Benefits of Alosetron: Serious Problem with Irritable Bowel Syndrome Studies Because of Very High Placebo Response Rate
In a review of 27 randomized placebo-controlled studies testing various treatments for irritable bowel syndrome (see below), the median placebo response rate was 47% (measured as % improved) with rates as high as 84% and 11 studies had placebo response rates of 60% or greater. The study concluded that this placebo response rate was approximately three times the size of the difference between placebo and drug response (median 16%).
That this problem of a large placebo response is applicable to alosetron can be seen in a reanalysis by Public Citizen’s Health Research Group of Glaxo data that we published in the Lancet, shown below. The mean pain and discomfort scores over a three-month period were quite similar in the alosetron and placebo groups even though there was a statistically significant difference between the groups as analyzed by Glaxo and the FDA.
The dose of alosetron used in this Lancet study (2 mg/day) is twice as much as the company is proposing (1 mg per day) in its new marketing plan being discussed at this meeting. According to the FDA Medical Officer review of the use of this lower dose in dose-ranging studies prior to approval, there is no adequate evidence from placebo-controlled trials that the 1-mg per day dose is significantly better than a placebo. However, there was evidence from those trials of an increased risk at the 1 mg dose a 4-fold increase in constipation severe enough to cause the patients to withdraw from the study compared with placebo. Thus, the Glaxo proposal for re-marketing Lotronex has a starting dose (1 mg per day) which lacks the proper evidence of efficacy required by the 1962 FDA Drug Efficacy Amendments but causes a significantly greater incidence of severe constipation.
Alosetron Risk Information
From our analysis of adverse event data from the FDA and the FDA briefing documents, as of 12/31/01 there have been 352 hospitalizations associated with the use of alosetron, the majority associated with gastrointestinal adverse reactions, 85 cases of ischemic colitis and 13 deaths, seven of which show a “strong association with alosetron” according to the FDA. Twenty-three patients required surgery because of complications from alosetron. That these reported cases are but the tip of the iceberg can be seen from clinical trial data included in an FDA memo by epidemiologist Dr. Zili Li who found that, in one large trial with adequate ascertainment of ischemic colitis, 10 out of 1819 women being treated with alosetron for diarrhea-predominant irritable bowel syndrome developed ischemic colitis over the 24 week duration of the trial (S3B30020). There were no cases in the 899 patients in that trial treated with traditional therapy. Given the results of this study and the 275,000 people who have used the drug, the 85 cases of ischemic colitis reported to the FDA after approval certainly represents the well-known under-reporting of hundreds of cases of ischemic colitis which may have actually occurred.
Glaxo has stated that the cases of ischemic colitis mainly occurred because the drug was not used properly. However, according to another FDA memo, of the first 70 cases of ischemic colitis reported to the FDA, in 80% of the cases the drug was prescribed as labeled. It should be noted that in 12% of these first 70 cases the patient was using the 1 mg/day dose being proposed for the new marketing plan.
The following table traces the development of incidence estimates for ischemic colitis from before approval until the present. Glaxo’s failure in its first estimate to factor duration of use accounts for that lower estimate. The recent Glaxo 1:700 estimate (12/01) of risk of ischemic colitis, based on 86 “studies,” is not believable because it dilutes the relatively small number of trials where there has been adequate duration of use and careful ascertainment of ischemic colitis with a large number of trials some of which involve a single dose and many of which do not have adequate ascertainment of ischemic colitis.
History of Incidence of Alosetron-induced Ischemic Colitis
Glaxo Estimates vs. FDA Estimates
Risk per 1,000 Person Years
6/99 FDA Medical officer
1/307 per 12 weeks
1/00 Glaxo label
1/700 (no time specified)
3/00 FDA epi
1/218 for 3 months
3/02 FDA epi
1/182 for 24 weeks
In two other instances of which we are aware, a drug that has been withdrawn from marketing was made available to patients under a restricted IND (Investigational New Drug) program. In the case of phenformin, about 3000 patients got the drug and more recently, for cisapride, approximately 1000 patients have obtained the drug after its withdrawal. We do not think that the necessary combination of safeguards to reduce the risk from alosetron can be accomplished under conditions of marketing; it needs to be done under a restricted IND. One of the ways being considered is the so-called Subpart H section of FDA regulations. In an April 10, 2002 internal FDA meeting, criteria for an ideal Subpart H (21 CFR 14) drug were listed:
Life-threatening disease; Disease is not prevalent; Disease well-defined Efficacy clear for clinical end-point; No therapeutic option (or H-drug more efficacious); Safety concern well-defined can lower risk.
None of these criteria are met in the case of alosetron.
In an excellent critique of Glaxo’s proposed marketing plan, the FDA pointed out that there were many elements missing from the Glaxo plan but present in a much more restricted marketing of the drug suggested by the FDA.
The items missing included: restricted prescribing by gastroenterologists only, patient registration, pharmacist registration, regular monitoring by physician, and exclusion of anyone without debilitating disease.
To this list we would add excluding people who have not previously used the drug and excluding those who had had constipation or other complications and requiring a thorough evaluation to exclude other gastrointestinal disease. We do not believe all of these restrictions are realistic for a marketed drug and, if the drug is to be made available, it needs to be under an IND.
With the exception of some drugs used to treat cancer, the frequency and severity of a life-threatening adverse reaction, in this case ischemic colitis in patients using alosetron, is amongst the highest I have seen for any other drug. This risk coupled with the marginal benefit beyond that seen with a placebo alone results in a risk benefit ratio clearly unfavorable to patients. The reintroduction of Lotronex into the market, even with the restrictions proposed by Glaxo, would be a serious public health mistake, likely, if not certain, to result in the need to ban the drug again.
 Spiller RC. Problems and challenges in the design of irritable bowel syndrome clinical trials: experience from published trials. Am J Med 1999;107:91s-97s.
 Barbehenn E, Lurie P, Wolfe SM. Alosetron for irritable bowel syndrome. Lancet 2000; 356(9246):2009-10.
 Medical Officer (John Senior, MD) Review of Alosetron, October 22, 1999 pages 1 and 20.
 Memo from Zili Li, MD, MPH, FDA, 3/15/02, Reevaluating the risk of ischemic colitis and its association with alosetron among female IBS patients in the United States.
 Zili Li and AC Mackey. OPDRA Postmarketing Safety Review. Ischemic colitis and severe constipation. February 23, 2001.
 Alosetron: Review of risk management plan. Toni Piazza-Hepp, Pharm. D. Associate Director, Division of Risk Evaluation, FDA (CDER). March 26, 2002.
My name is Paul Stolley and I was formerly the Chairman of the Department of Epidemiology and Preventive Medicine at the University of Maryland School of Medicine at Baltimore. I am co-author of Foundations of Epidemiology textbook and currently work half-time at the Public Citizen Health Research Group. During the academic year 2000/2001 I worked 80% time at the FDA as a consultant in Epidemiology for the group that collects and evaluates adverse drug reactions. I co-authored and signed the FDA memo of November 13,2000 that preceded the November 28th decision by Glaxo to withdraw Lotronex from the market.
In that memo, we argued that there were compelling reasons for the withdrawal of Lotronex from the market. Below are the main points we made in that memo. Because data accumulated since then have only made these points stronger, there are today no valid arguments for re-introducing alosetron as a marketed drug.
- The drug is minimally effective and for only the diarrhea-predominant form and only in women.
- The price paid for this gender-specific diarrhea-predominant efficacy is much too high: ischemic colitis that can result in surgery, colectomy and death; severe constipation that can require hospitalization and surgery; mesenteric artery thromboses requiring surgery and, rarely, causing death.
The rate of ischemic colitis associated with the drug is remarkably elevated and beyond dispute, as there were 16 cases in the alosetron-treated arms of the clinical trials and only 1 case in the placebo arm. While the drug is only approved for 12 weeks of use, in actual practice this chronic condition may be treated indefinitely. The rate of ischemic colitis associated with Lotronex may be as high as 1 per 300 users in just a 12-week period. While many of these colitis episodes have not led to serious damage, there have been perhaps 7 or more reported fatalities and numerous surgical interventions.
The specious argument has been made that ischemic colitis is a feature of Irritable Bowel Syndrome (IBS). However, when the FDA searched its own adverse drug reaction files for reports of ischemic colitis, no reports of ischemic colitis were found associated with loperamide or diphenoxylate, drugs much more widely used to treat IBS than alosetron. And the randomized studies are most convincing linking the adverse effect to the drug.
This drug cannot be “risk-managed” as the adverse-effects mimic the disease the drug is supposed to treat: abdominal pain, constipation and diarrhea. And the ischemic colitis and mesenteric artery thrombosis often have no warning symptoms. A patient attending a busy Emergency Room complaining of “constipation” while using Lotronex is likely to get triaged to the following week.
Therefore, the data produced by FDA and which led to the November withdrawal by Glaxo have not been seriously challenged. Lotronex is a drug that should never have been approved, was withdrawn because of its dangerous adverse reaction profile only 9 months after it was introduced, and should not return to market. If it is returned to the market, there are enough data available to confidently predict that serious ischemic colitis leading to surgery and death will again appear and the drug will be withdrawn once again, having produced yet a second needless mini-epidemic of drug-induced colitis and death in female patients with irritable bowel syndrome.