Review article: cardiac adverse effects of gastrointestinal prokinetics


Tonini Department of Internal Medicine and Therapeutics, Section of Experimental and Clinical Pharmacology, University of Pavia, Piazza Botta 10, I-27100 Pavia PV, Italy. E-mail:


Gastrointestinal prokinetics, such as metoclopramide, cisapride and levosulpiride, are widely used for the management of functional gut disorders. Recently, several studies have shown that cisapride (a partial 5-HT4 receptor agonist) can induce dose-dependent cardiac adverse effects, including lengthening of the electrocardiographic QT interval, syncopal episodes and ventricular dysrhythmias.

Until recently, it was not clear whether these effects were dependent on 5-HT4 receptor activation or related to peculiar characteristics in the molecular structure of single agents within the benzamide class. Experimental evidence now favours the second hypothesis: cisapride possesses Class III antiarrhythmic properties and prolongs the action potential duration through blockade of distinct voltage-dependent K+ channels, thus delaying cardiac repolarization and prolonging the QT interval.

Patients at risk of cardiac adverse effects are children, subjects with idiopathic, congenital or acquired long QT syndrome and, in particular, those receiving concomitant medication with Class III antiarrhythmic agents, some H1-receptor antagonists (e.g. terfenadine), or drugs such as azole antifungals (e.g. ketoconazole, itraconazole, miconazole and fluconazole) and macrolide antibacterials (e.g. erythromycin, clarithrod-mycin and troleandomycin), which can inhibit cisapride metabolism by interfering with the CYP3A4 isoenzyme.


Gastrointestinal prokinetics, such as the substituted benzamides cisapride, levosulpiride and metoclopramide, are commonly used in the clinical practice to stimulate propulsive motility with anal progression of the endoluminal contents1, 2 (Table 1). They are useful for the treatment of upper gastrointestinal motor disorders, including gastro-oesophageal reflux disease and functional dyspepsia.2, 3 However, the past 10 years have seen a revolution in our understanding of their mode of action as well as their ability to induce clinically relevant side-effects, as will be discussed in the following sections of this review.

Table 1.  . Pharmacological properties of currently available gastrointestinal prokinetics Thumbnail image of


It has long been assumed that blockade of the gastrointestinal dopaminergic inhibitory transmission was the main mechanism responsible for the motor stimulating effects of first-generation prokinetics such as metoclopramide, domperidone (a butyrophenone derivative), and levolsulpiride. Domperidone and levosulpiride are competitive antagonists at dopamine D2 receptors, whereas metoclopramide acts as an antagonist at both D1 and D2 receptors. D2 receptors are mainly distributed at neuronal level, where they inhibit the function of intrinsic (enteric) cholinergic pathways, through a negative feed-back mechanism on transmitter release.2 Dopamine receptor antagonists were proposed as gastrointestinal prokinetics because of the evidence that dopamine, which is present in the gastrointestinal wall of several mammals, had marked inhibitory effects on gastrointestinal motility. These effects, which include a reduction of the lower oesophageal sphincter pressure and gastric tone and inhibition of gastroduodenal coordination,3, 4 are ascribable to the blockade of peripheral dopaminergic receptors, which are located on both the effector cells and on cholinergic neurones, where they inhibit acetylcholine release. In addition, antagonism of central D2 receptors in the area postrema is responsible for the antiemetic properties of these compounds, especially when vomiting is associated with gastrointestinal motor disorders or is drug-induced (e.g. patients with Parkinson’s disease receiving L-DOPA may use domperidone, a drug which does not readily cross the blood–brain barrier, to relieve nausea and vomiting). The blockade of central D2 receptors may be associated with dystonic extrapyramidal reactions and an increased prolactin release, which are considered the most relevant side-effects of these prokinetics.2

It is now well established that metoclopramide, apart from being a potent dopamine receptor antagonist, may also act as a moderate antagonist at 5-HT3 receptors, as well as a moderate agonist at serotonin 5-HT4 receptors. The latter property is responsible, at least in part, for the gastrointestinal prokinetic action of the drug.1, 2, 5 Very recently, in a study aimed at reassessing the pharmacodynamic profile of levosulpiride in isolated tissues from the guinea-pig gastrointestinal tract, this drug was found to possess moderate 5-HT4 receptor agonist properties associated with mild 5-HT3 receptor antagonism (Table 1).6


As mentioned in the previous section, at least part of the prokinetic action of metoclopramide is probably related to the activation of 5-HT4 receptors. This mechanism is thought to play a pivotal role in mediating prokinesia by cisapride (a substituted benzamide partial 5-HT4 agonist), which, in contrast to metoclopramide and levosulpiride, has no relevant affinity for D2 receptors.1 Cisapride was found to promote peristalsis by reducing the threshold pressure necessary to trigger distension-evoked peristaltic contractions.7 These effects are probably sustained by 5-HT4 receptor-activated cholinergic transmission at both pre- and post-ganglionic level. Thus, integrity of the intrinsic excitatory innervation is required for substituted benzamides to exert their prokinetic effects in the gut. Recently, in the human isolated jejunum, 5-HT released by enterochromaffin cells following stroking of the mucosa was found to stimulate 5-HT4 receptors located on primary intrinsic sensory neurones with consequent activation of both ascending excitatory and descending inhibitory reflexes.8

5-HT4 receptors are not confined to the gastrointestinal tract ( Figure 1): they are also found in the central nervous system (especially in the limbic areas),9 and in various peripheral organs and systems, such as the heart and the vasculature, the adrenal cortex and the urinary bladder.10 In the upper gut, 5-HT4 receptors facilitate gastric emptying and peristalsis, whereas in the human isolated colon, 5-HT4 receptors modulate circular smooth muscle relaxation by increasing cAMP production.3, 11 In the small bowel and colon, they may also stimulate mucosal secretion of water and electrolytes.12, 13 In blood vessels, 5-HT4 receptors are involved in muscle relaxation, whereas in the heart they exert a positive chronotropic effect by an action on the atrium.14, 15 In the adrenal glands, they promote steroid hormone release,16, 17 and finally, in the human detrusor muscle, they facilitate cholinergic excitatory transmission.18, 19

Figure 1.

. Distribution and function of 5-HT4 receptors.


Because 5-HT4 receptor agonists are used for their gastrointestinal prokinetic action, any other effect mediated by these receptors in different target organs may be regarded as an untoward effect. 5-HT4 receptor agonists are known to exert their side-effects mainly on the lower urinary tract and the cardiovascular system. In particular, cisapride may evoke polyuria and, occasionally, urinary incontinence, especially in elderly people, or when the drug is used at high doses.20, 21 In fact, 5-HT4 receptor agonists have been proposed for the treatment of urinary disorders associated with hypomotility of the detrusor muscle.22

However, the most relevant extraintestinal effects may be observed in the cardiovascular system and theoretically may be due to two distinct mechanisms: (i) those mediated by stimulation of cardiac 5-HT4 receptors; and (ii) those caused by the Class III-antiarrhythmic properties displayed by some compounds.

Cisapride may indeed trigger tachycardia and supraventricular dysrhythmia through stimulation of 5-HT4 atrial receptors.14 However, the incidence of the atrial dysrhythmia is very low, probably because cisapride behaves as a partial agonist on the human atrium and because the density of 5-HT4 receptors in the atrium is rather low.15, 23

Class III antiarrhythmic effects

The cardiotoxic potential of cisapride stems from the drug’s ability to prolong the cardiac action potential duration (hence the lengthening of the electrocardiographic QT interval), which has been associated with the occurrence of torsades de pointes, a polymorphous ventricular dysrhythmia which may cause syncope and degenerate into ventricular fibrillation.24[25]–26 Apparently, these effects are not mediated by 5-HT4 receptor activation, as shown by the failure of the selective antagonist GR11380827 to modify cisapride-induced action potential prolongation in guinea-pig isolated papillary muscles, and by the observation that 5-HT4 receptors are generally absent from the ventricle.

Indeed, cisapride has been hypothesized to possess the pharmacophore for Class III antiarrhythmic properties28, 29 ( Figure 2). Other substituted benzamides, such as metoclopramide or levosulpiride ( Figure 2), as well as some of the newly developed 5-HT4 agonists, either benzamide such as mosapride29 or non-benzamide, such as HTF919 (tegaserod),30 prucalopride (Dr J. A. J. Schuurkes, personal communication) or ML10302,31 do not share Class III properties.

Figure 2.

. Class III antiarrhythmic pharmacophore described by Morgan and Sullivan and chemical structures of some substituted benzamide gastrointestinal prokinetics.28 The pharmacophore confers the ability to prolong cardiac action potential duration, hence the QT interval. The dotted line in the chemical structure of cisapride indicates the analogy with the Class III pharmacophore.

Several studies support the notion that lengthening of the QT interval by cisapride can be related to the blockade of potassium currents, which increases the duration of the action potentials originating from cardiac muscle cells.27, 32 This is a basic mechanism through which drugs such as amiodarone and d-sotalol exert their Class III antiarrhythmic properties.33, 34 Cisapride can dose-dependently prolong action potential duration of ventricular muscle and Purkinje fibres by blocking channels encoded by the human ether-a-go-go-related gene (HERG), hence reducing the rapid component of the cardiac delayed rectifier K+ current (IKr)27, 32, 35[36]–37 (Table 1).

However, when considering cisapride in the context of the whole class of substituted benzamides, it is important to point out that some substituted benzamide neuroleptics, such as sultopride and amisulpride, which do not possess affinity for 5-HT4 receptors, may alter, like cisapride, the QT interval.38[39]–40 For the sake of completeness, cardiac adverse reactions are also described with the benzamide derivative metoclopramide, which is devoid of the Class III pharmacophore described by Morgan and Sullivan,28 but whose chemical structure is related to the antiarrhythmic drug procainamide. Indeed, metoclopramide has sometimes been reported to induce sinus arrest after intravenous administration, possibly for its Class I antiarrhythmic properties.41 Further research into the specific ion channels involved in mediating these effects of substituted benzamides will clarify this aspect. In a more general context, it should be emphasized that proarrhythmic effects can be achieved through different mechanisms and that several potassium channels with different regional expression (in different species) are now identified.34, 42 Thus, it would be simplistic to relate the QT-prolonging effect of a compound solely to the Class III pharmacophore, as described in Figure 2.


Since molecular mechanisms responsible for the arrhythmogenic potential of cisapride are now sufficiently documented, the question arises whether the observed effects have clinical relevance. Although postmarketing surveillance has confirmed that cisapride, besides being an effective drug, is generally safe and well tolerated,43 the data gathered in the past few years suggest some caution at least in some patient populations. Caution stems from the observation that the IC50 values calculated for the Class III antiarrhythmic properties of cisapride in animal and human models are in the 6.5–44.5 n M range.32, 35, 44 After oral administration of cisapride at 10 mg t.d.s. (a therapeutic dosage often applied in clinical practice), total cisapride peak serum levels of 60–80 μg/L (120–170 n M) are achieved.45 Corinaldesi et al.46 found that in patients with idiopathic gastroparesis, peak concentrations of cisapride higher than 60 μg/L were invariably associated with acceleration of gastric emptying. If we consider that the protein bound fraction is approximately 98% (i.e. free cisapride plasma concentration 2.4–3.4 n M) then it becomes apparent that a Class III and the volume of distribution is 2.4 L/kg,45 antiarrhythmic effect is plausible, especially in cases of high dosage or pharmacokinetic interactions.47

The cardiotoxic potential of cisapride should be considered, particularly in newborns or children,48[49][50][51]–52 in patients with an idiopathic, congenital or acquired long QT interval, patients receiving Class III antiarrhythmic agents, phenothiazines, tricyclic antidepressants, H1-histamine receptor antagonists (e.g. terfenadine), those with renal or hepatic insufficiency, or, finally, patients concomitantly treated with drugs known to inhibit the CYP3A4 isoenzyme.32, 35, 48, 49, 53 CYP3A4 plays an important role in cisapride metabolism and is inhibited by azole antifungals (i.e. ketoconazole, itraconazole, miconazole, and fluconazole), macrolides (i.e. erythromycin, clarithromycin and troleandomycin),54[55]–56 and at least some protease inhibitors,57 whose concomitant use with cisapride may determine clinically relevant pharmacokinetic interactions with significant increases in cisapride plasma concentrations. In the case of erythromycin, which may be used in gastroenterology for its prokinetic effect58 (Table 1), the pharmacokinetic interaction may be further complicated by a pharmacodynamic mechanism, since erythromycin per se may prolong the QT interval by acting on potassium currents such as IKr and Kv 1.5 currents.44, 59, 60 Grapefruit juice should also be considered, since it contains active ingredients that inhibit the CYP3A4 isoenzyme.61 The possible occurrence of torsades de pointes with concomitant use of cisapride and clarithromycin62 is of particular interest to gastroenterologists, since this macrolide is widely used for the eradication of Helicobacter pylori.63 Several studies have already discussed possible pharmacokinetic interactions occurring with clarithromycin.64[65][66]–67


In conclusion, the benzamide gastrointestinal prokinetic cisapride, one of the most widely used compounds in functional digestive disorders, may induce cardiac adverse effects such as lengthening of QT interval, syncopal episodes and ventricular dysrhythmias (i.e torsades de pointes). Although part of the cardiac adverse reaction to cisapride can theoretically be ascribed to activation of 5-HT4 receptors, there is now consensus that this is not the main mechanism through which cisapride evokes its cardiotoxic potential. Recent evidence has demonstrated that this effect is related to the blockade of K+ currents, which leads to a prolonged repolarization phase. Knowledge of this side-effect is needed to minimize potential hazards and decide appropriate monitoring of the ECG along with rate-corrected QT interval (QTc) in patients at risk. Development of new 5-HT4 receptor agonists not sharing the Class III antiarrhythmic pharmacophore will be the next step to further improve our therapeutic armamentarium for gastrointestinal motor disorders.

Of course, another class of potential interest includes the erythromycin derivatives lacking antibacterial activity (motilides), which are under development as potent gastrointestinal prokinetics.68 This will hold true especially if they are proved to have no effect on cardiac K+ channels, unlike erythromycin.


This study was partly supported by a grant from the National Research Council (Italy) to M. Tonini (98.03156. CT04).