- © 2005 by American Society of Clinical Oncology
Phase I Trial of Capecitabine and Weekly Irinotecan in Combination With Radiotherapy for Neoadjuvant Therapy of Rectal Cancer
- Ralf-Dieter Hofheinz,
- Bolko von Gerstenberg-Helldorf,
- Frederik Wenz,
- Ulrike Gnad,
- Uta Kraus-Tiefenbacher,
- Albrecht Müldner,
- Rüdiger Hehlmann,
- Stefan Post,
- Andreas Hochhaus and
- Frank Willeke
- From the Onkologisches Zentrum, III Medizinische Klinik, Chirurgische Klinik und Poliklinik, Sektion Strahlentherapie, Institut für Klinische Radiologie, Fakultät für Klinische Medizin Mannheim der Universität Heidelberg, Germany.
- Address reprint requests to Ralf-Dieter Hofheinz, MD, Onkologisches Zentrum, III Medizinische Klinik, Fakultät für Klinische Medizin Mannheim der Universität Heidelberg, Theodor-Kutzer-Ufer, D-68167 Mannheim, Germany; e-mail: ralf.hofheinz{at}med3.ma.uni-heidelberg.de
Abstract
Purpose To establish the feasibility and efficacy of capecitabine in combination with weekly irinotecan (CAPIRI) with concurrent pelvic radiotherapy (RT) in patients with locally advanced rectal cancer.
Patients and Methods Nineteen patients with rectal cancer clinical stage T3-4, Nx received weekly irinotecan 50 mg/m2 (days 1, 8, 15, 22, 29) and two doses of capecitabine (days 1 through 38; dose level [DL] I, 500 mg/m2 bid; DL II, 625 mg/m2 bid) according to phase I methodology. Three-dimensional conformal RT was given to a dose of 50.4 Gy (45 Gy + 5.4 Gy).
Results On DL I, no dose-limiting toxicities occurred, whereas diarrhea grade 3 affected three of seven patients on DL II. Twelve patients were treated on DL I and received a median relative dose-intensity of 100% for both drugs. Grade 3 or 4 adverse events were observed in only one of these patients (asthenia grade 3). All patients underwent surgery and R0 resection was achieved in all patients. Pathologic complete remission was observed in four patients and another five patients had only microfoci of residual tumor.
Conclusion Preoperative chemoradiotherapy with CAPIRI is feasible and well tolerated. The preliminary efficacy is good, and the tolerability is at least comparable with data for fluorouracil plus irinotecan chemoradiotherapy. Larger phase II trials of the CAPIRI-RT schedule clearly are warranted.
INTRODUCTION
There have been a number of advances in surgical technique and radiotherapy (RT) for patients with localized rectal cancer during the last two decades. In particular, total mesorectal excision (TME) has reduced local failure rates to approximately 10%. Nevertheless, despite optimized surgery, preoperative RT can further improve local recurrence rates.1 Although postoperative chemoradiotherapy with bolus fluorouracil (FU) for Dukes' B and C rectal cancer is still regarded as one of the cornerstones of treatment,2,3 preliminary data from the German Intergroup trial, which compared pre- and postoperative chemoradiotherapy in resectable rectal carcinoma, challenge this practice.4 Both the reduction in local failure rates (11% v 7%) and superior tolerability profile seemed to favor the preoperative approach in this study. In addition, preoperative chemoradiotherapy is the treatment of choice for clinically staged T4 tumors and is widely used in low-lying T3 tumors with the aim of achieving sphincter preservation.
Although local recurrence is now less of a problem, distant metastasis is the most common form of treatment failure. Therefore, this raises the question of whether early administration of systemically effective doses of chemotherapy might improve long-term outcome. Chemoradiotherapy with biochemical modulation of FU (by folinic acid [FA], levamisole, or both) has not been proven to be superior to FU alone,5 nor has the use of combination chemotherapy with older cytostatic drugs (eg, semustine6) or prolongation of chemotherapy.7 However, a postoperative randomized trial using infusional instead of bolus FU throughout RT has showed improved disease-free and overall survival, probably due to a reduction in the rate of distant metastases.6 Nevertheless, this method of application requires indwelling catheters and portable pump systems, which are both inconvenient for patients and healthcare providers, and are associated with a number of complications, including thrombosis and infection.
To reduce the use of infusional devices and their associated morbidity, the oral fluoropyrimidine capecitabine (Xeloda; Hoffmann-La Roche Inc, Basel, Switzerland), which is preferentially activated in tumor tissues and mimics the pharmacokinetics of infusional FU,8 is replacing FU as part of the combined-modality treatment of a number of gastrointestinal (GI) cancers. Capecitabine has proven safety advantages over bolus FU in both the metastatic and adjuvant settings.9–11 In addition to simplifying chemoradiotherapy by avoiding the need for daily intravenous (IV) infusions, synergy between capecitabine and RT has been demonstrated, with capecitabine acting as a radiosensitizing agent.12 Moreover, capecitabine with or without oxaliplatin applied during the radiation period is well tolerated and active in patients with locally advanced rectal cancer.13–16
Irinotecan (Camptosar; Aventis, Bad Sooden, Germany) is another promising chemotherapeutic agent with radiosensitizing properties.17,18 It has now become standard therapy in the first-line treatment of metastatic colorectal cancer in combination with bolus or infusional FU.19,20 Several trials on the combination of irinotecan ± infusional FU with RT in rectal cancer have also been undertaken.21–26 Most of these studies evaluated a dose of 50 mg/m2 irinotecan weekly in combination with protracted infusional FU. Promising pathologic complete remission rates of approximately 20% to 25% were reported.
Combining capecitabine and irinotecan (CAPIRI) would be expected to improve the efficacy of preoperative radiotherapy in terms of local control and prevention of distant metastases via their systemic activity. Despite the combination of CAPIRI having promising activity in metastatic colorectal cancer,27,28 no data on the combination of both drugs with RT have been published to date. Therefore, this phase I study was initiated to establish a combination schedule of both drugs with standard pelvic RT (CAPIRI-RT) as preoperative treatment for locally advanced rectal adenocarcinoma. Irinotecan was given weekly as a short infusion 1 hour before irradiation,18 and capecitabine was administered continuously bid for the duration of RT in a dose-escalating scheme to exploit its radiosensitizing properties.
PATIENTS AND METHODS
The protocol was reviewed and approved by the local institutional review board and the study was performed according to the Declaration of Helsinki as amended in Somerset West in 1996. All patients provided written informed consent before entering onto the trial.
Eligibility Criteria
Patients with histologically confirmed, newly diagnosed, locally advanced nonmetastatic rectal adenocarcinoma (uT3-4, any N) were eligible for entry onto the study. Additional inclusion criteria were as follows: Eastern Cooperative Oncology Group performance score ≤ 2; age ≥ 18 years; adequate bone marrow function (leukocyte count > 3,000/μL, platelet count > 100,000/μL); and adequate renal (serum creatinine ≤ 1.4 mg/dL or creatinine clearance > 60 mL/min) and hepatic function (bilirubin ≤ 2 mg/dL). Patients of childbearing potential were required to be using appropriate contraception. Patients were excluded if they suffered from other cancers, ischemic heart disease, or had known hypersensitivity to FU and/or irinotecan.
Pretreatment Evaluation
Before study admission, all patients underwent a complete history, physical examination, biopsy, digital rectal examination, rectoscopy, transrectal ultrasonography, pelvic and abdominal computed tomographic scans, colonoscopy, and chest x-rays. A CBC with differential and serum chemistry (including electrolytes, creatinine, blood urea nitrogen, uric acid, liver aminotransferases, total bilirubin, alkaline phosphatase, and lactate dehydrogenase) was obtained within 14 days before the start of treatment. Weekly blood counts were obtained and serum chemistry was repeated every third week or whenever clinically indicated.
RT
RT was delivered with a linear accelerator using 23-MeV photons and a three-field box technique consisting of a posterior-anterior and two lateral fields. For three-dimensional treatment planning purposes, all patients had a computed tomographic scan in the treatment position (prone position) using a belly board. A total irradiation dose of 50.4 Gy was given in daily fractions of 1.8 Gy, 5 days a week. The clinical target volume (CTV) according to the International Commission on Radiation Units included the sacrum, the presacral space, and the posterior wall of the bladder and prostate or vagina. The common iliac lymph nodes were included in the CTV. The upper border of the CTV was at the L5-S1 interspace for cN0 and at L4-5 for cN-positive patients. The lower field border was 5 cm below the macroscopic tumor. After a dose of 45 Gy, the small bowel was excluded and an additional dose of 5.4 Gy was given to the boost volume using a shrinking-field technique (days 36 to 38).
Chemotherapy
Patients received weekly irinotecan at a dose of 50 mg/m2 beginning on day 1 of RT for 5 consecutive weeks (days 1, 8, 15, 22, and 29) one hour before irradiation (Fig 1). Owing to the experience with FU- plus irinotecan-based chemoradiotherapy schedules, irinotecan was given weekly rather than every third week. All patients received standard antiemetic prophylaxis to avoid any bias relating to GI adverse events. Patients received capecitabine orally bid within 30 minutes after a meal (generally after breakfast and evening meal) for the whole time of irradiation (days 1 through 38), which was planned to start at three different dose levels (DLs): DL I (500 mg/m2 bid), DL II (625 mg/m2 bid), and DL III (750 mg/m2 bid). To exploit the radiosensitizing effects of capecitabine, a continuous-application schedule (ie, including weekends), starting in the morning of day 1, was chosen.
Preoperative concomitant chemoradiotherapy with capecitabine and irinotecan in patients with locally advanced rectal cancer.
If a patient reported one of the following adverse events according to National Cancer Institute Common Toxicity Criteria (version 2.0) on the day of planned irinotecan application, chemotherapy was to be interrupted until adverse events had resolved to grade 0 or 1: leucopenia ≥ grade 2, thrombocytopenia ≥ grade 1, diarrhea ≥ grade 1, mucositis or stomatitis ≥ grade 1, skin adverse events ≥ grade 2, other adverse events more than grade 2. Treatment was resumed with a dose reduction of 25% for either drug. In the event of adverse events ≥ grade 2 in the interval between two irinotecan applications, we suggested interruption of chemotherapy and reduction of both drugs to 75% of the starting dose. On the second occurrence of an adverse event ≥ grade 2, the capecitabine dose was to be reduced to 50% of the starting dose.
Every effort was made to continue daily RT without interruption, unless the patient experienced a persistent or severe adverse event. The RT schedule was not modified or interrupted unless an adverse event (such as a skin adverse event ≥ grade 2) was clearly related to radiation. In these occurrences, RT was withheld until adverse events resolved to grade 0 or 1.
Surgery and Histopathologic Examination of Resected Specimen
Surgical resection (either low anterior or abdominoperineal resection) with TME was performed 4 to 6 weeks after completion of the CAPIRI-RT in all patients. Pathologic examination of the surgical specimen involved opening the specimen, identifying the former tumor-bearing area, and providing a macroscopic description. For obvious residual tumor, at least four paraffin blocks were processed, and an additional large-area block was embedded. If only an ulcer persisted in the region of the tumor bed, the entire former tumor-bearing area was sliced and embedded. The completion of TME, lateral margins, proximal and distal tumor-free margins, tumor mass, fibrotic changes, and irradiation vasculopathy were evaluated and semiquantitatively described where feasible.
Study Design, Definitions, and End Points
The primary objective of this phase I study was to determine the dose-limiting toxicities (DLTs) and the maximum-tolerated dose (MTD) of continuously applied capecitabine in combination with weekly irinotecan and standard pelvic RT for rectal adenocarcinoma. At least three patients were enrolled per dose level, with this number increased to six if DLTs occurred in more than one patient. Dose escalation was halted if DLT occurred in two or more patients. The MTD was defined as the highest dose at which fewer than two of six patients experienced DLTs during the first course of chemotherapy. Escalation beyond DL III was not foreseen and intraindividual dose escalation was not permitted. Adverse events were monitored weekly during treatment. DLT was defined by the occurrence of one of the following adverse events: grade 4 leucopenia/neutropenia or thrombocytopenia, symptomatic thrombocytopenia (hemorrhage), grade 3 or 4 febrile neutropenia, or any ≥ grade 3 nonhematologic adverse events except renal adverse events (≥ grade 2) and nausea or vomiting. Secondary end points included preliminary assessment of antitumor activity (pathologic downstaging rate) and surgical complication rate.
RESULTS
Nineteen patients entered onto this single-center trial between June 2002 and August 2003. Patient characteristics are listed in Table 1. All patients had newly diagnosed rectal adenocarcinoma. The median distance from the lower border of the primary to the anal verge was 5 cm.
Patient and Tumor Characteristics
Dose Escalation and DLTs
Three patients were treated at DL I (capecitabine 500 mg/m2 bid). In a 62-year-old male, therapy had to be discontinued at the end of week 4 after the patient received 39.6 Gy because of grade 3 asthenia in combination with grade 2 anorexia. Chemotherapy was omitted and RT was stopped. The patient recovered within 2 weeks. Although the other two patients treated on DL I had no DLTs, we accrued an additional three patients at this DL. In these patients, no DLTs were recorded and dose escalation progressed to level II (capecitabine 625 mg/m2 bid). Of seven patients treated with this dose, three developed grade 3 diarrhea, which was accompanied by grade 3 vomiting in a 78-year-old female. This patient had to be hospitalized and required IV fluids. All three patients recovered within 10 days without sequelae. Consequently, capecitabine dose escalation stopped at a dose of 625 mg/m2 bid and six additional patients were treated at DL I. No DLTs were observed in these patients and, therefore, capecitabine 500 mg/m2 bid was determined as the MTD.
Safety and Dose Intensity
Table 2 lists all adverse events separately for the 12 patients treated at DL I and the seven patients treated on DL II. No grade 4 adverse events were observed and only one patient was hospitalized for treatment-related adverse events. Hematologic adverse events were generally mild at both dose levels, with only one patient developing grade 3 leucopenia at DL II lasting for 2 weeks. At DL I, seven of 12 patients (58%) had grade 1 or 2 leucopenia lasting for a maximum of 7 days in all but one patient. Granulocyte colony-stimulating factors were not administered prophylactically during the study. Grade 1 or 2 anemia was observed frequently. The mean decrease of hemoglobin during CAPIRI-RT therapy was 1.6 g/dL (range, 0.5 to 3.4 g/dL). No prophylactic measures (eg, recombinant human epoetin) were in place and no blood transfusions were required during chemoradiotherapy.
Incidence and Severity of Adverse Events According to Capecitabine Dose Levels
Gastrointestinal adverse events were observed more frequently. Three of seven patients at DL II suffered from grade 3 diarrhea, which was accompanied by grade 3 nausea or vomiting in one patient. Diarrhea was considered a DLT. Gastrointestinal adverse events at the recommended dose level (DL I) comprised grade 1 or 2 diarrhea in 75% and nausea or vomiting in approximately 50% of the patients. No grade 3 GI adverse events were seen in patients treated at DL I. Apart from one patient experiencing grade 3 asthenia in combination with grade 2 anorexia, which led to the withdrawal from chemoradiotherapy, additional grade 3 adverse events at the recommended dose level were not observed. Hand-foot syndrome was restricted to one patient. Of note was a mild hepatic event with elevated transaminases in four of seven patients treated at the capecitabine 625 mg/m2 bid dose level. At DL I, no elevation of transaminases was noted.
The median dose-intensity of both irinotecan and capecitabine delivered to patients treated at capecitabine 500 mg/m2 bid was 100% (mean, 94% and 95%, respectively). Only three patients required dose adjustment and received less than 95% of the assigned doses of irinotecan and capecitabine. Apart from the earlier mentioned patient, RT was delivered as scheduled in all patients.
Surgery and Pathologic Features of the Resected Specimen
Radical resection was scheduled and carried out 4 to 6 weeks after CAPIRI-RT treatment in all 19 patients. Of these, 15 were treated with low anterior and four with abdominoperineal resection. The patients without sphincter-sparing surgery had tumors between 2 and 4 cm from the anal verge. Fourteen patients were lymph node-negative despite 16 having positive lymph nodes at baseline. When the ultrasonographic stage was compared with the pathologic stage, nodal status downstaging (uN-positive to ypN0) was detected in 12 (75%) of 16 patients previously classified as uN-positive. T category was downstaged in 13 of 19 patients (Table 3). A pathologic complete regression of the tumor (pCR; ypT0, N0) was found in four (21%), and microfoci (few tumor cells scattered within fibrotic tissue) were found in another five (26%) of 19 patients.
Pathologic Staging Compared With Clinical Staging at Baseline
There were no intra- or postoperative deaths. The median hospital stay was 14 days and the median time to complete oral food intake was 8 days. Surgical procedures lasted a median of 230 minutes (range, 192 to 308 minutes) and resulted in a median blood loss of 700 mL (range, 200 to 1,900 mL). Nine patients (47%) experienced postoperative complications including wound dehiscence (n = 2), bowel atonia (n = 4), bladder dysfunction (n = 1), recto-vaginal fistula in need of revision (n = 1), and uncomplicated bladder tamponade due to catheter-associated bleeding (n = 1). In one patient a presacral abscess required percutaneous drainage. One male patient developed a major complication and suffered from anastomotic insufficiency and presacral abscess in need of surgical revision. Complicated secondary wound healing required additional surgical revision in this and in another male patient.
DISCUSSION
Combined-modality treatment with FU and pelvic RT is one of the cornerstones of managing patients with locally advanced rectal cancer. It is commonly used postoperatively but theoretical and practical advantages led to a growing interest in the administration of preoperative chemoradiotherapy. Although preoperative chemoradiotherapy resulted in an improved local control rate in comparison with standard adjuvant chemoradiotherapy, no differences in the rates of distant metastases were observed in the German Intergroup trial.4 This highlighted the need for the implementation of more effective chemotherapeutic drugs for application at an earlier stage during treatment to eradicate micrometastases.
When used as first-line treatment in metastatic colorectal cancer, the oral fluoropyrimidine capecitabine produces a higher response rate than bolus FU/FA.9 The improvement in safety has been demonstrated in the metastatic and adjuvant colon cancer setting.10,11 Following the trend for replacing FU with capecitabine in the metastatic setting, capecitabine is also beginning to replace infusional FU as a radiosensitizer during chemoradiation for rectal cancer. Importantly, capecitabine has similar pharmacokinetic properties and avoids the need for inconvenient and troublesome IV catheters. Capecitabine is activated by the enzyme thymidine phosphorylase, which is upregulated by irradiation and thus improves the likelihood of capecitabine-mediated radiosensitizing. Capecitabine in combination with radiotherapy produced tumor downstaging when used as preoperative therapy in patients with locally advanced rectal cancer.13,14,29 In a study by Kim et al,29 45 patients with locally advanced rectal cancer received preoperative conventionally fractionated radiotherapy (50.4 Gy) and capecitabine (825 mg/m2 bid, days 1 to 14 and 29 to 42) in combination with FA (20 mg/m2 days 1 to 14 and 29 to 42). Of note, only 4% patients experienced grade 3 or 4 diarrhea. Of the resected patients (n = 38), 31% demonstrated pCR of the tumor. Recently, Dunst et al14 reported a multicenter phase II trial on patients with locally advanced rectal cancer using radiotherapy plus capecitabine (825 mg/m2 bid, from the first to last day of radiotherapy as determined in the previous phase I trial). In accordance with the data reported by Kim et al,29 only 4% of the patients suffered grade 3 or 4 diarrhea. In contrast, in only two of 50 patients was pCR observed.14
Irinotecan, a topoisomerase I inhibitor, is effective as a single agent as well as in combination with FU or capecitabine in metastatic colorectal cancer, both as first- and second-line therapy.27,28,30,31 In addition, preclinical and clinical studies have shown the effectiveness of irinotecan as radiosensitizer.17 Phase II clinical trial data indicate that combinations of capecitabine and either weekly or every third week irinotecan are feasible, effective, and well-tolerated options for the first-line treatment of patients with metastatic colorectal cancer.27,28 Consequently, there is a strong rationale for testing the CAPIRI combination in conjunction with preoperative RT in patients with locally advanced rectal cancer.
This study demonstrated the feasibility of combining capecitabine and irinotecan with pelvic RT for locally advanced rectal cancer. Our schedule allows for the continuous exploitation of the radiosensitizing effects of capecitabine by its daily administration. Irinotecan was scheduled weekly 1 hour before irradiation on the basis of preclinical studies, which showed maximum radiosensitizing when irinotecan was administered 1 hour before irradiation.18
The more frequent administration of irinotecan and the continuous application of capecitabine bear a higher risk for the development of diarrhea. Nonetheless, no patient treated at the recommended dose level of capecitabine 1,000 mg/m2 developed grade 3 or 4 diarrhea, whereas three of seven patients treated at DL II were compromised by diarrhea grade 3. Therefore, despite the limited number of patients treated so far, we believe that DL I may be considered as safe for additional testing in phase II trials. Hematologic toxicities were rather uncommon, and hand-foot syndrome was seen in only one patient. Overall, our findings compare favorably with those reported for FU, irinotecan, and RT combinations in locally advanced rectal carcinoma.21–26
Nevertheless, the most important concern in using topoisomerase inhibitors with or without fluoropyrimidines in combination with RT is the development of severe diarrhea. There is preclinical evidence from human tumor xenograft models that jejunal toxicity is enhanced by the addition of 9-amino-20(S)-camptothecin (a topoisomerase I inhibitor) to RT, owing to a decreased survival of jejunal crypt cells.32 That the intestine is the critical target tissue becomes evident from several clinical trials on concomitant irinotecan, FU, and radiotherapy. Studies conducted at the Thomas Jefferson University (Philadelphia, PA) and at Stanford University (Stanford, CA) using infusional FU (225 to 300 mg/m2) throughout radiotherapy and weekly irinotecan (30 to 60 mg/m2) reported severe diarrhea in 27% and 28% of the patients, respectively.21,24
Conversely, in a randomized phase II trial presented at the Annual Meeting of the American Society of Clinical Oncology in 2004, preoperative combined chemoradiotherapy using infusional FU plus irinotecan did not result in a significantly higher rate of GI toxicity in comparison with FU alone.33 Patients in this trial received either infusional FU (225 mg/m2/d as continuous infusion until completion of RT) or infusional FU (225 mg/m2 for 120 hours of infusion per week until completion of RT) plus irinotecan (50 mg/m2 once weekly for 4 weeks). In the FU plus RT group, 13 (28%) of 47 patients suffered from grade 3 to 4 GI toxicity, whereas the FU plus irinotecan plus RT treatment resulted in severe GI toxicity in 19 (37%) of 52 patients.
To date, it is unknown if the narrow therapeutic window might be opened by using oxaliplatin instead of irinotecan for chemoradiotherapy in combination with FU or capecitabine. Although Rödel et al15 observed a rate of severe diarrhea in only 8% of patients treated with chemoradiotherapy using capecitabine (825 mg/m2 bid days 1 to 14 and 22 to 35) and oxaliplatin (50 mg/m2 days 1, 8, 22, 29), Duck et al34 more recently reported a 22% diarrhea rate using a different treatment schedule: capecitabine (825 mg/m2 bid, days 1 to 5, repeated day 7) and oxaliplatin (50 mg/m2 weekly for 5 weeks). Randomized studies are awaited to define the most suitable cytostatic drug in combination with fluoropyrimidines during chemoradiotherapy for rectal cancer.
An alternative approach to avoid combined small-bowel toxicity of CAPIRI-RT could be the use of intensity-modulated radiation therapy, which allows a significant reduction of dose to the small bowel.35
Retrospective analyses have shown that achieving a pCR improves outcome.36 Randomized combination studies with FU plus irinotecan in metastatic colorectal cancer have consistently shown significant higher remission rates for the combination chemotherapy.19,20,37 Therefore, it is logical to assume that two-drug combinations may also accomplish higher complete remission rates when used as part of chemoradiotherapy regimens. Indeed, pCR rates of between 19% and 37% have been reported by several study groups using weekly irinotecan (40 to 50 mg/m2) plus daily protracted infusional FU (200 to 300 mg/m2). Although the pCR rates were not the primary end point of the present study, we observed promising preliminary efficacy with complete or almost complete remissions of the primary tumor in about half of our patients. A high rate of lymph node sterilization was observed and perioperative morbidity did not seem to be increased compared with previous findings with FU-based chemoradiotherapy.
On the basis of the current findings, capecitabine at 500 mg/m2 bid in combination with weekly irinotecan seems to be safe and effective. Clearly, larger phase II trials are warranted to assess the real magnitude of the pCR rate of the CAPIRI-RT schedule as well as the tolerability of combination chemotherapy in relation to FU-based single-agent therapy. Although significant improvements in RT and surgical technique have been reported in the last decade, new and highly active agents, such as capecitabine and irinotecan, are needed in the perioperative management of locally advanced rectal cancer.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
Footnotes
-
R-D.H. and B.v.G-H. contributed equally to this study.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
- Received April 29, 2004.
- Accepted September 22, 2004.
