Bevacizumab Assessed in Three Safety Studies of Bleeding, Wound Healing Complications in CRC

September 1, 2004

The 30 reports in this special supplement to Oncology News International represent highlights of ongoing major clinical trials and new research presented at ASCO 2004 regarding state-of-the-art chemotherapeutic management of gastrointestinal and other cancers. Important developments in capecitabine as adjuvant therapy, novel targeted agents, and new combinations are discussed.

SOUTH SAN FRANCISCO-Three related safety studies assessingbleeding and wound healing complicationsin colorectal cancer patientswhose chemotherapy included the recombinanthumanized monoclonalantibody bevacizumab (Avastin) haveyielded some promising results (abstracts3528, 3529, and 3530), accordingto investigators.Investigators found no increasedrisk of venous thromboembolism inpatients with metastatic colorectal cancer,no increase in hemorrhagic complicationsin metastatic colorectal cancerpatients receiving bevacizumabconcurrent with full-dose anticoagulanttherapy during chemotherapy,and no increase in wound healing orbleeding complications when bevacizumabwas given with chemotherapy28 to 60 days following primary surgeryfor colon or rectal cancer.However, an increase in arterialthromboembolism was seen with thedrug. Further, caution must beexercised in colorectal cancer patientsalready on bevacizumab who thenneed to undergo surgery, a fourth studyhas concluded (abstract 3702).Careful Scrutiny
"Since Avastin represents the firstsurvival benefit ever shown with anantiangiogenic drug, we wanted to seeif it's really safe to block angiogenesis,"said Susan Desmond-Hellmann,MD, MPH, president of Product De-velopment, Genentech, and adjunctassociate professor of Epidemiologyand Biostatistics at the University ofCalifornia, San Francisco.Participating in the studies wereresearchers from Genentech and frommultiple institutions, including DukeUniversity Cancer Center, UCLA,Vanderbilt University, Kaiser Perma-nente, US Oncology, Sarah CannonCancer Center, and the Hematology/Oncology Association of NortheastPennsylvania. In discussing the resultsin a poster presentation, the investigatorsconcluded that the findings areencouraging for planned further evaluationof bevacizumab in the treatmentof colorectal cancer.Concurrent Anticoagulation
One of the safety studies (Hambletonet al, abstract 3528) was done toevaluate the effect of bevacizumab(BV) on bleeding complications inpatients with metastatic colorectal cancerreceiving full-dose anticoagulationtherapy (FDAC); owing to the bleedingseen in NSCLC patients in phase IIstudies, patients on FDAC had beenexcluded from a major phase III trialof BV that showed improved survivalwhen it was added to IFL (irinotecan[CPT-11, Camptosar]/5-FU/leucovorin[LV]), with the hazard of deathreduced by 34% vs IFL alone (mediansurvival increased from 15.6 to 20.3months;P < .001).To analyze the outcome of patientswith metastatic colorectal cancer whohad a thrombotic event while on BVor placebo and remained on studyfollowing FDAC, the investigators randomizedpatients with previously untreatedmetastatic colorectal cancer totwo study arms: arm 1 patients (n =411) received IFL plus placebo, withirinotecan 125 mg/m2, plus 5-FU500 mg/m2, and LV 20 mg/m2 onceweekly for 4 weeks, with a placebo IVgiven every 2 weeks; and arm 2 patients(n = 402) received the same IFLregimen but with 5 mg/kg BV intravenouslyevery 2 weeks. Chemotherapywas continued for 96 weeks or untildisease progression. Warfarin was usedas FDAC therapy. Nearly half (43%)of the patients had an ECOG (EasternCooperative Oncology Group) performancestatus greater than 0.Venous Events Comparable
The overall rate of arterial andvenous thrombolic events was about16% in arm 1 (n = 396) and 19% inarm 2 (n = 392). Venous events werecomparable in both arms, at about15% in arm 1 and nearly 17% in arm 2.Overall incidence of grade 3/4bleeding was about 3% in both arms.Concomitant FDAC in patients whohad a thrombotic event did not increasethe incidence of grade 3/4 bleeding,which occurred in 2 of 30 (about7%) patients in arm 1 and in 2 of 53(about 4%) patients in arm 2.Two Patient Populations
Researchers conducted two subsequentrandomized, double-blind,multicenter, placebo-controlled trials,one phase II and the other phaseIII, in patients with metastatic colorectalcancer (Novotny et al, abstract3529).The phase II study patients were amedian of 10 years older than patientsin the phase III study, and 72% had anECOG performance status greaterthan 0, compared with 43% of thephase III study group.Phase II and III Studies
The design and results of the phaseIII trial are as described in abstract3528, with arm 1 (n = 396) patientsreceiving IFL plus placebo, and arm 2patients (n = 392) receiving IFL plusBV. Deep venous thrombophlebitisoccurred in about 6% of patients inarm 1 vs 9% of those in arm 2. Athrombotic event resulted in death infour patients in arm 1 and four patientsin arm 2, with time to first onsetof a thromboembolic event duringfirst-line therapy similar in both arms.Arterial events occurred in about 1%of patients in arm 1 and in 3% of thosein arm 2.Phase II PatientsUnsuited to Irinotecan
The phase II trial tested the efficacyand safety of adding BV to 5-FU/LVchemotherapy in patients with metastaticcolorectal cancer who were unsuitedto receive first-line treatmentwith irinotecan. Arm 1 patients (n =104) and arm 2 patients (n = 100)received the Roswell Park regimen of5-FU/LV weekly for 6 weeks with eitherplacebo or BV, respectively, givenevery 2 weeks). Treatment was continueduntil disease progression, unacceptabletoxicity, or a maximum of 96weeks.The overall incidence of thromboembolicadverse events in this olderpatient population was comparable,at about 18% for both arms. Venousevents occurred in about 14% of patientsin arm 1 vs 9% in arm 2, andarterial events occurred in about 5%of patients in arm 1 vs 10% in arm 2.The investigators concluded that"the high background rate of thromboembolicdisease in patients withmetastatic CRC highlights the needfor well-designed, placebo-controlledclinical trials to assess the safety profileof new oncologic agents in thispatient population."Safety Post Ca Surgery
The third report was a randomized,double-blind, placebo-controlledphase III trial, to assess wound healingand bleeding complications in patientswho had undergone colonic or rectalcancer surgery 28 to 60 days prior toinitial use of BV (Scappaticci et al,abstract 3530).A total of 155 patients who receivedonly bolus IFL plus placebo postsurgerywere compared with 150 patientswho received IFL plus BV (with regimensas described in abstract 3528) vsa third arm of 37 patients who receivedpostsurgery therapy with 5-FU/LV/BV (5-FU 500 mg/m2 IV, LV 20mg/m2 once a week for 6 weeks, withthe cycle repeated every 8 weeks, andBV 5 mg/kg IV every 2 weeks).Incidence of grade 3/4 wound healingor bleeding complications was 1patient (0.65%) receiving only IFL, 3patients (2.0%) receiving bolus IFL/BV, and zero patients in the 5-FU/LV/BV group. All four patients were ableto resume study treatment. "These resultssuggest that initating BV in patientswithin a 28- to 60-day time periodafter surgery does not lead to asignificantly increased incidence ofwound healing/bleeding complications,"the investigators wrote.Surgery While on BV: Caution
Caution is warranted in cases ofsurgery on patients who are alreadyreceiving bevacizumab, noted Genentechresearcher Julie S. Hambleton,MD. One study (Hurwitz et al, abstract3702), she said, shows "4 of 40patients who had surgery while onAvastin [plus bolus IFL] had postoperativewound healing complications[as well as 1 of 15 on BV/5-FU/LV], sowe need to be cautious about takingpatients to the operating room ifthey're actually on an antiangiogenicdrug." Prior to their surgery, Hurwitzet al. reported, the five patients hadbeen off BV for 13 to 89 days, with fouroff therapy for at least 1 month. Inconclusion, they advised that "investigatorsshould consider the risks andbenefits of surgery when subjects arebeing treated with BV."