- © 2000 by American Society of Clinical Oncology
Phase I and Pharmacologic Study of Irinotecan Administered as a 96-Hour Infusion Weekly to Adult Cancer Patients
- Chris H. Takimoto,
- Geraldine Morrison,
- Nancy Harold,
- Mary Quinn,
- Brian P. Monahan,
- Roger A. Band,
- Jeff Cottrell,
- Aida Guemei,
- Victor Llorens,
- Hilary Hehman,
- Abdel Salam Ismail,
- Donald Flemming,
- David M. Gosky,
- Haruyo Hirota,
- Sosamma J. Berger,
- Nathan A. Berger,
- Alice P. Chen,
- Jeremy D. Shapiro,
- Susan G. Arbuck,
- John Wright,
- J. Michael Hamilton,
- Carmen J. Allegra and
- Jean L. Grem
- From the Developmental Therapeutics DepartmentMedicine Branch, Division of Clinical Sciences, and Investigational Drug Branch, Cancer Therapy Evaluation Program, Division of Cancer Treatment and Diagnosis, National Cancer Institute; Department of Radiology, National Naval Medical Center, Bethesda, MD; and Cancer Research Center, Case Western Reserve University, School of Medicine, Cleveland, OH.
- Address reprint requests to Chris Takimoto MD, PhD, Medicine Branch, National Cancer Institute, Bldg 8, Rm 5101, National Naval Medical Center, 8901 Wisconsin Ave, Bethesda, MD 20889-5105; email ctakim{at}helix.nih.gov
Abstract
PURPOSE: We conducted a phase I and pharmacologic study of a weekly 96-hour infusion of irinotecan to determine the maximum-tolerated dose, define the toxicity profile, and characterize the clinical pharmacology of irinotecan and its metabolites.
PATIENTS AND METHODS: In 26 adult patients with solid tumors, the duration and dose rate of infusion were escalated in new patients until toxicity was observed.
RESULTS: In 11 patients who were treated with irinotecan at 12.5 mg/m2/d for 4 days weekly for 2 of 3 weeks, dose-limiting grade 3 diarrhea occurred in three patients and grade 3 thrombocytopenia occurred in two patients. The recommended phase II dose is 10 mg/m2/d for 4 days given weekly for 2 of 3 weeks. At this dose, the steady-state plasma concentration (Css) of total SN-38 (the active metabolite of irinotecan) was 6.42 ± 1.10 nmol/L, and the Css of total irinotecan was 28.60 ± 17.78 nmol/L. No patient experienced grade 3 or 4 neutropenia during any cycle. All other toxicities were mild to moderate. The systemic exposure to SN-38 relative to irinotecan was greater than anticipated, with a molar ratio of the area under the concentration curve (AUC) of SN-38 to irinotecan of 0.24 ± 0.08. One objective response lasting 12 months in duration was observed in a patient with metastatic colon cancer.
CONCLUSION: The recommended phase II dose of irinotecan of 10 mg/m2/d for 4 days weekly for 2 of 3 weeks was extremely well tolerated. Further efficacy testing of this pharmacologic strategy of administering intermittent low doses of irinotecan is warranted.
IRINOTECAN (CPT-11) IS a camptothecin derivative and topoisomerase I poison with single-agent activity in advanced colorectal cancer and other solid tumors.1 As a prodrug, it has little intrinsic activity and must first be enzymatically converted to the active metabolite, SN-38, by the action of irinotecan carboxylesterase-converting enzyme.1 SN-38 noncovalently interacts with the nuclear enzyme, topoisomerase I, to form a drug-stabilized cleavable complex in which the protein is covalently bound to DNA at the site of a single-strand break in the phosphodiester backbone. In the presence of ongoing DNA synthesis, these drug-stabilized cleavable complexes are converted into irreversible cytotoxic DNA damage.1 In some in vitro systems, camptothecins can induce potentially mutagenic and oncogenic chromosomal changes, including sister chromatid exchanges, gene deletions, and/or rearrangements.2 However, the carcinogenic risk associated with the use of these agents in humans is unknown.
The most commonly used schedule of irinotecan administration in North America is a 90-minute infusion of 125 mg/m2 weekly for 4 of 6 weeks.3 In previously treated patients with advanced colorectal cancer, this regimen produces objective responses in 13% to 27% of patients.3,4 However, systemic toxicity is frequent, with severe dose-limiting delayed diarrhea occurring in 36.4% of patients, even when high-dose loperamide antidiarrheal therapy is used.4 Pronounced myelosuppression is also common, with grade 3 or 4 neutropenia occurring in 22% to 32% of patients.3-5 In Europe, a 30-minute infusion of irinotecan at 350 mg/m2 every 3 weeks is the preferred schedule, but this regimen also produces substantial severe delayed diarrhea and myelosuppression.6 Therefore, modulations that can decrease the systemic toxicity of irinotecan while maintaining its antitumor efficacy would clearly be of clinical benefit.
There are several rationales for testing prolonged continuous infusions of irinotecan. First, because camptothecins require ongoing DNA synthesis to generate a cytotoxic drug effect, prolonged drug exposure times allow for an increased number of tumor cells to enter S phase of the cell cycle, thereby increasing the fractional cell kill. The clinical relevance of this schedule-dependent cytotoxicity was recently reviewed by Gerrits et al.7 Second, because longer infusions tend to produce lower peak-plasma drug concentrations, saturation of the carboxylesterase-converting enzyme might be avoided, leading to more efficient drug activation. Lower peak concentrations of the active metabolite, SN-38, could also theoretically lead to more efficient hepatic glucuronidation and biliary excretion, thereby reducing systemic toxicity.
In animal studies, low-dose repeated administration of irinotecan was shown to be active with low systemic toxicity in nude mice bearing human tumor xenografts.8 Furthermore, clinical administration of prolonged continuous infusions of other camptothecin analogs has been shown to be tolerable by Hochster et al9 and others.10,11 However, at the molecular level, Danks et al12 has shown that intermittent exposure to topotecan allows for maximal cleavable complex formation and maximal cytotoxic potency compared with continuous drug exposure. Consequently, we postulated that prolonged continuous exposure to irinotecan with periodic “breaks” or drug-free periods could potentially maximize clinical efficacy and decrease toxicity.
Therefore, we initiated the following phase I study of weekly prolonged 96-hour infusions of irinotecan in adult cancer patients weekly for 2 of 3 weeks. Our goals were to determine the recommended phase II dose and toxicity profile of this schedule of drug administration. We also examined the plasma pharmacokinetics of irinotecan and its metabolites, SN-38 and SN-38 glucuronide (SN-38G), in patients receiving this treatment.
PATIENTS AND METHODS
Patients
Eligibility requirements included age older than 18 years, refractory solid tumors, Eastern Cooperative Oncology Group (ECOG) performance status ≤ 2, absolute granulocyte count greater than 2,000/μL, platelet count greater than 100,000/μL, total bilirubin levels ≤ 1.5 mg/mL, AST level ≤ two times the upper limits of normal, and serum creatinine concentration ≤ 1.5 mg/dL. Objectively measurable disease was not required and prior topoisomerase I–targeting therapy was allowed, with the exception of prior irinotecan chemotherapy. At least 4 weeks had to have elapsed since the completion of any previous chemotherapy, and the patient had to be fully recovered from the toxicities of previous therapies. The protocol was approved by the institutional review boards of the National Cancer Institute (NCI) and the National Naval Medical Center, and all patients gave written informed consent. Patients were enrolled from April 1996 through March 1998.
Dosage and Dose Escalation
The starting dose of irinotecan was 10 mg/m2/d given as a 96-hour infusion weekly for 1.5 weeks (96 hours of drug infusion, 72 hours off, and 48 hours of infusion, repeated every 3 weeks). This initial dose rate was selected because it produced no grade 3 toxicity in patients when administered over a 120-hour period every 3 weeks.13 The initial dose escalation plan was to maintain a rate of infusion of irinotecan of 10 mg/m2/d and to increase the duration of treatment in successive cohorts of new patients until the initial targeted schedule of 3 of 4 weeks was achieved (Table 1). However, during the study the protocol was modified to target a 2 of 3 weeks treatment schedule to match the preclinical animal data and to increase patient convenience and accrual. Once the targeted duration of treatment was achieved, further escalation of the rate of infusion was planned (12.5 mg/m2/d, 15 mg/m2/d, and so on). This pattern of phase I dose escalation was originally described by Hochster et al.9 Individual patient dose escalations were permitted every four cycles provided that no toxicity worse than grade 1 was experienced in the previous three cycles. At least three irinotecan-naive patients were entered at each dose level. The toxicity at a given irinotecan dose level was considered to be dose-limiting toxicity (DLT) if any of the following were observed during cycle 1: (1) absolute neutrophil count (ANC) less than 500/μL for longer than 5 days or febrile neutropenia, (2) platelet count less than 50,000/μL, (3) grade 3 or 4 nonhematologic toxicity according to the NCI common toxicity criteria (version 1),14 or (4) a greater than 2-week delay in treatment as a result of drug-related toxicity. If DLT was observed in any patient, the number of patients treated at that dose level was expanded from three up to a maximum of six new patients. The recommended phase II and maximum-tolerated dose (MTD) was defined as one dose level below the dose that induced DLT during cycle 1 in at least 33% of new patients. Once the recommended phase II dose was determined, as many as eight additional patients were entered to confirm the tolerability of the regimen. Routine use of colony-stimulating factors was not permitted. Treatment was resumed when the ANC had recovered to greater than 2,000/μL and the platelet count was greater than 100,000/uL. Dose reductions were made by decreasing the rate of drug infusion by 15% to the next lower dose level rather than shortening the duration of drug administration for any grade 3 or greater nonhematologic toxicity, grade 4 hematologic toxicity (except for an ANC < 500/μL, which had to persist for longer than 5 days), or for a failure of blood counts to recover by day 28.
Dose Levels
In the event of any diarrhea, patients were instructed to start high-dose loperamide15 at 4 mg initially followed by 2 mg orally every 2 hours during the day and 4 mg every 4 hours at night at the first sign of any loosening bowel movements. Patients continued on this regimen until all bowel movements ceased for at least 12 hours.
Drug Administration
Irinotecan was supplied by the Cancer Therapy Evaluation Program, Division of Cancer Treatment and Diagnosis, NCI (Bethesda, MD), in vials that contained irinotecan concentrate (100 mg/5 mL). After reconstitution in 5% dextrose for injection, the drug was administered over 24 hours via central venous access device using CADD I or CADD-PLUS pumps (Pharmacia Deltec, St Paul, MN) with fresh solution prepared at 24-hour intervals. Antiemetic premedications were not routinely administered.
Pretreatment and Follow-Up Studies
Complete blood cell counts, laboratory chemistries, urinalysis, and a history and physical examination were performed at every cycle. In addition, complete blood cell counts were obtained twice weekly. If the ANC decreased to less than 1,000/μL, then complete blood cell counts were obtained at least three times per week until recovery to greater than this value. Formal tumor measurements were performed after the first two cycles of therapy and then every three cycles thereafter. Treatment was continued indefinitely in the absence of disease progression, provided that all toxicities were acceptable and that the patient remained willing to continue on study. A complete response was defined as the disappearance of all disease on two measurements separated by a time interval of at least 4 weeks. A partial response was defined as at least a 50% decrease in the sum of the products of the bidimensional measurement of the measurable lesions on two separate occasions, again at least 4 weeks apart. Time to treatment failure was defined as the time from study entry to the time of documented disease progression, if available. If the disease progression data was not available, then time to study withdrawal for any reason was considered as a treatment failure.
Pharmacologic Analyses
Plasma samples were obtained from 11 patients treated at 10 mg/m2/d and 12.5 mg/m2/d during cycle 1 of therapy at 24, 48, 72, and 96 hours after the start of the infusion to determine the plasma concentrations of irinotecan, SN-38, and SN-38G. Clinical sample processing and quantitation of total (lactone and carboxylate) concentrations of irinotecan and SN-38 were analyzed using a reversed-phase high-performance liquid chromatography (HPLC) method modified from Rivory et al.16 Only total drug plasma concentrations were measured. Briefly, 0.4-mL plasma samples were deproteinated by precipitation in 1.1 mL of acetonitrile and stored at −80°C before subsequent analysis. Stored samples were allowed to thaw, evaporated to dryness, and then resuspended in 0.25 mL of acidified methanol immediately before HPLC injection. Samples were analyzed using an HP 1050 HPLC system (Hewlett-Packard, Wilmington, DE), a Beckman Ultrasphere ODS column (Beckman Coulter, Fullerton, CA), and a Waters 470 fluorescence detector (Waters Corp, Millford, MA; excitation wave length, 355 nm; emission wavelength, 425 nm; bandwidth, 18 nm). An isocratic mobile phase consisting of 25% (volume to volume ratio) acetonitrile/50 mmol/L potassium phosphate buffer (pH 3.0) containing 5 mL of PIC-B reagent per liter (Waters Corp) ran at a flow rate of 1 mL/min. The lower limit of quantitation for irinotecan was 1.25 nmol/L with a coefficient of variation of 15.4%, and for SN-38 it was 0.1 nmol/L with a coefficient of variation of 17%. Plasma SN-38G concentrations were assayed by incubating 0.5 mL of plasma with and without 1,000 units of beta-glucuronidase (Sigma Chemical Co, St Louis, MO). Paired samples were then analyzed for SN-38 content and the amount of the glucuronide was calculated by subtracting the amount of SN-38 in the control sample from the enzyme-treated sample.17
Pharmacokinetic parameters of area under the concentration versus time curve to 96 hours (AUC96 hours) and total-body clearance were calculated using noncompartmental methods using WinNonlin (SCI, Inc, Cary, NC). Steady-state plasma concentrations (Css) for irinotecan were determined by averaging the plasma concentrations measured for each patient at 48, 72, and 96 hours. The total-body clearance was calculated by dividing the rate of infusion by the steady-state plasma irinotecan concentration.18
Pharmacodynamics
The biliary index was previously shown to be correlated with the incidence of irinotecan-associated diarrhea when administered as an intravenous (IV) bolus.17,19 The biliary index was defined as the (AUCSN-38/AUCSN-38G) × AUCCPT-11 and was calculated for our patients using the AUC values obtained out to 96 hours.
RESULTS
Patient Demographics
Twenty-six patients were enrolled onto the study and 24 patients were assessable for toxicity. Overall, 101 courses of irinotecan therapy over six dose levels were assessable for toxicity (Table 1). One unevaluated patient withdrew from the study during cycle 1 for personal reasons after receiving treatment with irinotecan at 10 mg/m2/d for only 2 days, and the other patient enrolled at 12.5 mg/m2/d was found after only 3 days of treatment to have a new bone metastasis that required interruption of the infusion for radiation therapy. Neither patient exhibited any toxicity that was considered to be drug-related. The median number of courses administered per assessable patient was two and ranged from one to 19. Patient demographics are listed in Table 2. All but two of the patients had an ECOG performance status of either 0 or 1. All patients enrolled except one had received prior chemotherapy, and nine patients received both prior chemotherapy and radiation therapy.
Patient Demographics
DLTs
The DLTs were grade 3 diarrhea and grade 3 thrombocytopenia; both occurred at the dose level of 12.5 mg/m2/d for 4 days for 2 weeks. At this dose level, one new patient of six experienced grade 3 delayed diarrhea during cycle 1 despite administration of high-dose loperamide, and two additional new patients experienced grade 3 diarrhea during cycles 2 or 3 (Table 3). Furthermore, because grade 3 thrombocytopenia was also seen during cycle 1 in two of six patients at this dose level, our recommended phase II dose for this regimen is 10 mg/m2/d for 4 days for 2 of 3 weeks. One of two patients who were treated with the longest infusion schedule, which consisted of 10 mg/m2/d for 4 days for 2.5 weeks repeated every 4 weeks, experienced dose-limiting grade 3 diarrhea. The duration of infusion was shortened for this patient to 10 mg/m2/d for 4 days for 2 of 3 weeks, which was tolerated without further dose-limiting diarrhea. The second patient at this dose level did not experience DLT but was removed from the study for disease progression after completing two cycles. No additional patients were treated on the 2.5-week schedule because the decision was made to shorten the duration of treatment to 2 of 3 weeks per cycle to improve patient convenience and to better match the preclinical schedules that were used in animal studies.8
Severity of Toxicity by Patient: Delayed Diarrhea and Nausea/Vomiting Toxicity
Seventeen patients received 41 cycles of therapy at the recommended phase II dose level of 10 mg/m2/d for 4 days for 2 of 3 weeks. Although one of these patients experienced grade 3 diarrhea and two patients experienced grade 2 diarrhea, overall this dose regimen was generally well tolerated, and no hospitalizations were required for dehydration. High-dose loperamide was generally effective in controlling diarrhea. Three of the 17 patients who were treated at 10 mg/m2/d for 4 days for 2 of 3 weeks complained of constipation. When diarrhea did occur, it was delayed in onset, generally started by day 10, and lasted for 3 to 5 days. No symptoms of a cholinergic syndrome (early onset diarrhea, flushing, or diaphoresis) were observed.
Other Gastrointestinal Toxicities
Nausea and vomiting were mild in most patients, with grade 2 nausea and/or vomiting occurring in 18% of patients who were treated at the highest rate of drug infusion (12.5 mg/m2/d for 4 days for 2 of 3 weeks). At the recommended phase II dose, 12% of patients experienced grade 2 or 3 nausea or vomiting. A single patient (6%) at this dose level developed grade 4 nausea and vomiting (Table 4); however, this individual subsequently developed disease-related small-bowel obstruction. In most cases, nausea and vomiting responded well to oral ondansetron therapy, and patients were able to maintain good oral intake. No patient required termination of their therapy for nausea and vomiting, and prophylactic antiemetics were not required. Early onset diarrhea did not occur. Abdominal pain was uncommon; however, constipation was noted in 36% of patients overall. Mild anorexia was noted in 29% of patients who were treated at 10 mg/m2/d for 4 days for 2 of 3 weeks; this was generally more pronounced during the third and fourth days of the infusion. Anorexia generally dissipated over 24 hours after stopping the infusion. Two patients overall complained of changes in taste sensation during the drug infusions.
Severity of Toxicity by Patient: Hematologic Toxicity
Hematologic Toxicity
Hematologic toxicity was surprisingly mild, with no patient experiencing grade 3 or 4 neutropenia; grade 3 leukopenia was seen in one patient who was treated with 10 mg/m2/d for 4 days for 2 of 3 weeks (Table 4). Granulocyte colony-stimulating factor support was not required in any patient. In contrast, at the highest dose level of 12.5 mg/m2/d for 4 days for 2 of 3 weeks, two of 11 patients experienced grade 3 thrombocytopenia during cycle 1. At the recommended phase II dose of 10 mg/m2/d for 4 days for 2 of 3 weeks, no patients experienced grade 3 or 4 thrombocytopenia. No bleeding complications were associated with any of these transient episodes of thrombocytopenia, although one patient with a normal platelet count and tumor-related chronic vaginal bleeding did experience grade 4 anemia after seven cycles of chemotherapy. Baseline grade 1 anemia was observed in 10 patients, grade 2 in three patients, and grade 3 in a single patient before any chemotherapy was administered. Treatment-related grade 3 or 4 anemia only occurred in two patients, and overall, only three patients required blood transfusions. Cumulative hematologic toxicity was not observed.
Other Toxicities
Other toxicities of this regimen were minimal. At the recommended phase II dose of 10 mg/m2/d for 4 days for 2 of 3 weeks, grade 1 fatigue occurred during drug infusion in 95% of patients, mild alopecia was noted in 14%, and headaches occurred in 14%. Pulmonary toxicity, which has been previously reported with irinotecan therapy,1 did not occur. One patient experienced a transient urticarial skin rash that required termination of therapy after 17 cycles; however, no other allergic or hypersensitivity reactions were identified. Hemorrhagic cystitis was not noted, although one patient had intermittent mild hematuria that predated the initiation of therapy and was attributed to preexisting benign obstructive prostatic disease. No documented drug-related toxic deaths occurred, but one patient with rapidly progressive disease who refused further treatment and evaluation after completing one cycle of chemotherapy at 12.5 mg/m2/d for 4 days for 2 of 3 weeks died in hospice care within 2 weeks of being removed from the study.
All patients had central venous catheters while on study, the majority of which were port devices. One patient experienced a central venous catheter line infection that necessitated treatment interruption. Another patient’s treatment cycle was stopped 3 days early because of needle dislodgement from the central venous port during the infusion. This raised the concern about possible extravasation complications; however, no local reaction developed and the patient was able to continue on protocol therapy.
Administered Doses of Therapy
Of the 14 patients who started at the recommended phase II dose of 10 mg/m2/d for 4 days weekly for 2 of every 3 weeks, only one required a dose reduction (for grade 3 diarrhea), and three had their doses escalated to 12.5 mg/m2/d for 4 days for 2 of 3 weeks because no toxicities worse than grade 1 occurred. Thus we were successfully able to administer irinotecan at the recommended phase II dose or higher in the majority of cycles. Of 41 courses administered to 17 patients at the recommended phase II dose, 93% were administered on schedule.
Responses
Twenty-one of the 26 patients were evaluated for response. Three of the nonassessable patients developed medical complications from their underlying tumor before an initial restaging, including a tracheoesophageal fistula, bowel obstruction, and a newly recognized bone lesion that required radiation therapy. Two additional patients withdrew from therapy before restaging and refused further therapy. Of the 21 assessable patients, a partial response was observed in a 63-year-old woman with metastatic colon cancer who had previously been treated with fluorouracil and leucovorin and with paclitaxel combined with PSC-388, an experimental multidrug resistance inhibitor. Multiple small pulmonary nodules completely disappeared and periaortic adenopathy decreased in size, with the best response noted after 4 months of irinotecan therapy. The carcinoembryonic antigen (CEA) gradually decreased from 500 ng/mL to 46 ng/mL, and the CEA nadir was reached after 15 cycles of chemotherapy. This patient was initially treated at 10 mg/m2/d for 4 days for 2.5 of 4 weeks for one cycle; however, this duration was then decreased to 10 mg/m2/d for 4 days for 2 of 3 weeks for the next six cycles because of grade 3 diarrhea. During cycle 7, this patient developed grade 3 vomiting, which led to a dose reduction to 8.5 mg/m2/d for 4 days for 2 of 3 weeks for her remaining 13 cycles, and these cycles were well tolerated. However, this patient’s therapy was stopped after a total of 19 cycles because of an urticarial skin rash. Her duration of response was greater than 12 months, and during the last 4 months she received no systemic treatment. One other patient with metastatic colon cancer who had received prior fluorouracil/leucovorin chemotherapy achieved a 39% decrease in measurable disease, with the best response noted after eight cycles of chemotherapy at the recommended phase II dose. Of the remaining 19 assessable patients, four had stable disease lasting ≥ 3 months, whereas 15 experienced progressive disease during the first 3 months on study. The median time on study was 1.8 months (range, 0.5 to 14.5 months), and the overall median time to treatment failure was 1.5 months (range 0.5 to 17.5 months).
Pharmacokinetics Analysis
Pharmacokinetic blood sampling in 11 patients was instituted at 0, 24, 48, 72, and 96 hours during weeks 1 and 2 of the initial cycle of chemotherapy. Measurements of the blood samples at 48, 72, and 96 hours were used to determine the steady-state concentrations of total plasma SN-38 and SN-38G and the total-body clearance and Css of irinotecan (Table 5). The clearance of irinotecan was 24.86 ± 12.85 L/h/m2 for five patients treated at 10 mg/m2/d for 4 days and 19.68 ± 4.56 L/h/m2 for six patients treated at 12.5 mg/m2/d for 4 days. These values for total-body clearance are similar to the 15 to 26 L/h/m2 previously reported for other schedules of irinotecan administration.20-23 No evidence of drug accumulation was found in patients between weeks 1 and 2 (data not shown).
Steady-State Plasma Pharmacokinetics
The relative rate of conversion of irinotecan to its active metabolite was quite high, with a mean molar ratio of AUCSN-38 to AUCCPT-11 of 0.24 ± 0.08. This was much higher than the 0.03 to 0.04 ratio previously reported for short infusions of irinotecan.20-25 The metabolic ratio, defined as the (AUCSN-38 + AUCSN-38G)/AUCCPT-11,26 was also quite high, with a mean value of 1.13 ± 0.34. Likewise, the AUC ratio of unconjugated SN-38 to the glucuronidated form was 0.30 ± 0.12. These parameters suggest a much greater relative systemic exposure to SN-38 for a given dose of irinotecan when the drug is administered as a prolonged infusion. In agreement with other studies of irinotecan, there was marked interpatient variability in the relative exposure of SN-38 in individual patients, with a coefficient of variation for the SN-38/irinotecan AUC ratio of 33.2% and a coefficient of variation for the SN-38/SN-38G AUC ratio of 40.3%.
Pharmacodynamic correlations were examined for the biliary index during the first 96 hours of infusion and the degree of diarrhea. No clear correlations were observed (data not shown), but the power of this analysis was limited by the minimal toxicity of the regimen.
DISCUSSION
Infusional chemotherapy of irinotecan was extremely well tolerated in this study. The principal DLTs were diarrhea and thrombocytopenia. At the recommended phase II dose of 10 mg/m2/d for 4 days weekly for 2 of 3 weeks, however, the regimen produced only modest, easily controlled, delayed diarrhea, and clinically significant myelosuppression was not seen. This experience is in marked contrast to the high incidence of severe grade 3 and 4 diarrhea that has been seen with the weekly or every 3 week short-infusion regimens,3,6 despite the use of high-dose loperamide. Likewise, severe myelosuppression was uncommon in our study, in contrast to the grade 3 or 4 neutropenia seen in 18% of irinotecan cycles using the every 3 week schedule6 and in 32% of patients on the weekly schedule.3 Thus our infusional regimen was associated with much less clinical toxicity than the shorter infusion regimens. One previous phase I study examined a prolonged 5-day infusion of irinotecan at doses of 5 to 40 mg/m2/d for 5 days every 3 weeks.13 The principal DLTs in this study were diarrhea, nausea and vomiting, and leukopenia, which were not substantially different from the IV short-infusion regimens. However, because treatments were only administered every 3 weeks, this study did not mimic the well-tolerated, low-dose, intermittent treatment schedule used in animal experiments that provided the rationale for our current regimen.8
A major concern in our study is the low amount of irinotecan of 40 mg/m2/wk (80 mg/m2/cycle) administered at our recommended phase II dose. This irinotecan dose-intensity is much less than that achieved during the weekly or every 3 week short-infusion schedules (Table 6). However, a more fair comparison is to examine the relative exposures to the active metabolite, SN-38, and not to the inactive parental prodrug. When the weekly SN-38 dose-intensity is compared, the 96-hour infusion schedule produces a dose-intensity that is much closer to the range achieved with the other commonly used schedules of irinotecan administration (Table 6). Furthermore, for an S-phase cell cycle–specific cytotoxic agent, prolonged low-dose exposures may be theoretically more effective than equivalent AUC exposures for a shorter period of time.7,12 Finally, although an accurate assessment of therapeutic efficacy cannot be made in a phase I study, our regimen was active, generating a partial and minor response in two patients with advanced colorectal cancer who had previously received fluorouracil. Thus further efficacy testing of our well-tolerated regimen seems to be indicated.
Dose-Intensity Comparison
Pharmacokinetic monitoring of systemic AUC drug exposure also allows for a detailed comparison of our clinical trial results and preclinical animal models. In nude mice bearing the relatively sensitive HC1 human colon cancer xenografts, the lowest daily SN-38 AUC associated with complete responses on a daily times 5 days for 2 of 3 weeks schedule was 942.6 ng/mL·h.27 This corresponds to a weekly SN-38 AUC dose-intensity of 8,015 nmol/L·h/wk, which is six to 10 times greater than the SN-38 exposures that are achievable in humans using any schedule of irinotecan administration (Table 6). Thus it may be extremely difficult to achieve the same degree of antitumor efficacy in humans compared with preclinical animal models, largely because mice are able to tolerate much greater exposures to SN-38.
The most unexpected finding in this study was the markedly increased systemic exposure to SN-38 achieved for a given dose of irinotecan administered as a prolonged infusion. The ratio of AUCSN-38/AUCCPT-11 of 0.24 was much higher than the 0.03 to 0.04 molar ratios that were reported in earlier studies of weekly or every 3 week irinotecan infusions.20-25 The reason for this increased exposure to SN-38 is not known. One possible explanation is that the prolonged infusions may increase the efficiency of irinotecan activation to SN-38. In man, the irinotecan carboxylesterase-converting enzyme that is responsible for this reaction is relatively inefficient compared with that of other species, such as the rat.28,29 One of our hypotheses for initiating this study was that lower peak doses of irinotecan during an infusion may prevent saturation of the carboxylesterase enzyme, leading to more efficient enzymatic conversion of irinotecan to SN-38. Our pharmacokinetic observations are consistent with this hypothesis; however, Haaz et al28 recently reported that the Michaelis constant, Km, for human liver carboxylesterase for irinotecan is 50 μmol/L, which is considerably greater than the concentrations of irinotecan achieved in our study. More recent evidence suggests that other carboxylesterases may be important for the production of SN-38 in humans.30 Thus enzyme saturation may still be a possible explanation for our findings. Interestingly, a similar increased efficiency of exposure to SN-38 was reported in a recent pediatric phase I trial conducted by Stewart et al31 using low-dose daily IV bolus administration of irinotecan at 20 mg/m2/d for 5 days, in which a relatively high AUCSN-38/AUCCPT-11 ratio of 0.22 was also seen. This later observation supports our theory that the increased relative ratio of SN-38 to irinotecan seen with low-dose repeated administration of irinotecan may be a schedule-dependent phenomenon.
Other theoretical factors may contribute to higher relative exposures to SN-38, including decreased SN-38 clearance resulting from either impaired glucuronidation, impaired biliary excretion, or enhanced enterohepatic circulation. However, we currently have no direct evidence to support the decreased clearance of SN-38 in our patients who received prolonged infusions of SN-38. To better define these drug pharmacokinetics, we are measuring the full distribution and elimination kinetics of irinotecan, SN-38, SN-38G, and the oxidative metabolite of irinotecan, 7-ethyl-10-[4-N-(5-aminopentanoic acid)-1-piperidino] carbonyloxycamptothecin in a cohort of additional new patients treated at our recommended phase II dose of 10 mg/m2/d for 4 days weekly for 2 of 3 weeks. In this study, only the total (lactone plus carboxylate) forms of irinotecan and SN-38 were measured in plasma. Sasaki et al32 have previously reported that total drug plasma concentrations are highly correlated with lactone concentrations for irinotecan and SN-38 in patients who receive short drug infusions. However, because the fraction of total drug present as the lactone may vary during more prolonged 96-hour infusions of irinotecan, we are now collecting samples to measure the lactone and carboxylate species of both irinotecan and SN-38 in patients who are currently receiving this regimen.
The high interpatient variability in drug AUC observed here is in agreement with other pharmacokinetic studies of irinotecan.26 Because interpatient differences in carboxylesterase activity may explain some of the variability in SN-38 exposure, we are currently studying methods to assess the activity of this enzyme in individual patients in an effort to help to predict drug kinetics. Finally, no clear pharmacodynamic relationships were identified in our study, but as previously described, the overall toxicity of this regimen was quite limited.
The greatest potential drawback of using prolonged infusions of irinotecan was the inconvenience of this schedule. Patients returned to the clinic daily to reload their ambulatory infusion pumps. However, with additional sterility and stability testing, it may be possible to lengthen the infusion time for each loading of the ambulatory pump, thereby eliminating the need for daily, every-24-hours clinic visits during therapy. This would substantially improve the convenience of this regimen. Perhaps the greatest significance of this study is that it provides a pharmacokinetic proof of principle demonstrating that prolonged low-dose irinotecan exposures are well tolerated, clinically active, and associated with the enhanced conversion of irinotecan to SN-38. If the pharmacokinetic profile achieved here could be reproduced by administering an oral preparation of irinotecan, then the clinical relevance of our observations would be substantially enhanced.
In conclusion, we have determined that irinotecan infused at 10 mg/m2/d for 4 days weekly for 2 of 3 weeks is well tolerated, with mild to moderate gastrointestinal toxicity and minimal myelosuppression. Furthermore, it is clinically active in previously treated patients with colorectal cancer, which suggests a possible improvement in therapeutic index compared with more commonly used weekly or every 3 week short-infusion regimens. Prolonged infusions of irinotecan are also associated with the increased relative systemic exposure to the active metabolite, SN-38, which is consistent with increased efficiency of metabolic activation of irinotecan by the carboxylesterase-converting enzyme. Further clinical testing of this pharmacologic strategy is warranted.
Acknowledgments
ACKNOWLEDGMENT
We thank the National Cancer Institute and the United States Department of Defense medical staff for the excellent medical care they provided for the patients who were enrolled on this clinical study.
- Received March 18, 1999.
- Accepted October 6, 1999.
