Gastroenterology
Volume 131, Issue 6 , Pages 2003-2005, December 2006

Is the Demonstration of Adenoma Reduction With Rofecoxib a Pyrrhic Victory?

  • Patrick M. Lynch

      Affiliations

    • Corresponding Author InformationAddress requests for reprints to: Patrick M. Lynch, JD, MD, Associate Professor of Medicine, The University of Texas M.D. Anderson Cancer Center, P.O. Box 301402, Houston, TX 77230-1402. fax: (713) 563-4398.

The University of Texas, M.D. Anderson Cancer Center, Houston, Texas

Article Outline

 

See “A randomized trial of rofecoxib for the chemoprevention of colorectal adenomas” by Baron JA, Sandler RS, Bresalier RS, Quan H, Riddell R, Lanas A, Bolognese JA, Oxenius B, Horgan K, Loftus S, Morton DG on page 1674.

In 279 bc, Pyrrhus, king of Epirus, defeated the Romans at Asculum. However, the cost was great, as he had lost irreplaceable troops and comrades; the Romans would eventually be able to regroup in greater numbers than ever. A quote related by Plutarch was to the effect that “another such victory would utterly undo him.” Since that time, a Pyrrhic victory has been a metaphor for any undertaking in which the costs proved to be greater than the benefits. Clearly, such outcomes are seen only in retrospect; battles would not be undertaken in the first place if the unfortunate outcome could have been predicted.

How do the misfortunes of Pyrrhus relate to clinical chemopreventions trials? Those who have followed the story of the coxibs would likely consider the positive efficacy data from the several coxib chemoprevention trials in nonfamilial adenomas to have been such pyrrhic victories because of the cardiovascular toxicities reported, separately, from those same trials. But where did this all come from and what does it really mean?

Mounting evidence from observational and randomized clinical trials has shown that nonsteroidal anti-inflammatory drugs (NSAIDs) reduce risk of recurrence of colorectal adenomas. Large cohort studies and prospective trials have, with some exceptions, supported a reduction of colorectal cancer risk through long-term administration of aspirin.1, 2, 3

Because of the toxicities associated with use of existing NSAIDs, the new class of agents, selective cyclooxygenase (COX) inhibitors, was developed to take advantage of the fact that the COX-2 isoform appeared to be responsible for risk of neoplasms. By inhibiting COX-2 but not COX-1, the isoform mediating much of the gastrointestinal (GI) toxicity, inhibitors of COX-2 would have the benefits of NSAIDs without the drawbacks of toxicity. In this issue of Gastroenterology, Baron et al4 report the effects of 1 COX-2 inhibitor, rofecoxib (Vioxx), on recurrence of adenomas. These data correspond to data from 2 other trials of the COX-2 inhibitor celecoxib (Celebrex). These aggregated data could be used to support the use of COX-2 inhibitors in the treatment of patients with a history of adenomas.

In the APPROVe trial, eligibility was limited to subjects >40 years old and who had had ≥1 adenomas removed from a properly evaluated colon during the 3 months prior to enrollment. Subjects were excluded if there was evidence of familial syndromes, early onset adenomas (<35 years), previous colorectal resection, inflammatory bowel disease, cancer within past 5 years, expected need for regular NSAID use, evidence of occult GI bleeding, hypertension, angina, congestive heart failure, history of myocardial infarction, recent stroke, transient ischemic attack, or coronary revascularization.

Following a placebo run-in, subjects were randomized to placebo or rofecoxib 25 mg. A 50-mg arm was initially included but later dropped when a decision was made to allow use of low-dose (cardioprotective) aspirin. Colonoscopy follow-up was set at intervals of 1 and 3 years, with a 4-year examination added to assess the effect of drug withdrawal over 1 year postdosing.

In this trial, 2587 subjects were randomized. After all subjects were enrolled but shortly before planned study closure, the trial was terminated following a Data Safety and Monitoring Board recognition of what is now a well-known, well-publicized cardiovascular risk signal. At that time, 92% of the subjects had either completed their 3-year colonoscopy or had already dropped out. Adherence to drug dosing was >80% and to colonoscopy follow-up >90%. Significant reduction in risk of adenomas was seen in the rofecoxib arm, both at 1 year (relative risk [RR] = 0.6) and at 3 years (RR = 0.76). Invasive adenocarcinomas were seen in both groups, 6 on celecoxib, 11 on placebo. No major differences in adenoma recurrence rates were noted according to gender, colon site considered, or presence or absence of low-dose aspirin use. Analysis of data from year 4 (off treatment for 1 year) colonoscopy showed a possible rebound increase in adenomas in the group that had been taking rofecoxib during the trial, with an RR of 1.21.

As had been previously reported, toxicities in the rofecoxib group included thrombotic cardiovascular events. In addition, toxicities commonly associated with nonselective NSAIDs were noted to be greater in the rofecoxib group. These included upper GI perforation, obstruction, and symptomatic ulcer or bleeding, along with other, more nonspecific events, previously reported. Overall there were more deaths in the rofecoxib arm (n = 24) than in the placebo group (n = 18).

Although there are differences, the findings from this APPROVe trial harbor more similarities to the findings from the companion celecoxib Adenoma Prevention with Celecoxib (APC)5 and Prevention of Colorectal Sporadic Adenomatous Polyp (Pre-SAP)6 trials. Although a significant cardiovascular events signal was seen in only 1 of those 2 trials,5 a pooled data meta-analysis of the APC and Pre-SAP raised generalized concerns over a COX-2 inhibitor class effect with respect to cardiovascular disease risk.7 That somewhat similar cardiovascular disease risk may characterize even the nonselective NSAIDs is notable and it is likely that the signals arising from these COX-2 inhibitor trials will continue to focus attention on the questions of not only the breadth of the risk, but the mechanisms as well.

In an editorial critique of the companion Pre-SAP and APC celecoxib trials, Psaty and Potter8 provided a model or “virtual” trial in which observed toxicity and efficacy were projected over a population of screenees. Although such models can always be criticized on the basis of the validity of assumed or measured variable, their conservative estimate was that cardiovascular mortality would markedly exceed colon cancer prevention benefit. Thus, in a population of adenoma patients continuing to undergo periodic colonoscopy, treatment with celecoxib would be more detrimental than helpful. If aspirin were utilized instead of celecoxib in their rather simple model, a modest aspirin benefit would exist, even if colon cancer risk reduction were more modest, due to the cardiovascular benefits of aspirin. Recent meta-analyses have further complicated the picture. In one,9 rofecoxib was implicated in various adverse renal events (peripheral edema, hypertension, renal dysfunction) and arrhythmia, but no broader coxib class effect was consistently evident. In another, a rofecoxib dose effect was seen for cardiovascular events, with an RR of 1.33 at 25 mg/d and 2.19 at higher doses, and the effects were seen early in dosing.10 Again, no consistent coxib class effect was extracted from the meta-analysis. In an accompanying editorial, Graham remains highly skeptical of coxibs in general.11 He cautions that other analyses of pooled data have shown increased cardiac events for celecoxib and other selective and nonselective COX-2 inhibitors. Risks appeared to be dose dependent for celecoxib and the risks were most evident for those with a previous history of cardiac disease.

Regarding the recent data on both toxicity and efficacy of COX-2 inhibitors in prophylaxis of nonfamilial adenomas, we now have what should be the last piece of the puzzle, namely, the rofecoxib efficacy data from the APPROVe trial. Does this additional information enable us to arrive at a more confident conclusion about the total picture? Although the designs of the rofecoxib and celecoxib trials differed in a few details, they are really quite comparable. We have already seen that the magnitude of thrombotic cardiovascular risk is comparable between the rofecoxib and celecoxib trials. It is now evident that the magnitude of adenoma risk reduction is similar as well. This should not be surprising given the functional similarities of the agents and the similarities of the trial design, duration, and sample size.

If the coxib polyp prevention data and their various interpretations are valid, the conclusions must tilt toward the inappropriateness of considering these agents for adenoma prophylaxis in patients >50 years with a history of simple adenomas. So, is there any future for the coxibs in the field of colorectal neoplasia? I would argue that the answer is a very qualified yes. The risk of cardiovascular events seems real, but the data have come from trials in subjects >50 years old whose adenoma risk may, after all, be managed effectively by colonoscopic surveillance and polypectomy alone.

Although nonfamilial adenoma patients constitute the largest population at risk of colorectal cancer, it is not the only one. Recall that an important part of the foundation for the conduct of the APPROVe, APC, and Pre-SAP trials was the positive outcome from earlier trials in familial adenomatous polyposis (FAP). The first of these, in 80 subjects with mainly recurrent rectal polyposis postcolectomy, yielded a 30% reduction in adenoma burden following 6 months of treatment with high-dose (400 mg, twice daily) celecoxib.12 Very careful data and safety monitoring were present. The absence of demonstrated cardiovascular events in that trial can be readily attributed to the short course of therapy and to the much younger average age (37 years) of the subjects. Several small trials utilizing rofecoxib to treat FAP have also been published. These showed efficacy comparable to that of the celecoxib trial and no significant safety issues were noted. It must be emphasized that none of these FAP trials were specifically powered to detect cardiovascular signals.

We have since undertaken additional trials with celecoxib. One combines high-dose (400 mg twice daily) celecoxib with difluoromethylornithine (DFMO), again in a short-term, 6-month trial. Because the Food and Drug Administration (FDA) had approved celecoxib for use in FAP, it was not considered ethical to include a placebo arm, so the comparison arm is celecoxib alone. Following the report of a cardiovascular signal from the APC trial, our celecoxib/DFMO trial was suspended for several months, and only allowed to reinitiate enrollment and dosing following the addition of new cardiovascular exclusion criteria (untreated hypertension and hypercholesterolemia, family history of early onset cardiovascular events, current tobacco smoking). During the hiatus in which the trial was suspended, we interviewed FAP patients who had completed the trial to gauge their level of concern over cardiovascular risk. A majority expressed the opinion that risk of complications of their FAP outweighed the potential cardiovascular disease risk.13 The sample was likely biased by the fact that they had already completed the trial without the occurrence of any adverse events. The same survey demonstrated a surprising lack of familiarity with cardiovascular risk factors, so patient education will continue to be an important element in trial recruitment.

A second trial, requested by the FDA, is intended to address long-term clinical efficacy of celecoxib, the initial approval having been based on short-term data providing proof of principle rather than demonstrated clinical benefit. A more purely prevention, as opposed to regression trial, was planned, this time in children >10 who were carriers of APC mutations but who had not yet shown evidence of significant polyp burden. Before embarking on a full-scale Phase III trial, a very short-term (3 months of treatment) Phase I safety trial was conducted. This did demonstrate safety in the typical Phase I dose escalation by cohort design, up to a dose equivalent to the 400 mg twice daily dose used in the earlier adult trial.14 Although the sample size was small, reduction in adenoma burden paralleled the findings from the previous adult trial.12, 15 The Phase III trial was recently approved for enrollment.

The other important group targeted for treatment with coxibs is cancer patients. Several trials have continued in patients with colorectal and other malignancies. Among the more interesting are those attempting to determine if advantage can be taken of both the anti-inflammatory and antitumor attributes of the coxibs. Neoadjuvant trials are combining celecoxib with radiation therapy and 5-fluoracil to see if antitumor activity can be enhanced while at the same time reducing radiation-induced inflammation. Others are combining celecoxib with experimental agents in the treatment of advanced disease (www.clinicaltrials.gov/, the National Institutes of Health clinical trials website).16 According to the same website, the 2 rofecoxib studies, one for FAP and one colon adjuvant trial, were closed to accrual.

One may conclude several things from the status of clinical trials that have been conducted or are currently open:

1.The balance of cardiovascular risk versus adenoma prevention benefit does not support the use of coxibs in patients with a history of nonfamilial adenomas. It is unlikely that further adenoma prevention trials will be seen unless and until there are unexpected data showing a safety profile that is better than is now evident.

2.In conditions with greater underlying morbidity, such as FAP and invasive colorectal cancer, the balance may tilt in favor of a continued role for the coxibs and nonselective NSAIDs. Given the short-term nature of the FAP trials, longer term evaluation of potential toxicity in these relatively younger patients is warranted, as is careful attention to achieving the lowest possible therapeutic dose of COX-2 inhibitor. Unlike the brief dosing employed in adjuvant therapy for cancer (assuming some benefit is shown), chemopreventive treatment in FAP would be a long-term undertaking.

Following his “victory” over the Romans, Pyrrhus did live to fight another day. But we never really heard from him again. Time will tell whether there is life for the coxibs in colorectal neoplasia after the experiences of the sporadic adenoma trials. If so, it will likely be limited to those at highest risk of adenomas/cancer and at lowest risk of cardiovascular disease.

Back to Article Outline

References 

  1. Giovannucci E, Rimm EB, Stampfer MJ, Colditz GA, Ascherio A, Willett WC. Aspirin use and the risk for colorectal and adenoma in male health professionals. Ann Intern Med. 1994;121:241–246
  2. Chan AT, Giovannucci EL, Schernhammer ES, Colditz GA, Hunter DJ, Willett WC, et al. A prospective study of aspirin use and the risk for colorectal adenoma. Ann Intern. 2004;140:157–166Med
  3. Cook NR, Lee I-M, Gaziano JM, Gordon D, Ridker PM, Manson JE, et al. Low-dose aspirin in the primary prevention of cancer: the Women’s Health Study: a randomized controlled trial. JAMA. 2005;294:47–55
  4. Baron J, Sandler RS, Bresalier RS. A randomized trial of rofecoxib for the chemoprevention of colorectal adenomas. Gastroenterology. 2006;131:1674–1682
  5. Bertagnolli MM, Eagle CJ, Zauber AG, Redston M, Solomon SD, Kim K, et al. Celecoxib for the prevention of sporadic colorectal adenomas. N Engl J Med. 2006;355:873–884
  6. Arber N, Eagle CJ, Spicak J, Racz I, Dite P, Hajer J, et al. Celecoxib for the prevention of colorectal adenomatous polyps. N Engl J Med. 2006;355:885–895
  7. Kearney PM, Baigent C, Godwin J, Halls H, Emberson JR, Patrono C. Do selective cyclo-oxygenase-2 inhibitors and traditional non-steroidal anti-inflammatory drugs increase the risk of atherothrombosis? (Meta-analysis of randomized trials). BMJ. 2006;332:1302–1308
  8. Psaty BM, Potter JD. Risks and benefits of celecoxib to prevent recurrent adenomas. [editorial] N Engl J Med. 2006;355:950–952
  9. Zhang J, Ding EL, Song Y. Adverse effects of cyclooxygenase inhibitors on renal and arrhythmia events: meta-analysis of randomized trials. JAMA. 2006;296:1619–1632
  10. McGettigan P, Henry D. Cardiovascular risk and inhibition of cyclooxygenase: a systematic review of the observational studies of selective and nonselective inhibitors of cyclooxygenase 2. JAMA. 2006;296:1633–1644
  11. Graham DJ. Cox-2 inhibitors, other NSAIDs, and cardiovascular risk: the seduction of common sense. [editorial] JAMA. 2006;296:1653–1656
  12. Steinbach G, Lynch PM, Phillips RKS, Wallace MH, Hawk E, Gordon GB, et al. The effect of celecoxib, a cyclooxygenase-2 inhibitor, in familial adenomatous polyposis. N Engl J Med. 2000;342:1946–1952
  13. Peterson SK, Watts BG, McGivern B, Burke S, Latchford A, Phillips R, et al. FAP-affected adults’ responses to reported cardiovascular risks associated with celecoxib. 2005;Presented at Collaborative Group of the Americas on Inherited Colon Cancer (CGA), Salt Lake City, Utah, October
  14. Burke C, Hawk E, Eagle C, Woloj M, Church J, Hasson H, et al. Design and primary safety outcome in a phase I trial of celecoxib in children with FAP. Abstract P29 Familial Cancer. 2005;4(Supp1):77
  15. Lynch P, Burke C, Hawk E, Eagle C, Woloj M, Church J, et al. Colonoscopy in a phase I safety trial of celecoxib in children with FAP. Abstract F42 Familial Cancer. 2005;4(Supp1):57
  16. Information on clinical trials and human research studies. 2006;Available: www.clinicaltrials.gov/. Accessed October 16

PII: S0016-5085(06)02286-4

doi:10.1053/j.gastro.2006.10.052

Gastroenterology
Volume 131, Issue 6 , Pages 2003-2005, December 2006

Article Tools