Copyright ©2014 Baishideng Publishing Group Co., Limited. All rights reserved.
World J Gastrointest Pharmacol Ther. Feb 6, 2014; 5(1): 40-49
Published online Feb 6, 2014. doi: 10.4292/wjgpt.v5.i1.40
Aspirin, cyclooxygenase inhibition and colorectal cancer
Carlos Sostres, Carla Jerusalen Gargallo, Angel Lanas
Carlos Sostres, Carla Jerusalen Gargallo, Angel Lanas, Department of Digestive Diseases, University Hospital Lozano Blesa, 50009 Zaragoza, Spain
Carlos Sostres, Carla Jerusalen Gargallo, Angel Lanas, Aragon Health Sciences Institute, 50009 Zaragoza, Spain
Angel Lanas, Centro de Investigación Biológica en Red de Enfermedades Hepáticasy Digestivas, 50009 Zaragoza, Spain
Angel Lanas, Department of Gastroenterology, University of Zaragoza, 50009 Zaragoza, Spain
Author contributions: All the authors contributed equally to the design, drafting and reviewing process of this paper.
Supported by Funds from FIS (P108/1301); Dr. Lanas A has received speaking and consultancy fees from AstraZeneca, Pfizer and Bayer
Correspondence to: Angel Lanas, MD, DSc, Clinical Chief, Professor, Department of Digestive Diseases, University Hospital Lozano Blesa, c/Domingo Miral s/n, 50009 Zaragoza, Spain.
Telephone: +34-976-765786 Fax: +34-976-765787
Received: July 30, 2013
Revised: November 13, 2013
Accepted: December 9, 2013
Published online: February 6, 2014


Colorectal cancer (CRC) is the third most common type of cancer worldwide. Screening measures are far from adequate and not widely available in resource-poor settings. Primary prevention strategies therefore remain necessary to reduce the risk of developing CRC. Increasing evidence from epidemiological studies, randomized clinical trials and basic science supports the effectiveness of aspirin, as well as other non-steroidal anti-inflammatory drugs, for chemoprevention of several types of cancer, including CRC. This includes the prevention of adenoma recurrence and reduction of CRC incidence and mortality. The detectable benefit of daily low-dose aspirin (at least 75 mg), as used to prevent cardiovascular disease events, strongly suggests that its antiplatelet action is central to explaining its antitumor efficacy. Daily low-dose aspirin achieves complete and persistent inhibition of cyclooxygenase (COX)-1 in platelets (in pre-systemic circulation) while causing a limited and rapidly reversible inhibitory effect on COX-2 and/or COX-1 expressed in nucleated cells. Aspirin has a short half-life in human circulation (about 20 minutes); nucleated cells have the ability to resynthesize acetylated COX isozymes within a few hours, while platelets do not. COX-independent mechanisms of aspirin have been suggested to explain its chemopreventive effects but this concept remains to be demonstrated in vivo at clinical doses.

Key Words: Aspirin, Colorectal cancer, Cyclooxygenase inhibition, Mechanisms, Risk, Benefits

Core tip: Colorectal cancer (CRC) is a major cause of morbidity and mortality worldwide. Currently, CRC screening programs are not widely available and need to be improved. New prevention strategies are therefore necessary. Daily low-dose aspirin, as given for the prevention of cardiovascular disease events, has demonstrated benefits in clinical and basic studies in terms of preventing adenoma recurrence and decreasing the incidence of CRC and attributable mortality. These findings indicate that the antiplatelet action of aspirin plays a central role in its antitumor effect. Cyclooxygenase-dependent and independent mechanisms have been suggested to explain this effect. Extensive translational medical research is mandatory for future progress in CRC prevention.


Colorectal cancer (CRC) is the third most common cancer worldwide, accounting for an estimated 9.8% of all new cancers (1.2 million cases annually) and 8.1% of all cancer mortality[1]. It arises through the cumulative effects of inherited genetic predisposition and environmental factors. Genomic instability is an integral part of the transformation of normal colonic or rectal mucosa into carcinoma. Three molecular pathways have been identified: chromosomal instability, microsatellite instability and CpG island methylator phenotype pathways. These pathways are not mutually exclusive, with some tumors exhibiting features of multiple pathways. Germline mutations are responsible for hereditary CRC syndromes (accounting for less than 5% of all CRC), while a stepwise accumulation of genetic and epigenetic alterations results in sporadic CRC.

Today it is well known that screening reduces CRC mortality and is recommended, beginning at age 50, for average risk individuals, although compliance is far from adequate and screening is not widely available in resource-poor settings[2,3]. Primary prevention strategies are therefore still necessary to reduce the risk of CRC, especially because of the limitations of population-based secondary prevention programs that rely on detection and removal of adenomas.

Aspirin has demonstrated its efficacy in the prevention of adverse events related to cardiovascular disease (CVD). It is one of the most widely used drugs in the world. One survey suggested that over one-third of the United States adult population use low-dose aspirin (LDA) regularly[4]. In England in 2007, over 30 million primary care prescriptions were issued for aspirin[5]. Hence, both physicians and patients are largely familiar with the long-term use of aspirin for chronic disease management. In addition, CRC and CVD share the same risk factors, such as older age, being overweight/obesity and physical inactivity.

Today, a large body of clinical and experimental evidence indicates that aspirin can protect against different types of cancer, in particular CRC[6]. A role of the antiplatelet effect of aspirin in its anti-cancer effect is also supported by several studies.

In this review we will discuss clinical results related to the impact of aspirin on the risk of CRC. Then, we will explain the pharmacology of aspirin at low doses in order to provide a mechanistic interpretation of aspirin action as a chemopreventive agent for CRC, in particular the selective inhibition of platelet cyclooxygenase (COX)-1 activity.

Evidence from epidemiological studies

Most case-control and cohort studies have found that regular aspirin use was associated with reduced risk of CRC[7]. A systematic review of case-control studies published in 2012 showed a statistically significant reduction of long-term risk of developing CRC (OR = 0.62, 95%CI: 0.58-0.67) in regular aspirin users compared with non-users, as well as a significant reduction in the proportion of cancers with distant metastasis at diagnosis (OR = 0.69, 95%CI: 0.57-0.83)[8] (Table 1). An analysis of 662424 men and women enrolled in the Cancer Prevention Study II cohort showed that daily use of aspirin for at least 5 years was associated with a 32% reduction in risk of CRC[9]. Two cohort studies of United States health professionals (47363 men and 82911 women) showed that regular aspirin users (≥ 2 times/wk) had 21% and 23% lower risk of CRC, respectively, during follow-up periods of 18 and 20 years respectively[10,11]. Moreover, in a separate analysis of a Nurses Health Study cohort, regular aspirin use also reduced the risk of death from CRC by 28% and risk of death from any type of cancer by 12%[12].

Table 1 Summary of the associations between regular use of aspirin and risk of colorectal cancer in case-control and cohort studies.
Study typenAspirinControlsOR (95%CI)P value
Any ASA2610464/2561828300/478340.67 (0.60-0.74)< 0.0001
Maximum reported ASA171551/126592664/181530.62 (0.58-0.67)< 0.0001
ASA ≥ 5 yr10971/76821534/100290.68 (0.63-0.75)< 0.0001
Daily ASA4165/1254349/15230.49 (0.40-0.60)< 0.0001
Daily ASA ≥ 5 yr166/1668121/19730.63 (0.46-0.86)0.004
Standard cohort
Any Aspirin113791/27644143623/25146520.85 (0.82-0.89)< 0.0001
Maximum reported ASA8661/6644751858/13749050.78 (0.71-0.84)< 0.0001
ASA ≥ 5 yr4889/1 0221921311/13047600.76 (0.70–0.82)< 0.0001
Daily ASA5741/6585361115/8192880.80 (0.73-0.88)< 0.0001
Daily ASA ≥ 5 yr160/38302420/2321160.68 (0.52-0.90)0.0060
Nested case-control
Any ASA62215/892665 099/1095260.87 (0.75-1.00)0.0700
Maximum reported ASA5206/44578302/409480.67 (0.58-0.77)< 0.0001
ASA ≥ 5 yr1116/22823704/379350.62 (0.48-0.81)< 0.0001
Daily ASA153/1658744/229750.77 (0.55-1.07)0.1400
Daily ASA ≥ 5 yr129/1417274/215050.51 (0.34-0.76)0.0120
Evidence from clinical trials

In 2010, Rothwell et al[13] obtained long-term follow-up data on cancer outcomes from four randomised trials that were originally designed to evaluate the effect of aspirin on the prevention of CVD events (Table 2). These trials studied diverse populations with CVD, including men at low risk (n = 10,224) and men and women at high risk (n = 3809). Dosage ranged from 75-1200 mg/d, median treatment duration was 6 years and median follow-up was 18.3 years. Treatment with aspirin (75-500 mg/d) reduced the 20-year risk of CRC by 24% and CRC-associated mortality by 35%. The benefit increased with longer duration of treatment and seemed to be higher for proximal CRC compared to distal CRC. An absolute reduction of 1.76% (P = 0.001) in 20-year risk of any fatal CRC after 5 years of daily treatment with aspirin (75-300 mg) was observed. Subsequently, the same authors published a study that examined the effects of daily aspirin on long-term risk of death due to all cancers. They included data from eight randomised trials (25570 patients, 674 cancer deaths) and concluded that aspirin use reduced the risk of death due to cancer (pooled OR = 0.79, P = 0.003), but the benefit was only apparent after 5 years of treatment. Absolute reduction reached 7% in 20-year risk of death due to cancer for patients aged ≥ 65 years. In the 3 trials reporting data on the specific site of cancer occurrence with treatment duration of 5 years or longer and long-term follow-up, patients randomized to aspirin showed a statistically significant 20 year risk reduction of death due to CRC of 40% (HR = 0.60; 95%CI: 0.45-0.81, P = 0.0007)[14].

Table 2 Characteristics of trials included in Rothwell et al study and details of post-trial follow-up.
Thromobosis prevention trialSwedish aspirin low dose trialUK-TIA aspirin trialBritish doctors aspirin trial
ASA comparison75 mg/d vs placebo75 mg/d vs placebo300 mg vs 1200 mg/d vs placebo500 mg/d vs placebo
Recruitment period1989-19921984-19891979-19851978-7199
Median duration of scheduled treatment in original trial (yr)
Year post-trial follow up extended to2009200720062002

Although these data are compelling, it should be taken into account that these studies were secondary analyses of CVD prevention trials and therefore they were not originally designed to examine CRC incidence or mortality. In addition, there are two large randomized trials of alternate-day aspirin treatment in healthy subjects: the Physician’s Health Study (PHS)[15] and Women’s Health Study (WHS), which showed no effect of aspirin on the incidence of CRC over a 10-year follow-up period[16]. The PHS determined the effect of aspirin 325 mg every other day on CVD in 22,071 healthy male physicians. In this study, the relative risk of CRC over a 10-year follow up was 1.03 (95%CI: 0.83-1.28). The WHS examined the effect of 100 mg every other day in 39876 healthy women. The relative risk of CRC was 0.97 (95%CI: 0.77-1.24). There are several plausible explanations for the discrepancy in results between the meta-analyses performed by Rothwell and incidence data of the PHS and WHS trials. Firstly, both trials used alternate-day dosing regimens in contrast to daily dosing used in the studies included in both meta-analyses. Secondly, in the PHS and WHS trials, the duration of follow-up was shorter and may have been insufficient to detect the aspirin effect. Finally, in the WHS trial, the equivalent daily dose of aspirin was 50 mg, lower than the 75 mg/d shown to be the minimum effective dose in the Rothwell meta-analyses[13] (Table 3).

Table 3 Clinical effects of aspirin on sporadic colorectal cancer (clinical trials).
Rothwell et al meta analysisPhysician’s health studyWomen’s health study
ASA dosage75-1200 mg/d325 mg per every other day100 mg per every other day
Duration of follow up (yr)≥ 201010
Relative risk of CRC over follow up (HR)0.76 (95%CI: 0.60-0.96)1.03 (95%CI: 0.83-1.28)0.97 (95%CI: 0.77-0.24)

The precursors of CRC are colorectal adenomas in most cases. It would be expected that the chemopreventive effects of aspirin should begin before the development of CRC. The long duration of aspirin treatment required to show a preventive effect against invasive CRC probably reflects the time required for the development of cancer from precursor lesions (5-10 years). To date, four randomized double-blind placebo-controlled trials with 2967 participants have evaluated aspirin versus placebo for the secondary prevention of colorectal adenomas (in patients who had had colorectal adenomas or CRC)[17-20]. Doses ranged from 81 to 325 mg/d and median follow-up was 33 mo. The meta-analysis of these randomized trials[21] showed a statistically significant 17% reduction of the risk of developing adenoma with any dose of aspirin vs placebo (RR = 0.83; 95%CI: 0.72-0.96). This corresponded to a significant absolute risk reduction of 6.7%. For any advanced lesion, a significant relative risk reduction of 28% for aspirin at any dose was observed. This preventive effect emerged rather quickly (1 year) after the initiation of aspirin use (Table 4).

Table 4 Clinical effects of aspirin in incidence of sporadic colorectal adenomas (clinical trials).
StudyPatientsTreatmentRR (95%CI)
AFPPS trialPatients with a recent history of histologically documented (removed) adenomasASA (81 or 325 mg/d) or folic acid (1 mg/d) or placebo for 2.7 yearsAny adenoma 0.81 (0.69-0.96), ASA 81mg vs non ASA 0.96 (0.81-1.13), ASA 325 mg vs non ASA Advanced lesion 0.59 (0.38-0.92), ASA 81 mg vs non ASA 0.83 (0.55–1.23), ASA 325 mg vs non ASA
CAPS trialPatients with a histologically documented colon or rectal cancer with a low risk of recurrent diseaseASA 325 mg/d or placebo for 2.6 years0.65 (0.46-0.91)
APACC trialPatients with a history of colorectal adenomasASA 160 or 300 mg/d or placebo for 1 and 4 years0.73 (0.52-1.04) for both doses, after 1 year 0.96 (0.75-1.22), for both doses, after 4 years
ukCAP trialPatients with an adenoma removed in the 6 mo before recruitmentASA (300 mg/d) plus placebo or ASA plus folic acid (0.5 mg/d) or folic acid plus placebo or double placebo for about 2.6 yearsAny adenoma 0.79 (0.63-0.99), ASA vs non ASA, Advanced adenoma 0.63 (0.43-0.91), ASA vs non ASA
J-CAPP trialPatients with previous sporadic colorectal tumorsASA 100 mg/d or placebo for 2 yearsOngoing

To date, there are two controlled randomized trials that primarily evaluated the efficacy of aspirin in high-risk CRC patients: the Colorectal Adenoma/Carcinoma Programme (CAPP)1[22], which included 206 young individuals with a diagnosis of familial adenomatous polyposis (FAP), and CAPP2[23], which studied 1009 patients with Lynch syndrome. Both studies compared aspirin (600 mg/d), with or without resistant starch or resistant starch placebo. Data from CAPP1 patients were only analyzed if they had received treatment for at least 1 year. CAPP2 patients received aspirin for a mean of 29 mo.

The CAPP1 trial showed that the mean size of the largest polyps was significantly reduced in aspirin users. Despite a trend to fewer polyps in the rectum and sigmoid colon in aspirin versus non-aspirin users at the end of intervention (from 1 to 12 years), the difference was not significant.

CAPP2 was the first clinical trial that had cancer prevention as a primary endpoint. At the end of the intervention phase, analysis showed that aspirin treatment did not reduce the risk of developing new adenomas (RR = 1.03; 95%CI: 0.7-1.4) or CRC. The study design involved post-intervention follow-up[24]. Over a mean follow-up of 55.7 mo, 48 aspirin users had developed 53 primary CRC, whereas in intention-to-treat analysis of time to first CRC there were no differences (HR = 0.63; P = 0.12) and the per-protocol analysis of patients completing 2 years of intervention yielded a HR of 0.41 (0.19-0.86, P = 0.02) (Table 5).

Table 5 Clinical effects of aspirin in high risk population (clinical trials).
StudyPatientsTreatmentRR or HR (95%CI)Ref.
CAPP1 trialFAP young patients (10 to 21 years of age)ASA (600 mg/d) plus placebo or resistant starch (30 g daily) plus placebo or double placebo for 17 yearsRR = 0.77 (0.54-1.10), ASA vs non ASA[22]
CAPP2 trialHereditary non-polyposis colon cancer or HNPCCASA (600 mg/d) or ASA placebo or resistant starch (30 g daily) or starch placebo for up to 4 yearsHR = 0.63 (0.35-1.13), for the entire post-randomization period (ASA vs placebo) HR = 0.41 (0.19-0.86), for ≥ 2 years of treatment (ASA vs placebo)[23]
J-FAPP II trialFAP patients (≥ 16 years of age)Placebo vs enteric coated ASA (100 mg/d ) for 6-10 moOngoing[25]

In Japan, two chemoprevention studies are currently being performed: one in patients with previous sporadic colorectal tumors [Japan Colorectal Aspirin Polyps Prevention (J-CAPP study)] and the second in patients with familial adenomatous polyposis (J-FAPP study II). Both are double-blind randomized controlled trials with low-dose aspirin (100 mg/d) and study the effect of aspirin in colorectal carcinogenesis[25].


It has been suggested that aspirin may prevent recurrence or death in CRC patients. In a placebo-controlled randomized trial of patients with a history of non-metastatic CRC after resection, daily treatment with LDA was associated with a 35% reduction in risk of recurrent adenoma or carcinoma at 36 mo[18]. In a cohort study of health professionals diagnosed with stage I-III CRC, regular use of aspirin after diagnosis was associated with higher CRC specific survival compared with non-users[26].


Because most aspirin-related adverse effects are dose-dependent, to find the minimum effective dose required for CRC prevention remains a critically important issue. The Rothwell meta-analysis found that daily LDA regimens for the prevention of CVD-related events (75-325 mg) were as effective as daily high-dose aspirin[13]. However, the short-term follow up data from the PHS[15] (aspirin 325 mg every other day) and the WHS[16] (aspirin 100 mg every other day) did not show a reduction in risk of CRC. These negative findings could be attributed to alternate day dosage and/or short follow up and/or the lower dose, especially in the WHS trial. The adenoma trials (REFS) also indicate that LDA (81-325 mg/d) reduces the risk of developing adenomas and advanced adenomas.

Although follow-up of the randomized trial of daily aspirin in CVD prevention and adenoma prevention trials demonstrated that daily LDA of 75-81 mg may be sufficient for CRC prevention, the results of observational studies are controversial. Some suggested that 300-325 mg may be necessary for CRC prevention but most provided incomplete information regarding the dose and duration of aspirin treatment[7,8,10,11,27].

Therefore, taking the clinical trial and observational information together, there is very strong evidence that long-term LDA (75-325 mg/d) reduces the risk of CRC. Importantly, for the prevention of CVD-related events, LDA (75-81 mg/d) seems to be as effective as high-dose aspirin (300-325 mg/d) and, moreover, LDA has a better safety profile. However, daily aspirin at any dose may show greater benefit in patients with CVD than in those at risk of CRC.


Based on current evidence, treatment with LDA for 5 years in patients at risk of CVD-related events will probably prevent between 12 and 40 myocardial infarctions per 1000 patients treated, assuming an overall 10% risk of CVD-related events in this population[28]. Unfortunately, LDA use is also associated with 2-4 upper gastrointestinal bleeding events per 1000 patients[28]. However, the risk of adverse events differs according to patient characteristics (gender, age, history of ulcer, etc.). It is of course possible to reduce gastrointestinal risk with proton pump inhibitors, but we cannot reduce the risk of intracranial bleeding. Given the risk of bleeding, clinical guidelines (2007) recommended against the routine use of aspirin for CRC prevention in average-risk individuals[29]. However, the accumulating evidence from randomized clinical trials provides an exciting opportunity to reconsider the potential role of aspirin in cancer prevention; therefore, future practice guidelines recommendation for primary prevention in average-risk individuals for aspirin prophylaxis may also consider the prevention of cancer and not only the benefits of aspirin for the prevention of CVD-related events (Figure 1).

Figure 1
Figure 1 Balancing risk-benefits for the use of low dose aspirin. CRC: Colorectal cancer; NSAID: Nonsteroidal anti-inflammatory drug; CV: Cardiovascular; PPI: Proton pump inhibitor; ASA: Aspirin.

Aspirin, like other nonsteroidal anti-inflammatory drugs (NSAIDs), has the capacity to reduce prostanoid generation by inhibiting the activity of COX isozymes. Prostanoids are biologically active derivatives of arachidonic acid (AA) released from membrane phospholipids through the activity of different phospholipases[30,31]. There are two isoforms of COX, named COX-1 and COX-2[32]. Both COX isozymes are differently regulated catalytically, transcriptionally and post-transcriptionally, but they share the same catalytic activities.

COX-1 gene is considered a “housekeeping gene” and the protein is highly expressed in platelets where it is responsible for the generation of thromboxane A2 (TXA2), which promotes platelet activation and aggregation, vasoconstriction and proliferation of vascular smooth muscle cells[31,33]. In addition, COX-1 is highly expressed in gastric epithelial cells where it plays an important role in cytoprotection through the generation of prostanoids, such as prostaglandin E2 (PGE2)[31,33]. In contrast, COX-2 gene, a primary response one with many regulatory sites[34], is constitutively expressed in some tissues in physiological conditions, such as the endothelium, kidney and brain, and in pathological conditions, such as in cancer[35]. In cancer cells, the major prostanoid produced through COX-2 is PGE2, which plays important roles in modulating motility, proliferation and resistance to apoptosis[36,37].

Unlike other NSAIDs, aspirin is able to produce an irreversible inactivation of COX isozymes through the acetylation of a specific serine moiety (Ser529 of COX-1 and Ser516 of COX-2)[38]. Acetylation of the allosteric subunit of COX-1 by aspirin causes an irreversible inhibition of COX activity and, in turn, of the generation of PGG2 from AA. Acetylated COX-2 is not able to form PGG2 but it generates 15R-hydroxyeicosapentaenoic acid (15R-HETE) from AA[39]. However, there is no convincing evidence that these lipid mediators triggered by aspirin are generated in vivo in humans.


Aspirin has a short half-life when administered in vivo and it is rapidly inactivated by plasma and tissue esterases into salicylic acid, which is a weak inhibitor of COXs (in the millimolar range)[40-42]. The inhibitory effects of aspirin have been found to be > 100-fold more potent in inhibiting platelet COX-1 than monocyte COX-2[17-20]. Aspirin at low doses (75-100 mg daily) is able to cause nearly complete inhibition of the capacity of platelet COX-1 to generate TXA2[43,44]. Due to irreversible inhibition of COX-1 and the limited capacity of platelets for de novo protein synthesis[45], the profound inhibitory effect of platelet function by aspirin persists throughout the dose interval (i.e., 24 h).

The major part of the inhibitory effect of platelet COX-1 by the oral administration of low-dose aspirin occurs in the presystemic circulation where the drug reaches higher concentrations[46,47]. The impact of low-dose aspirin, administered once daily, on COX-2 activity in vivo is marginal. In summary, the pharmacokinetics and pharmacodynamics of low-dose aspirin support the fact that the drug acts mainly by modifying platelet function as a consequence of COX-1 inhibition. At higher doses, aspirin may affect COX-2 in a dose-dependent fashion.


Randomised clinical trials have shown that once daily LDA provides a chemopreventive effect against atherothrombosis[24] and CRC[13,14]. This finding suggests that enhanced platelet activation is involved in the development of these two pathological conditions. In fact, these aspirin doses and dosing intervals are consistent with a selective inhibitory effect of aspirin on platelet COX-1 activity and on TXA2-dependent platelet function.

Transcriptional upregulation of the COX-2 gene has been observed in nearly half of human colorectal adenomas and 80%-90% of CRC, probably related to the disturbed function of the APC gene. However, COX-1 gene and protein expression are not affected[48-52] and the role of this enzyme in CRC carcinogenesis remains unclear. In colonic mucosa, COX-2 is localized predominantly in tumor tissue, including epithelial cells, mononuclear cells, endothelial and stromal cells, but not in nearby normal tissue. Upregulation of COX-2 is associated with increased cell adhesion, phenotypic changes, resistance to apoptosis and tumor angiogenesis[53-57]. COX-2 expression does not always correlate with survival and/or with Duke’s stage of the disease[58-60]. This suggests a role of upregulated COX-2 for the initial stages of colon carcinogenesis but not for clinical outcome at advanced stages. The best studied consequence of upregulated COX-2 in CRC is enhanced prostaglandin production[48]. Prostaglandin levels in CRC tissue are 3-4 fold higher than in healthy tissue in the vicinity, with PGE2 being the predominant product[61]. PGE2 inhibits apoptosis and stimulates tumor growth and angiogenesis via stimulation of b-catenin/T-cell factor dependent transcription[62]. In addition, PGE2 acts as an immunosuppressant in patients with CRC[53,63]. The clearest clinical evidence for COX-2 as a pharmacological target for the chemopreventive action of aspirin was the finding that aspirin reduced the risk of CRC exclusively in individuals with elevated COX-2 expression but not in those without[64]. This was associated with a reduction in mortality[65]. Although these findings were from observational studies, they confirmed experimental data that prostaglandins and non-prostaglandin COX-2 products are central to the pathogenesis of CRC. The vast majority of published experimental studies have reported beneficial antitumor effects for aspirin, celecoxib and non-aspirin NSAIDs in a variety of experimental models[65-67]. These data strongly suggest a central role of COX-2 in CRC and its inhibition is an effective chemopreventive measure (Figure 2).

Figure 2
Figure 2 Cyclooxygenase-dependent mechanisms for antitumoral effects of low dose aspirin. CRC: Colorectal cancer; COX: Cyclooxygenase; PGE2: Prostaglandin E2.

The generation of TXA2, a major product of platelet COX-1 which promotes platelet aggregation and vasoconstriction[68], represents another important mechanism by which platelets can affect tumorigenesis. One study has shown that enhanced TXA2 generation into murine colon-26 adenocarcinoma cell line (C26) stimulated tumor angiogenesis, tumor growth in vivo[69] and promoted the interaction between metastasizing tumor cells and the host hemostatic system[69], thus suggesting a role of TXA2 in promoting angiogenesis and the development of tumor metastasis[70].

In one study, the authors demonstrated PGE2 inhibition in rectal biopsies performed 1 mo after treatment with three different doses of aspirin (81, 325 and 650 mg) versus placebo[71]. Unexpectedly, the 81 mg daily aspirin dose suppressed PGE2 levels to the same extent as the 650 mg dose. In another study, treatment with 81 mg of aspirin per day for 3 mo reduced mucosal PGE2 and transforming growth factor-α expression in apparently normal rectal mucosa of individuals with a history of adenomatous polyps[72]. Further studies using more appropriate methodologies are required to definitively clarify whether LDA affects COX-1 activity in the gastrointestinal tract.

Some investigators have proposed that both COX-1 and COX-2 pathways are involved in intestinal tumorigenesis and that they operate sequentially. This is strongly supported by the findings of experimental animal studies in which the loss of either COX-1 or COX-2 genes blocks intestinal polyposis in mouse models of FAP by about 90%[66,67].


Evidence from different lines of research indicates that COX-2 independent mechanisms may also affect apoptosis and cell proliferation in CRC and are sensitive to both aspirin and non-aspirin NSAIDs. Nearly but not all human colon cancer cells express COX-2 and produce prostaglandins[53,73]. Today, several COX-independent mechanisms of aspirin have been reported that might contribute to its chemopreventive effects in tumorigenesis[73]. Most of these effects have been found in vitro using supra-therapeutic concentrations of aspirin which cannot be obtained in systemic circulation with low doses of the drug. However, no convincing evidence has been obtained to demonstrate that these mechanisms are operative in vivo, particularly with low doses of aspirin which have been associated with chemopreventive benefits in randomised clinical trials. In any case, currently available evidence clearly points to the existence of further cellular targets of NSAIDs, in addition to COX-2 inhibition, which may contribute to their antitumor effects. Further studies are needed to completely understand the mechanisms involved.


A large body of clinical evidence supports the protective action of aspirin as a chemopreventive agent for different types of cancer, in particular CRC[6]. Also, increasing indirect evidence has led to the hypothesis that the antiplatelet effect of aspirin is a central mechanism for its antitumor effect[34,41]. The finding of an apparent maximum chemopreventive efficacy against cancer and atherothrombosis by low-dose aspirin lends support to this hypothesis[6]. At low doses every 24 h, aspirin acts as a complete and persistent inhibitor of COX-1 in platelets (in pre-systemic circulation)[47], while causing a limited and rapidly reversible inhibitory effect on COX-2 and/or COX-1 expressed in nucleated cells[39]. Despite uncertainty about the precise mechanisms that underlie aspirin’s anticancer benefit, the evidence supporting its effectiveness for the prevention of CRC is substantial; daily aspirin for at least 5 years has been shown to reduce the 20-year risk of CRC by 32% and 20-year mortality by 43%[13]. Therefore, the potential benefit of aspirin in both CVD and prevention of cancer at multiple sites may favor its use for broader chronic disease prevention. It is likely that the benefits in terms of morbidity and mortality will outweigh concerns about gastrointestinal bleeding, which is rarely life threatening, and cerebral bleeding, which is extremely uncommon. Health authorities should consider the possibility of extending recommendations on the routine use of aspirin, taking into account its beneficial effects in cardiovascular disease and cancer prevention.

Extensive translational medical research is required to confirm the hypothesis of platelet-mediated colon tumorigenesis. Importantly, these studies will need to address the current uncertainty concerning the optimal aspirin dose, the dosing regimen for cancer prevention, the possible contribution of individual genetic cancer susceptibility to aspirin response[74] and also the target population most likely to benefit from daily aspirin use.


P- Reviewer: Roesler R S- Editor: Zhai HH L- Editor: Roemmele A E- Editor: Wu HL

1.  Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM.  GLOBOCAN 2008 v1. 2, Cancer Incidence and Mortality Worldwide: IARC CancerBase No. 10 [Internet]. Lyon, France: International Agency for Research on Cancer 2011; .  [PubMed]  [DOI]
2.  Shapiro JA, Seeff LC, Thompson TD, Nadel MR, Klabunde CN, Vernon SW. Colorectal cancer test use from the 2005 National Health Interview Survey. Cancer Epidemiol Biomarkers Prev. 2008;17:1623-1630.  [PubMed]  [DOI]
3.  United States Preventive Services Task Force. Screening for colorectal cancer: U.S. Preventive Services Task Force recommendation statement. Ann Intern Med. 2008;149:627-637.  [PubMed]  [DOI]
4.  Ajani UA, Ford ES, Greenland KJ, Giles WH, Mokdad AH. Aspirin use among U.S. adults: Behavioral Risk Factor Surveillance System. Am J Prev Med. 2006;30:74-77.  [PubMed]  [DOI]
5.  National Health Service Information Centre. Prescription cost analysis for England 2007.  Available from: http://www.ic.nhs. uk/webfiles/publications/PCA publication/PCA 2007 complete V2.pdf.  [PubMed]  [DOI]
6.  Thun MJ, Jacobs EJ, Patrono C. The role of aspirin in cancer prevention. Nat Rev Clin Oncol. 2012;9:259-267.  [PubMed]  [DOI]
7.  Flossmann E, Rothwell PM. Effect of aspirin on long-term risk of colorectal cancer: consistent evidence from randomised and observational studies. Lancet. 2007;369:1603-1613.  [PubMed]  [DOI]
8.  Algra AM, Rothwell PM. Effects of regular aspirin on long-term cancer incidence and metastasis: a systematic comparison of evidence from observational studies versus randomised trials. Lancet Oncol. 2012;13:518-527.  [PubMed]  [DOI]
9.  Jacobs EJ, Thun MJ, Bain EB, Rodriguez C, Henley SJ, Calle EE. A large cohort study of long-term daily use of adult-strength aspirin and cancer incidence. J Natl Cancer Inst. 2007;99:608-615.  [PubMed]  [DOI]
10.  Chan AT, Giovannucci EL, Meyerhardt JA, Schernhammer ES, Wu K, Fuchs CS. Aspirin dose and duration of use and risk of colorectal cancer in men. Gastroenterology. 2008;134:21-28.  [PubMed]  [DOI]
11.  Chan AT, Giovannucci EL, Meyerhardt JA, Schernhammer ES, Curhan GC, Fuchs CS. Long-term use of aspirin and nonsteroidal anti-inflammatory drugs and risk of colorectal cancer. JAMA. 2005;294:914-923.  [PubMed]  [DOI]
12.  Chan AT, Manson JE, Feskanich D, Stampfer MJ, Colditz GA, Fuchs CS. Long-term aspirin use and mortality in women. Arch Intern Med. 2007;167:562-572.  [PubMed]  [DOI]
13.  Rothwell PM, Wilson M, Elwin CE, Norrving B, Algra A, Warlow CP, Meade TW. Long-term effect of aspirin on colorectal cancer incidence and mortality: 20-year follow-up of five randomised trials. Lancet. 2010;376:1741-1750.  [PubMed]  [DOI]
14.  Rothwell PM, Fowkes FG, Belch JF, Ogawa H, Warlow CP, Meade TW. Effect of daily aspirin on long-term risk of death due to cancer: analysis of individual patient data from randomised trials. Lancet. 2011;377:31-41.  [PubMed]  [DOI]
15.  Gann PH, Manson JE, Glynn RJ, Buring JE, Hennekens CH. Low-dose aspirin and incidence of colorectal tumors in a randomized trial. J Natl Cancer Inst. 1993;85:1220-1224.  [PubMed]  [DOI]
16.  Cook NR, Lee IM, Gaziano JM, Gordon D, Ridker PM, Manson JE, Hennekens CH, Buring JE. Low-dose aspirin in the primary prevention of cancer: the Women’s Health Study: a randomized controlled trial. JAMA. 2005;294:47-55.  [PubMed]  [DOI]
17.  Baron JA, Cole BF, Sandler RS, Haile RW, Ahnen D, Bresalier R, McKeown-Eyssen G, Summers RW, Rothstein R, Burke CA. A randomized trial of aspirin to prevent colorectal adenomas. N Engl J Med. 2003;348:891-899.  [PubMed]  [DOI]
18.  Sandler RS, Halabi S, Baron JA, Budinger S, Paskett E, Keresztes R, Petrelli N, Pipas JM, Karp DD, Loprinzi CL. A randomized trial of aspirin to prevent colorectal adenomas in patients with previous colorectal cancer. N Engl J Med. 2003;348:883-890.  [PubMed]  [DOI]
19.  Logan RF, Grainge MJ, Shepherd VC, Armitage NC, Muir KR. Aspirin and folic acid for the prevention of recurrent colorectal adenomas. Gastroenterology. 2008;134:29-38.  [PubMed]  [DOI]
20.  Benamouzig R, Deyra J, Martin A, Girard B, Jullian E, Piednoir B, Couturier D, Coste T, Little J, Chaussade S. Daily soluble aspirin and prevention of colorectal adenoma recurrence: one-year results of the APACC trial. Gastroenterology. 2003;125:328-336.  [PubMed]  [DOI]
21.  Cole BF, Logan RF, Halabi S, Benamouzig R, Sandler RS, Grainge MJ, Chaussade S, Baron JA. Aspirin for the chemoprevention of colorectal adenomas: meta-analysis of the randomized trials. J Natl Cancer Inst. 2009;101:256-266.  [PubMed]  [DOI]
22.  Burn J, Bishop DT, Chapman PD, Elliott F, Bertario L, Dunlop MG, Eccles D, Ellis A, Evans DG, Fodde R. A randomized placebo-controlled prevention trial of aspirin and/or resistant starch in young people with familial adenomatous polyposis. Cancer Prev Res (Phila). 2011;4:655-665.  [PubMed]  [DOI]
23.  Burn J, Bishop DT, Mecklin JP, Macrae F, Möslein G, Olschwang S, Bisgaard ML, Ramesar R, Eccles D, Maher ER. Effect of aspirin or resistant starch on colorectal neoplasia in the Lynch syndrome. N Engl J Med. 2008;359:2567-2578.  [PubMed]  [DOI]
24.  Burn J, Gerdes AM, Macrae F, Mecklin JP, Moeslein G, Olschwang S, Eccles D, Evans DG, Maher ER, Bertario L. Long-term effect of aspirin on cancer risk in carriers of hereditary colorectal cancer: an analysis from the CAPP2 randomised controlled trial. Lancet. 2011;378:2081-2087.  [PubMed]  [DOI]
25.  Ishikawa H, Nakamura T, Kawano A, Gondo N, Sakai T. Chemoprevention of colorectal cancer in Japan: a brief introduction to current clinical trials. J Gastroenterol. 2009;44 Suppl 19:77-81.  [PubMed]  [DOI]
26.  Chan AT, Ogino S, Fuchs CS. Aspirin use and survival after diagnosis of colorectal cancer. JAMA. 2009;302:649-658.  [PubMed]  [DOI]
27.  Mahipal A, Anderson KE, Limburg PJ, Folsom AR. Nonsteroidal anti-inflammatory drugs and subsite-specific colorectal cancer incidence in the Iowa women’s health study. Cancer Epidemiol Biomarkers Prev. 2006;15:1785-1790.  [PubMed]  [DOI]
28.  Hayden M, Pignone M, Phillips C, Mulrow C. Aspirin for the primary prevention of cardiovascular events: a summary of the evidence for the U.S. Preventive Services Task Force. Ann Intern Med. 2002;136:161-172.  [PubMed]  [DOI]
29.  United States Preventive Services Task Force. Routine aspirin or nonsteroidal anti-inflammatory drugs for the primary prevention of colorectal cancer: U.S. Preventive Services Task Force recommendation statement. Ann Intern Med. 2007;146:361-364.  [PubMed]  [DOI]
30.  Funk CD. Prostaglandins and leukotrienes: advances in eicosanoid biology. Science. 2001;294:1871-1875.  [PubMed]  [DOI]
31.  Smyth EM, Grosser T, Wang M, Yu Y, FitzGerald GA. Prostanoids in health and disease. J Lipid Res. 2009;50 Suppl:S423-S428.  [PubMed]  [DOI]
32.  Smith WL, DeWitt DL, Garavito RM. Cyclooxygenases: structural, cellular, and molecular biology. Annu Rev Biochem. 2000;69:145-182.  [PubMed]  [DOI]
33.  Patrono C, Patrignani P, García Rodríguez LA. Cyclooxygenase-selective inhibition of prostanoid formation: transducing biochemical selectivity into clinical read-outs. J Clin Invest. 2001;108:7-13.  [PubMed]  [DOI]
34.  Kang YJ, Mbonye UR, DeLong CJ, Wada M, Smith WL. Regulation of intracellular cyclooxygenase levels by gene transcription and protein degradation. Prog Lipid Res. 2007;46:108-125.  [PubMed]  [DOI]
35.  Harper KA, Tyson-Capper AJ. Complexity of COX-2 gene regulation. Biochem Soc Trans. 2008;36:543-545.  [PubMed]  [DOI]
36.  Dixon DA, Tolley ND, King PH, Nabors LB, McIntyre TM, Zimmerman GA, Prescott SM. Altered expression of the mRNA stability factor HuR promotes cyclooxygenase-2 expression in colon cancer cells. J Clin Invest. 2001;108:1657-1665.  [PubMed]  [DOI]
37.  Wang D, Dubois RN. Eicosanoids and cancer. Nat Rev Cancer. 2010;10:181-193.  [PubMed]  [DOI]
38.  Patrono C, Baigent C, Hirsh J, Roth G. Antiplatelet drugs: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8th Edition). Chest. 2008;133:199S-233S.  [PubMed]  [DOI]
39.  Sharma NP, Dong L, Yuan C, Noon KR, Smith WL. Asymmetric acetylation of the cyclooxygenase-2 homodimer by aspirin and its effects on the oxygenation of arachidonic, eicosapentaenoic, and docosahexaenoic acids. Mol Pharmacol. 2010;77:979-986.  [PubMed]  [DOI]
40.  Dovizio M, Bruno A, Tacconelli S, Patrignani P. Mode of action of aspirin as a chemopreventive agent. Recent Results Cancer Res. 2013;191:39-65.  [PubMed]  [DOI]
41.  Patrignani P. Nonsteroidal anti-inflammatory drugs, COX-2 and colorectal cancer. Toxicol Lett. 2000;112-113:493-498.  [PubMed]  [DOI]
42.  Warner TD, Giuliano F, Vojnovic I, Bukasa A, Mitchell JA, Vane JR. Nonsteroid drug selectivities for cyclo-oxygenase-1 rather than cyclo-oxygenase-2 are associated with human gastrointestinal toxicity: a full in vitro analysis. Proc Natl Acad Sci USA. 1999;96:7563-7568.  [PubMed]  [DOI]
43.  Ricciotti E, Dovizio M, Di Francesco L, Anzellotti P, Salvatore T, Di Francesco A, Sciulli MG, Pistritto G, Monopoli A, Patrignani P. NCX 4040, a nitric oxide-donating aspirin, exerts anti-inflammatory effects through inhibition of I kappa B-alpha degradation in human monocytes. J Immunol. 2010;184:2140-2147.  [PubMed]  [DOI]
44.  Patrignani P, Filabozzi P, Patrono C. Selective cumulative inhibition of platelet thromboxane production by low-dose aspirin in healthy subjects. J Clin Invest. 1982;69:1366-1372.  [PubMed]  [DOI]
45.  Evangelista V, Manarini S, Di Santo A, Capone ML, Ricciotti E, Di Francesco L, Tacconelli S, Sacchetti A, D’Angelo S, Scilimati A. De novo synthesis of cyclooxygenase-1 counteracts the suppression of platelet thromboxane biosynthesis by aspirin. Circ Res. 2006;98:593-595.  [PubMed]  [DOI]
46.  Pedersen AK, FitzGerald GA. Dose-related kinetics of aspirin. Presystemic acetylation of platelet cyclooxygenase. N Engl J Med. 1984;311:1206-1211.  [PubMed]  [DOI]
47.  Patrono C, García Rodríguez LA, Landolfi R, Baigent C. Low-dose aspirin for the prevention of atherothrombosis. N Engl J Med. 2005;353:2373-2383.  [PubMed]  [DOI]
48.  Eberhart CE, Coffey RJ, Radhika A, Giardiello FM, Ferrenbach S, DuBois RN. Up-regulation of cyclooxygenase 2 gene expression in human colorectal adenomas and adenocarcinomas. Gastroenterology. 1994;107:1183-1188.  [PubMed]  [DOI]
49.  Kutchera W, Jones DA, Matsunami N, Groden J, McIntyre TM, Zimmerman GA, White RL, Prescott SM. Prostaglandin H synthase 2 is expressed abnormally in human colon cancer: evidence for a transcriptional effect. Proc Natl Acad Sci USA. 1996;93:4816-4820.  [PubMed]  [DOI]
50.  Kargman SL, O’Neill GP, Vickers PJ, Evans JF, Mancini JA, Jothy S. Expression of prostaglandin G/H synthase-1 and -2 protein in human colon cancer. Cancer Res. 1995;55:2556-2559.  [PubMed]  [DOI]
51.  Marnett LJ, DuBois RN. COX-2: a target for colon cancer prevention. Annu Rev Pharmacol Toxicol. 2002;42:55-80.  [PubMed]  [DOI]
52.  Sano H, Kawahito Y, Wilder RL, Hashiramoto A, Mukai S, Asai K, Kimura S, Kato H, Kondo M, Hla T. Expression of cyclooxygenase-1 and -2 in human colorectal cancer. Cancer Res. 1995;55:3785-3789.  [PubMed]  [DOI]
53.  Marnett LJ. Aspirin and the potential role of prostaglandins in colon cancer. Cancer Res. 1992;52:5575-5589.  [PubMed]  [DOI]
54.  Wang D, Dubois RN. The role of COX-2 in intestinal inflammation and colorectal cancer. Oncogene. 2010;29:781-788.  [PubMed]  [DOI]
55.  Nie D, Honn KV. Cyclooxygenase, lipoxygenase and tumor angiogenesis. Cell Mol Life Sci. 2002;59:799-807.  [PubMed]  [DOI]
56.  Tsujii M, Kawano S, DuBois RN. Cyclooxygenase-2 expression in human colon cancer cells increases metastatic potential. Proc Natl Acad Sci USA. 1997;94:3336-3340.  [PubMed]  [DOI]
57.  Fujita T, Matsui M, Takaku K, Uetake H, Ichikawa W, Taketo MM, Sugihara K. Size- and invasion-dependent increase in cyclooxygenase 2 levels in human colorectal carcinomas. Cancer Res. 1998;58:4823-4826.  [PubMed]  [DOI]
58.  Ogino S, Kirkner GJ, Nosho K, Irahara N, Kure S, Shima K, Hazra A, Chan AT, Dehari R, Giovannucci EL. Cyclooxygenase-2 expression is an independent predictor of poor prognosis in colon cancer. Clin Cancer Res. 2008;14:8221-8227.  [PubMed]  [DOI]
59.  Sheehan KM, Sheahan K, O’Donoghue DP, MacSweeney F, Conroy RM, Fitzgerald DJ, Murray FE. The relationship between cyclooxygenase-2 expression and colorectal cancer. JAMA. 1999;282:1254-1257.  [PubMed]  [DOI]
60.  Fux R, Schwab M, Thon KP, Gleiter CH, Fritz P. Cyclooxygenase-2 expression in human colorectal cancer is unrelated to overall patient survival. Clin Cancer Res. 2005;11:4754-4760.  [PubMed]  [DOI]
61.  Pugh S, Thomas GA. Patients with adenomatous polyps and carcinomas have increased colonic mucosal prostaglandin E2. Gut. 1994;35:675-678.  [PubMed]  [DOI]
62.  Shao J, Jung C, Liu C, Sheng H. Prostaglandin E2 Stimulates the beta-catenin/T cell factor-dependent transcription in colon cancer. J Biol Chem. 2005;280:26565-26572.  [PubMed]  [DOI]
63.  Balch CM, Dougherty PA, Cloud GA, Tilden AB. Prostaglandin E2-mediated suppression of cellular immunity in colon cancer patients. Surgery. 1984;95:71-77.  [PubMed]  [DOI]
64.  Chan AT, Ogino S, Fuchs CS. Aspirin and the risk of colorectal cancer in relation to the expression of COX-2. N Engl J Med. 2007;356:2131-2142.  [PubMed]  [DOI]
65.  Fischer SM, Hawk ET, Lubet RA. Coxibs and other nonsteroidal anti-inflammatory drugs in animal models of cancer chemoprevention. Cancer Prev Res (Phila). 2011;4:1728-1735.  [PubMed]  [DOI]
66.  Oshima M, Dinchuk JE, Kargman SL, Oshima H, Hancock B, Kwong E, Trzaskos JM, Evans JF, Taketo MM. Suppression of intestinal polyposis in Apc delta716 knockout mice by inhibition of cyclooxygenase 2 (COX-2). Cell. 1996;87:803-809.  [PubMed]  [DOI]
67.  Chulada PC, Thompson MB, Mahler JF, Doyle CM, Gaul BW, Lee C, Tiano HF, Morham SG, Smithies O, Langenbach R. Genetic disruption of Ptgs-1, as well as Ptgs-2, reduces intestinal tumorigenesis in Min mice. Cancer Res. 2000;60:4705-4708.  [PubMed]  [DOI]
68.  Grosser T, Fries S, FitzGerald GA. Biological basis for the cardiovascular consequences of COX-2 inhibition: therapeutic challenges and opportunities. J Clin Invest. 2006;116:4-15.  [PubMed]  [DOI]
69.  Pradono P, Tazawa R, Maemondo M, Tanaka M, Usui K, Saijo Y, Hagiwara K, Nukiwa T. Gene transfer of thromboxane A(2) synthase and prostaglandin I(2) synthase antithetically altered tumor angiogenesis and tumor growth. Cancer Res. 2002;62:63-66.  [PubMed]  [DOI]
70.  Honn KV. Inhibition of tumor cell metastasis by modulation of the vascular prostacyclin/thromboxane A2 system. Clin Exp Metastasis. 1983;1:103-114.  [PubMed]  [DOI]
71.  Sample D, Wargovich M, Fischer SM, Inamdar N, Schwartz P, Wang X, Do KA, Sinicrope FA. A dose-finding study of aspirin for chemoprevention utilizing rectal mucosal prostaglandin E(2) levels as a biomarker. Cancer Epidemiol Biomarkers Prev. 2002;11:275-279.  [PubMed]  [DOI]
72.  Barnes CJ, Hamby-Mason RL, Hardman WE, Cameron IL, Speeg KV, Lee M. Effect of aspirin on prostaglandin E2 formation and transforming growth factor alpha expression in human rectal mucosa from individuals with a history of adenomatous polyps of the colon. Cancer Epidemiol Biomarkers Prev. 1999;8:311-315.  [PubMed]  [DOI]
73.  Hanif R, Pittas A, Feng Y, Koutsos MI, Qiao L, Staiano-Coico L, Shiff SI, Rigas B. Effects of nonsteroidal anti-inflammatory drugs on proliferation and on induction of apoptosis in colon cancer cells by a prostaglandin-independent pathway. Biochem Pharmacol. 1996;52:237-245.  [PubMed]  [DOI]
74.  Liao X, Lochhead P, Nishihara R, Morikawa T, Kuchiba A, Yamauchi M, Imamura Y, Qian ZR, Baba Y, Shima K. Aspirin use, tumor PIK3CA mutation, and colorectal-cancer survival. N Engl J Med. 2012;367:1596-1606.  [PubMed]  [DOI]