Cisapride in cats and dogs: regulatory status, dosing, and cardiac safety

Bottom line

Cisapride has the broadest prokinetic reach of the drugs in common use for feline constipation and early megacolon, increasing colonic motility as well as oesophageal, gastric and small-intestinal motility [1]. It is unapproved and compounded-only, has no controlled clinical trial in cats, and the published feline figures disagree by roughly fifteenfold per dose (arithmetic across the sourced table below, not any single source's figure). The human torsades disaster was a hERG/IKr effect [2] separate from 5-HT4 agonism [3]; QT prolongation appears in conscious cats only at grossly supratherapeutic doses [4]. And prokinetics "are not effective if enteric nerves have degenerated or become nonfunctional (as in cats with end-stage megacolon)" [1] — that cat needs a surgical conversation, not a dose increase.

Drug facts

Class	"A 5-HT4 receptor agonist with some 5-HT3 receptor antagonist activity" [1]
Cardiac liability	hERG blockade — IC50 44.5 nmol/L for tail currents at −40 mV, 6.70 nmol/L after prolonged depolarising steps [2]; unrelated to 5-HT4 agonism [3]
Approval status	No approved veterinary product; withdrawn from the global human market in 2000 [3]. Obtained "through compounding veterinary pharmacies from imported active pharmaceutical ingredients" [1]
Dosing	See the sourced table below — for a 4 kg cat the published per-dose figures span 0.4 mg to 6 mg [5][6]; dogs 0.1–0.5 mg/kg, PO, q 8–12 h [1]
Chief limitation	Ineffective once enteric nerves have degenerated, as in end-stage megacolon [1]

Regulatory status: why it is compounded, and what that means

Cisapride left the market on cardiac grounds, not efficacy grounds. Between 1993 and 1999 the FDA received reports of 341 patients with cisapride-associated cardiac events — 107 of them torsades de pointes — of whom 80 (23%) died; "in most individuals, the arrhythmia occurred in the presence of risk factors (other drugs and/or medical conditions)" [7]. A 2012 systematic review records that cisapride "was withdrawn from the global market in 2000" [3].

Two consequences follow, and clinicians conflate them. Cisapride appears at 21 CFR 216.24 as "All drug products containing cisapride" among products found "unsafe or not effective" — but that section sits in Part 216, Human Drug Compounding [8], and does not govern animal use. Animal compounding runs separately under FDA's GFI #256, whose patient-specific prescription conditions are a VCPR meeting the federal definition at 21 CFR 530.3(i); compounding "under the direct supervision of a veterinarian or pharmacist"; specified labeling, including "a statement indicating that the product is a compounded drug and not an FDA-approved or indexed drug"; compliance with state law and USP-NF monographs; adverse-event reporting within 15 business days; and the condition most often forgotten — the product must not be a "copy" of an FDA-approved or indexed product "unless the product produces a clinical difference for the specific patient(s)" [9]. Because no approved veterinary cisapride product exists, that copy test is moot here.

What this framework does not do is convert a compounded preparation into an approved one. "While drugs compounded from bulk substances are considered to be unapproved, the FDA CVM recognizes that veterinarians need these compounded drugs" [9]. Consent owners on those terms.

Mechanism: reaches the colon, is not an antiemetic

The decisive property is anatomical reach. "Metoclopramide has little or no effect on colonic motility, so it is not useful in cats with megacolon" [1]; cisapride "is more potent than metoclopramide and has broader prokinetic activity, increasing the motility of the colon as well as that of the esophagus, stomach, and small intestine" [1]. That is the whole reason it persists in feline practice. It is not, however, the only colon-active agent — ranitidine and nizatidine "also stimulate colonic smooth muscle contraction in cats through a cholinergic mechanism" [1].

It is not an antiemetic in the metoclopramide sense. Metoclopramide's antiemesis comes from "dopaminergic receptor antagonism at the chemoreceptor trigger zone (CRTZ)", while cisapride "does not cross the blood-brain barrier or have antidopaminergic effects" [1] — no CRTZ antiemesis, but no extrapyramidal signs either. It is a moderate feline P-glycoprotein substrate [10]. Arrhythmogenicity is a third, separable pharmacology: hERG blockade [2], "unrelated to 5-HT4 receptor agonism" [3].

What the efficacy evidence actually is

Ex vivo evidence is strong; controlled clinical evidence does not exist.

In smooth muscle strips from healthy cats, cumulative cisapride (10⁻⁹ to 10⁻⁶ M) produced contractions comparable to substance P and neurotensin, with nifedipine inhibiting maximal contraction by approximately 80% — a smooth-muscle-mediated, calcium-influx-dependent effect [11]. In tissue from cats with idiopathic megacolon, cisapride "stimulated contractions of longitudinal smooth muscle from both the ascending and the descending colon", "only partially inhibited by tetrodotoxin and atropine, but… virtually abolished by removal of extracellular calcium" [12]. Read the magnitude carefully, because it is not uniform along the colon: those contractions "were similar in magnitude to those induced by substance P and acetylcholine in the ascending colon, but were less than those observed in the descending colon" [12] — the weaker response is in the segment where obstipated faeces actually sit.

Note also that this ex vivo persistence does not license optimism in end-stage disease. The muscle strips still contract, and largely independently of enteric cholinergic nerves [12], yet the clinical position is that prokinetics "are not effective if enteric nerves have degenerated or become nonfunctional (as in cats with end-stage megacolon)" [1]. An isolated strip under an isometric transducer is not a colon that has to generate propulsion.

What is missing is any randomised or prospective clinical efficacy study in constipated cats. Merck grades its own claim: "Anecdotal experience suggests that cisapride (0.1–0.5 mg/kg, PO, every 8–12 hours) effectively stimulates colonic propulsive motility in cats with mild to moderate idiopathic constipation" [6]. A 2026 review notes that "effective drug combinations remain challenging" [13].

In dogs the best controlled data are upper-GI: in 6 awake Beagles, a single oral 0.5 mg/kg dose raised median lower oesophageal sphincter pressure from 29.1 to 50.7 mm Hg at 4 hours, while metoclopramide 0.5 mg/kg PO did not differ from placebo [14].

Dosing: the published figures, unreconciled

Every figure below is quoted from the source beside it. They disagree, and averaging them would invent a number nobody published. Note before you weigh them that three of the five rows are chapters of the same reference work — much of this spread is one manual disagreeing with itself, not four independent authorities converging.

Source	Species	Dose as published
Feline PK study, 7 healthy cats [5]	Cat	"1 mg/kg body weight per os (PO) every eight hours or 1.5 mg/kg body weight every 12 hours"
Merck, prokinetics chapter [1]	Cat	2.5 mg/cat for cats < 5 kg, and 5 mg/cat for cats > 5 kg, PO, q 8 h
Merck, constipation chapter [6]	Cat	0.1–0.5 mg/kg PO q 8–12 h (mild–moderate); "up to 1 mg/kg… in cats with moderate to severe constipation"
Merck, megacolon chapter [15]	Cat	0.5–1 mg/kg, PO, every 8–12 hours
Merck, prokinetics chapter [1]	Dog	0.1–0.5 mg/kg PO q 8–12 h; up to 1 mg/kg in some dogs

Applying those to a 4 kg cat — our arithmetic, not any source's — the per-dose figures run from 0.4 mg (0.1 mg/kg) to 6 mg (1.5 mg/kg): a fifteenfold spread, and the same fifteenfold spread on a daily basis (0.8 mg/day to 12 mg/day).

The feline backing is one crossover study in seven healthy cats: after a 2 mg/kg oral capsule, Cmax was 73.32 ± 16.59 ng/mL, bioavailability 29.0 ± 22.6% and oral half-life 5.27 ± 3.16 h, with no adverse effects observed [5]. That bioavailability SD is why a fixed mg/cat and a mg/kg schedule can both be defensible. Print formularies such as Plumb's carry their own ranges; use the edition in front of you rather than blending it with the table. Titrate to stool consistency, starting low in cardiac disease or on an azole.

Adverse effects and cardiac safety in cats and dogs

The veterinary safety record is thin rather than reassuring. Merck states that "in veterinary medicine, adverse reactions to clinical use of cisapride have not been reported" [1]. A 2025 feline P-glycoprotein study mentions anecdotal reports of neurological toxicosis in treated cats, but cites no primary case literature for them, so they cannot be weighed against Merck as contradicting evidence [10]. What that study does contribute is primary data: in a competitive efflux assay using a feline P-gp-expressing cell line, cisapride returned an MFI ratio of 6 and was classed a moderate P-gp substrate [10]. Read alongside the fact that cisapride "does not cross the blood-brain barrier" in normal animals [1], that gives a coherent mechanistic prediction: cats homozygous for the ABCB1 deletion, or cats on a P-gp inhibitor, are the plausible exposure subpopulation. That is an inference, not an observed signal.

Both cardiac datasets are supratherapeutic-dose studies:

• Conscious cats. 30 mg/kg PO twice daily for 7 days prolonged QT by 11–35% and QTc by 11–32%, with ST-segment depression in two of five cats [4] — roughly 20–30 times the per-dose amount in the published feline PK regimen [5].
• Anaesthetised dogs. IV cisapride escalated from 2 to 8 mg/kg, "four times the recommended therapeutic dose" per the authors, raised heart rate and prolonged QTc, but "no dogs developed spontaneous arrhythmias, and arrhythmias were not inducible by PES" [16]. That "four times" is an intravenous-to-intravenous comparison; measured against the 0.1–0.5 mg/kg oral dose in the table above, the systemic exposure multiple is far larger, and the two routes are not directly convertible because canine oral bioavailability is not established in any source used here.

So cisapride prolongs QT in both species if you give enough of it. That does not establish that clinical doses are safe — that study has not been done, and "effects appeared at ≥ X" never licenses "below X is safe".

Drug interactions

CYP3A4 is "the dominant isoform" in cisapride metabolism once intrinsic clearances are corrected for hepatic abundance, and 1 µM ketoconazole inhibited norcisapride formation from 10 µM cisapride by 51 ± 9% [17].

• Azole antifungals (ketoconazole, itraconazole, fluconazole) — direct CYP3A inhibition [17]; long-course itraconazole is the classic setup.
• Macrolides — erythromycin and clarithromycin "inhibit the hepatic metabolism of other drugs, including… cisapride" [1]. For balance, erythromycin pretreatment "failed to enhance the effect of cisapride on either HR or QTc" in the anaesthetised dog model [16].
• P-glycoprotein inhibitors — a plausible route to raised CNS exposure [10].

Add other QT-prolonging drugs, hypokalaemia and structural cardiac disease — the human signal concentrated in patients with exactly those cofactors [7].

Monitoring and stop criteria

No validated veterinary protocol exists, so this is a defensible position rather than a sourced one: baseline ECG in any cat with a murmur, gallop, cardiomyopathy or concurrent azole/macrolide therapy; potassium checked in the dehydrated or CKD patient; stop for new syncope, collapse or arrhythmia.

The more important stop criterion is therapeutic, and it is a statement about nerves, not muscle: prokinetics "are not effective if enteric nerves have degenerated or become nonfunctional (as in cats with end-stage megacolon)" [1], and "cats with long-standing obstipation and megacolon are not likely to improve with cisapride therapy" [6]. There is no bedside test for enteric nerve degeneration, so the usable proxies are duration and response: "the presence of clinical signs of megacolon for longer than 6 months has been associated with irreversible colonic changes" [15], and a cat still obstipating on an escalated dose with adequate stool softening and hydration has answered the question. That cat needs referral, not another 2.5 mg.

Alternatives and adjuncts

Osmotic laxatives. "Lactulose liquid 0.5 mL/kg, or crystals ¼–½ teaspoonful, PO, or in food, every 8–12 hours" — "the most effective agent of this group" [6]. Polyethylene glycol 3350 is the alternative: "1/8–1/4 teaspoonful, PO or in food, every 12 hours, adjusted to achieve soft stool consistency" [6], or "¼ teaspoon/cat, mixed in food, every 12 hours" [15].

Fibre and hydration. Fibre-supplemented diets, or psyllium (1–4 teaspoon/meal, lower for cats), wheat bran or pumpkin added to canned food; and hydration is a hard constraint, because "laxatives should be avoided in dehydrated animals" [6].

Other prokinetics. Ranitidine and nizatidine "also stimulate colonic smooth muscle contraction in cats through a cholinergic mechanism", their prokinetic activity being "due to acetylcholinesterase inhibition, with the greatest activity in the proximal GI tract" [1]. Metoclopramide is not an option here [1].

Surgery. "Subtotal colectomy, with or without ileocolic resection, is considered the procedure of choice" [15] for megacolon refractory to medical management [13]. Across 166 client-owned cats, major perioperative complications occurred in 9.9% (15/151), 14% (12/87) died as a direct result of treatment or complications of megacolon, constipation recurred in 32% (24/74) at a median of 344 days, and median survival was not reached; removing the ileocolic junction meant more liquid faeces (OR 3.45) and worse owner assessment (OR 3.6) [18]. Preserve the ICJ where possible.

Compounding quality

"While drugs compounded from bulk substances are considered to be unapproved, the FDA CVM recognizes that veterinarians need these compounded drugs" [9] — and the label must carry "a statement indicating that the product is a compounded drug and not an FDA-approved or indexed drug" [9]. The closest hard data on what that unapproved status costs come from another drug: in compounded fluconazole from four US pharmacies, capsules were acceptably accurate (median 96.3%, range 87.3% to 135.2% of labelled strength) but suspensions were not (median 73.8%, range 53.9% to 95.2%) [19]. Expect the same in cisapride suspensions — extrapolation, not cisapride data. Prefer capsules, stay with one pharmacy, and treat lost efficacy after a refill as a formulation question.

Frequently Asked Questions

Does cisapride actually work for feline megacolon, or is it just tradition? Both. The mechanism is real — in cats with idiopathic megacolon, cisapride "stimulated contractions of longitudinal smooth muscle from both the ascending and the descending colon", though those contractions "were less than those observed in the descending colon" relative to substance P and acetylcholine [12]. But no randomised or prospective clinical efficacy study exists in constipated cats, and Merck calls the supporting clinical experience anecdotal [6].

What dose should I start a constipated cat on? There is no single right answer. The feline PK study gives 1 mg/kg PO every eight hours or 1.5 mg/kg PO every 12 hours [5]; Merck's prokinetics chapter gives 2.5 mg/cat under 5 kg and 5 mg/cat over 5 kg PO every 8 hours [1]; its constipation chapter gives 0.1–0.5 mg/kg PO every 8–12 hours, up to 1 mg/kg in moderate to severe cases [6]. For a 4 kg cat that is 0.4 mg to 6 mg per dose — a fifteenfold spread, and that arithmetic is ours, not any source's. Pick one published scheme and titrate to stool consistency.

Is the human QT and torsades risk a real problem in cats? Not at doses anyone uses clinically, but the reassurance is thin. Conscious cats given 30 mg/kg orally twice daily for seven days showed QT prolonged 11–35% and QTc 11–32%, with ST depression in two of five [4] — some 20–30 times the per-dose PK regimen [5]. No clinical-dose safety study exists in either species, so the position is low concern rather than no concern, especially alongside other QT-prolonging drugs, hypokalaemia or structural heart disease [7].

Can I use cisapride in a cat on itraconazole or fluconazole? Cautiously, at a reduced dose, ideally with a baseline ECG. CYP3A4 is the dominant enzyme in cisapride metabolism, and 1 µM ketoconazole cut formation of the main metabolite by 51 ± 9% in human liver microsomes [17] — the mechanism behind the human fatalities [7].

Why not just use metoclopramide? Because it does not work where you need it: metoclopramide "has little or no effect on colonic motility, so it is not useful in cats with megacolon" [1]. It also crosses the blood-brain barrier and can cause extrapyramidal signs, while cisapride does neither — and so gives no CRTZ antiemesis [1].

Where do I get cisapride, and is prescribing it legal? From a compounding pharmacy, using imported active pharmaceutical ingredient [1]. Cisapride is on the FDA's withdrawn-or-removed list at 21 CFR 216.24, but that governs human compounding [8]. Animal compounding runs under GFI #256, whose patient-specific conditions include a 21 CFR 530.3(i) VCPR, direct supervision, prescribed labeling, and that the preparation not be a copy of an approved or indexed product unless it produces a clinical difference [9] — moot here, since no approved veterinary cisapride exists. It stays an unapproved drug.

When do I stop escalating the dose and refer for colectomy? When enteric nerve function is gone — prokinetics "are not effective if enteric nerves have degenerated or become nonfunctional (as in cats with end-stage megacolon)" [1], and cats with long-standing obstipation and megacolon "are not likely to improve" [6]. There is no bedside test for that, so use duration and response: signs beyond 6 months are associated with irreversible colonic changes [15], and failure on an escalated dose with adequate softening and hydration is your answer. In 166 cats having subtotal colectomy, major perioperative complications were 9.9% and median survival was not reached [18].

Is cisapride useful in dogs, or is this a cat drug? Yes, mostly upper-GI. In six awake Beagles a single 0.5 mg/kg oral dose raised median LES pressure from 29.1 to 50.7 mm Hg at four hours, while metoclopramide did not differ from placebo [14]. Merck lists it for gastric stasis, postoperative ileus and idiopathic megaoesophagus at 0.1–0.5 mg/kg PO every 8–12 hours [1].