Irinotecan(Drug information on irinotecan) (CPT-11 [Camptosar]) is an important new drug in the chemotherapeutic armamentarium. Irinotecan is active against a broad spectrum of malignancies, including carcinomas of the colon, stomach, and lung. Unfortunately, frequent and often severe gastrointestinal toxicities, particularly diarrhea, have limited its more widespread use.
Clinical observations, preclinical models, and pharmacokinetic studies have yielded some critical insights into the pathophysiology of these side effects. Early recognition and treatment of these toxicities have resulted in a reduction in patient morbidity. Despite these new pathophysiologic insights and advances in treatment, gastrointestinal toxicities remain a major problem with the clinical use of irinotecan. The gastrointestinal side effects of irinotecan administration can be divided into two distinct syndromes: early and late.
Early toxicity occurs within the first 24 hours of irinotecan administration and is characterized by diarrhea, emesis, diaphoresis, abdominal cramping, and, less commonly, hyperlacrimation and rhinorrhea. Various components of the syndrome have been reported in up to 80% of patients receiving the drug.[2,3]
The spectrum of symptoms appears to be dose-related. Patients who receive lower doses experience abdominal cramping, diarrhea, and diaphoresis, while those given over 300 mg/m2 also complain of salivation, hyperlacrimation, and visual changes. In a study of patients treated with 250 mg/m2 of irinotecan every 2 weeks, Petit et al found that most symptoms occurred within the first 2 hours, and median duration was approximately 30 minutes.
The constellation of early symptoms is consistent with cholinergic hyperstimulation. Irinotecan has been shown to mimic the effects of acetylcholine in various in vitro preparations. This action is mediated by inhibition of acetylcholinesterase and, perhaps less importantly, by direct binding to and stimulation of muscarinic receptors (Figure 1).
In both animal models and humans, symptoms are inhibited by administration of the anticholinergic drug atropine(Drug information on atropine). Atropine as needed is now routinely used in patients treated with irinotecan, and the incidence of cholinergic symptoms severe enough to interfere with treatment is quite low. In the series of Petit et al, the use of atropine was required in 34% of patients, and only one case of grade 3 early diarrhea occurred despite the use of a relatively high dose of irinotecan.
In contrast to early diarrhea, late diarrhea, defined as that occurring more than 24 hours after irinotecan administration, is a common and often serious and dose-limiting side effect. Although usually controllable with nonspecific and supportive measures, late diarrhea can be particularly dangerous in elderly or debilitated patients who experience other toxicities, such as neutropenia.
The overall incidence of late diarrhea in most US and European phase I and II trials of irinotecan ranges from 60% to 87% and appears to be dose-dependent.[1,9-11] The incidence of severe (National Cancer Institute [NCI] grade 3 or 4) diarrhea in these studies varies from 20% to 40%.[1,9-12] The incidence in Japanese studies is somewhat lower; however, most of these studies used less intensive dosing regimens.
The onset and duration of late diarrhea may vary with the dosing schedule. In European studies in which patients received 350 mg/m2 every 3 weeks, the median time to onset was 5 days and the median duration was 5 days. In the pivotal American trials, in which patients received 125 mg/m2/wk for 4 out of 6 weeks, the median time to onset was 11 days, and the median duration was 2 days (Figure 2).
Identification of patient characteristics that predispose to diarrhea may allow for the identification of patients who require close monitoring and early treatment. Unfortunately, this approach has not proven to be clinically useful.
In European studies, age older than 65 years, prior pelvic irradiation, and low performance status were significantly associated with an increased incidence of severe diarrhea.[1,3,11] In American studies, however, the results are conflicting. Von Hoff reported that ³ 65 years was a risk factor for grade 3/4 diarrhea, whereas Pazdur found no significant increase in these patients. Schaaf et al uncovered no differences in irinotecan pharmacokinetics between patients older and younger than 65 years of age.
In normal states, intestinal fluids remain in homeostasis, maintaining a finely regulated balance between fluid secretion and absorption. The intestines receive 8 to 9 L of ingested and secreted fluids each day and absorb all but 100 to 200 mL. Alteration of this balance by increased secretion or reduced absorption may result in the clinical symptom of diarrhea.
One potential mechanism by which drugs may induce diarrhea is direct damage to the intestinal epithelium. The resulting denuded mucosa is leaky and unable to absorb fluid. This is the probable mechanism of fluorouracil(Drug information on fluorouracil)-induced diarrhea, which results from diffuse mucosal injury.[17,18] Alternatively, compounds may increase secretion or decrease absorption of fluid by intestinal epithelial cells.
Irinotecan Pharmacology and Metabolism--Clues to the etiology of irinotecan-induced late diarrhea may be found in the complex pharmacology and metabolism of the drug (Figure 3). Irinotecan is a prodrug that is converted to an active form, SN-38, by carboxyl-esterases, which in humans are found predominantly in the liver.[19-21] The SN-38 metabolite is 250 to 1,000 times as potent an inhibitor of topoisomerase I as irinotecan.
Inhibition of topoisomerase I, which correlates with antitumor activity, results in the formation of cleavable complexes in DNA, which induce strand breaks. This DNA damage is thought to lead to cell death by apoptosis, which may be mediated through the interleukin-1 beta-converting enzyme (ICE) pathway. Both irinotecan and SN-38 require an intact lactone ring for topoisomerase I inhibition. They are inactivated by pH-dependent hydrolysis of the ring to the hydroxy acid.[22,26]
SN-38 is further metabolized by glucuronidation to SN-38 glucuronide (SN-38G), which is inactive.[27,28]. Glucuronidation is specifically performed by the UGT*1.1 isoform of hepatic uridine diphosphate glucuronosyltransferase, which also glucur- onidates bilirubin and is deficient in Gilberts syndrome. The SN-38G metabolite can also be deconjugated back to SN-38 in the gut by bacterial glucuronidases, which may result in increased exposure of the intestinal epithelium to toxic products. Takasuna et al found a correlation between intestinal bacterial beta-glucuronidase activity and the site of epithelial damage in rats exposed to irinotecan.
Biliary excretion is an important mechanism in the elimination of irinotecan and its metabolites, with 25%, 2%, and 1% of a dose excreted in the bile as irinotecan, SN-38G, and SN-38, respectively. Levels of irinotecan and SN-38 in the bile are up to 113- and 40-fold higher than levels in plasma. Normal excretion of these compounds into the bile is via the canalicular multispecific organic anion transporter (cMOAT), as well as other less-well characterized transporters.[34,35] Treatment with cyclosporine (Neoral, Sandimmune), which decreases biliary flow and inhibits MOAT, increases the areas under the curve (AUCs) of irinotecan, SN-38, and SN-38G severalfold.
The relationship between the pharmacokinetics of irinotecan and its metabolites and diarrhea is also quite complex. Both the parent compound and metabolites undergo enterohepatic circulation, and the concentration of SN-38 has been correlated with diarrhea in mice and in humans.[10,38,39] Ratains group found that a calculated biliary index [AUCCPT-11 × (AUCSN-38/AUCSN-38G)] was predictive of diarrhea.[40,41] However, Conti et al did not find the biliary index or other pharmacokinetic parameters to correlate with diarrhea in patients dosed with irinotecan weekly. The clinical usefulness of any of these measures is limited by the large overlap between putative high- and low-risk groups. Wasserman et al reported severe irinotecan toxicity in two patients with Gilberts syndrome, in which glucuronida- tion is deficient. This finding indicates the importance of glucuronidation in the detoxification of irinotecan and its metabolites. Patients with Gilberts syndrome, which may be found in up to 6% of the general population,[43-45] may constitute a group at high risk for late diarrhea.
In rats, treatment with valproic acid, which competes for glucuronidation with SN-38, reduced SN-38G by 99% and increased the AUC of SN-38 by 270%. Phenobarbital(Drug information on phenobarbital), an inducer of glucuronidation, increased the AUC of SN-38G while decreasing the AUCs of irinotecan and SN-38.
Unresolved Questions--Several questions remain, however: Why is the gastrointestinal tract preferentially affected by irinotecan or its metabolites, and by what mechanism do they induce diarrhea? The excretion of biliary irinotecan and SN-38 may expose the intestinal mucosa to high levels of the compound responsible for late diarrhea.
Further investigation of late diarrhea has been hampered by the lack of a pathologic correlation with symptoms. There have been several animal studies revealing bowel injury,[31,47] but no comparable human reports, though human studies are under way. Ikuno et al reported that mice treated with irinotecan exhibited intestinal wall thinning with epithelial vacuolation, vascular dilatation, and an inflammatory cell infiltrate. There was evidence of apoptosis in the ileum, as well as epithelial cell hyperplasia with goblet cell metaplasia in the cecum.
Interspecies variation may exist with respect to susceptibility to irinotecan toxicity. Guffroy and Hodge observed villous atrophy in the small intestine but not cecal lesions in their mouse studies. Takasuna et al found characteristic intestinal changes that appeared in a time-dependent fashion in rats treated with irinotecan. Gross thinning of both the intestines and cecum was seen. Histologically, there was cell death and apoptosis with crypt dropout, followed by the development of severe submucosal edema and an inflammatory infiltrate.
These results indicate that, at least in these models, diarrhea may result from a direct toxic action of irinotecan on the intestinal mucosa. One small human study of irinotecan-induced late diarrhea, reported in abstract form only, found normal d-xylose absorption, indicating a relatively intact intestinal mucosa but increased clearance of alpha-1-antitrypsin, which is associated with protein-losing enteropathy.
Other Effects of Irinotecan and Its Metabolites on the Intestinal Mucosa--Irinotecan and its metabolites may have additional effects on the intestinal mucosa that may induce diarrhea. In the normal intestine, secretion of fluid is driven by active secretion of chloride. Chloride is actively transported into the cell across the basolateral membrane by the Na+:K+:2Cl- cotransporter and then exits the cell via chloride channels along an electrochemical gradient.
In the rat, Sakai et al demonstrated that colonic chloride secretion, as measured with Ussing chambers as a short-circuit current, is stimulated by irinotecan. This stimulation is mimicked by stable analogs of the unstable thromboxane A2 and is blocked by inhibitors of cyclooxygenase, thromboxane synthase, and thromboxane A2 receptors. Our unpublished data show that SN-38 stimulates the short-circuit current in human colonic mucosa, indicating that a metabolite of irinotecan can induce chloride secretion. This effect is abrogated by cyclooxygenase inhibition.
Irinotecan may induce inflammation and intestinal secretion by paracrine mechanisms as well. Exposure of both mouse and human mononuclear cells to irinotecan induces secretion of tumor necrosis factor (TNF). Expression of TNF is associated with AIDS-related diarrhea and induces chloride secretion in colonocytes. Tumor necrosis factor also induces inflammation and may be important in the pathogenesis of inflammatory conditions, such as Crohns disease.