;342:1946–1952). Baron et al. (N Engl J Med 2003;348:891–899) have now shown that low-dose aspirin has a moderate but significant chemopreventive effect on sporadic adenomas in the large bowel, whereas Sander et al. (N Engl J Med 2003;348:883–890) have demonstrated that daily use of aspirin is associated with a significant reduction in the incidence of colorectal adenomas in patients with previous colorectal cancer.Baron and the Polyp Prevention Study Group randomly assigned 1121 patients with a recent history of histologically documented adenomas to receive placebo (372 patients), 81 mg of aspirin (377 patients), or 325 mg aspirin (372 patients)daily. Eligible patients had one or more adenomas removed within 3 months before recruitment, one or more adenomas removed within 16 months before recruitment, and a lifetime history of 2 or more confirmed adenomas, or an adenoma at least 1 cm in diameter removed within 16 months before recruitment. Each patient was required to undergo a complete colonoscopy within 3 months before recruitment and to have a “clean colon” with no residual polyps. Patients with a history of familial colorectal cancer or a history of invasive large bowel cancer were excluded. Although the trial was initially designed to investigate aspirin only, the study was extended soon after recruitment began to also compare daily folic acid with placebo. The current report focuses solely on aspirin because the folate intervention is ongoing. Patients were counseled regarding avoidance of aspirin and NSAIDs, and patients received questionnaires every 4 months to monitor adherence to study treatment and compliance. The study design called for patients to undergo a complete surveillance colonoscopy 34–40 months after the qualifying examination. The primary outcome was the proportion of patients in whom one or more colorectal adenomas were detected during the period from 1 year after randomization through the anticipated surveillance follow-up examination. Prespecified secondary outcomes were the numbers of colorectal adenomas and advanced lesions. Advanced lesions were defined as tubulovillous adenomas (25%–75% villous features), villous adenomas (>75% villous), large adenomas (at least 1 cm in diameter), severe dysplasia, or invasive cancer.Compliance in this study was excellent, with the vast majority of patients taking virtually all of their study medications, and few taking prohibited NSAIDs. Follow-up colonoscopy was performed in 97% of eligible individuals. One or more adenoma was diagnosed at follow-up in 47.1% of patients in the placebo group, 38.3% of patients given 81 mg of aspirin per day, and 45.1% of those taking 325 mg of aspirin per day (P = 0.04). Unadjusted relative risks of any adenoma (compared with the placebo group) were 0.81 in the 81 mg group (95% confidence intervals, 0.69–0.96), but only 0.96 in the 325 mg group (95% confidence intervals, 0.8–1.13). Results were more dramatic for advanced neoplasms. The relative risks were 0.59 (95% confidence intervals, 0.38–0.92) in the low-dose aspirin group, and 0.83 (95% confidence intervals, 0.55–1.23) in the group taking the higher aspirin dose. Findings were similar with respect to adenomas in the right side and left side of the colorectum. There were a small number of serious medical events. The risk of death and serious bleeding were similar among the groups. Seven patients had nonfatal strokes: 2 in the group given 81 mg of aspirin per day, and 5 in the group given 325 mg of aspirin per day (compared with none in the placebo group; P = 0.06). One stroke was judged to be hemorrhagic.Sandler et al. studied a group at higher risk for adenoma recurrence. Six hundred thirty-five patients with previous colorectal cancer resected for cure were randomly assigned to receive either 325 mg of aspirin per day or placebo. Eligible patients included those with Dukes’ stage A or B1 colon or rectal cancer, or those with Dukes’ stage B2 or C cancers who had been free of disease for more than 5 years after curative surgery. Patients with familial polyposis, high likelihood of NSAID use, or a history of stroke, transient ischemic attacks, angina, myocardial infarction, or atherosclerotic peripheral vascular disease were excluded. Of the 635 patients who underwent randomization, 10 were subsequently found to be ineligible, 8 withdrew consent, and 2 never started treatment. Each patient’s own physician performed endoscopic examinations. Endoscopy was not performed solely for the purposes of the study, but as part of usual follow-up for patients with colorectal cancer. The primary endpoint of the study was the detection of adenomas by either colonoscopy or sigmoidoscopy after randomization.Five hundred seventeen of the 635 randomized patients had at least one colonoscopy after randomization (258 in the placebo group and 259 in the aspirin group). The median time to first colonoscopy was 11.3 months in the placebo group and 15.5 months in the aspirin group (P = 0.25). There was no significant difference between the 2 groups in adherence rates over time. The median duration of follow-up was 30.9 months in the aspirin group and 31.6 months in the placebo group. One or more adenomas were found at follow-up in 17% of the aspirin group and 27% of the placebo group (P = 0.004). The mean number of adenomas was also lower in the aspirin group than in the in placebo group (P = 0.003), and the time to development of a first adenoma was longer in the aspirin group than in the placebo group. The adjusted relative risk of any recurrent adenoma in the aspirin group compared with the placebo group was 0.65 (95% confidence intervals, 0.46–0.91). There was no significant difference in adverse events between the 2 groups after randomization.CommentA substantial amount of epidemiologic and preclinical animal data suggests that NSAIDs, particularly aspirin, reduce the risk of colorectal neoplasia (N Engl J Med 1995;333:605–614, Cancer 2002;94:950–963, N Engl J Med 2000;342:1960–1968, Gastroenterology 2002;122:641–645, N Engl J Med 2000;342:1946–1952, Cancer Res 2002;61:1733–1740, Cancer Res 2000;60:293–297). The precise mechanisms for cancer prevention are unknown, but seem to be related, in part, to altered synthesis of arachidonic acid metabolites that include prostaglandins, thromboxanes, leukotrienes, and hydroxyeicosatetraenoic acids. These compounds modulate signal transduction pathways that affect cellular and lesion growth, differentiation, angiogenesis, and programmed cell death (apoptosis) (N Engl J Med 2000;342:196–198, Gastroenterology 2001;121:1339–1397, Cancer Res 2002;62:1567–1572, Gastroenterology 2001;120:1713–1719, Cancer Res 2002;62:1178–1183, N Engl J Med 2000;342:651–653). Cyclooxygenase (COX) is the principal enzyme responsible for synthesis of prostaglandins and other eicosanoids. NSAIDs may inhibit both major isoforms of cyclooxygenase, COX-1 and COX-2, or may selectively inhibit the COX-2 isoform. The relative importance of each isoform in carcinogenesis remains to be determined, but the inducible form COX-2 has recently gained the most attention because its expression is markedly increased in colorectal adenomas and carcinomas (Cancer Res 1995;55:2556–2559, Gastroenterology 2001;121:1335–1347). Knockout of COX-2 through the use of specific pharmacologic inhibitors, or through genetic manipulation, significantly reduces tumor formation in animal models (Cancer Res 2001;61:1733–1740) and causes adenoma regression in individuals with familial adenomatous polyposis (FAP) (N Engl J Med 2000;372:1946–1952). Aspirin may affect development of neoplastic lesions in the colon through COX-dependent and -independent mechanisms (N Engl J Med 2000;342:1960–1968).Most randomized clinical trials of adenoma prevention using NSAIDs have targeted patients at very high risk for adenoma occurrence, principally patients with familial adenomatous polyposis (FAP). These agents have been convincingly shown to cause regression of established adenomas in patients with FAP (N Engl J Med 1993;328:1313–1316, Gastroenterology 2002;122:641–645, N Engl J Med 2000;372:1946–1952), but may not be effective for primary prevention of adenoma development in genetically disposed individuals with FAP (N Engl J Med 2002;346:1054–1059) who have not yet developed macroscopic adenomas. Few trials have specifically targeted individuals with a history of sporadic adenomas, or those with moderately elevated risk for development of neoplasia such as those with a prior history of colorectal cancer.Both case-control and cohort studies suggest a substantial reduction in colorectal cancer in patients taking aspirin or other NSAIDs. Case-control studies compare the use of these drugs in patients with colorectal cancer with their use in selected controls. Case-control studies are subject to various forms of bias because they are retrospective in nature, but show remarkably similar results for the most part. Individuals who take NSAIDs are approximately 50% as likely to develop colorectal cancer as those who do not (J Natl Cancer Inst 1991;83:355–358, Cancer 2002;94:950–963). Cohort studies that are prospective in nature have, in all but one case, shown a similar protective effect of aspirin and NSAIDs (N Engl J Med 1995;333:609–614).There are few randomized trials that have examined the potential chemopreventive effect of NSAIDs in humans. A randomized controlled trial of low-dose aspirin in health professionals was designed to study cardiovascular disease as its primary endpoint (Ann Intern Med 1998;128:713–720). Follow-up of greater than 12 years’ duration demonstrated no significant difference with regard to colorectal cancer between the aspirin and placebo groups, but the study was not designed to study this as a primary endpoint, aspirin was delivered every other day, and the study duration was short when considering cancer as an endpoint. The studies reviewed here provide strong evidence that the conclusions from the aforementioned epidemiologic studies are valid, and that aspirin has a chemopreventive effect on neoplasia in the colon.Clinical randomized trials of chemoprevention of colorectal neoplasia have most often concentrated on prevention of colorectal adenomas. The natural history of colorectal cancer is protracted and the duration of studies required, sample sizes necessary, cost, and ethical considerations make the use of cancer as an endpoint impractical. Adenomas are reasonable surrogate endpoints for chemoprevention trials because they represent a form of intraepithelial neoplasia (Clin Cancer Res 2002;8:314–346) from which the vast majority of colorectal cancers arise. Achieving regression of adenomas or preventing adenoma recurrence should, therefore, reduce development of colorectal cancer and related mortality. Few prospective randomized trials of potential chemopreventive agents for prevention of adenoma recurrence have been positive, however, strengthening the significance of the results obtained in the studies reviewed here.Several questions are raised by the results of these studies and should be discussed. The magnitude of response in the study by Baron et al. was modest. The risk-reduction for recurrence of any adenoma was only 19% and was limited to those taking low-dose (81 mg) aspirin, whereas the risk reduction in the Sandler study was 35%. It should be pointed out, however, that the risk reduction for development of advanced neoplasms (large adenomas, those with villous features, dysplasia, invasive cancer) in the former study was 41%. Furthermore, the populations studied were quite different. It is probable that selection of higher risk patients with a history of cancer made it easier to demonstrate an effect in the Sandler study. It is also possible that a longer study duration would have demonstrated a greater effect.The effective dose of aspirin was different in the 2 studies. In the Baron study, 325 mg aspirin had no significant effect, whereas 81 mg was effective in reducing adenoma recurrence. A 325 mg dose in the Sandler study had a significant effect in the higher risk group studied. The differences in the 2 studies may be explained by the difference in study populations, differences in the study endpoints, or even chance because both doses of aspirin used suppressed colorectal prostaglandin levels to a similar extent (Cancer Epidemol Biomarkers Prev 2002;11:275–279). Different doses of aspirin could also have differential effects on biological processes that are relevant to chemoprevention, including processes independent of cyclooxygenase.What is the clinical relevance of these studies? Certainly reducing adenoma recurrence should reduce the incidence of colorectal cancer, yet the overall chemopreventive effect of aspirin seems modest. Nonetheless, if one can reduce the incidence of “high-risk” adenomas (Baron) or adenomas in higher risk patients (Sandler), the clinical implications are substantial. Although some would argue that the potential risks of regular aspirin use (gastrointestinal hemorrhage, hemorrhagic stroke) make colonoscopy a potentially more cost-effective alternative (N Engl J Med 2003;348:879–880), the two are not mutually exclusive. Colonoscopy does miss adenomas (Gastroenterology 1997;112:24–28), and interval development of neoplasia, especially in high-risk patients, does occur. It is also possible that chemoprevention could reduce the number of recurrent polyps as in the Sandler study, or increase the time to adenoma development, leading to an increased interval between surveillance colonoscopies. These studies provide strong evidence that chemoprevention of colorectal neoplasia is possible, and open the door for further exploration of the role of chemopreventive agents in prevention of colorectal cancer. Baron JA, Cole BF, Sandler RS, Haile RW, Ahnen D, Bresalier R, McKeown-Eyssen G, Summers, RW, Rothstein R, Burke CA, Snover DC, Church TR, Allen JI, Beach M, Beck GJ, Bond JH, Byers T, Greenberg ER, Mandel JS, Marcon N, Mott LA, Pearson L, Saibil F, van Stolk RU (Norris Cotton Cancer Center, Dartmouth-Hitchcock Medical Center, Lebanon, New Hampshire; Dartmouth Medical School, Hanover, New Hampshire; University of North Carolina School of Medicine, Chapel Hill; University of Southern California School of Medicine, Los Angeles; Veterans Affairs Medical Center, Denver; University of Colorado School of Medicine, Denver, Colorado; Henry Ford Health Sciences Center, Detroit, Michigan; University of Toronto, Toronto, Ontario, Canada; the Roy J. and Lucille A. Carver College of Medicine, University of Iowa, Iowa City; Cleveland Clinic Foundation, Cleveland, Ohio; Fairview Southdale Hospital, Minneapolis; University of Minnesota, Minneapolis; University of Minnesota School of Medicine, Minneapolis; Minnesota Gastroenterology, Minneapolis; the Veterans Affairs Medical Center, Minneapolis, Minnesota; and Rollins School of Public Health, Emory University, Atlanta, Georgia). A randomized trial of aspirin to prevent colorectal adenomas. N Engl J Med 2003;348:891–899. Sandler RS, Halabi S, Baron JA, Budinger S, Paskett E, Keresztes R, Petrelli N, Pipas JM, Karp DD, Loprinzi CL, Steinbach G, Schilsky R (Departments of Medicine and Epidemiology, University of North Carolina, Chapel Hill; Cancer and Leukemia Group B Statistical Center and Department of Biostatistics and Bioinformatics, Duke University Medical Center, Durham, North Carolina; Departments of Medicine and Community and Family Medicine, Dartmouth Medical School, Hanover, New Hampshire; Wake Forest University School of Medicine, Winston-Salem, North Carolina; Weill Medical College of Cornell University, New York; Roswell Park Cancer Institute, Buffalo, New York; Eastern Cooperative Oncology Group, M.D. Anderson Cancer Center, Houston, Texas; North Central Cancer Treatment Group, Mayo Clinic, Rochester, Minnesota; and Cancer and Leukemia Group B, Central Office of the Chairman, Chicago, Illinois). A randomized trial of aspirin to prevent colorectal adenomas in patients with previous colorectal cancer. N Engl J Med 2003:348:883–890. Colorectal cancer is a major cause of cancer-associated mortality worldwide. Globally, it is the fourth most common cancer in males and third most common in females with mortality paralleling incidence. It is the second most common cause of cancer-associated mortality in the United States. Colorectal cancers arise through complex interactions involving environmental and genetic factors. Chemoprevention involves the use of natural or synthetic agents to reverse, suppress, or prevent the occurrence of cancer. Few prospective randomized trials, however, have shown a significant effect of any given agent in altering the natural history or occurrence of colorectal cancer or its precursor lesion, the adenomatous polyp. Two recently published prospective randomized trials (N Engl J Med 2003;398:891–899, 2003;398:883–890) have now demonstrated that aspirin can reduce the recurrence of colorectal adenomas in at-risk individuals. Long-term use of aspirin and other nonsteroidal anti-inflammatory drugs (NSAIDs) has been associated with a 40%–50% reduction in colorectal cancer-associated morality in clinical case-control and cohort studies (N Engl J Med 1995;333:605–614, Cancer 2002;94:950–963). Likewise, NSAIDs have been shown to be associated with regression of adenomas in short-term prospective randomized trials in high-risk patients with familial adenomatous polyposis (Gastroenterology 1991;101:635–639, Br J Surg 1993;80:1618–1619, N Engl J Med 1993;328:1313–1316, N Engl J Med 2000;342:1946–1952). Baron et al. (N Engl J Med 2003;348:891–899) have now shown that low-dose aspirin has a moderate but significant chemopreventive effect on sporadic adenomas in the large bowel, whereas Sander et al. (N Engl J Med 2003;348:883–890) have demonstrated that daily use of aspirin is associated with a significant reduction in the incidence of colorectal adenomas in patients with previous colorectal cancer. Baron and the Polyp Prevention Study Group randomly assigned 1121 patients with a recent history of histologically documented adenomas to receive placebo (372 patients), 81 mg of aspirin (377 patients), or 325 mg aspirin (372 patients)daily. Eligible patients had one or more adenomas removed within 3 months before recruitment, one or more adenomas removed within 16 months before recruitment, and a lifetime history of 2 or more confirmed adenomas, or an adenoma at least 1 cm in diameter removed within 16 months before recruitment. Each patient was required to undergo a complete colonoscopy within 3 months before recruitment and to have a “clean colon” with no residual polyps. Patients with a history of familial colorectal cancer or a history of invasive large bowel cancer were excluded. Although the trial was initially designed to investigate aspirin only, the study was extended soon after recruitment began to also compare daily folic acid with placebo. The current report focuses solely on aspirin because the folate intervention is ongoing. Patients were counseled regarding avoidance of aspirin and NSAIDs, and patients received questionnaires every 4 months to monitor adherence to study treatment and compliance. The study design called for patients to undergo a complete surveillance colonoscopy 34–40 months after the qualifying examination. The primary outcome was the proportion of patients in whom one or more colorectal adenomas were detected during the period from 1 year after randomization through the anticipated surveillance follow-up examination. Prespecified secondary outcomes were the numbers of colorectal adenomas and advanced lesions. Advanced lesions were defined as tubulovillous adenomas (25%–75% villous features), villous adenomas (>75% villous), large adenomas (at least 1 cm in diameter), severe dysplasia, or invasive cancer. Compliance in this study was excellent, with the vast majority of patients taking virtually all of their study medications, and few taking prohibited NSAIDs. Follow-up colonoscopy was performed in 97% of eligible individuals. One or more adenoma was diagnosed at follow-up in 47.1% of patients in the placebo group, 38.3% of patients given 81 mg of aspirin per day, and 45.1% of those taking 325 mg of aspirin per day (P = 0.04). Unadjusted relative risks of any adenoma (compared with the placebo group) were 0.81 in the 81 mg group (95% confidence intervals, 0.69–0.96), but only 0.96 in the 325 mg group (95% confidence intervals, 0.8–1.13). Results were more dramatic for advanced neoplasms. The relative risks were 0.59 (95% confidence intervals, 0.38–0.92) in the low-dose aspirin group, and 0.83 (95% confidence intervals, 0.55–1.23) in the group taking the higher aspirin dose. Findings were similar with respect to adenomas in the right side and left side of the colorectum. There were a small number of serious medical events. The risk of death and serious bleeding were similar among the groups. Seven patients had nonfatal strokes: 2 in the group given 81 mg of aspirin per day, and 5 in the group given 325 mg of aspirin per day (compared with none in the placebo group; P = 0.06). One stroke was judged to be hemorrhagic. Sandler et al. studied a group at higher risk for adenoma recurrence. Six hundred thirty-five patients with previous colorectal cancer resected for cure were randomly assigned to receive either 325 mg of aspirin per day or placebo. Eligible patients included those with Dukes’ stage A or B1 colon or rectal cancer, or those with Dukes’ stage B2 or C cancers who had been free of disease for more than 5 years after curative surgery. Patients with familial polyposis, high likelihood of NSAID use, or a history of stroke, transient ischemic attacks, angina, myocardial infarction, or atherosclerotic peripheral vascular disease were excluded. Of the 635 patients who underwent randomization, 10 were subsequently found to be ineligible, 8 withdrew consent, and 2 never started treatment. Each patient’s own physician performed endoscopic examinations. Endoscopy was not performed solely for the purposes of the study, but as part of usual follow-up for patients with colorectal cancer. The primary endpoint of the study was the detection of adenomas by either colonoscopy or sigmoidoscopy after randomization. Five hundred seventeen of the 635 randomized patients had at least one colonoscopy after randomization (258 in the placebo group and 259 in the aspirin group). The median time to first colonoscopy was 11.3 months in the placebo group and 15.5 months in the aspirin group (P = 0.25). There was no significant difference between the 2 groups in adherence rates over time. The median duration of follow-up was 30.9 months in the aspirin group and 31.6 months in the placebo group. One or more adenomas were found at follow-up in 17% of the aspirin group and 27% of the placebo group (P = 0.004). The mean number of adenomas was also lower in the aspirin group than in the in placebo group (P = 0.003), and the time to development of a first adenoma was longer in the aspirin group than in the placebo group. The adjusted relative risk of any recurrent adenoma in the aspirin group compared with the placebo group was 0.65 (95% confidence intervals, 0.46–0.91). There was no significant difference in adverse events between the 2 groups after randomization. CommentA substantial amount of epidemiologic and preclinical animal data suggests that NSAIDs, particularly aspirin, reduce the risk of colorectal neoplasia (N Engl J Med 1995;333:605–614, Cancer 2002;94:950–963, N Engl J Med 2000;342:1960–1968, Gastroenterology 2002;122:641–645, N Engl J Med 2000;342:1946–1952, Cancer Res 2002;61:1733–1740, Cancer Res 2000;60:293–297). The precise mechanisms for cancer prevention are unknown, but seem to be related, in part, to altered synthesis of arachidonic acid metabolites that include prostaglandins, thromboxanes, leukotrienes, and hydroxyeicosatetraenoic acids. These compounds modulate signal transduction pathways that affect cellular and lesion growth, differentiation, angiogenesis, and programmed cell death (apoptosis) (N Engl J Med 2000;342:196–198, Gastroenterology 2001;121:1339–1397, Cancer Res 2002;62:1567–1572, Gastroenterology 2001;120:1713–1719, Cancer Res 2002;62:1178–1183, N Engl J Med 2000;342:651–653). Cyclooxygenase (COX) is the principal enzyme responsible for synthesis of prostaglandins and other eicosanoids. NSAIDs may inhibit both major isoforms of cyclooxygenase, COX-1 and COX-2, or may selectively inhibit the COX-2 isoform. The relative importance of each isoform in carcinogenesis remains to be determined, but the inducible form COX-2 has recently gained the most attention because its expression is markedly increased in colorectal adenomas and carcinomas (Cancer Res 1995;55:2556–2559, Gastroenterology 2001;121:1335–1347). Knockout of COX-2 through the use of specific pharmacologic inhibitors, or through genetic manipulation, significantly reduces tumor formation in animal models (Cancer Res 2001;61:1733–1740) and causes adenoma regression in individuals with familial adenomatous polyposis (FAP) (N Engl J Med 2000;372:1946–1952). Aspirin may affect development of neoplastic lesions in the colon through COX-dependent and -independent mechanisms (N Engl J Med 2000;342:1960–1968).Most randomized clinical trials of adenoma prevention using NSAIDs have targeted patients at very high risk for adenoma occurrence, principally patients with familial adenomatous polyposis (FAP). These agents have been convincingly shown to cause regression of established adenomas in patients with FAP (N Engl J Med 1993;328:1313–1316, Gastroenterology 2002;122:641–645, N Engl J Med 2000;372:1946–1952), but may not be effective for primary prevention of adenoma development in genetically disposed individuals with FAP (N Engl J Med 2002;346:1054–1059) who have not yet developed macroscopic adenomas. Few trials have specifically targeted individuals with a history of sporadic adenomas, or those with moderately elevated risk for development of neoplasia such as those with a prior history of colorectal cancer.Both case-control and cohort studies suggest a substantial reduction in colorectal cancer in patients taking aspirin or other NSAIDs. Case-control studies compare the use of these drugs in patients with colorectal cancer with their use in selected controls. Case-control studies are subject to various forms of bias because they are retrospective in nature, but show remarkably similar results for the most part. Individuals who take NSAIDs are approximately 50% as likely to develop colorectal cancer as those who do not (J Natl Cancer Inst 1991;83:355–358, Cancer 2002;94:950–963). Cohort studies that are prospective in nature have, in all but one case, shown a similar protective effect of aspirin and NSAIDs (N Engl J Med 1995;333:609–614).There are few randomized trials that have examined the potential chemopreventive effect of NSAIDs in humans. A randomized controlled trial of low-dose aspirin in health professionals was designed to study cardiovascular disease as its primary endpoint (Ann Intern Med 1998;128:713–720). Follow-up of greater than 12 years’ duration demonstrated no significant difference with regard to colorectal cancer between the aspirin and placebo groups, but the study was not designed to study this as a primary endpoint, aspirin was delivered every other day, and the study duration was short when considering cancer as an endpoint. The studies reviewed here provide strong evidence that the conclusions from the aforementioned epidemiologic studies are valid, and that aspirin has a chemopreventive effect on neoplasia in the colon.Clinical randomized trials of chemoprevention of colorectal neoplasia have most often concentrated on prevention of colorectal adenomas. The natural history of colorectal cancer is protracted and the duration of studies required, sample sizes necessary, cost, and ethical considerations make the use of cancer as an endpoint impractical. Adenomas are reasonable surrogate endpoints for chemoprevention trials because they represent a form of intraepithelial neoplasia (Clin Cancer Res 2002;8:314–346) from which the vast majority of colorectal cancers arise. Achieving regression of adenomas or preventing adenoma recurrence should, therefore, reduce development of colorectal cancer and related mortality. Few prospective randomized trials of potential chemopreventive agents for prevention of adenoma recurrence have been positive, however, strengthening the significance of the results obtained in the studies reviewed here.Several questions are raised by the results of these studies and should be discussed. The magnitude of response in the study by Baron et al. was modest. The risk-reduction for recurrence of any adenoma was only 19% and was limited to those taking low-dose (81 mg) aspirin, whereas the risk reduction in the Sandler study was 35%. It should be pointed out, however, that the risk reduction for development of advanced neoplasms (large adenomas, those with villous features, dysplasia, invasive cancer) in the former study was 41%. Furthermore, the populations studied were quite different. It is probable that selection of higher risk patients with a history of cancer made it easier to demonstrate an effect in the Sandler study. It is also possible that a longer study duration would have demonstrated a greater effect.The effective dose of aspirin was different in the 2 studies. In the Baron study, 325 mg aspirin had no significant effect, whereas 81 mg was effective in reducing adenoma recurrence. A 325 mg dose in the Sandler study had a significant effect in the higher risk group studied. The differences in the 2 studies may be explained by the difference in study populations, differences in the study endpoints, or even chance because both doses of aspirin used suppressed colorectal prostaglandin levels to a similar extent (Cancer Epidemol Biomarkers Prev 2002;11:275–279). Different doses of aspirin could also have differential effects on biological processes that are relevant to chemoprevention, including processes independent of cyclooxygenase.What is the clinical relevance of these studies? Certainly reducing adenoma recurrence should reduce the incidence of colorectal cancer, yet the overall chemopreventive effect of aspirin seems modest. Nonetheless, if one can reduce the incidence of “high-risk” adenomas (Baron) or adenomas in higher risk patients (Sandler), the clinical implications are substantial. Although some would argue that the potential risks of regular aspirin use (gastrointestinal hemorrhage, hemorrhagic stroke) make colonoscopy a potentially more cost-effective alternative (N Engl J Med 2003;348:879–880), the two are not mutually exclusive. Colonoscopy does miss adenomas (Gastroenterology 1997;112:24–28), and interval development of neoplasia, especially in high-risk patients, does occur. It is also possible that chemoprevention could reduce the number of recurrent polyps as in the Sandler study, or increase the time to adenoma development, leading to an increased interval between surveillance colonoscopies. These studies provide strong evidence that chemoprevention of colorectal neoplasia is possible, and open the door for further exploration of the role of chemopreventive agents in prevention of colorectal cancer. A substantial amount of epidemiologic and preclinical animal data suggests that NSAIDs, particularly aspirin, reduce the risk of colorectal neoplasia (N Engl J Med 1995;333:605–614, Cancer 2002;94:950–963, N Engl J Med 2000;342:1960–1968, Gastroenterology 2002;122:641–645, N Engl J Med 2000;342:1946–1952, Cancer Res 2002;61:1733–1740, Cancer Res 2000;60:293–297). The precise mechanisms for cancer prevention are unknown, but seem to be related, in part, to altered synthesis of arachidonic acid metabolites that include prostaglandins, thromboxanes, leukotrienes, and hydroxyeicosatetraenoic acids. These compounds modulate signal transduction pathways that affect cellular and lesion growth, differentiation, angiogenesis, and programmed cell death (apoptosis) (N Engl J Med 2000;342:196–198, Gastroenterology 2001;121:1339–1397, Cancer Res 2002;62:1567–1572, Gastroenterology 2001;120:1713–1719, Cancer Res 2002;62:1178–1183, N Engl J Med 2000;342:651–653). Cyclooxygenase (COX) is the principal enzyme responsible for synthesis of prostaglandins and other eicosanoids. NSAIDs may inhibit both major isoforms of cyclooxygenase, COX-1 and COX-2, or may selectively inhibit the COX-2 isoform. The relative importance of each isoform in carcinogenesis remains to be determined, but the inducible form COX-2 has recently gained the most attention because its expression is markedly increased in colorectal adenomas and carcinomas (Cancer Res 1995;55:2556–2559, Gastroenterology 2001;121:1335–1347). Knockout of COX-2 through the use of specific pharmacologic inhibitors, or through genetic manipulation, significantly reduces tumor formation in animal models (Cancer Res 2001;61:1733–1740) and causes adenoma regression in individuals with familial adenomatous polyposis (FAP) (N Engl J Med 2000;372:1946–1952). Aspirin may affect development of neoplastic lesions in the colon through COX-dependent and -independent mechanisms (N Engl J Med 2000;342:1960–1968). Most randomized clinical trials of adenoma prevention using NSAIDs have targeted patients at very high risk for adenoma occurrence, principally patients with familial adenomatous polyposis (FAP). These agents have been convincingly shown to cause regression of established adenomas in patients with FAP (N Engl J Med 1993;328:1313–1316, Gastroenterology 2002;122:641–645, N Engl J Med 2000;372:1946–1952), but may not be effective for primary prevention of adenoma development in genetically disposed individuals with FAP (N Engl J Med 2002;346:1054–1059) who have not yet developed macroscopic adenomas. Few trials have specifically targeted individuals with a history of sporadic adenomas, or those with moderately elevated risk for development of neoplasia such as those with a prior history of colorectal cancer. Both case-control and cohort studies suggest a substantial reduction in colorectal cancer in patients taking aspirin or other NSAIDs. Case-control studies compare the use of these drugs in patients with colorectal cancer with their use in selected controls. Case-control studies are subject to various forms of bias because they are retrospective in nature, but show remarkably similar results for the most part. Individuals who take NSAIDs are approximately 50% as likely to develop colorectal cancer as those who do not (J Natl Cancer Inst 1991;83:355–358, Cancer 2002;94:950–963). Cohort studies that are prospective in nature have, in all but one case, shown a similar protective effect of aspirin and NSAIDs (N Engl J Med 1995;333:609–614). There are few randomized trials that have examined the potential chemopreventive effect of NSAIDs in humans. A randomized controlled trial of low-dose aspirin in health professionals was designed to study cardiovascular disease as its primary endpoint (Ann Intern Med 1998;128:713–720). Follow-up of greater than 12 years’ duration demonstrated no significant difference with regard to colorectal cancer between the aspirin and placebo groups, but the study was not designed to study this as a primary endpoint, aspirin was delivered every other day, and the study duration was short when considering cancer as an endpoint. The studies reviewed here provide strong evidence that the conclusions from the aforementioned epidemiologic studies are valid, and that aspirin has a chemopreventive effect on neoplasia in the colon. Clinical randomized trials of chemoprevention of colorectal neoplasia have most often concentrated on prevention of colorectal adenomas. The natural history of colorectal cancer is protracted and the duration of studies required, sample sizes necessary, cost, and ethical considerations make the use of cancer as an endpoint impractical. Adenomas are reasonable surrogate endpoints for chemoprevention trials because they represent a form of intraepithelial neoplasia (Clin Cancer Res 2002;8:314–346) from which the vast majority of colorectal cancers arise. Achieving regression of adenomas or preventing adenoma recurrence should, therefore, reduce development of colorectal cancer and related mortality. Few prospective randomized trials of potential chemopreventive agents for prevention of adenoma recurrence have been positive, however, strengthening the significance of the results obtained in the studies reviewed here. Several questions are raised by the results of these studies and should be discussed. The magnitude of response in the study by Baron et al. was modest. The risk-reduction for recurrence of any adenoma was only 19% and was limited to those taking low-dose (81 mg) aspirin, whereas the risk reduction in the Sandler study was 35%. It should be pointed out, however, that the risk reduction for development of advanced neoplasms (large adenomas, those with villous features, dysplasia, invasive cancer) in the former study was 41%. Furthermore, the populations studied were quite different. It is probable that selection of higher risk patients with a history of cancer made it easier to demonstrate an effect in the Sandler study. It is also possible that a longer study duration would have demonstrated a greater effect. The effective dose of aspirin was different in the 2 studies. In the Baron study, 325 mg aspirin had no significant effect, whereas 81 mg was effective in reducing adenoma recurrence. A 325 mg dose in the Sandler study had a significant effect in the higher risk group studied. The differences in the 2 studies may be explained by the difference in study populations, differences in the study endpoints, or even chance because both doses of aspirin used suppressed colorectal prostaglandin levels to a similar extent (Cancer Epidemol Biomarkers Prev 2002;11:275–279). Different doses of aspirin could also have differential effects on biological processes that are relevant to chemoprevention, including processes independent of cyclooxygenase. What is the clinical relevance of these studies? Certainly reducing adenoma recurrence should reduce the incidence of colorectal cancer, yet the overall chemopreventive effect of aspirin seems modest. Nonetheless, if one can reduce the incidence of “high-risk” adenomas (Baron) or adenomas in higher risk patients (Sandler), the clinical implications are substantial. Although some would argue that the potential risks of regular aspirin use (gastrointestinal hemorrhage, hemorrhagic stroke) make colonoscopy a potentially more cost-effective alternative (N Engl J Med 2003;348:879–880), the two are not mutually exclusive. Colonoscopy does miss adenomas (Gastroenterology 1997;112:24–28), and interval development of neoplasia, especially in high-risk patients, does occur. It is also possible that chemoprevention could reduce the number of recurrent polyps as in the Sandler study, or increase the time to adenoma development, leading to an increased interval between surveillance colonoscopies. These studies provide strong evidence that chemoprevention of colorectal neoplasia is possible, and open the door for further exploration of the role of chemopreventive agents in prevention of colorectal cancer.