FDA Briefing Document

 

 

 

 

 

Gastrointestinal Drugs Advisory Committee

 

 

 

 

 

July 14, 2004

 

 

 

 

 

 Zelnorm^® for Treatment of Chronic Constipation
 Gastrointestinal Drugs Advisory Committee on Zelnorm^® for Treatment of Chronic Constipation

 

FDA Briefing Document

 

 

Novartis Pharmaceuticals Corporation submitted a Supplemental New Drug Application (21-200/S-005) on October 20, 2003 seeking approval of Zelnorm (6 mg bid) for the treatment of chronic constipation.  Zelnorm is a 5-HT₄ partial agonist with moderate affinity for the 5-HT₁ receptor.  It was first approved in July 2002 for the short-term treatment (4-6 weeks) of women with constipation predominant irritable bowel syndrome (c-IBS).  The therapeutic mechanism of action is based primarily on its agonist action on 5-HT₄ receptors, resulting in augmented bowel motility, increased intestinal secretion and inhibition of visceral sensitivity.  Two clinical studies were submitted in support of the chronic constipation indication.

 

This briefing document for the Gastrointestinal Drugs Advisory Committee meeting consists of three sections:

 

1.    Clinical Summary of Efficacy  (pages 4-16)

2.    Clinical Summary of Safety (pages 17-64)

3.    Draft Statistical Review and Evaluation (pages 65-118)

 

This document contains information from IMS Health National Prescription Audit Plus and is not to be used outside of the FDA without prior clearance by IMS Health.

 

 

Issues for Discussion:

 

1.    Efficacy

a.    Discuss the appropriateness of a primary efficacy endpoint of an increase of ≥1 complete spontaneous bowel movement per week vs. ≥3 complete spontaneous bowel movements per week.

b.    Only 9 to 16% of subjects were ≥65 years of age and the treatment effect was significantly smaller in older patients.  Are these data adequate for an indication that is common in the elderly?

c.     Only 9 to 14% of the subjects were male and the treatment effect was smaller in males than females.  Are these data adequate to support approval of Zelnorm for use in the treatment of chronic constipation in males?

d.    Is the population studied representative of patients with chronic idiopathic constipation?

e.    Are the clinical trial data adequate with respect to the population with chronic constipation that is likely to be treated with Zelnorm?

f.      Is Zelnorm effective for the treatment of chronic constipation?

2.    Safety

a.    Post-marketing cases of ischemic colitis and serious complications of diarrhea were not limited to patients with IBS.  What are the implications of these adverse events for patients with chronic constipation?

b.    The incidence of diarrhea and discontinuations due to diarrhea was higher in patients ≥65 years of age.  Is there sufficient information that Zelnorm is safe for use in this age group?

c.     Do the adverse event data from the clinical trials and post-marketing surveillance provide adequate evidence of safety of Zelnorm for the treatment of chronic constipation?

d.    Should the information on the post-marketing cases of ischemic colitis and intestinal ischemia be moved from the PRECAUTIONS section to the WARNINGS section of the package insert?

 

Clinical Summary of Efficacy

 

 

Robert Prizont, M.D., Division of Gastrointestinal and Coagulation Drug Products

 

I.               Background.

 

Zelnorm^® (tegaserod maleate tablets), is a drug with affinity for 5-HT₄ receptors located in the smooth muscle, particularly bowel smooth muscle.  The agonist action of Zelnorm on these 5-HT₄ receptors results in augmented bowel motility.  On July 2002, Zelnorm was approved for the short-term (4-6 weeks) treatment of women with irritable bowel syndrome (IBS), in whom the primary bowel symptom was constipation (IBS-C).  The approved dosage was a 6 mg tablet taken twice a day.  To support safety and efficacy of Zelnorm in IBS-C patients, Novartis conducted three randomized, double-blind, placebo-controlled, multi-center studies. Eligible patients were required to have no less than a 3-month history of abdominal pain, bloating and constipation, the classical symptoms of IBS-C according to the Rome Criteria¹. Approval of Zelnorm^® for IBS-C was based on a higher proportion of responders in the Zelnorm^® groups compared to the placebo groups.  In patients treated with Zelnorm^® there was an increase (5%) in diarrhea associated with use of the drug.  There was also a small numerical increase in the rate of abdominal surgeries, particularly cholecystectomies. Post-marketing, there have been reports of ischemic colitis, death, hypotension, syncope, and serious complications of diarrhea. The Zelnorm^® label has been updated to include these serious adverse events.

 

In this submission, Novartis proposes the use of Zelnorm^® (tegaserod maleate) for the treatment of chronic constipation at a dose of 12 mg/day (6 mg bid).  Chronic constipation is a common, benign, functional disorder of the lower gastrointestinal tract, affecting primarily populations in Western countries². The presently accepted definition of constipation includes number of bowel movements as the main objective evidence to assess constipation, plus the subjective symptoms of straining, lumpy stools, and sensation of incomplete evacuation.  The Rome II Criteria summarizes the consensus on the definition of functional or idiopathic constipation (taken from the Lembo and Camilleri article).

 

 

Functional or idiopathic constipation is the most common form of constipation.  It affects equally men and women, particularly elderly, and the main complaint is infrequency of BMs.  Outlet delay or outlet obstruction constipation, and slow-transit constipation, affect women and represent, in epidemiological studies, a smaller proportion of the constipation population^3,4.  IBS-C is more predominant in younger and middle age women, and manifests by abdominal pain, abdominal distention or bloating as main associated symptoms.  There is a plethora of available over-the-counter laxatives, enemas, and bulk-forming agents⁵.  Women with outlet constipation are more refractory and less responsive to treatment. Women with IBS-C respond similarly to women with outlet obstruction constipation.  Of relevance, improvement of the abdominal symptoms in IBS-C, and outlet obstruction, may parallel improvement in constipation. 

 

 

II.             Efficacy.

 

A.    Relevant Points of the Prospective Protocol.

 

• 2-week baseline, 12-week randomized treatment, 4 week withdrawal period. 

 

• To be randomized, screened patients had to fulfill the constipation criteria, i.e., (a) less than three spontaneous bowel movements per week that result in a feeling of complete evacuation, (b) at least 25% of stools are very hard and/or hard stools, (c) sensation of incomplete evacuation in at least 25% of the bowel movements, and (d) straining on at least 25% of the defecations.

 

• Three treatment groups: placebo, tegaserod 2 mg bid, tegaserod 6 mg bid. 

 

• Eligible patients had to meet the Inclusion Criteria:

 

ü     Males and females 18 years of age (no upper limit)

ü     History of constipation, as defined above, for at least 6 months before screening.  Patients were required to have negative structural bowel disease as demonstrated by a radiology/endoscopy performed during the 5 years prior to the trial.

 

• Excluded were patients who had the following history or diagnosis:

 

ü     Evidence of cathartic colon or evidence of laxative abuse,

ü     Chronic constipation due to bowel, gynecological surgery, neurological diseases, connective tissue disorders affecting muscle. Clinical evidence of (including ECG, lab tests) of endocrine/metabolic disorders (including insulin-dependent diabetes), cardiovascular, respiratory, liver, gastrointestinal, hematology, or any disease that may have interfered with patients completing the study.

ü     Organic gastrointestinal diseases affecting the colon

(There was no specific exclusion of IBS patients, by the Rome I or II Criteria).

 

• The Primary Efficacy Endpoint: the response for the first 4 weeks of double-blind treatment period using the following criterion:

 

ü     A mean increase of I or more complete spontaneous bowel movement (CSBM) compared to baseline

ü     At least 7 days in study for the first 4 weeks of double-blind treatment period

 

• Secondary efficacy endpoints included (1) response rate throughout the 12 weeks of treatment, (2) evaluations of bowel habit (3) QOL.

 

• Bisacodyl was used as rescue medication. Bulk-forming agents were allowed.

 

B.    Summary of Demographics and Efficacy Results Submitted by Novartis.

 

To support the use of Zelnorm^® (tegaserod) 6 mg bid in chronic constipation, Novartis conducted two multi-center, randomized, double-blind, dose-ranging, placebo-controlled clinical studies.  The studies, coded by Novartis as Study HTF919E2301 and Study HTF9192302, will be identified here as Studies 2301 and 2302.  Study 2301 enlisted 128 centers from European countries, Turkey, Australia and South Africa.  Study 2302 enlisted 101 centers from the USA (71), Canada, Argentina, Brazil, Colombia, Venezuela and Chile.

 

The prospective protocol planned a randomization of 1185 patients in each trial.  Study 2301 randomized a total of 1264 patients. Study 2302 randomized a total of 1348 patients. The trials enrolled a majority of women (86-91%),with a mean age of 46-47 years of age, who were Caucasian (84-98%).   

 

1.    Study 2301. Relevant Efficacy Results.

 

The primary efficacy endpoint was an increase in ≥1 CSBM/week during the first month of study.  The increase in the CSBM/week represents the average CSBMs for the 4 weeks, e.g., a patient with 8 CSBMs during the first week of study and no other CSBMs would have averaged 2 CSBM/week during the first month. The primary efficacy results are illustrated in the next table.  As shown in Novartis Table 9-1, 40% of patients treated with 6 mg tegaserod (teg 6 mg), and 36% of patients treated with 2 mg tegaserod (teg 2 mg) were responders to the therapy.  The placebo (P) response of 27% was rather high. These results revealed a therapeutic gain relative to placebo of 13.5%  for the teg 6 mg, during the first month.  The therapeutic gain was lower for the teg 2 mg (9%).  Differences between tegaserod doses and placebo were statistically significant.

 

 

The difference between the teg 6 mg and placebo remained statistically significant over the 12 weeks of study, although therapeutic gain relative to placebo was only 12.6%.  Assessment of drug responses for the entire length of study, revealed no significant difference between placebo and the 2 mg tegaserod dose.  The loss of a significant therapeutic gain in the 2 mg teg was largely due to a 4% increase in placebo response over the 12 weeks of study. The weekly response over the 12 week study period is graphically illustrated in the Novartis Figure 9.1 shown below.

 

 

The degree of therapeutic gain is decreased if responders are defined as those patients who no longer meet the definition of constipation (<3 BM/week).  As seen in Novartis Table 9-3, in the first month, only 22% of the 6 mg teg patients were responders who no longer met the definition of constipation.  Although the difference against placebo is significant, the therapeutic gain is 9.3%.

 

 

Most of the patients perceived marked improvement in the number of stools per week.  Stool characteristics, percentage of SBM with a sensation of complete evacuation, and bothersome bowel habits, abdominal discomfort/pain, and abdominal distention/bloating were the relevant secondary efficacy data gathered from patients’ daily diaries.  Among the patients treated in Study 2301,  the 12 week diaries did not reveal a significant difference between placebo and the 6 mg teg in the percentage of bowel movements with a sensation of “complete” evacuation; patient perception of completeness being the qualifying assessment for relief of constipation added by Novartis.  Although there was improvement from baseline in relief of abdominal symptoms, the difference between placebo and the tegaserod in the mean (%) weekly improvement was not significant.

 

2.    Study 2302. Relevant Efficacy Results.

 

Differences between the tegaserod 6 mg bid doses and placebo noted in Study 2301 were replicated in Study 2302.  For the primary efficacy endpoint, the therapeutic gain for either tegaserod dose relative to placebo was 15-17% during the first 4 weeks.  The average CSBM/week remained stable over the 12 week study period and was significantly higher for both tegaserod doses relative to placebo.  Notable is the absence of dose response observed in this study compared to Study 2301.  The response rate for the 2 mg teg group was just over 40%, persisted during the study, and was basically similar to the response rate depicted in the 6 mg teg group.  The results during the first month are shown in the next Novartis Table 9-1.

 

 

As in study 2301, the therapeutic gain for tegaserod decreased markedly if responders were defined by the absence of constipation (≥3BM/week).  The response rate for the 6 mg teg treatment group went down to 21.5%, and the therapeutic gain over placebo narrowed to 8.9% (see next Novartis Table 9.4).  Again, no dose response was noted.

 

 

Analyses of secondary endpoints were favorable to the 6 mg tegaserod group relative to placebo.

 

3.    Use of Rescue Laxative Use.

 

According to the protocol, investigators were allowed to administer bisacodyl tablets as rescue laxative to patients who had a period of 4 days without BMs. This 4 day period was not adhered to in Study 2301.  In this study, patients were given the rescue laxative after a period of 3 days without BMs.  In both trials, laxative use was higher during baseline than the double-blind period.  However, during the double-blind period, 50-58% of patients treated in Study 2301 used rescue laxative sometime during the trial.  The proportion of patients who used the rescue laxative was higher in Study 2302, 60-64%.  In Study 2301, placebo patients had a significantly higher mean number of days of laxative consumption compared to tegaserod –treated patients, though the difference was not significant, as shown in the next figure (Novartis Figure 8-1).

 

 

The difference in laxative use between the tegaserod doses and placebo observed in Study 2301, was not replicated in Study 2302.  As seen in the next Novartis Figure 8-1, the mean number of days that patients used the laxative bisacodyl, was similar in the tegaserod and placebo treatment groups.

 

 

C.    Reviewer Comments.

 

This reviewer acknowledges Novartis analyses and results.  The submitted studies include fundamental deficiencies related to study design, including issues regarding patient representation and choice of primary efficacy endpoint, and issues regarding the clinical significance of the results.  These issues are the topics of discussion included in my next comments.

 

1.    Patient Representation for the Proposed Indication.

 

The prospective study protocol defined the aim for conducting the trials, i.e., to demonstrate the effect of tegaserod on bowel habits in patients suffering from chronic idiopathic constipation.  The large epidemiological constipation study conducted in the USA (EPOC) defined the subtypes of constipation: functional or idiopathic, IBS-C, and outlet obstruction (outlet) or slow-peristalsis constipation.  Idiopathic or functional constipation, encompassing the largest subtype, is almost equally distributed among males and females, and actually represents the most common subtype among men⁶. The population studied by Novartis, 86-92% women, mean age of 46-47 years, does not appear to represent functional or idiopathic constipation, and strongly suggests a population largely comprised of IBS-C, outlet obstruction or slow peristalsis patients.

 

Inclusion of patients with IBS-C appears to be corroborated by the history of symptoms.  The main constipation complaint suffered by up to 30% of enrolled patients was abdominal distention/bloating and abdominal pain was reported by 15% (see next Novartis Table 7.6-1, Study 2301).  The proportion of subjects reporting such symptoms was similar for Study 2302. Although abdominal pain, abdominal distention and bloating may be present in idiopathic constipation, they are rarely main complaints. The Rome Criteria, universally accepted as the criteria for diagnosis of idiopathic constipation, does not include abdominal symptoms.

 

 

The lack of exclusion criteria for IBS-C patients allowed enrollment of patients with a prior confirmed diagnosis of IBS, including patients who appeared to meet the criteria of diarrhea-predominant IBS.  The next Novartis table revealed that at least 595 patients (23%) enrolled in the two controlled studies had IBS-like symptoms. Included in this group were patients who had a diagnosis of IBS prior to study entry. 

 

 

In this table, Novartis excluded patients who had as their main complaint abdominal distention/bloating and who appeared to have characteristics of either IBS-C or outlet obstruction constipation.

 

Further assessment of the majority of these young to middle aged women revealed characteristics of patients affected by outlet obstruction or slow transit constipation.  The main characteristic of outlet obstruction or slow transit constipation is severity of constipation (one bowel movement per week and weeks with absence of bowel movements).  The examination of baseline frequency of CSBM/week revealed that a total of 1638 patients (63%) randomized to treatments in these studies had shown complete absence of CSBM at baseline (0 CSBM).

 

2.    The Primary Efficacy Endpoint.

 

Analyses of efficacy based on the primary endpoint established in the protocol, i.e., ≥1 CSBM than at baseline, showed that treatment with 12 mg/d Zelnorm resulted in ≥13% superiority over placebo. In considering these numbers, we should more closely examine the relevance of the chosen primary efficacy endpoint, as it relates to the definition of constipation.  Does the chosen efficacy endpoint encompass relief of constipation (defined by ≥3 CSBM per week) or does the chosen endpoint simply refer to an increase in one BM/week, but without actual relief of constipation?  The latter appears to be the case.  Taking this a setp further, we examined the proportion of patients exhibiting zero BMs at baseline and declared responders with an average of just one single bowel movement per week during the initial month of treatment.  As seen in the comparisons of data below, 13% to 18% of patients were considered responders with an average of one single BM/week during weeks 1-4 of the twelve week study period (analyses performed by the FDA statistician reviewer, Dr. Joy Mele).  Minimal differences across treatment groups were observed in this exploratory analysis.  The relationship between a change from 0 to 1 CSBM/week and symptom improvement is under review.

 

Study 2301

   Increase=1 CSBM/wk.  Wks 1-4 for patients with 0 CSBM at baseline

                    PLA =   13%   34/266

                    ZEL 2 = 14%  36/253

                    ZEL 6 = 13%  35/273

Study 2302

   Increase=1 CSBM/wk.  Wks 1-4 for patients with 0 CSBM at baseline

                    PLA =    13%   35/274

                    ZEL 2 = 19%  55/289

                    ZEL 6 = 18%  51/283

 

To further ascertain this point, we analyzed the number of weeks, out of the 12 week study period, in which patients were declared responders, and, met the definition of non-constipation, i.e., ≥3 BM per week.  The next table (prepared by Dr. Joy Mele) shows that responders met the definition of complete relief from constipation (≥3 BM/week) less than 25% of the twelve week study period.

Number of Weeks with 3 or More CSBM

	Study 2301			Study 2302
	PLA n=416	ZEL 2 n=417	ZEL 6 n=431	PLA n=447	ZEL 2 n=450	ZEL 6 n=451
All patients   Mean (SD)   Median    Range	2.2 (3.3) 0 0-12	2.6 (3.5) 1 0-12	3.2 (3.9) 1 0-12	2.2 (3.2) 0 0-12	3.1 (3.9) 1 0-12	3.3 (3.7) 2 0-12
Completers*   Mean (SD)   Median    Range	2.4 (3.3) 0 0-12	2.9 (3.7) 1 0-12	3.6 (4.1) 2 0-12	2.5 (3.4) 1 0-12	3.5 (4.0) 2 0-12	3.8 (3.8) 3 0-12
# of wks w/ 3 or more CSBM    0    1    2    3    4    5    6    7    8    9   10   11   12	52.6% 12.2% 7.3% 4.6% 4.6% 3.9% 1.5% 1.7% 2.2% 2.9% 2.4% 1% 3.2%	46.7% 10% 7.3% 6.1% 5.6% 3.9% 3.7% 3.4% 2.7% 3.4% 1.2% 2.4% 3.4%	40.2% 10.9% 6.4% 6.9% 4.5% 4.0% 3.5% 4.5% 3.1% 3.3% 4.3% 3.8% 4.7%	51.6% 10.8% 6.5% 7.1% 5.1% 3.7% 1.8% 3.0% 2.5% 2.8% 1.2% 2.1% 1.8%	38.3% 14.5% 8.2% 5.7% 3.2% 5.2% 4.3% 2.7% 4.1% 1.6% 3.4% 3.9% 5%	34.8% 13.6% 5.9% 7% 5.7% 5.4% 6.8% 4.3% 4.3% 2.5% 3.2% 3.6% 2.9%

          *received study drug treatment for 12 weeks

 

3.    Clinical Relevance.

 

It has been estimated that between 4 million to 55 million people in the U.S. are affected by constipation⁷.  Idiopathic constipation, with almost equal prevalence in men and women appears to be the most prevalent of the subtypes of constipation. Laxatives, whether bulk-forming agents, osmotic products, or stimulants of the intestinal mucosa are easily available OTC.  A few, like PEG-3350 and non-absorbable disaccharides, require a physician prescription.  The population affected by constipation, in its majority, self-medicate with laxatives.  Many of the habitual laxative consumers, become laxative abusers (as mentioned in Novartis protocol).  In its 2000 technical review, the American Gastroenterological Association (AGA) recommended to initiate the medical treatment of constipation with a slow increase in the content of dietary fiber.  If drugs are required, the AGA⁸ recommends starting with a saline laxative, such as milk of magnesia. Only later, stimulant drugs or osmotic agents should be added to the therapy. Randomized trials with high fiber diets, or comparing laxatives, have been carried out, albeit many of them, under deficient designs⁹.  Trials revealed little difference between laxatives, and modest improvement over placebo¹⁰.  Hence, the wide interest in the results of these tegaserod trials in chronic constipation.  Specifically, the Novartis aim was to assess safety and efficacy of tegaserod maleate in idiopathic constipation.

 

4.    Conclusion

 

At first glance, the results revealed significant therapeutic gain for tegaserod 6 mg over placebo ranging from as high as13% to as low as 9% depending on the methodology applied for analysis. Dose response was shown only in one trial.  Careful examination reveals deficiencies in study design, in study execution, and robustness of results. The design of the studies excluded patients considered laxative abusers, and lacked a provision to exclude patients with IBS-C, a subtype of constipation for which tegaserod is already approved for use under prescription.  This lack of provision to exclude IBS-C led to contamination of the total enrolled patient population with almost 600 patients who met the criteria of IBS-C (a few of them met the criteria of IBS-diarrhea predominant).  In the execution of the studies, men and the elderly were underrepresented (discussed in the draft statistical review in greater detail), and the studied patient population was young or middle age women, 46 years old.  A large proportion (≥63%) of these women exhibited severe constipation at the run-in baseline period (0 CSBM) coupled with abdominal symptoms.  This latter clinical picture is reminiscent of the clinical picture encountered in outlet obstruction or slow transit constipation.  It appears, therefore, idiopathic constipation patients, if present, constituted a minority 9379 or only 15%) of enrollees. A further fundamental deficiency in the design, i.e. choice of primary efficacy endpoint, was subsequently manifested in the results.  About 18% of patients with 0 CSBM/wk at baseline were declared responders with only 1 CSBM per week.  Responders to treatment were non-constipated for approximately 42% of the 12-week study treatment.

 

In acknowledging the favorable statistics toward tegaserod, this reviewer ponders about the clinical significance of these efficacy results, in the lifelong treatment of chronic constipation, and rather pointedly, in the lifelong treatment of idiopathic, outlet obstruction or slow transit constipation. 

 

References

 

1.    American Gastroenterological Association Medical Position Statement: Irritable Bowel Syndrome.  Gastroenterology; 112:2118-2119, 1997.

 

2.    Lembo A and Camilleri M. Chronic Constipation. NEJM; 349:1360-1368, 2003.

 

3.    Wald A.  Approach to the Patient with Constipation. Pages 894-910, Textbook of Gastroenterology; Ed T. Yamada, 2003.

 

4.    Lennard-Jones JE.  Constipation.  Pages 181-210, In Sleisenger & Fordtran’s Gastrointestinal and Liver Disease; 7^th Edition, 2002.

 

5.    Constipation and the Use of Laxatives; www.terrubins.com

 

6.    Epidemiology of Constipation (EPOC) Study in the United States: Relation of Clinical Subtypes to Sociodemographic Features. American J Gastroenterol; 94, 3530-3540, 1999.

 

7.    Everhart JE et al. A longitudinal study of self-reported bowel habits in the United States. Dig Dis Sci; 43:1153-1162,1989.

 

8.    AGA Technical Review on Constipation. Gastroenterology; 119:1766-1778, 2000.

 

9.    Dettmar PW, Sykes J. A Multi-Centre, General Practice Comparison of Ispaghula Husk versus Lactulose and other Laxatives in the Treatment of Simple Constipation.  Curr Med Res Opin; 14:227-233, 1998.

 

10. Jones MP et al. Lack of Objective Evidence of Efficacy of Laxatives in Chronic Constipation.  Dig Dis Sci; 47:2222-2230, 2002.

 

Clinical Summary of Safety

 

 

Gary Della’Zanna, D.O., M.S., Division of Gastrointestinal and Coagulation Drug Products

 

Ann Corken Mackey, R.Ph., M.P.H. and Allen Brinker, M.D., M.S., Division of Drug Risk Evaluation, Office of Drug Safety

 

I.      Background

 

Zelnorm was first approved in July 2002 for the short-term treatment (4-6 weeks) of women with constipation predominant irritable bowel syndrome (c-IBS).  Since approval, several adverse events of special interest have been identified; these include ischemic colitis, rectal bleeding, and serious complications of diarrhea, including hypotension and syncope.  In response to post marketing reports, Novartis revised the Zelnorm package insert on April 27, 2004.  These revisions included a WARNINGS section about the serious consequences of diarrhea, including hypovolemia, hypotension, and syncope and a PRECAUTIONS section describing ischemic colitis and other forms of intestinal ischemia.  In addition to this, Novartis mailed a Dear Health Care Professional Letter outlining these changes (see Appendix 1 and 2). 

 

With the proposed new treatment indication, chronic constipation, and the recent labeling change, the Division is asking the Gastrointestinal Drugs Advisory Committee to discuss the risk:benefit profile of Zelnorm for the proposed indication and to address the adequacy of the labeling for adverse events of special interest.  This review will discuss safety data from the constipation clinical trials, from pooled clinical trials, and from postmarketing surveillance.       

 

II.     Chronic Constipation Clinical Trials

 

The chronic constipation studies consisted of two clinical trials (E2301, E2302) lasting 12 weeks each (Key Safety Population) and a long-term extension of study E2301, (E2301E1) lasting an additional 10 months (Key Long-term Safety Population).  The type and incidence of adverse events reported during the chronic constipation trials were similar to the c-IBS trials and to what is printed in the current label.  Table 1 lists the most frequent adverse events in the chronic constipation trials in the key safety population.

 

Table 1

Most Frequent Adverse Event (³ 2% of patients in any group)

Key Safety Population

 

 

During the uncontrolled long-term extension study (E2301E1), adverse events followed a similar pattern as seen in the key safety population, although the incidence rates were generally higher.  Interestingly, constipation was reported more frequently as an adverse event during long-term extension. Table 2 lists the most frequent adverse events in the chronic constipation trials in the key long-term safety population.

 

Table 2

Most frequent Adverse Event (³ 2% of patients in any group)

Key Long-Term Safety Population

 

This safety review will focus on selected adverse events of special interest.  The review will include post-marketing data and the safety data from the chronic constipation trials, as well as pooled data from 33 additional clinical trials for other indications (Pooled Indication Population).

 

III.     Drug Use*

 

IMS Health (projected data): There were a total of ------------ prescriptions for tegaserod 6 mg and ---------- prescriptions for tegaserod 2 mg dispensed by retail pharmacies (chain, independent, food stores, and mail order) in the U.S from August 1, 2002 through April 30, 2004. The chart below depicts quarterly totals of prescriptions dispensed (tegaserod 2 mg and 6 mg combined). (Drug use data for August 2002 was minimal and is not included in the chart.)

 

 

 

 

Drug Use Demographics^†

 

-------------------------------------------------------------

 

 

* Drug use data provided by Yoon Kong, Pharm.D., Drug Utilization Specialist, Division of Surveillance, Research and Communication Support, ODS. 

^† Projected data per IMS. Note that prescriber specialty data represent the August 2002 through April 2004 time period; indication for use data and gender/age data represent the August 2002 through March 2004 time period.

 

 

IV.  Adverse Events of Special Interest

 

The Adverse Event Reporting System (AERS) is a passive surveillance system that is subject to under-reporting; normally only 1 to 10% of adverse events are reported to FDA (Physician knowledge, attitudes, and behavior related to reporting of adverse drug events. Arch Intern Med 1998; 148: 1596-1600 and Rhode Island physicians’ recognition and reporting of adverse drug reactions. R I Med J 1987; 70: 311-6).

 

The post-marketing cases discussed in this document were reported through the Adverse Event Reporting System (AERS) between the initiation of marketing in the US in August 2002 through April 15, 2004. The applicant’s data lock point was March 31, 2004, so the April 15 date allows for the applicant’s reports to be received and processed by the agency. Note that a paragraph in section B (Ischemic Colitis and Rectal Hemorrhage) discusses cases of ischemic colitis and intestinal ischemia received between April 15 and June 1, 2004.   These reports were included in order to capture reports subsequent to dissemination of the Dear Health Care Professional letter by the sponsor on April 26, 2004.

 

When evaluating spontaneous reports, it is important to keep the following limitations in mind. The main utility of a spontaneous reporting system, such as AERS, is to detect signals of potential drug safety issues that are rare. It should be realized that accumulated case reports cannot be used to calculate incidence or estimates of drug risk for a particular product because under-reporting of adverse events exists. Some of the factors that influence reporting include the length of time a drug is marketed, the market share, size and sophistication of the sales force, publicity about an adverse reaction and regulatory actions. It should also be noted that in some of these cases, the reported clinical data were incomplete, and there is no certainty that these drugs caused the reported reactions.  A given reaction may actually have been due to an underlying disease process or to another coincidental factor.

 

Some of the reports received through AERS were submitted by patients; in general the quality and completeness of the data are not as good as reports received from health care professionals. ODS has included these reports in our analysis because the actual occurrence of these events could not be ruled out. The absence of supporting documentation does not imply that the patient did not have the event, only that documentation was not obtainable.

 

A.   Fatalities

As of April 15, 2004, there were a total of 22 deaths from all causes in patients receiving tegaserod in AERS (note that some of these cases also are included in the Ischemic Colitis and Rectal Hemorrhage section below). This number represents unduplicated patient cases, not individual reports. The cases are described below.

 

Causes of death (mutually exclusive)

Total (n=22)

 

Sepsis related to ischemic bowel disease (n=1)

Bowel infarction/peripheral vascular disease (n=1)

Intestinal gangrene/bowel perforation (n=1)

Intestinal ischemia (n=1)

 

Other causes (n=18) (e.g., cardiac arrest, suicide, coma/diabetic neuropathy, bulbar palsy, renal failure, MI, bowel impaction, complications of anorexia, cancer)

 

Eighteen of the patients were female and four patients were male, ranging in age from 32 to 90 years, with a mean age of 67 years. These patients were taking tegaserod for the following indications: IBS constipation predominant (7), IBS predominance not specified (2), IBS alternating predominance (2), constipation (6), paralytic ileus (1), and unknown indication (4).

 

B.     Ischemic Colitis and Rectal Bleeding

 

1. Post-Marketing Surveillance

 

Ischemic colitis, and other forms of intestinal ischemia, were identified as adverse events of special interest.  The first reported post-marketing case of intestinal ischemia in a patient receiving Zelnorm was identified by the Office of Drug Safety (ODS) in March 2003, when a search of the Adverse Event Reporting System (AERS) database for rectal bleeding was performed.

 

As of April 15, 2004, the Agency received 20 reports of ischemic colitis and 4 reports of intestinal ischemia through AERS.  The definition used by ODS to identify potential cases of ischemic colitis for epidemiological risk assessment was based on either of the following: (1) the term ischemic colitis was explicitly used in the AERS report as a possible diagnosis, or (2) any endoscopic or histologic evidence of ischemic change or necrosis.  The search criteria were extended to include other forms of intestinal ischemia.  The definition for intestinal ischemia included cases where an occlusive process of the proximal large vessels of the bowel was suggested.

 

A summary of the ischemic colitis and intestinal ischemia cases is provided in Appendix 3.  Of the 20 cases of ischemic colitis, 19 were female, ranging in age from 26 to 82 years, with a mean age of 55.  The majority of the patients were treated for IBS constipation predominant (n=10), IBS predominance unspecified (5), and IBS alternating predominance (1).  The remaining four patients were treated off label [constipation (n=2), postoperative ileus (n=1), unknown indication (n=1)].  Five of the 20 reported cases of ischemic colitis had no documented risk factors.  The remaining 15 patients had one or more identifiable risk factors (i.e., hormone therapy, tobacco use, and vascular disease).  Three of the 20 reported cases occurred on the first day of therapy, with two of the three cases occurring in patients with no known risk factors.  The other times to onset were: 2 to 20 days (6), 21 to 122 days (7), 230 to 398 days (3), and unknown (1).  Thirteen of the 20 patients required hospitalization (with one of these thirteen required surgery) and one died.

 

The four cases defined as intestinal ischemia included the following diagnoses: intestinal ischemia (n=1), intestinal gangrene (n=1), mesenteric ischemia (n=1), and abdominal compartment syndrome with intestinal ischemia (n=1).  All patients were female, ranging in age from 41 to 67 years.  Three of the patients were treated for IBS; one was treated off label for constipation.  The times to onset were 6, 56, and 105 days (1 case unknown).  Three patients were treated with surgery.  One patient required a bowel resection; the other two had exploratory laparotomies.  Three of the four patients died.    

 

The majority of these cases have been adjudicated with Novartis and for the most part there is agreement that the cases represent some form of bowel ischemia. Many of the post-marketing cases of ischemic colitis and intestinal ischemia had confounding factors that may have contributed to the development of intestinal ischemia.  Of the 20 cases of ischemic colitis, six were receiving hormonal therapy, which can be associated with vascular thrombosis and coagulopathies.  Several had complicated medical histories.

 

ODS has received 7 cases of ischemic colitis, 1 case of bowel infarction, and 1 case of ischemic colitis secondary to small vessel ischemia from April 15, 2004 through June 1, 2004 (ODS is waiting for additional information on some of these cases). All 9 patients were female, ranging in age from 31 to 78 years, with a mean age of 45 years. The patients were treated for the following indications: IBS constipation predominant (3), IBS predominance unspecified (2), IBS alternating predominance (1), constipation (1), “bloating” (1), and unknown indication (1). They were receiving the following daily doses:  4 mg (1), 6 mg (2), 12 mg (3), 6 mg every other day (1), and unknown (2). Times to onset were: 1 to 8 days (3), 16 to 41 (3), and 156 to 293 (3). Three of the seven patients had risk factors (i.e., marathon running [1] and vascular disease [2]). Six patients required hospitalization, with one of those patients requiring a small bowel resection.

 

Rectal bleeding was also analyzed using the post-marketing data and the safety data from the current application.  Rectal bleeding is difficult to assess using post-marketing data.  Spontaneous reporting systems are designed for detection of rare and serious adverse events.  The definition for epidemiological risk assessment included any AERS report using the terms rectal bleeding, rectal hemorrhage, bloody stool, hematochezia, lower gastrointestinal bleeding, or melena.  

 

As of April 15, 2004, ODS received 40 AERS reports of rectal bleeding (note: cases of rectal bleeding resulting from other processes [e.g., inflammatory bowel disease, and ischemic colitis] have been excluded from this number).  The quality of the information for many of the reports was poor.  Fourteen (35%) reports were submitted by patients, most of which provided very little information.

 

Almost half of the post marketing cases of rectal bleeding (n=19) originated from foreign sources.  Eighteen (18) of the 40 cases had diagnostic workups that demonstrated the following: normal exam (n=9), hemorrhoids (n=2), polyp and hemorrhoids (n=1), rectal irritation (n=2), diverticulum (n=3), and angiodysplasia (1). 

 

2.  Clinical Trials

 

A thorough review of the safety data from the chronic constipation trials did not identify any cases suspicious of ischemic colitis.  The incidence and severity of rectal bleeding in the key safety population were balanced across treatment groups (Table 3).  Two patients discontinued from the study due to rectal bleeding, one in the placebo group and one in the tegaserod 2 mg BID group. 

 

Table 3

 

Gastrointestinal Bleeding Key Safety Population
	Tegaserod 2 mg bid (N=861)	Tegaserod 6 mg bid (N=881)	Placebo (N=861)	Tegaserod Any dose (N=1742)
GI bleeding and Related Symptoms	10 (1.2)	10 (1.1)	11 (1.3)	20 (1.1)
Rectal hemorrhage	5 (0.6)	6 (0.7)	5 (0.6)	11 (0.6)
Blood in stool	3 (0.3)	3 (0.3)	5 (0.6)	6 (0.3)
Anal hemorrhage	0	1 (0.1)	0	1 (0.1)
Gastrointestinal hemorrhage NOS	1 (0.1)	0	0	1 (0.1)
Melena	1 (0.1)	0	0	1 (0.1)
Occult blood NOS positive	0	0	1 (0.1)	0
Discontinuations due to GI bleeding	1 (0.1)	0	1 (0.1)	1 (0.1)
(Ref: Table 8-3  Summary of Clinical Safety)

 

 

As part of the safety review for the present application, the Division also reviewed the safety data from completed trials of similar design.  This review included the original IBS application and safety data from completed studies through September 2003.  This database included over 12,000 patients in randomized trials.  To identify potential cases of bowel ischemia, these data were analyzed by the Novartis, at the request of the Division, using search criteria for all forms of rectal bleeding that resulted in any diagnostic work-up or therapeutic intervention (endoscopy or x-ray).  A detailed review of the case report forms and source data from this search did not identify a case that appeared suspicious for ischemic colitis.

 

C.            Diarrhea

 

1.  Post-Marketing Surveillance

 

The current label states that diarrhea was reported as an adverse event in 9% of the patients receiving Zelnorm during the IBS trials, compared to 4% in the placebo group.  During the post-marketing period, ODS received 22 AERS reports of serious complications of diarrhea.  The definition for epidemiological risk assessment was diarrhea or suspected diarrhea that led to an ER visit, serious outcome (i.e., death, life-threatening, hospitalization), or complications, including but not limited to, dehydration, hypokalemia, and/or the need for intravenous fluid replacement  (note: cases of serious diarrhea were excluded from this analysis if the diarrhea was caused by another process [e.g., infection]).  

 

Of the 22 cases of serious complications of diarrhea, 20 were reported by health care professionals and two were reported by the consumer.  Consistent with prescribing patterns, the majority of the cases occurred in female patients (Female=20, Male=2).  These patients ranged in age from 24 to 82 years (Median=59, Mean=56).  Patients were taking the following daily doses: 6 mg (3), 12 mg (13), 6 mg tapered to 2mg (1), and dose unspecified (5). Times to onset were: 1 day (5), 2 to 7 days (6), 21 days (1), 72 to 210 days (4), and unknown (6). In addition to diarrhea, the complications included the following (not mutually exclusive): dehydration (n=12), abdominal pain (n=8), hypotension (n=3), hypokalemia (n=2), nausea/vomiting (n=3), hyponatremia (1), hypothermia/shock (n=1), atrial flutter/fibrillation (n=1), hypovolemic shock/loss of consciousness (n=1).  Fifteen of the cases of diarrhea required hospitalization and three were described as life threatening.

 

2. Clinical Trials

 

During the chronic constipation trials the frequency and severity of diarrhea were dose-related (Table 4).  Four percent of patients in the tegaserod 2 mg b.i.d. group and 7% of patients in the 6 mg b.i.d. group reported diarrhea as an adverse event.  Diarrhea was reported as an adverse event in only 3% of the patients in the placebo group.  Diarrhea was reported as severe in three patients in the tegaserod 2 mg b.i.d. group, 7 patients in the 6 mg b.i.d. group and 2 patients in the placebo group. 

 

Table 4

 

 

 

The chronic constipation trials enrolled a total of 213 (12.2%) patients ³ 65 years of age (Table 5).  For the proposed dose, patients 65 years and older had a higher incidence of diarrhea (12.5%) and discontinuations due to diarrhea (3.4%) than patients younger than 65 years of age [diarrhea (5.9%), discontinuations due to diarrhea (0.6%)].  This is relevant considering the potential number of elderly people who may be treated for constipation.

 

 

Table 5

 

 

 

There was also an increased incidence of diarrhea during the long-term extension portion of the study.  Diarrhea was reported in 9.5% of patients receiving tegaserod during the long-term extension, compared to 6.6% in the core part of the study lasting 12 weeks (Tegaserod 6 mg bid group).  Although this is similar to what is reported in the label, it is relevant considering the indication is for chronic therapy and the potential number of elderly people that will be treated for constipation chronically. 

 

In the pooled indication population, the frequency of diarrhea was analyzed by treatment indication (Table 6).  The incidence of diarrhea was similar to the current label (10%) in patients treated for a lower GI indication.  The incidence was much higher in patients treated for an upper GI indication in other clinical trials (22%).

 

 

Table 6

 

 

 

D.     Hypotension

 

1. Post-Marketing Surveillance

 

As part of the recent labeling changes, hypotension is now listed in the WARNINGS section of the current label as one of the serious complications of diarrhea.  During the post-marketing period, the ODS received 15 AERS reports of hypotension.  Many of these cases were confounded by underlying medical conditions (i.e., myocardial infarction, drug allergy, and small bowel obstruction).  Hypotension was reported in three of the cases of serious complications of diarrhea and in two of the cases of ischemic colitis.

 

2. Clinical Trials

 

The development of hypotension may not be limited to complications of diarrhea. During Phase I development of Zelnorm, rare cases of hypotension were reported in healthy subjects.  Because of this, Phase II and Phase III studies paid close attention to the effects of tegaserod on blood pressure and pulse.  In the c-IBS trials, adverse events suggestive of orthostatic hypertension were reported with similar frequency in the placebo and tegaserod groups.  The most common adverse event suggestive of orthostatic hypertension was dizziness, which had a similar frequency in all the treatment groups.  However, syncope was more frequent in the tegaserod group compared with the placebo group (0.5% vs. 0.1%) p=0.16.

 

In the chronic constipation trials, orthostatic hypotension was defined as a reduction in systolic blood pressure of at least 20 mm Hg or a reduction in diastolic blood pressure of at least 10 mm Hg immediately after standing (or 3 min after standing) compared to the measurements taken in the sitting position.  The incidence of orthostatic hypotension in the key safety population was balanced across treatment groups with no appreciated dose relationship.  Orthostatic hypotension occurred in 14.6% of patients on tegaserod 2 mg b.i.d., 10.9% on tegaserod 6 mg b.i.d. and 12.0% on placebo. 

 

 

E.      Syncope

 

1. Post-Marketing Surveillance

 

As part of the recent labeling changes, syncope is now listed in the WARNINGS section of the current label as one of the serious complications of diarrhea.  As of April 15, 2004, ODS received eight post-marketing reports of syncope/loss of consciousness.  Most of those patients had other factors that may have contributed to the events; however, the role of tegaserod could not be completely ruled out.

 

 

2. Clinical Trials

 

In the chronic constipation trials, Novartis reports that none of the severe cases of diarrhea developed syncope.  During the c-IBS trials the incidence of syncope was low, but it was more frequent in the tegaserod group compared with the placebo group  (0.5% vs. 0.1%) p=0.16.

 

 

F.      Abdominal and Pelvic Surgery

 

1. Post-Marketing Surveillance

 

At the time of the original approval, there were questions about whether the use of tegaserod resulted in an increase in abdominal and pelvic surgery.  Between August 2002 and April 15, 2004, ODS received 28 AERS reports of patients who experienced adverse events involving the gallbladder while receiving tegaserod.  The definition for epidemiological risk assessment was any case reported as cholecystectomy, cholelithiasis, or cholecystitis.  Of the 28 reported cases involving the gallbladder, five were excluded from analysis for the following reasons: cholecystectomy planned before tegaserod therapy initiated (n=4) and one patient had a cholecystectomy while having a colon resection for colon cancer.  For the remaining 23 cases, 6 had very little information (e.g., medical history, concomitant medications) and 8 had a prior history of gallbladder disease. 

 

During the same period, ODS received 13 AERS reports of patients who experienced adverse events involving the ovary or fallopian tube.  The definition for epidemiological risk assessment was any adverse event reported as ovarian or fallopian tube cyst or ovarian surgery.  Five reports were excluded for the following reasons: underlying ovarian cancer (n=3); underlying colon cancer leading to removal of gall bladder, ovaries, and colon (n=1); and patient had pain "suggestive" of ovarian cyst rupture 3 months after tegaserod was discontinued (n=1).  For the eight remaining cases, the adverse events were reported as: ovarian cyst (7), hematosalpinx cyst (n=1), oophorectomy (n=1), hysterectomy (n=1) (not mutually exclusive). 

 

 

2. Clinical Trials

 

In the original application nine cases of ovarian cysts were reported.  Eight of the nine cysts were in tegaserod-treated patients; only one occurred in the placebo group.  Five of the eight cases required surgery, all from the tegaserod group.  There was also an imbalance in the number of cholecystectomies performed in Zelnorm-treated patients [Zelnorm (5/2,965; 0.17%) vs. placebo (1/1,740; 0.06%)], this difference was not statistically significant.  To determine whether the use of tegaserod resulted in an increase in abdominal and pelvic surgery, Novartis created an adjudication board consisting of independent consultants with expertise in IBS, GI motility, and evidence-based medicine.  This board reviewed all surgeries in a blinded manner.

 

The number of abdominal and pelvic surgeries performed during the chronic constipation trials were too small to identify an imbalance.  In the key safety population, the incidence of any abdominal and pelvic surgeries in tegaserod-treated patients was lower than in the placebo group [Zelnorm 0.5% (9 cases), Placebo 0.9% (8 cases)].  Only one cholecystectomy was reported in the key safety population.  This occurred in a patient receiving tegaserod 6 mg b.i.d.

 

Six patients (0.7%) in the key long-term safety population required abdominal/pelvic surgery (non-placebo-controlled study).  Two of these surgeries were for removal of ovarian cysts; one was detected on day 1 of the extension period.  Prior to enrolling in the extension study, this patient was in the placebo group during the core trial.  The other case occurred in a patient treated with tegaserod and was detected during the core period and removed on day 5 of the extension study.  The other four surgeries occurred between 210 and 392 days after start of the extension phase and included an inguinal hernia repair in the tegaserod 2 mg b.i.d. group, one hysterectomy and curettage due to increased menorrhagia in the tegaserod 6 mg group, and bladder surgery to correct a preexisting urinary stress incontinence and an appendectomy in the placebo-tegaserod 6 mg b.i.d. group.

 

In the pooled indication population, 27 surgeries were adjudicated in a blinded fashion by the independent board and were judged to be unrelated to study drug and were excluded from analysis [Tegaserod (n=15), Placebo (n=12)].  Three cases in the tegaserod group were not adjudicated because they were identified after the review.  These cases were included in the number of cases defined as possibly related to study drug.  Table 7 lists the frequency of abdominal and pelvic surgeries in the pooled indication population and shows the results of the independent board’s assessment.

 

Table 7

 

Frequency of Abdominal and Pelvic Surgeries  Pooled Indications Population Placebo-Controlled Trials
Population	Tegaserod (N = 6864)	Placebo (N = 3915)	Treatment Difference (95% CI)	p- value	Relative Risk (95% CI)
All cases	0.42% (29 cases)	0.41% (16 cases)	0.01 (-0.24, 0.27)	0.790	1.08 (0.60, 1.97)
Cases adjudicated as unrelated to study drug	15 cases	12 cases
Cases adjudicated as at least possibly related to study drug.	0.20% (14 cases)	0.10% (4 cases)	0.10 (-0.04, 0.25)	0.206	2.08 (0.65, 6.61)
Uncontrolled trials All cases	N = 4614 0.72% (33 cases)	NA
Frequency corresponds to number of patients with surgeries (including cholecystectomies)/number of patients treated. The p-value was calculated using the Mantel-Haenszel test. (Ref. Post-text tables 4.26-1, 4.26-3, 4.26-5)

 

 

The incidence of abdominal and pelvic surgeries was comparable across treatment arms.  However, a higher proportion of surgeries in the tegaserod group was adjudicated as at least possibly related to study drug. 

 

As described earlier, only one cholecystectomy was reported in the chronic constipation trials.  In the pooled indication population, the frequency of cholecystectomy (all cases) was higher in the tegaserod group than in the placebo group [Tegaserod 0.12% (8/6864), Placebo 0.03% (1/3915).  Novartis calculated exposure-adjusted frequency of unadjudicated and adjudicated cholecystectomies (Table 8).

 

The blinded adjudication excluded four cases of cholecystectomy from the analysis of risk of cholecystectomy (all in the tegaserod group); this resulted in a smaller difference between groups (0.06% on tegaserod vs. 0.03% on placebo).  Also, in the pooled indications population the frequency of hepatobiliary disorders reported as serious adverse events was higher in the tegaserod group (0.09% (6/6864)) compared to placebo (0.03% (1/3915)).

Table 8

 

Cholecystectomy Incidence in Placebo-controlled Trials Pooled Indications Population
	Treatment	% (n/N)	Exposure (Days)	Estimated Frequency Per 100 patient-years exposure	p-value vs placebo
All cases	Tegaserod	0.12 (8/6864)	491402	0.59 (0.18, 1.01)	0.111
	Placebo	0.03 (1/3915)	284777	0.13 (0.00, 0.38)
Cases adjudicated as related	Tegaserod	0.06 (4/6864)	491402	0.30 (0.01, 0.59)	0.438
	Placebo	0.03 (1/3915)	284777	0.13 (0.00, 0.38)
Cases adjudicated as unrelated	Tegaserod	4 / 6864
	Placebo	0 / 3915
Uncontrolled Trials	Tegaserod	0.13 (6/4614)	770215	0.28 (0.06, 0.51)
(Ref: Table 4-16  Summary of Clinical Safety) Frequency corresponds to number of patients with cholecystectomies/number of patients treated. The p-value was calculated using the Mantel-Haenszel test and refers to the exposure-adjusted frequency.

 

 

These data are difficult to interpret.  It is uncertain how adjudicated cases were handled.  It is generally accepted that approximately 10% of the adult population have cholelithiasis, but less than half of these patients develop symptoms.  

 

III.  CONCLUSIONS

 

The chronic constipation trials did not identify any new safety concerns, and the incidence and type of adverse events were similar to what is already included in the current label.  Many of the Division’s safety concerns that were identified during the post-marketing period have been addressed with the inclusion of serious consequences of diarrhea in the WARNINGS section of the label and ischemic colitis and other forms of intestinal ischemia in the PRECAUTIONS section (Appendix 1). 

 

The Agency is seeking the committee’s advice about whether ischemic colitis and other forms of intestinal ischemia should be moved to the WARNINGS section of the package insert.  The regulations [21 CFR 201.57(e)] state that “The labeling shall be revised to include a warning as soon as there is reasonable evidence of an association of a serious hazard with a drug; a causal relationship need not have been proved.”  Seven of the 20 cases of ischemic colitis presented were less than 49 years of age, with two of the patients aged 20 and 29 years.  Five of the 20 reported cases had no documented risk factors.  Three cases occurred on the first day of therapy, with two of the three cases occurring in patients with no reported risk factors. 

 

The appearance of ischemic colitis in young patients, in close temporal association with the drug is concerning.  Ischemic colitis is generally considered a disease of the elderly.  A recent study reported that the crude, age-stratified incidence of ischemic colitis differ by two orders of magnitude between the youngest strata [0.5 per 100,000 person years in individuals aged <20 years] and the oldest [97 per 100,000 person-years for individuals aged 70-79 years] (Occurrence of colon ischemia in relation to irritable bowel syndrome.  Am J Gastroenterol 2004;99(3):486-91).  Thus, the appearance of ischemic colitis in association with tegaserod in young patients is unexpected.  This suggests, but does not prove, that tegaserod caused the ischemic colitis.  Reports of IC in older patients could be attributed to the elevated rate of IC in that segment of the population or to misdiagnosis.  Further accumulation of case reports such as these, including reports of clinical severity (i.e., hospitalization, surgery, death) may suggest that the Agency consider new labeling for tegaserod to exclude organic diseases that mimic IBS.

 

Novartis believes that there is no causal relationship between the use of Zelnorm and the development of ischemic colitis.  It is their position that there is already a higher background incidence of ischemic colitis in IBS patients. To support this position, Novartis references a claims data study describing a higher incidence of ischemic colitis in IBS patients. According to Novartis’ interpretation of the data, Zelnorm’s post-marketing report rate is actually lower than the anticipated background rate.  They also reference a study by the American Society of Gastrointestinal Endoscopy that reports the background rate of ischemic colitis in the general population, found during asymptomatic screening, as 20/100,000 patients.  Novartis reports the incidence of ischemic colitis in patients treated with Zelnorm as 6/100,000.  The Division has requested the complete ASGE study report to review.

 

After reviewing the available data, it appears that the data supporting an association between ischemic colitis and IBS may be attributable to the significant limitations in the assessment and classification of ischemic colitis based on ICD9 codes.  The studies employed the ICD9 code 564.1 (irritable colon) as a surrogate for a diagnosis of IBS, as there is no unique ICD9 code that is limited to ischemic colitis alone.  In review of data submitted for the re-evaluation of alosetron, a strong temporal association was found between the index appearance of ICD9 code 564.1 within patient records and a follow-up (subsequent) diagnostic claim for ischemic colitis.  This suggests ICD9 code 564.1 may have been an interim diagnosis or in some instances a misdiagnosis.  Therefore, there does not appear to be compelling evidence to suggest that a clinically robust diagnosis of IBS is associated with any increased risk for ischemic colitis in comparison to age-matched peers.

 

There were no cases of ischemic colitis observed in the clinical trials.  An analysis of patients randomized to tegaserod among placebo-controlled trials of at least 3-months duration (n=7,000) was performed by the Agency.  Based on application of a Poisson distribution, this would suggest, with 95% confidence, that ischemic colitis occurs no more frequently in the population studied than approximately 1 in 2,000.  While this estimate could be viewed by some as too high given the large utilization/exposure of tegaserod, it should be noted that patients in clinical trials were subjected to inclusion, exclusion, and follow-up criteria that are not applicable to general clinical practice.  On average, the patients with ischemic colitis as reported to FDA, are both older and carry more co-morbid conditions than those in the tegaserod clinical trials.  Thus, generalizabilty of a rate, even an upper bound, for tegaserod-associated ischemic colitis from clinical trials to the population at large is problematic.

 

Novartis also states that no mechanism of action has been identified in animal models.  It is the Division’s opinion that a mechanism of action has not been ruled out and that there may be cross reactivity with other receptors and ligands that have not been identified.  Zelnorm is a 5-HT₄ partial agonist with moderate affinity for the 5-HT₁ receptor.  There is recent medical literature proposing a link between Zelnorm and the development of Raynaud’s phenomenon. The article presents a case history of a 21-year-old female, with no prior history of Raynaud’s who developed painful discoloration of the fingers after exposure to cold, two days after initiating tegaserod (12 mg/day).  Symptoms disappeared completely after drug therapy was stopped.  The patient was not on any concomitant medication during this period (Pharmacoepidemiology and Drug Safety 2002; 11: 231-294).  Another article discusses the potential risk of Zelnorm-induced myocardial infarction.  The article, titled “Tegaserod-induced myocardial infarction: case report and hypothesis,” proposes that since tegaserod has moderate affinity for the 5-HT₁ receptor, it is plausible that tegaserod could cause coronary artery contraction and spasm similar to other 5-HT₁ receptor agonists, such as those used for treating migraine (Pharmacotherapy 2004 Apr; 24 (4):526-31).  Although these two articles are not conclusive, they do support the Division’s position that a mechanism of action explaining an association between Zelnorm and ischemic colitis has not been ruled out.  

 

Appendix 1

Dear Health Care Professional Letter.  See end of this briefing document.
Appendix 2

Zelnorm Package Insert (April 2004).  See end of this briefing document.
Appendix 3       

Case Summaries