Dr M. Camilleri, Mayo Clinic, Enteric Neuroscience Program, Charlton 7-154, 200 First St SW, Rochester, MN 55905, USA. E-mail: camilleri.michael@mayo.edu


Tegaserod (Zelmac), an aminoguanidine indole derivative of serotonin, is a selective partial agonist highly selective for 5-HT4 receptor with an affinity constant in the nanomolar range. Tegaserod does not cause adverse pharmacodynamic effects, is absorbed rapidly after oral administration and distributes widely into tissues. Pharmacokinetics of oral tegaserod are linear in the 2–12 mg dose range. After oral administration tegaserod is metabolized mainly pre-systemically; when absorbed, intact tegaserod is excreted as N-glucuronides mainly via the bile. No clinically relevant drug–drug interactions were identified. Tegaserod has proven safe in toxicity studies.

In pharmacodynamic studies, tegaserod stimulated the peristaltic reflex in vitro, increased canine intestinal and colonic motility and transit, reduced visceral afferent firing or sensation in response to distension in animals, and accelerated gastric, small bowel and colonic transit in healthy patients, and small bowel transit in patients with constipation-predominant irritable bowel syndrome.

Three large phase III randomized, double-blinded, and placebo-controlled trials were performed predominantly in females (≈ 85%) with constipation-predominant irritable bowel syndrome. Overall, phase III results support efficacy as assessed by the subject’s global assessment of relief with significant improvement in secondary endpoints such as abdominal pain, bowel frequency and consistency. Tegaserod was well-tolerated; the most frequent adverse event was transient diarrhoea.


Irritable bowel syndrome is an exceedingly common condition in all societies; it is characterized by abdominal discomfort or pain in association with altered bowel habit or incomplete stool evacuation. The estimated prevalence in the community is about 10%, and the incidence is 1–2% per annum.1, 2 Irritable bowel syndrome and its variants, collectively called functional gastrointestinal disorders, constitute 40–50% of all the patients seen by gastroenterologists in Western countries.3 It has been estimated that about 10% of patients seen in general practice also have complaints consistent with irritable bowel syndrome. This is a heterogeneous disorder; the three most important mechanisms are hypersensitivity of the gut, altered motility, and psychosocial disorder.4 Gastrointestinal infection may serve as a triggering or antecedent factor.5

The current therapy for irritable bowel syndrome focuses on the major symptom experienced by the patient.4, 6 In general, patients may present with either diarrhoea, constipation, or a combination of pain, gas, and bloating. However, patients’ symptoms are not constant over time; apart from temporal fluctuation, the bowel complaint and predominance of specific symptoms may vary over time. Thus, patients may present with diarrhoea on one occasion, and constipation on another. The mainstays of pharmacotherapy at present are loperamide for diarrhoea, fibre and an osmotic laxative (such as magnesium salts or lactulose) for constipation, and anticholinergic or calcium channel blocker and muscle relaxants for pain, gas, and bloating. A systematic review of pharmacological treatment of irritable bowel syndrome suggests that smooth muscle relaxants are beneficial for abdominal pain, although only one of seven trials accepted for the analysis involved more than 100 patients; the remainder were small studies.7 Recently, alosetron was approved for the treatment of female patients with diarrhoea-predominant irritable bowel syndrome. Among patients with a significant psychological or psychiatric component of the irritable bowel syndrome, low dose antidepressants have also been effective in relief of pain, depression, and diarrhoea.4, 8

Current medical therapies for constipation-predominant irritable bowel syndrome have been insufficiently effective, and there is a need for novel approaches to treatment. Some of the deficiencies with current therapies are: first, fibre intake, particularly if it is over 20 g per day, often results in aggravation of gas and bloating; second, tricyclic antidepressants, which are effective for pain and diarrhoea, tend to make constipation in irritable bowel syndrome worse;8 and third, laxatives may relieve the constipation but have never been demonstrated to ameliorate the abdominal pain component of the syndrome. The results of controlled trials with selective serotonin re-uptake inhibitors for irritable bowel syndrome are still awaited; the hypothesis is that the serotonin selective re-uptake inhibitors result in improved pain relief and the known side-effect of diarrhoea may present a potential advantage for patients whose predominant bowel complaint is constipation. Thus, a medication that results in improvement in the relief of pain and discomfort from the irritable bowel syndrome associated with predominant constipation, while at the same time improving bowel function, is not generally available. This sometimes leads physicians to prescribe analgesics, including those of the opioid variety, with their consequent impact on drug dependence, aggravation of constipation, and sometimes worsening of the quality of life.

In recent years, there has been a resurgence of interest in the development of compounds for the treatment of irritable bowel syndrome. Candidate compounds which are in late development include the following: dexloxiglumide (a cholecystokinin antagonist); cilansetron (a 5-HT3 receptor antagonist with in vitro properties that are quite similar to those of the recently approved alosetron); tegaserod (a selective 5-HT4 receptor partial agonist); prucalopride (a 5-HT4 agonist which, like tegaserod, is intended primarily for the relief of constipation-predominant irritable bowel syndrome); fedotozine (a kappa opioid peripheral agonist which, in experimental animals, relieves visceral pain without central side-effects); and neurokinin receptor subtype antagonists.

In this review, the goal is to summarize the basic science, mechanism of action, efficacy and safety profile of a novel 5-HT4 partial agonist, tegaserod.


Tegaserod (Zelmac) is a selective partial agonist at the serotonin type-4 receptors (5-HT4 receptors), which mediate physiological functions in the gastrointestinal tract. 5-HT4 receptors are located on intrinsic primary afferent neurones, cholinergic interneurones that activate excitatory and inhibitory neurones involved in the peristaltic reflex, and conceivably extrinsic visceral afferents. 5-HT4 agonists have been used for the treatment of motility disorders of the upper gastrointestinal tract including gastro-oesophageal reflux disease, gastroparesis and chronic intestinal pseudo-obstruction. However, the available 5-HT4 agonists have been generally ineffective in the treatment of lower bowel motility disorders such as constipation-predominant irritable bowel syndrome and constipation.9[10]–11

5-HT4 receptors belong to the family of seven transmembrane domain receptors coupled to G-protein translation;12 they are responsible for eliciting the depolarizing action of serotonin, which results in the release of neurotransmitter, such as acetylcholine, from enteric neurones. 5-HT4 receptors are located in the central nervous system where they modulate dopamine release and have a direct role in cognition and memory. In the heart, these receptors are located in the atria, not the ventricles, and have a chronotropic effect. In the adrenal cortex, activation of 5-HT4 receptors transiently stimulates aldosterone secretion;13 in the urinary bladder, activation of the receptors increases detrusor tone.14, 15

In vitro studies on intestinal tissues, and in vivo animal and human studies of motor and sensory functions (discussed below) provide the scientific rationale to suggest that medications activating 5-HT4 receptors should provide relief of pain and discomfort and constipation in irritable bowel syndrome.


The chemical name of tegaserod is 3-(5-methoxy-1H-indol-3ylmethylene)-N-pentyl-carbazimidamide. Tegaserod is an aminoguanidine-indole derivative structurally similar to serotonin; its molecular structure is shown in Figure 1.

Figure 1.

The molecular structure of tegaserod.


Tegaserod is a partial agonist at the 5-HT4 receptor. Compared to serotonin, a full agonist with 100% intrinsic activity, tegaserod possesses 21% intrinsic activity in the electrically stimulated guinea pig ileum.16

As a partial agonist at the 5-HT4 receptor, tegaserod is less likely to induce receptor desensitization, compared to full agonists; it is also likely to produce a normalizing effect, augmenting the insufficient endogenous stimulation, and preventing the exaggerated affects associated with full agonists. Thus, it is expected that as a partial agonist, tegaserod should provide a balanced modulation of the 5-HT4 receptors.17 The lower likelihood to desensitize the 5-HT4 receptors is particularly relevant since this class of G-protein coupled receptor, which possesses seven transmembrane domains, is particularly prone to desensitization leading to tachyphylaxis or tolerance.12


The absorption, distribution, metabolism, and excretion of tegaserod have been investigated in several species (rat, mouse, rabbit and dog) in both pregnant and non-pregnant animals.18, 19 The metabolic profile of the rat is representative of that in humans. In the rat, no major accumulation of tegaserod or its metabolites in organs was observed. Tegaserod showed the potential to inhibit cytochrome P 450 1 A2 and 2D6 in human liver microsomes in vitro (Ki approximately 1 μM). However, no significant drug-to-drug interactions were observed in clinical studies with the prototype substrates theophylline (CYP1 A2) and dextromethorphan (CYP2D6) (see below). In vitro, the major human metabolite (a glucuronide) did not show relevant inhibition of any of these isoenzymes. Toxicology studies in animals showed no relevant effects on reproductive or hormonal functions, or embryofoetal or neonatal development. Although tegaserod was detected at low levels in foetuses, radiotracer studies show that radioactivity was about three times higher in milk than in plasma, suggesting that a suckling infant might ingest a significant dose of tegaserod. Tegaserod had no mutagenic or clastogenic potential and did not induce any DNA damage.18


Tegaserod is rapidly absorbed following oral administration, with peak plasma concentration after approximately 1 h. Tegaserod is hydrolysed by approximately 20% in the stomach (acidic pH under fasted conditions) and absolute bioavailability is about 10%. Concomitant administration with food reduces systemic exposure by 40–65%. It is recommended that tegaserod be administered before the morning and evening meals; this was the mode of administration in the Phase III clinical trials. Tegaserod is approximately 98% bound to plasma proteins, primarily to α1-acid glycoprotein. The volume of distribution at steady state is 368 ± 223 L. After oral administration, tegaserod undergoes acid hydrolysis in the stomach (pre-systemic metabolism), subsequent oxidation and glucuronidation. The main metabolite is a glucuronide which has negligible affinity for the 5-HT4 receptors and has no effects on gastrointestinal motor activity in the dog. Systemically available tegaserod is excreted as N-glucuronides mainly via the bile. The estimated terminal half-life is 11 ± 5 h. Blood clearance is approximately 48 L/h after intravenous administration, which represents approximately 50% of hepatic blood flow. Two-thirds of an orally administered radiolabelled dose is excreted unchanged in faeces, and the remainder is excreted in urine as metabolites.

The pharmacokinetics of oral doses of tegaserod (b.d. for 5 days) are linear in the 2–12 mg dose range, and are comparable in patients with irritable bowel syndrome and healthy subjects. The pharmacokinetic profiles of tegaserod are also similar between genders with no relevant changes in the elderly compared to young adults. Thus, dose adjustment is not needed in elderly subjects or in subjects with mild to moderate hepatic or renal impairment. No clinically relevant pharmacokinetic and pharmacodynamic drug–drug interactions have been identified; specifically, no clinically relevant interactions occur with theophylline (a prototype substrate for cytochrome P450 1 A2), dextromethorphan (a prototype substrate for cytochrome P450 2D6), digoxin, warfarin and oral contraceptives.


Actions of tegaserod on 5-HT4 receptors in vitro

There are four main 5-HT receptor subtypes in the human gastrointestinal tract: 5-HT1, 5-HT2, 5-HT3 and 5-HT4.23, 24 The diverse effects of serotonin are due to stimulation of different receptor subtypes. Like 5-HT3 receptors, 5-HT4 receptors are involved in motor, sensory and secretory functions. 5-HT4 receptors are located on enterocytes, enterochromaffin cells, neurones, and smooth muscle cells. Tegaserod is a partial agonist that displays high affinity for human 5-HT4 receptors; the affinity constant is 18 nM.25 Tegaserod has negligible affinity for the 5-HT3 receptor, with KB of 7200 nM; tegaserod has an almost 10-fold greater affinity for 5HT-4 receptors over dopamine receptors.

The partial agonist activity of tegaserod was demonstrated in the guinea pig ileum. Thus, whereas serotonin demonstrates 100% intrinsic activity and other 5-HT4 agonists, cisapride and renzapride, induce parallel dose–response curves to serotonin at 1–2 orders of magnitude lower EC50, tegaserod demonstrated lower intrinsic activity (21%), consistent with a partial agonist.16 This observation may have important implications, in particular that the partial agonist may be less likely to result in receptor down-regulation and hence there would be less likelihood of tachyphylaxis or tolerance. Another potential advantage is that the partial agonist may induce a second receptor state that reduces the potential ‘overdrive’ provided by the endogenous ligand, serotonin.

Tegaserod mimics the action of endogenous serotonin released from enterochromaffin cells, stimulating the intrinsic sensory neurones in the intestinal mucosa, and activating the peristaltic reflex with an EC50 of 5 nM. Tegaserod also inhibits visceral afferents involved in pain sensation (discussed below under in vivo actions).

The actions of tegaserod on the peristaltic reflex and propulsion have been thoroughly evaluated by Grider et al.26, 27 Endogenous serotonin is released from the enterochromaffin cells in the epithelial layer of the mucosa in response to chemical or mechanical stimuli; luminal application of tegaserod activates the 5-HT4 receptors on intrinsic primary afferent neurones. These neurones release transmitters such as calcitonin gene-related peptide, activating interneurones (e.g. cholinergic) which in turn stimulate an excitatory neurone-producing contraction on the orad side of the mucosal stimulus, and an inhibitory neurone on the caudad side. This results in the familiar ‘peristaltic reflex’ with contraction above and relaxation below the site of origination of the stimulus. The excitatory neurones mediate their effects through acetyl choline and substance P. The transmitters for the inhibitory arm of the reflex are vasoactive intestinal peptie (VIP), pituitory adenylate cyclase activating peptide (PACAP) and nitric oxide. Grider et al. demonstrated, in a series of elegant studies (Figure 2), that the application of tegaserod resulted in the release of calcitonin gene-related peptide, substance P and VIP; this release was inhibited by 5-HT4 receptor antagonist (SDZ 205–557), not by a 5-HT3 receptor antagonist (LY 278–584), suggesting considerable specificity in the activation by tegaserod.27 Importantly, the same laboratory replicated these effects in guinea pig colon, rat colon and human jejunum.

Figure 2.

Evidence-based model of the action of the partial 5-HT4 agonist, tegaserod, which activates intrinsic primary afferent neurones and the peristaltic reflex. Reproduced from Grider et al.26

In vitro studies of propulsion in isolated segments of guinea pig colon demonstrated that tegaserod stimulated the velocity of propulsion of artificial faecal pellets.28

Tegaserod’s effects on intestinal secretion have been tested in crypt cells from rat distal colon; at low nanomolar concentrations, tegaserod stimulated chloride and water secretion through an adenylate cyclase dependent pathway.29


Motor actions of tegaserod on 5-HT4 receptors in vivo

In vivo studies show that tegaserod enhances motility at all levels of the gastrointestinal tract. It stimulated gastric emptying in rats and dogs under normal or perturbed conditions.30, 31 During acoustic stress (a model that retards gastric emptying) in dogs, tegaserod 0.1 and 0.3 mg/kg restored normal gastric emptying rate to pre-stress levels; in contrast saline and vehicle had no such effect.31 Tegaserod stimulated small bowel and colonic motility in dogs (estimated as a motility index) in a dose-dependent manner.31 In dogs, Nguyen et al.30 demonstrated that intravenous tegaserod activated small intestinal rhythmic contractions, and induced prolonged colonic contractions (Figure 3) that were associated with activation of the transit of a radiotracer instilled into the proximal colon immediately before the intravenous medication. In another model of colonic transit, tegaserod reversed to normal the retardation of transit induced by the α2 adrenergic agonist, lidamidine.32 These effects of tegaserod in vivo suggest that the medication may also activate gastrointestinal motility by a mechanism other than the luminal route activation of the peristaltic reflex. This alternative activation mechanism is likely to involve 5-HT4 receptors on enteric cholinergic neurones.

Figure 3.

Small bowel and colonic contractions in response to intravenous tegaserod (0.3 mg/kg). Reproduced from Nguyen A et al.29

Activity of tegaserod on noxious colonic and rectal hypersensitivity in animal models

Distention of the colorectum performed in a stepwise fashion using a latex balloon (0.5–2.5 mL volume or 0–60 mmHg pressure) is an established model to assess anti-nociceptive actions of medications; the level of the stimulus can be sub-noxious to noxious. This results in reflex hypotension or increased abdominal contractions or a pseudo-affective behaviour such as withdrawal or writhing, which are considered to be indices of pain. Tegaserod (1 mg/kg i.p.) inhibited the abdominal contractions in response to colorectal distention of the rat.33 In contrast, the compound did not inhibit the hypotensive response to acetylcholine, which does not itself cause pain. Other studies assessed the effects of tegaserod on visceral sensitivity in cats.34 Recordings from single afferent units in dorsal S2 showed that distension results in firing rates proportional to the pressure in the distending balloon; the firing rate was sub-maximal at 50 mmHg distension, which was used as the standard stimulus to assess the effects of tegaserod on sensation (Figure 4). Tegaserod dose-dependently inhibited the firing of these rectal afferents following rectal distension, without altering the rectal compliance. This suggests that the effect was not the result of relaxation or increased rectal compliance with tegaserod and implies an effect on nociceptive afferents. The effects of tegaserod were inhibited by a selective 5-HT4 antagonist, SB203186, consistent with the interpretation that the effects of tegaserod on sensation are mediated through 5-HT4 receptors.19

Figure 4.

Firing of visceral afferents in cats in response to noxious (50 mmHg) colorectal distension is dose-dependently reduced by tegaserod. Plotted from Schikowski et al.33

Pharmacodynamic studies in humans

In order to explore the mode of action of the 5-HT4 agonist, tegaserod, clinical pharmacological studies have been performed in healthy subjects and irritable bowel syndrome patients.

Effect of tegaserod on whole gut transit in healthy subjects.

Degen et al. have demonstrated that intravenous (0.6 mg) and oral (6 mg) tegaserod accelerate gastric emptying, small bowel and colonic transit.35

Effect of orally administered tegaserod on whole gut transit in constipation-predominant irritable bowel syndrome.

Prather et al. evaluated the effect of 2 mg b.d. of tegaserod for 1 week on whole gut transit using a scintigraphic method; colonic filling, a surrogate of orocecal transit time, was significantly accelerated by tegaserod relative to placebo (Figure 5). The colonic transit time was also accelerated in the tegaserod group relative to pre-treatment (Figure 6), although this effect did not reach statistical significance relative to placebo treatment.36

Figure 5.

Effect of tegaserod and placebo on orocecal transit measured as colonic filling at 6 h in patients with constipation-predominant irritable bowel syndrome. Reproduced with permission from Prather et al.35

Figure 6.

Effect of tegaserod on orocecal and colonic transit demonstrated on scintiscan from a patient with constipation-predominant irritable bowel syndrome. Reproduced with permission from Prather et al.35


Pre-clinical and pharmacodynamic studies

Phase I studies have been reviewed in detail under the section on pharmacodynamics. In summary, tegaserod increases the secretion of water and sodium from colonocytes and enhances gastric, small bowel and colonic transit in healthy subjects.29, 35 Tegaserod accelerates small bowel transit in patients with constipation-predominant irritable bowel syndrome.36 Although studies to date do not demonstrate any effect of tegaserod on rectal compliance or sensation in humans, the data from validated animal models of visceral sensation suggest that it possesses properties that reduce extrinsic afferent firing, and thus visceral sensitivity. Specifically, tegaserod inhibits firing of rectal afferents during ramp distension of the rectum, and reduces the pseudo-affective response to noxious colonic distensions. Tegaserod is devoid of effects on gastric acid secretion, significant central nervous system or cardiovascular effects.

Other in vitro and pharmacodynamic studies have been undertaken to explore the effects of this partial 5-HT4 agonist on cardiac rhythms in animal models (the isolated Langendorff perfused rabbit hearts).37 5-HT4 receptors are located on atrial muscle and the affinity constants for cisapride and tegaserod for human 5-HT4 receptors are 59 nM and 18 nM, respectively. In vitro studies of the delayed rectifier potassium current (IKr) were performed on the HERG channel, guinea pig papillary muscles, and isolated rabbit heart. In vivo studies were also undertaken in conscious dogs. Neither in vitro studies, nor in vivo studies demonstrated any relevant QTc (corrected QT interval) prolongation by tegaserod.19

Thus, these data suggest that the effects on cardiac repolarization observed with the benzamide 5-HT4 agonist/5-HT3 antagonist cisapride were not observed with tegaserod, suggesting that the new chemical class of tegaserod (an aminoguanidine indole) does not share the latter’s potential to inhibit the IKr in cardiac myocytes. Specifically, IKr inhibition by cisapride occurs with an IC50 of 44 nM, and with tegaserod at 13 000 nM; similarly, QTc prolongation in the Langendorff model in isolated rabbit heart occurs with 100 nM cisapride and > 10 000 nM tegaserod. These data also suggest that the arrhythmogenic potential of cisapride is not related to activation of the 5-HT4 receptors but to the chemical structure of the benzamide.

Phase III studies

Three phase III trials were submitted to the FDA in the application for approval of tegaserod.38, 39 Two studies had identical designs. Following a 4-week placebo-free baseline period, eligible patients were randomized in equal allocation of approximately 250 patients per group to placebo, 2 mg, or 6 mg b.d. tegaserod. In the third study, following a 4-week baseline period, patients were randomized in equal numbers to either placebo, or a fixed dose of 2 mg b.d., or a dose-titration regimen with 2 mg b.d. for the first 4 weeks, followed by dose titration at the end of week 4. The dose was increased to 6 mg b.d. if the response was not complete or considerable relief for at least 50% of the time during the first 4 weeks.

The studies performed to date with tegaserod involved large numbers of patients in placebo-controlled studies; however, interpretation of results is complicated for a number of reasons:

1 The decision to change study endpoints after the first trial was completed.

2 The lack of uniformity of study design in the third titration trial with the clear increase in placebo responsiveness after the 4th week of this titration trial.

3 The study endpoint pertained to the last 4 weeks of the 12 week trial, not taking into consideration the qualitative evidence of responsiveness in the first 8 weeks of the trials.

4 The complicated nature of the criteria characterizing responsiveness with the initial endpoint based on complete or considerable global relief being associated with a relatively low response rate for both active and placebo groups.

Defining study endpoints.

It is unnecessary to document the sequence of observations that led to the change in study endpoints; it is sufficient to note that the final definition of responder for the subject global assessment of relief was: (i) at least 50% of the subject’s global assessments at study endpoint (i.e. at least two of the weekly questions during the final 4 weeks in the 12 week trial) were answered positively for complete or considerable relief—this is identical to the original study endpoint and served to assess the magnitude of the effect; (ii) all of the subject’s global assessments at endpoint (final 4 weeks of trial) with at least somewhat relief—this criterion was added to capture a persistent positive response to medication at study endpoint.

The inclusion of the ‘somewhat’ relieved category was based on the documentation of the associations between the several efficacy variables and the ‘somewhat’ relieved response. These associations were evaluated in the first phase III trial in which changes from baseline were evaluated for ‘responders’ and ‘non-responders’ to either tegaserod or placebo. Thus, significant differences (all P < 0.0001) between responders and non-responders were identified in several parameters. These included the visual analogue scale scores for the subject’s global assessment of abdominal discomfort/pain and bowel habit, the number of days with significant pain and bloating and the number of days with no bowel movement or hard/very hard stools. Importantly, this analysis also demonstrated that the ‘somewhat relieved’ category was associated numerically (e.g. visual analogue scale scores, daily pain or bloating scores, number of days with specific symptoms) about midway between considerable relief and unchanged (Table 1).40

Table 1.  Associations between subject’s global assessment of relief and other efficacy variables (B351) Thumbnail image of

The all-important question on which efficacy was evaluated was the weekly question posed to patients: ‘Please consider how you felt this past week in regard to your irritable bowel syndrome, in particular your overall well-being, and symptoms of abdominal discomfort, pain and altered bowel habit. Compared to the way you usually felt before entering the study, how would you rate your relief of symptoms during the past week?’

• completely relieved;

• considerably relieved;

• somewhat relieved;

• unchanged;

• worse.

The nature of the question was clearly intended to provide a global endpoint that included overall well-being as a surrogate for quality of life and the individual contributions of pain and altered bowel function. This is in agreement with the recently published consensus guidelines on the assessment of drugs for irritable bowel syndrome.41

In the analysis, the permissible number of days with laxative use (other than fibre or bulk laxatives) during the 12-week treatment period was ≤ 5 and no laxative use was permitted during the last 28 days of the trial, that is during the period in which the study drug was being evaluated. Patients exceeding this amount of laxative use were considered non-responders. This restriction was imposed because of the need to evaluate the drug separately from the influence of laxatives which were permitted as a ‘rescue’ for participants. However, the restriction added considerably to the difficulty in interpretation of the data on study efficacy and contributed to a lack of clarity as to the potential effectiveness of tegaserod in the context of laxative use, a commonly prescribed class of medications for patients with constipation-predominant irritable bowel syndrome.

Demographics and baseline characteristics of intention-to-treat populations.

The demographics and duration of irritable bowel syndrome in the study participants are shown in Table 2.

Table 2.  Demographics and duration of irritable bowel syndroms in the study participants Thumbnail image of

In all three studies, there were more females than males, the majority were Caucasian and less than 65-years-old.

Subject global assessment of relief in the intention-to-treat population including adjustment for use of laxatives.

In the first study, both treatment groups had higher response rates compared to placebo, with 9% therapeutic gain for the 4 mg and 8% for the 12 mg group (Figure 7). This study was analysed after the new study endpoints were instituted (i.e. including the ‘somewhat relieved’ for four of the 4 weeks at study endpoint). In the second study, which included dose titration after 4 weeks, the therapeutic gains were not significant. In the third study, which compared placebo, 4 mg tegaserod per day and 12 mg tegaserod per day for 12 weeks, the therapeutic gains were 6% and 12%, respectively, in the 4 mg and 12 mg groups.

Figure 7.

Effect of 4 mg and 12 mg per day tegaserod and placebo on monthly responders for subjective global assessment of relief of irritable bowel syndrome symptoms. Reproduced from Mueller-Lissner et al.37

Summaries of the response rates and therapeutic gains are shown in Table 3.

Table 3.  Summaries of the response rates and therapeutic gains Thumbnail image of

However, further analyses were performed to appraise the overall effects of the study drug without adjustment for use of laxatives, potentially providing a better estimate of the drug’s effectiveness in clinical practice. In the studies of identical design, response rates at endpoint in one study (B301) for placebo and tegaserod 12 mg/day were 33% and 45%, respectively, and in a second study (B351), the response rates for placebo and tegaserod 12 mg per day were 37% and 49%, respectively. Thus, the therapeutic gain for the 12 mg/day group was 12% in both studies. In addition, in female patients, the therapeutic gains for the 12 mg/day group were 15% (study B301) and 13% (study B351). Gender effects are discussed further below.

Week by week assessment of the subject’s global assessment of relief (primary endpoint) and secondary efficacy endpoints also demonstrated beneficial effects of tegaserod, particularly the 6 mg b.d. dose, over placebo, in both the B301 and B351 parallel-group design studies.

Efficacy analysis by gender.

Tegaserod was effective in female patients, but no evidence for efficacy was demonstrated for males in the individual studies. When data were pooled across studies, efficacy in males was not observed at study endpoint. Although some improvements in the pooled assessment of efficacy were observed during the first month of therapy, this was not significant. However, the sample sizes were considerably smaller for males, and a type II statistical error cannot be excluded.

Female patients in the tegaserod treatment groups also demonstrated a higher response rate for the subject global assessment of abdominal pain and discomfort, with therapeutic gains over placebo of 9% in the 4 mg group and 10% in the 12 mg group in study B301. Similarly, the therapeutic gain was 9% for the 12 mg group in study B351, but was not significant for the 4 mg group in this study. The dose titration study showed no improvements for either gender at either dose of tegaserod at study endpoint. It is unclear whether this represented a true lack of efficacy or the result of an inappropriate study design.

Efficacy in sub-group with constipation demonstrated during run-in.

Although it was intended that patients with constipation-predominant irritable bowel syndrome be recruited for this study, the analysis of the stool parameters recorded in the daily diaries demonstrated that up to a fifth of patients across these studies actually experienced diarrhoea or alternating bowel habit during the 1-month placebo-free baseline. The study design did not a priori exclude these participants from the trial, since they still qualified as constipation-predominant irritable bowel syndrome using Rome criteria, and hence they were randomized to receive therapy. However, if one restricts the efficacy analysis to those subjects without loose bowel movements or a low (i.e. loose) mean stool consistency during the run-in period, the therapeutic gain over placebo is greater. Indeed, in a pooled analysis across the three studies, the therapeutic gain over placebo averages about 15% for those with normal consistency or hard stools. These data suggest that the drug is more efficacious in those whose current symptoms of irritable bowel syndrome include normal consistency stool or hard stools associated with constipation.

Weekly response rates in subjects’ global assessment of relief.

An analysis of weekly response rates suggests that the tegaserod treatment groups were often at a significant advantage over placebo treated patients during the first 4–6 weeks of therapy (e.g. in trial B351). In these analyses, amelioration of bowel function was frequently noted from the first week of treatment. The primary study analysis, however, discounted these initial observations, focusing the study endpoint on the last 4 weeks of therapy. This was intended to demonstrate the advantageous position over placebo in the last 4 weeks of therapy and to confirm the maintenance of the response to therapy. However, as has been demonstrated in several studies of medications applied to the treatment of functional or motility disorders of the gastrointestinal tract (e.g. cisapride in chronic intestinal pseudo-obstruction or alosetron in diarrhoea-predominant irritable bowel syndrome), the placebo effect is still strong, even in the 3rd month of therapy.42, 43 Hence, the study analysis used was particularly disadvantageous in the context of the exercise of assessing efficacy, although the week by week responsiveness suggests that the drug may indeed be effective if used when the current bowel dysfunction is constipation.


Over 3500 healthy subjects or patients with irritable bowel syndrome have received tegaserod. The maximum daily dose was 200 mg in healthy subjects. Almost 1900 patients with irritable bowel syndrome received the drug for > 85 days and more than 300 patients for > 330 days. Approximately 12% of tegaserod-treated patients developed diarrhoea in the phase III trials, compared with 5% of placebo-treated patients. When observed, diarrhoea usually occurred during the first week of the study, was transient in nature, occurred as a single episode, and lasted a median of 2 days. In the group of patients (18%) who had experienced at least one report of diarrhoea during baseline, the occurrence of diarrhoea during the 12-week studies was only slightly higher than in the overall study population, and only rarely (1.6% of tegaserod-treated patients) did patients discontinue medication as a result of this adverse event. More than 10 000 electrocardiograms from participants receiving tegaserod have been analysed to date, with no increased frequency of QTc interval prolongations or dysrhythmias recorded.

The only significant adverse events noted were ovarian ‘cysts’ and abdominal or pelvic surgery. Careful assessment of each report of ovarian ‘cysts’ suggested first that many of these reports did not actually demonstrate cyst formation and no causal relationship to tegaserod could be identified. Moreover, pelvic surgery was performed exclusively for intercurrent gynaecological problems, and abdominal surgery was clinically indicated for unrelated surgical emergencies, such as acute appendicitis or drainage of a pancreatic cyst.19


Tegaserod is a potent 5-HT4 receptor partial agonist which has shown efficacy in the treatment of abdominal pain and discomfort, and constipation in female patients with irritable bowel syndrome. To date, it has been demonstrated to be safe and well-tolerated except for the adverse effect of transient diarrhoea in about 10% of recipients. This is a known side-effect of this class of compounds and the majority of patients chose not to stop the drug during the trials. The exact mechanism of its action in humans is as yet incompletely understood. Tegaserod appears to have an effect on visceral afferent function, reducing pain and discomfort associated with noxious distensions of the colorectum in animals and increasing motility in the dog as well as accelerating transit in the small bowel and colon. Further studies are necessary to more fully characterize the efficacy of the medication and its mode of action.

It is likely that the drug will find a place in the market because available alternative strategies are sub-optimal. The alternative competing classes of therapy for pain and constipation in irritable bowel syndrome are laxatives or serotonin re-uptake inhibitors, but the evidence basis for their widespread use in practice is weak or non-existent. Current treatment of constipation-predominant irritable bowel syndrome often requires the use of more that one medication to control the multiple symptoms; the clinical trials suggest that tegaserod may be efficacious on multiple symptoms as well as overall well-being. There has been no trial comparing tegaserod to an ‘established’ irritable bowel syndrome therapy (although this is clearly a challenge in the absence of approved medications for the constipation-predominant group), or to a laxative. Nevertheless, its safety profile suggests that physicians are likely to consider it when patients do not respond to simpler approaches such as fibre and osmotic laxatives, or when there are multiple symptoms including discomfort or pain or impaired well-being associated with irritable bowel syndrome. It is also clear that tegaserod will not be indicated for all patients with irritable bowel syndrome, and specifically that it may aggravate diarrhoea in patients with other forms of irritable bowel syndrome. Tegaserod has not yet been shown to be efficacious in male patients. Further studies in phase IV will be indicated to evaluate the effectiveness of this medication in clinical practice.

Based on the studies performed to date, an application for the approval of tegaserod in the treatment of irritable bowel syndrome was submitted to the United States Food and Drug Administration. It is anticipated that, when the drug is approved, the package insert will stress the following features: first, that the drug results in relief of abdominal pain and discomfort in female patients with irritable bowel syndrome; second, that the drug is not proven to be effective in diarrhoea-predominant patients; third, that the drug has only been demonstrated to be efficacious in female patients and that the optimal dose is 6 mg b.d.


This study was supported in part by grants RO1-DK54681–02 and K24-DK02638–02 (Dr M. Camilleri) from the National Institutes of Health.

The author thanks Mrs Cindy Stanislav for excellent secretarial assistance.