FDA Executive Summary
Prepared for the
October 13, 2011 meeting of the
Circulatory System Devices Advisory Panel
Zilver® PTX® Drug-Eluting Peripheral Stent System
This is the FDA Executive Summary for a first-of-a-kind drug-eluting peripheral stent, the Cook Zilver PTX Drug-Eluting Stent, indicated for use in the femoropopliteal arteries. This device has been reviewed by the Division of Cardiovascular Devices within the Center for Devices and Radiological Health of the Food and Drug Administration under Premarket Approval (PMA) application P100022, which is the subject of this Advisory Panel meeting.
This memorandum will summarize the FDA’s review of the PMA up to this point, highlighting the particular areas for which we are seeking your expertise and input. These topics will include the proposed indications for use, pre-clinical study findings, the results of the randomized clinical study conducted by the sponsor, and the proposed post-approval study. At the conclusion of your review and discussion of the data presented, the Agency will ask for your recommendation regarding whether or not the data demonstrate a reasonable assurance of safety and effectiveness.
Table of Contents
1. PROPOSED INDICATIONS FOR USE 8
2. DEVICE DESCRIPTION 9
3. REGULATORY HISTORY 11
COOK FIRST BEGAN DISCUSSIONS WITH FDA REGARDING THE EVALUATION OF THE ZILVER PTX DRUG-ELUTING STENT UNDER PRE-IDE I020209 IN JULY 2002. COOK SUBMITTED THE IDE FOR THE PIVOTAL STUDY IN NOVEMBER 2003, AND THE STUDY WAS GRANTED CONDITIONAL APPROVAL TO BEGIN THE FIRST PHASE OF ENROLLMENT IN JUNE 2004. IN THE FIRST PHASE OF THE STUDY, ONLY SUBJECTS WITH LESION LENGTHS < 7 CM WERE ENROLLED, AND SUBJECTS COULD ONLY RECEIVE A SINGLE ZILVER PTX STENT. PHASE I ENROLLMENT WAS COMPLETED IN FEBRUARY 2006 AND INCLUDED 60 SUBJECTS. 11 COOK RECEIVED FULL APPROVAL OF THE IDE IN AUGUST 2007, AND STARTED PHASE 2 OF THE STUDY. PHASE 2 CONSISTED OF PATIENTS WITH LESION LENGTH 0 – 14 CM. ENROLLMENT WAS COMPLETED IN AUGUST 2008. LONG-TERM FOLLOW-UP IS ONGOING. 11 IN ADDITION TO THE PIVOTAL STUDY, FDA REQUESTED THAT COOK DEVELOP A PLAN TO EVALUATE THE SAFETY OF THE ZILVER PTX IN ADDITIONAL SUBJECTS, AS THE SIZE OF THE PIVOTAL STUDY MIGHT NOT BE SUFFICIENT TO ADEQUATELY DETECT AND EVALUATE RARE ADVERSE EVENTS RELATED TO THE STENT OR ITS DRUG COATING. IN RESPONSE, COOK CONDUCTED A GLOBAL REGISTRY INTENDED TO FURTHER ASSESS THE SAFETY OF THE PRODUCT. BECAUSE THE REGISTRY WAS CONDUCTED COMPLETELY OUTSIDE THE UNITED STATES, FDA REVIEW AND APPROVAL OF AN IDE WAS NOT NECESSARY AND FDA DID NOT REVIEW THE PROTOCOL FOR THE REGISTRY PRIOR TO INITIATION OF ENROLLMENT. ENROLLMENT IN THE REGISTRY, CONSISTING OF 787 SUBJECTS FROM EUROPE, CANADA, AND SOUTH KOREA, TOOK PLACE FROM APRIL 2006 – JUNE 2008. FOLLOW-UP ON THE REGISTRY SUBJECTS IS COMPLETE. 11 COOK SUBMITTED THE FIRST MODULE OF THEIR PMA (M090010) IN APRIL 2009. THE FINAL PMA MODULE (P100022) WAS SUBMITTED IN JUNE 2010. IN AUGUST 2010, FDA ISSUED A MAJOR DEFICIENCY LETTER TO THE SPONSOR THAT REQUESTED ADDITIONAL INFORMATION REGARDING CLINICAL AND NON-CLINICAL DATA. COOK’S FORMAL RESPONSE WAS RECEIVED IN JUNE 2011. 12 4. PRE-CLINICAL STUDIES 12 4.1. IN VITRO TESTING 12 4.1.1 STENT FUNCTIONAL TESTING 12 FDA RECOMMENDS STENT FUNCTIONAL TESTING AS OUTLINED IN OUR GUIDANCE DOCUMENT “NON-CLINICAL TESTS AND RECOMMENDED LABELING FOR INTRAVASCULAR STENTS AND ASSOCIATED DELIVERY SYSTEMS.” THE PURPOSE OF THIS TESTING IS TO ENSURE THAT THE STENT PLATFORM (I.E., METAL STENT) FUNCTIONS AS INTENDED. THE FOLLOWING TESTS WERE PERFORMED ON THE ZILVER PTX STENT AND RESULTS WERE SUBMITTED FOR REVIEW: 12 O MATERIAL ANALYSIS OF NITINOL TUBING 12 O STENT CORROSION RESISTANCE 12 O DIMENSIONAL VERIFICATION 12 O PERCENT SURFACE AREA 12 O UNIFORMITY OF EXPANSION 12 O FORESHORTENING 12 O RADIAL (HOOP) STRENGTH 12 O STENT INTEGRITY 12 O MECHANICAL PROPERTIES 12 O FLEX/KINK EVALUATION 12 O CRUSH RESISTANCE 12 O STRESS ANALYSIS 12 O FINITE ELEMENT ANALYSIS (FEA) 12 O FATIGUE/DURABILITY TESTING 12 O MRI SAFETY 12 O RADIOPACITY 12 THE REVIEW TEAM HAS COMPLETED ITS REVIEW OF THE STENT FUNCTIONAL TESTING AND HAS NO OUTSTANDING CONCERNS. 12 4.1.2 STENT COATING TESTING 12 FOR DRUG-ELUTING STENTS, FDA RECOMMENDS ADDITIONAL TESTING OF THE STENT COATING TO ENSURE THAT THE INTEGRITY OF THE COATING WILL BE MAINTAINED OVER THE SIMULATED USE OF THE DEVICE. THE FOLLOWING STENT COATING TESTS WERE PERFORMED ON THE ZILVER PTX STENT AND RESULTS WERE SUBMITTED FOR REVIEW: 12 O COATING THICKNESS AND UNIFORMITY 12 O CONTENT UNIFORMITY ALONG THE STENT LENGTH 12 O MICROSTRUCTURE ANALYSIS 12 O IN VITRO DRUG RELEASE 13 O IN VITRO ELUTION 13 O POTENCY 13 O IMPURITIES/DEGRADANTS 13 O PARTICULATE ANALYSIS 13 THE REVIEW TEAM HAS COMPLETED ITS REVIEW OF THE STENT FUNCTIONAL TESTING AND HAS NO OUTSTANDING CONCERNS RELATED TO BASELINE CHARACTERIZATION OF THE PRODUCT. THERE ARE OUTSTANDING ISSUES RELATED TO THE STABILITY TESTING OF THE DRUG COATING OVER TIME. FDA IS WORKING INTERACTIVELY WITH THE SPONSOR TO ENSURE THAT THESE ISSUES ARE RESOLVED IN A TIMELY FASHION. 13 4.1.3 STENT DELIVERY SYSTEM TESTING 13 FDA RECOMMENDS DELIVERY SYSTEM FUNCTIONAL TESTING AS OUTLINED IN OUR GUIDANCE DOCUMENT “NON-CLINICAL TESTS AND RECOMMENDED LABELING FOR INTRAVASCULAR STENTS AND ASSOCIATED DELIVERY SYSTEMS.” THE PURPOSE OF THIS TESTING IS TO ENSURE THAT THE STENT DELIVERY SYSTEMS FUNCTION AS INTENDED. THE FOLLOWING STENT DELIVERY SYSTEM TESTS WERE PERFORMED ON THE ZILVER PTX STENT AND RESULTS WERE SUBMITTED FOR REVIEW: 13 O DELIVERY SYSTEM PROFILE 13 O EASE OF ACCESS 13 O DEPLOYMENT ACCURACY 13 O DEPLOYMENT FORCE 13 O FLEX/KINK EVALUATION 13 O DELIVERY SYSTEM TENSILE STRENGTH 13 THE REVIEW TEAM HAS COMPLETED ITS REVIEW OF THE STENT FUNCTIONAL TESTING AND HAS NO OUTSTANDING CONCERNS. 13 4.1.4 ANIMAL STUDIES 13 COOK HAS CONDUCTED SIX ANIMAL STUDIES OF THE ZILVER PTX STENT IN ORDER TO DEMONSTRATE THE SAFETY OF THE ZILVER PTX DRUG-ELUTING STENT SYSTEM. STUDIES WERE CARRIED OUT IN A PORCINE NON-ATHEROSCLEROTIC MODEL FOR DURATIONS OF 1, 3, AND 6 MONTHS TO DETERMINE THE SAFETY OF THE CLINICAL DOSE OF PACLITAXEL THAT SHOULD BE INCORPORATED INTO THE ZILVER PTX STENT, TO EVALUATE THE DELIVERABILITY AND THE REGIONAL AND SYSTEMIC SAFETY OF THE ZILVER PTX STENT, AND TO DETERMINE THE SHORT- AND LONG-TERM PHARMACOKINETICS OF THE ZILVER PTX STENT. ANIMAL STUDIES WERE ALSO UTILIZED TO ASSESS THE SAFETY OF OVERLAPPING ZILVER PTX STENTS, AND TO ASSESS A MAXIMUM DOSE BASED ON A DETERMINED DRUG SAFETY MARGIN. 13 OVERALL, THE ANIMAL STUDIES HAVE DEMONSTRATED THE NON-CLINICAL SAFETY OF THE ZILVER PTX STENT IN THE ANIMAL MODEL AT MULTIPLE TIME POINTS. THE ZILVER PTX INCORPORATES A DRUG LOADING OF 3 UG/MM2 AND APPEARS TO ELICIT SIMILAR BIOLOGIC RESPONSES AS THE BARE METAL ZILVER STENT WHEN IMPLANTED IN THE PIG MODEL. BECAUSE THE ANIMAL STUDIES WERE LIMITED TO OBTAINING SUFFICIENT OVERDOSE DATA TO SUPPORT THE SAFETY OF MAXIMUM OF TWO OVERLAPPED 80 MM ZILVER PTX STENTS PER LIMB WITH A MAXIMUM STENTED LENGTH OF 280 MM PER PATIENT, THE REVIEW TEAM BELIEVES THAT CONCLUSIONS REGARDING THE NON-CLINICAL SAFETY OF THE ZILVER PTX STENT CANNOT BE DRAWN FOR LESION LENGTH GREATER THAN 140 MM PER LIMB. THE STENTED LENGTHS IN THE ANIMAL STUDIES ARE CONSISTENT WITH THE LESION LENGTHS SPECIFIED FOR TREATMENT IN THE PIVOTAL CLINICAL TRIAL PROTOCOL. 13 4.1.5 CHEMISTRY, MANUFACTURING, & CONTROLS (CMC) 14 THE DRUG SUBSTANCE, PACLITAXEL, IS PURCHASED FROM YUNNAN HANDE BIO-TECH CO., LTD., OF THE PEOPLE’S REPUBLIC OF CHINA. CMC INFORMATION DESCRIBING THE MANUFACTURE AND CONTROLS OF PACLITAXEL DRUG SUBSTANCE IS INCORPORATED IN THE ZILVER PTX PMA BY AUTHORIZED REFERENCE TO DRUG MASTER FILE (DMF) #11909. THE PACLITAXEL DRUG SUBSTANCE MEETS THE QUALITY SPECIFICATIONS WHICH WERE DEVELOPED BASED ON THE US PHARMACOPOEIA (USP) AND EUROPEAN PHARMACOPOEIA (EP). THE CMC INFORMATION FOR THE DRUG SUBSTANCE WAS REVIEWED AND FOUND TO BE ACCEPTABLE. 14 FOR THE FINISHED PRODUCT, COOK HAS PROVIDED THE DETAILS OF THE MANUFACTURING PROCESS, THE QUANTITATIVE COMPOSITION OF THE PRODUCT, AND CONTROLS FOR THE PACLITAXEL DRUG SUBSTANCE USED IN THE MANUFACTURE OF THE ZILVER PTX STENT, INCLUDING THE ANALYTICAL TEST METHODS AND SUPPORTING VALIDATION DATA. 14 COOK HAS ALSO PROPOSED COMPLETE FINISHED PRODUCT SPECIFICATION, INCLUDING THE ANALYTICAL PROCEDURES AND SUPPORTING VALIDATION DATA. THERE ARE OUTSTANDING ISSUES RELATED TO THESE SPECIFICATIONS. FDA IS WORKING INTERACTIVELY WITH THE SPONSOR TO ENSURE THAT THESE ISSUES ARE RESOLVED IN A TIMELY FASHION. 14 COOK IS COLLECTING STABILITY DATA WHICH WILL BE SUBMITTED TO FDA FOR CONSIDERATION OF A YET-TO-BE DETERMINED SHELF LIFE. THE SHELF LIFE CANNOT BE ESTABLISHED UNTIL THE SPECIFICATIONS ARE FINALIZED AND THE SUPPORTING STABILITY DATA ARE REVIEWED. 14 4.1.6 STERILIZATION 14 THE ZILVER PTX STENT AND DELIVERY SYSTEMS ARE PROVIDED STERILE. AFTER REVIEWING THE INFORMATION SUBMITTED BY THE SPONSOR, THE REVIEW TEAM HAS CONCLUDED THAT UNDER THE STATED EXPOSURE CONDITIONS, THE ETHYLENE OXIDE CYCLE WILL RENDER THE ZILVER PTX STENT AND DELIVERY SYSTEMS STERILE AT A STERILITY ASSURANCE LEVEL OF 10-6, OR THE PROBABILITY OF ONE SURVIVOR IN ONE MILLION PRODUCTS STERILIZED. ETHYLENE OXIDE AND ETHYLENE CHLOROHYDRIN RESIDUAL ANALYSIS WAS PERFORMED TO CONFIRM THAT RESIDUAL LEVELS ARE BELOW STATED ACCEPTANCE CRITERIA. PACKAGING STUDIES WERE PERFORMED TO DEMONSTRATE THAT THE CURRENT PACKAGING CONFIGURATION WILL MAINTAIN A STERILE BARRIER TO SUPPORT THE FORTHCOMING SHELF-LIFE CLAIM. 14 4.1.7 BIOCOMPATIBILITY 14 COOK HAS IDENTIFIED ALL MATERIAL COMPONENTS OF THE STENT AND DELIVERY SYSTEMS. 14 COOK HAS PROVIDED DATA FROM THE FOLLOWING BIOCOMPATIBILITY EVALUATIONS OF THE FINISHED DRUG-ELUTING STENT: CYTOTOXICITY, SENSITIZATION, INTRACUTANEOUS REACTIVITY, ACUTE SYSTEMIC TOXICITY, IN VITRO HEMOLYSIS, IN VIVO THROMBOGENICITY, COMPLEMENT ACTIVATION, PLASMA RECALCIFICATION, SUBCHRONIC TOXICITY, MUSCLE IMPLANTATION, MATERIAL-MEDIATED PYROGENICITY, IN VIVO MOUSE BONE MARROW MICRONUCLEUS ASSAY, BACTERIAL REVERSE MUTATION ASSAY, AND CHROMOSOMAL ABERRATION ASSAY. 15 COOK HAS PROVIDED DATA FROM THE FOLLOWING BIOCOMPATIBILITY EVALUATIONS OF THE FINISHED DRUG-ELUTING STENT AND DELIVERY SYSTEM: CYTOTOXICITY, PYROGENICITY, SENSITIZATION, INTRACUTANEOUS REACTIVITY, ACUTE SYSTEMIC TOXICITY, DIRECT COMPLEMENT ACTIVATION, PLASMA RECALCIFICATION, IN VIVO MOUSE BONE MARROW MICRONUCLEUS ASSAY, BACTERIAL REVERSE MUTATION ASSAY, CHROMOSOMAL ABERRATION ASSAY, IN VITRO HEMOLYSIS, AND IN VIVO THROMBOGENICITY. 15 THE REVIEW TEAM HAS REQUESTED CLARIFICATIONS REGARDING THIS TESTING. FDA IS WORKING INTERACTIVELY WITH THE SPONSOR TO ENSURE THAT THESE ISSUES ARE RESOLVED IN A TIMELY FASHION. 15 4.1.8 MANUFACTURING 15 FDA HAS REVIEWED THE MANUFACTURING INFORMATION. REVIEWS OF THE REPORTS FROM THE FACILITY INSPECTIONS ARE NOT YET COMPLETE. 15 4.2. STUDIES OF THE DRUG SUBSTANCE 15 THE ACTIVE PHARMACEUTICAL INGREDIENT IN THE ZILVER PTX STENT IS PACLITAXEL PROVIDED BY YUNNAN HANDE BIO-TECH CO., LTD. PACLITAXEL, A DITERPENOID WITH A TAXANE SKELETON, IS EXTRACTED FROM THE BARK, BRANCHES, OR NEEDLES OF THE YEW TREE, THEN PURIFIED AND CONCENTRATED BY COLUMN CHROMATOGRAPHY, CRYSTALLIZATION AND RECRYSTALLIZATION. THE DRUG SUBSTANCE IS A WHITE TO OFF-WHITE CRYSTALLINE POWDER AND IS HIGHLY LIPOPHILIC, INSOLUBLE IN WATER, AND FREELY SOLUBLE IN: METHANOL, ETHANOL, CHLOROFORM, ETHYL ACETATE AND DIMETHYLSULFOXIDE. PACLITAXEL HAS A MOLECULAR FORMULA OF C47H51NO14. PACLITAXEL STABILIZES THE MICROTUBULE POLYMER AND INHIBITS ITS DISASSEMBLY, LEADING TO MITOTIC BLOCK WITH SUBSEQUENT CELL CYCLE ARREST. PACLITAXEL (TAXOL®) INJECTION IS USED AS AN ANTINEOPLASTIC, AND IS FDA-APPROVED TO TREAT ADVANCED CANCER OF THE OVARIES, BREAST AND LUNG, AND KAPOSI SARCOMA. PACLITAXEL DELIVERED LOCALLY USING DEVICES, INCLUDING SOME CURRENTLY MARKETED CORONARY DRUG-ELUTING STENTS, IS USED AS AN ANTI-PROLIFERATIVE AGENT FOR THE PREVENTION OF ARTERIAL RESTENOSIS. 15 THE SPONSOR HAS PROVIDED A LETTER FROM YUNNAN HANDE BIO-TECH CO., LTD. AUTHORIZING FDA THE RIGHT TO ACCESS THE DMF FOR PACLITAXEL IN SUPPORT OF THIS PMA APPLICATION. 15 4.2.1 SAFETY PHARMACOLOGY 15 REVIEW OF THESE DATA DID NOT INDICATE ANY SAFETY CONCERNS AND SUPPORTED INITIATION OF THE HUMAN CLINICAL STUDIES OF THE ZILVER PTX STENT. 15 4.2.2 TOXICOLOGY 15 TOXICOLOGY INFORMATION ON PACLITAXEL WAS INCORPORATED BY REFERENCE TO THE MANUFACTURER’S DMF. FDA HAS REVIEWED THIS INFORMATION AND THERE ARE NO REMAINING CONCERNS. 15 BASED ON THIS INFORMATION, THE LABELING PROPOSED BY THE APPLICANT CONTAINS THE FOLLOWING CONTRAINDICATION: 15 “WOMEN WHO ARE PREGNANT, BREASTFEEDING, OR PLAN TO BECOME PREGNANT IN THE NEXT 5 YEARS SHOULD NOT RECEIVE A ZILVER PTX DRUG-ELUTING PERIPHERAL STENT. IT IS UNKNOWN WHETHER PACLITAXEL WILL BE EXCRETED IN HUMAN MILK, AND THERE IS A POTENTIAL FOR ADVERSE REACTION IN NURSING INFANTS FROM PACLITAXEL EXPOSURE.” 15 5. ZILVER PTX RANDOMIZED, CONTROLLED TRIAL 15 5.1. CLINICAL EVALUATION STRATEGY 15 5.2. ZILVER PTX RANDOMIZED CONTROLLED TRIAL (IDE G030251) 15 5.3. PATIENT SELECTION PROCESS AND ENROLLMENT CRITERIA 15 5.4. KEY STUDY DEFINITIONS 15 O TYPE A FEMOROPOPLITEAL LESIONS: SINGLE STENOSES UP TO 3 CM IN LENGTH, NOT AT THE ORIGIN OF THE SUPERFICIAL FEMORAL ARTERY (SFA) OR THE DISTAL POPLITEAL ARTERY. 15 O TYPE B FEMOROPOPLITEAL LESIONS: SINGLE STENOSES OR OCCLUSIONS 3-5 CM LONG, NOT INVOLVING THE DISTAL POPLITEAL ARTERY; HEAVILY CALCIFIED STENOSES UP TO 3 CM IN LENGTH; MULTIPLE LESIONS, EACH LESS THAN 3 CM (STENOSIS OR OCCLUSIONS); SINGLE OR MULTIPLE LESIONS IN THE ABSENCE OF CONTINUOUS TIBIAL RUNOFF TO IMPROVE INFLOW FOR DISTAL SURGICAL BYPASS. 15 O TYPE C FEMOROPOPLITEAL LESIONS: SINGLE STENOSES OR OCCLUSIONS LONGER THAN 5 CM; MULTIPLE STENOSES OR OCCLUSIONS, EACH 3-5 CM, WITH OR WITHOUT HEAVY CALCIFICATION. 15 O TYPE D FEMOROPOPLITEAL LESIONS: COMPLETE COMMON FEMORAL ARTERY OR SFA OCCLUSIONS OR COMPLETE POPLITEAL AND PROXIMAL TRIFURCATION OCCLUSIONS. 15 O GRADE A: SMALL RADIOLUCENT AREA WITHIN THE LUMEN OF THE VESSEL DISAPPEARING WITH THE PASSAGE OF THE CONTRAST MATERIAL. 15 O GRADE B: FILLING DEFECT PARALLEL TO THE LUMEN OF THE VESSEL DISAPPEARING WITH THE PASSAGE OF CONTRAST MATERIAL. 15 O GRADE C: DISSECTION PROTRUDING OUTSIDE THE LUMEN OF THE VESSEL PERSISTING AFTER PASSAGE OF CONTRAST MATERIAL. 15 O GRADE D: SPIRAL SHAPED FILLING DEFECT WITH DELAYED RUNOFF OF THE CONTRAST MATERIAL IN THE DISTAL VESSEL. 15 O GRADE E: PERSISTENT LUMINAL FILLING DEFECT WITH DELAYED ANTEROGRADE FLOW. 15 O GRADE F: FILLING DEFECT ACCOMPANIED BY TOTAL OCCLUSION. 15 O CLASS I: THE PATIENT HAS CARDIAC DISEASE BUT WITHOUT RESULTING LIMITATIONS OF ORDINARY PHYSICAL ACTIVITY. ORDINARY PHYSICAL ACTIVITY (I.E., WALKING SEVERAL BLOCKS OR CLIMBING STAIRS) DOES NOT CAUSE UNDUE FATIGUE, PALPITATION, DYSPNEA, OR ANGINAL PAIN. LIMITING SYMPTOMS MAY OCCUR WITH MARKED EXERTION. 15 O CLASS II: THE PATIENT HAS CARDIAC DISEASE RESULTING IN SLIGHT LIMITATION OF ORDINARY PHYSICAL ACTIVITY. PATIENT IS COMFORTABLE AT REST. ORDINARY PHYSICAL ACTIVITY SUCH AS WALKING MORE THAN TWO BLOCKS OR CLIMBING MORE THAN ONE FLIGHT OF STAIRS RESULTS IN LIMITING SYMPTOMS (E.G., FATIGUE, PALPITATION, DYSPNEA, OR ANGINAL PAIN). 15 O CLASS III: THE PATIENT HAS CARDIAC DISEASE RESULTING IN MARKED LIMITATION OF PHYSICAL ACTIVITY. PATIENT IS COMFORTABLE AT REST. LESS THAN ORDINARY PHYSICAL ACTIVITY (I.E., WALKING 1-2 BLOCKS OR CLIMBING ONE FLIGHT OF STAIRS) CAUSES FATIGUE, PALPITATION, DYSPNEA, OR ANGINAL PAIN. 15 O CLASS IV: THE PATIENT HAS DYSPNEA AT REST THAT INCREASES WITH ANY PHYSICAL ACTIVITY. PATIENT HAS CARDIAC DISEASE RESULTING IN INABILITY TO PERFORM ANY PHYSICAL ACTIVITY WITHOUT DISCOMFORT. SYMPTOMS OF CARDIAC INSUFFICIENCY OR ANGINAL SYNDROME MAY BE PRESENT EVEN AT REST. IF ANY PHYSICAL ACTIVITY IS UNDERTAKEN, DISCOMFORT IS INCREASED. 15 O CLASS 0: ASYMPTOMATIC, NO HEMODYNAMICALLY SIGNIFICANT OCCLUSIVE DISEASE. 15 O CLASS 1: MILD CLAUDICATION. 15 O CLASS 2: MODERATE CLAUDICATION. 15 O CLASS 3: SEVERE CLAUDICATION. 15 O CLASS 4: ISCHEMIC REST PAIN. 15 O CLASS 5: MINOR TISSUE LOSS. 15 O CLASS 6: MAJOR TISSUE LOSS. 15 O TYPE 0: NO STRUT FRACTURES. 15 O TYPE I: SINGLE STRUT FRACTURE ONLY. 15 O TYPE II: MULTIPLE SINGLE STRUT FRACTURES THAT CAN OCCUR AT DIFFERENT SITES. 15 O TYPE III: MULTIPLE STRUT FRACTURES RESULTING IN COMPLETE TRANSECTION OF THE STENT, WITHOUT DISPLACEMENT OF THE STENT SEGMENTS. 15 O TYPE IV: MULTIPLE STRUT FRACTURES RESULTING IN DISPLACEMENT OF SEGMENTS OF THE STENT (INCLUDES SPIRAL FRACTURES THAT COULD RESULT IN STENT DISPLACEMENT WITHOUT COMPLETE TRANSACTION). 15 5.5. STUDY ENDPOINTS, HYPOTHESES, AND SUCCESS CRITERIA 15 O = KAPLAN-MEIER ESTIMATED EVENT-FREE SURVIVAL AT 12 MONTHS POST-PROCEDURE FOR THE PTA PER-PROTOCOL COHORT 15 O = KAPLAN-MEIER ESTIMATED EVENT-FREE SURVIVAL AT 12 MONTHS POST-PROCEDURE FOR THE ZILVER PTX PER-PROTOCOL COHORT 15 O = MINIMUM DIFFERENCE OF PRACTICAL INTEREST (IN THIS CASE 10%) 15 O SE = STANDARD ERROR = , WHERE AND ARE THE ESTIMATES OF THE VARIANCES OF THE KAPLAN-MEIER ESTIMATES USING THE GREENWOOD FORMULA. 15 O PROCEDURAL SUCCESS 15 O 12-MONTH CLINICAL SUCCESS BASED ON AN IMPROVEMENT OF 2 RUTHERFORD CATEGORIES 15 O 12-MONTH CLINICAL IMPROVEMENT BASED ON AN IMPROVEMENT OF 1 RUTHERFORD CATEGORY 15 O 12-MONTH SECONDARY PATENCY 15 O 12-MONTH RESTENOSIS RATE 15 O 12-MONTH THROMBOSIS RATE 15 O 12-MONTH FUNCTIONAL STATUS IMPROVEMENT AS MEASURED BY ABI AND THE WALKING IMPAIRMENT QUESTIONNAIRE 15 5.6. SUBJECT DISPOSITION AND DATA ANALYSIS SETS 15 5.7. SUBJECT DEMOGRAPHICS AND OTHER BASELINE CHARACTERISTICS 15 5.8. SAFETY RESULTS 15 5.8.1 PRIMARY SAFETY ENDPOINT 15 5.8.2 ADVERSE EVENTS 15 5.8.3 STENT THROMBOSIS 15 5.9. EFFECTIVENESS RESULTS 15 5.9.1 PRIMARY EFFECTIVENESS ENDPOINT ANALYSIS 15 5.9.2 PRIMARY PATENCY OF PROVISIONAL BARE ZILVER VS. PROVISIONAL ZILVER PTX 15 5.9.3 PRIMARY PATENCY OF OPTIMAL PTA + PROVISIONAL BARE METAL ZILVER VS. PRIMARY ZILVER PTX 15 5.9.4 OPTIMAL PTA + PROVISIONAL BARE ZILVER + PROVISIONAL ZILVER PTX VS. PRIMARY ZILVER PTX 15 5.9.5 SECONDARY EFFECTIVENESS ENDPOINTS 15 5.10. SEX/GENDER ANALYSIS 15 5.11. SUBSTUDIES 15 KEY SUBSTUDIES IN THE RCT INCLUDED ANALYSES OF DRUG PHARMACOKINETICS, ANGIOGRAPHIC/IVUS FOLLOW-UP, AND STENT FRACTURES. 15 5.11.1 PHARMACOKINETIC SUBSTUDY 15 5.11.2 ANGIOGRAPHIC/IVUS SUBSTUDY 15 5.11.3 STENT INTEGRITY 15 TABLE 14: STENT FRACTURE INCIDENCE 15 IMAGING TIMEPOINT 15 FRACTURE RATE, % (N/N STENTS) 15 PRE-DISCHARGE 15 0.0% (0/528) 15 12 MONTHS 15 0.9 % (4/457) 15 5.12. SUMMARY OF RCT RESULTS 15 THE ZILVER PTX RCT MET ITS PRE-SPECIFIED PRIMARY SAFETY AND EFFECTIVENESS ENDPOINTS OF EVENT-FREE SURVIVAL AND PRIMARY PATENCY, RESPECTIVELY. THE RESULTS FOR THE CONTROL ARM WERE CONSISTENT WITH THOSE TYPICALLY REPORTED FOR PTA PROCEDURES IN SYMPTOMATIC FEMOROPOPLITEAL ARTERIES. THERE WERE NO SIGNALS REGARDING ELEVATED RATES OF DEATH, STENT THROMBOSIS, OR STENT FRACTURE. HOWEVER, DUE TO THE RCT SAMPLE SIZE OF 479 SUBJECTS, LIMITED CONCLUSIONS CAN BE DRAWN REGARDING THE DETECTION OF RARE ADVERSE EVENTS. 15 6. GLOBAL REGISTRY 15 6.1. STUDY ENDPOINTS 15 6.2. SUBJECT DISPOSITION 15 6.3. DEMOGRAPHICS AND OTHER BASELINE CHARACTERISTICS 15 6.4. RESULTS 15 6.4.1 EVENT-FREE SURVIVAL RATE 15 6.4.2 PRIMARY PATENCY 15 6.4.3 PROCEDURAL SUCCESS 15 OF THE 900 LESIONS TREATED, 97.5% (845/867) HAD < 30% RESIDUAL STENOSIS AT THE END OF THE PROCEDURE AND WERE THEREFORE PROCEDURAL SUCCESSES. 15 6.4.4 RUTHERFORD CLASSIFICATION 15 6.4.5 FUNCTIONAL STATUS 15 6.4.6 STENT INTEGRITY 15 7. SUMMARY OF CLINICAL EVIDENCE 15 7.1. PRIMARY PATENCY AND EVENT-FREE SURVIVAL AS A FUNCTION OF LESION LENGTH 15 7.2. STENT INTEGRITY 15 7.3. PARTICULATE EFFECTS 15 7.4. NEUTROPENIA 15 7.5. HYPERSENSITIVITY 15 7.6. STENT THROMBOSIS 15 8. FDA PERSPECTIVE REGARDING THE PRE-MARKET DATA 15 8.1. PROPOSED INDICATIONS FOR USE 15 8.2. STENT THROMBOSIS 15 8.3. BAILOUT STENTING 15 8.4. LONG-TERM FOLLOW-UP AVAILABILITY AND RESULTS 15 WHILE THE PRIMARY ENDPOINTS OF THE RCT WERE MEASURED AT ONE YEAR OF FOLLOW-UP, THE RCT PROTOCOL SPECIFIED FOLLOW-UP OUT TO FIVE YEARS AND AT THE TIME OF PMA SUBMISSION. ALL SUBJECTS WERE ELIGIBLE FOR THEIR TWO-YEAR FOLLOW-UP VISIT AND SUCH DATA WERE AVAILABLE FOR APPROXIMATELY 84 % OF SUBJECTS. THE DATA SUGGEST A SUSTAINED PRIMARY PATENCY BENEFIT FOR ZILVER PTX AS COMPARED TO PTA, BUT GIVEN THE PAUCITY OF HIGH-LEVEL DATA REGARDING THE LONG-TERM EFFECTS OF DRUG-ELUTING STENTS IN THE FEMOROPOPLITEAL ARTERIES, IT IS NOT CLEAR WHAT FOLLOW-UP DURATION IS ADEQUATE TO FULLY ASSESS THE SAFETY AND DURABILITY OF ZILVER PTX IMPLANTATION. 15 8.5. STENT FRACTURE 15 HIGH-RESOLUTION FLAT-PLATE X-RAYS WERE TAKEN TO DETECT STENT FRACTURE AT VARIOUS TIME POINTS FROM DISCHARGE TO 3 YEARS POST-PROCEDURE, DEPENDING ON THE STUDY. THE OVERALL RATE OF DETECTED STENT FRACTURES ACROSS BOTH THE RCT AND GLOBAL REGISTRY WAS 1.4%. HOWEVER, THE CLINICAL SIGNIFICANCE OF FEMOROPOPLITEAL STENT FRACTURES IS STILL BEING EXPLORED, AS IS IDENTIFICATION OF THE OPTIMAL TIME POINTS FOR FRACTURE ASSESSMENT. 15 9. POST-APPROVAL STUDY 15 9.1. OVERVIEW OF SPONSOR’S PROPOSED POST-APPROVAL STUDY 15 9.1.1 STUDY DESIGN DESCRIPTION AND HYPOTHESIS 15 O EFSPTA = KAPLAN-MEIER ESTIMATED EVENT-FREE SURVIVAL AT 60 MONTHS POST-PROCEDURE FOR THE PTA PER-PROTOCOL COHORT 15 O EFSPTX = KAPLAN-MEIER ESTIMATED EVENT-FREE SURVIVAL AT 60 MONTHS POST-PROCEDURE FOR THE ZILVER PTX PER-PROTOCOL COHORT 15 O Δ = MINIMUM DIFFERENCE OF PRACTICAL INTEREST (IN THIS CASE 10%) 15 O SE = STANDARD ERROR , WHERE V(EFSPTX) AND V(EFSPTA) ARE THE ESTIMATES OF THE VARIANCES OF THE KAPLAN-MEIER ESTIMATES USING THE GREENWOOD FORMULA. 15 9.1.2 DATA COLLECTION (ENDPOINTS) 15 THE PRIMARY ENDPOINT IS DESIGNED TO EVALUATE EVENT-FREE SURVIVAL (EFS) AT FIVE YEARS. EFS IS DEFINED AS IN THE RCT: FREEDOM FROM THE CEC- ADJUDICATED MAJOR ADVERSE EVENTS OF DEATH, TARGET LESION REVASCULARICATION, TARGET LIMB ISCHEMIA REQUIRING SURGICAL INTERVENTION (BYPASS OR AMPUTATION OF TOE, FOOT, OR LEG) OR SURGICAL REPAIR OF THE TARGET VESSEL (E.G., DISSECTION OR PERFORATION REQUIRING SURGERY), AND FREEDOM FROM WORSENING OF THE RUTHERFORD CLASSIFICATION BY 2 CLASSES OR TO CLASS 5 OR 6. 15 THE SECONDARY ENDPOINTS ARE THE RATES OF STENT FRACTURE AND LATE STENT THROMBOSIS. THESE ENDPOINTS WILL BE EVALUATED DESCRIPTIVELY AND WILL NOT BE SUBJECTED TO FORMAL HYPOTHESIS TESTING. 15 9.1.3 FOLLOW-UP VISITS AND LENGTH OF FOLLOW-UP 15 9.2. FDA PERSPECTIVE ON THE PROPOSED POST-APPROVAL STUDY 15 THE REVIEW TEAM BELIEVES THAT THE SPONSOR’S PROPOSAL TO FOLLOW-UP THE ZILVER PTX AND CONTROL ARMS OF THE RCT COHORT UP TO 5 YEARS WOULD PROVIDE A SUFFICIENT LENGTH OF LONG-TERM FOLLOW-UP. HOWEVER, THE PROPOSED STUDY DESIGN DOES NOT PROVIDE A SAMPLE SIZE ADEQUATE TO EVALUATE THE INCIDENCE OF LESS COMMON EVENTS SUCH AS STENT FRACTURE AND LATE STENT THROMBOSIS. THE REVIEW TEAM BELIEVES THAT THE PAS SHOULD BE STATISTICALLY POWERED TO ANALYZE THESE RARER EVENTS. WHILE THE SPONSOR CALCULATED A SAMPLE SIZE OF 113 SUBJECTS PER ARM TO ASSESS EFS AT 5 YEARS, A TOTAL OF 900 SUBJECTS WOULD BE NEEDED TO ADEQUATELY ASSESS THE INCIDENCE OF EVENTS SUCH AS STENT FRACTURE AND THROMBOSIS THAT MAY OCCUR AT A RATE OF 1 - 2%. 15 ENROLLMENT OF ADDITIONAL SUBJECTS FROM OUTSIDE THE RCT WOULD BE NECESSARY TO PROVIDE THIS NUMBER OF SUBJECTS, AND WOULD ALSO ALLOW FOR AN ASSESSMENT OF THE GENERALIZABILITY OF THE RCT DATA TO THE NON-RCT PATIENT POPULATION. ONE POTENTIAL APPROACH TO OBTAINING THE NECESSARY SAMPLE SIZE WOULD BE TO RE-CONSENT THE GLOBAL REGISTRY SUBJECTS FOR LONGER-TERM FOLLOW-UP. HOWEVER, RE-CONSENTING THESE SUBJECTS MAY BE CHALLENGING. IN ADDITION, ONE POTENTIAL RISK WITH THIS APPROACH WOULD BE THE INTRODUCTION OF SELECTION BIAS; THAT IS, THE SUBJECTS WHO CHOOSE TO RE-CONSENT FOR LONG-TERM FOLLOW-UP MAY NOT BE ENTIRELY REPRESENTATIVE OF THE GLOBAL REGISTRY COHORT AS A WHOLE. 15 THE PANEL WILL BE ASKED TO COMMENT ON THE DESIGN OF A PAS SHOULD THE PMA FOR THE ZILVER PTX BE APPROVED, PARTICULARLY WITH RESPECT TO THE NEED TO ENROLL NEW SUBJECTS FROM OUTSIDE THE RCT AND WHETHER RE-CONSENTING GLOBAL REGISTRY SUBJECTS WOULD PROVIDE VALID SCIENTIFIC EVIDENCE TO SATISFY THE PAS GOALS. 15 10. CONCLUSIONS 15 1. PROPOSED INDICATIONS FOR USE The Zilver PTX Drug-Eluting Stent is indicated for improving luminal diameter for the treatment of de novo or restenotic symptomatic lesions in vascular disease of the above-the-knee femoropopliteal arteries having reference vessel diameter from 4 mm to 9 mm and total lesion lengths up to 140 mm per limb and 280 mm per patient. To avoid involvement of the common femoral artery, the most proximal stent end should be placed at least 1 cm below the origin of the superficial femoral artery. To avoid involvement of the below-the-knee popliteal artery, the most distal stent end should be placed above the plane of the femoral epicondyles. The proposed indications for use specify a maximum lesion length of 140 mm per limb and 280 mm per patient. The pivotal clinical study conducted to support the effectiveness of this product was limited to patients with up to 2 lesions and a maximum individual lesion length of 140 mm. The panel will be asked to comment on the appropriateness of the lesion lengths specified in the proposed indications for use. 2. DEVICE DESCRIPTION The Zilver PTX Drug-Eluting Stent (Zilver PTX stent) is a self-expanding nitinol stent coated on its outer surface with the drug paclitaxel (without any polymer, binder, or excipient) at a dose density of 3 µg/mm2. The Zilver PTX stent is preloaded on a 6 Fr delivery system. Upon deployment, the Zilver PTX stent is designed to establish and maintain patency in the stented region. To facilitate fluoroscopic visualization of the stent, 4 radiopaque gold markers are positioned on each end of the device (Figure 1). Figure 1: Photograph of the Zilver PTX stent The delivery system is available in 80 cm and 125 cm lengths and is compatible with a 0.035 inch wire guide (Figure 2). The delivery system is identical to that used with the currently approved Zilver Vascular Stent (P050017), a bare metal self-expanding stent indicated for use in the iliac arteries. a. Handle b. Hub c. Safety lock d. Introducer catheter e. Introducer tip f. Side-arm flushing port g. Metal cannula h. Radiopaque marker on the delivery system i. Strain relief j. Gold radiopaque markers on the stent Figure 2: Schematic drawing of the Zilver PTX stent and delivery system Table 1 shows the proposed diameters and lengths of the Zilver PTX stent. Table 1: Zilver PTX stent size matrix Stent Outer Diameter (mm) Stent Length (mm) 20 30 40 60 80 5 • • • • • 6 • • • • • 7 • • • • • 8 • • • • • 9 • • • • • 10 • • • • • Zilver PTX stents are coated with paclitaxel API (active pharmaceutical ingredient) using a proprietary process. No excipients, polymers, carriers, binding agents, other materials, or other device modifications are involved. Paclitaxel is the same API used in some currently approved coronary drug-eluting stents. The chemical description of paclitaxel is provided in Figure 3. Paclitaxel • Synonyms: Taxol, Taxol A, Hunxol I, Paclitaxelum • IUPAC systematic name: β-(benzoylamino)-α-hydroxy-,6,12b-bis(acetyloxy)-12-(benzoyloxy) 2a,3,4,4a,5,6,9,10,11,12,12a,12b-dodecahydro-4,11-dihydroxy-4a,8,13,13-tetramethyl-5-oxo-7,11-methano-1H-cyclodeca(3,4)benz(1,2-b)oxet-9-yl ester,(2aR-(2a-α,4-β,4a-β,6-β,9-α(α-R*,β-S*),11-α,12-α,12a-α,2b-α))-benzenepropanoic acid • CAS registry number: 33069-62-4 • Chemical formula: C47H51NO14 • Structure of paclitaxel: Figure 3: Chemical description of paclitaxel For further device description details, please refer to the sponsor’s Executive Summary. 3. REGULATORY HISTORY Cook first began discussions with FDA regarding the evaluation of the Zilver PTX Drug-Eluting Stent under Pre-IDE I020209 in July 2002. Cook submitted the IDE for the pivotal study in November 2003, and the study was granted conditional approval to begin the first phase of enrollment in June 2004. In the first phase of the study, only subjects with lesion lengths < 7 cm were enrolled, and subjects could only receive a single Zilver PTX stent. Phase I enrollment was completed in February 2006 and included 60 subjects. Cook received full approval of the IDE in August 2007, and started Phase 2 of the study. Phase 2 consisted of patients with lesion length 0 – 14 cm. Enrollment was completed in August 2008. Long-term follow-up is ongoing. In addition to the pivotal study, FDA requested that Cook develop a plan to evaluate the safety of the Zilver PTX in additional subjects, as the size of the pivotal study might not be sufficient to adequately detect and evaluate rare adverse events related to the stent or its drug coating. In response, Cook conducted a global registry intended to further assess the safety of the product. Because the registry was conducted completely outside the United States, FDA review and approval of an IDE was not necessary and FDA did not review the protocol for the registry prior to initiation of enrollment. Enrollment in the registry, consisting of 787 subjects from Europe, Canada, and South Korea, took place from April 2006 – June 2008. Follow-up on the registry subjects is complete. Cook submitted the first module of their PMA (M090010) in April 2009. The final PMA module (P100022) was submitted in June 2010. In August 2010, FDA issued a major deficiency letter to the sponsor that requested additional information regarding clinical and non-clinical data. Cook’s formal response was received in June 2011. 4. PRE-CLINICAL STUDIES 4.1. In Vitro Testing 4.1.1 Stent Functional Testing FDA recommends stent functional testing as outlined in our guidance document “Non-Clinical Tests and Recommended Labeling for Intravascular Stents and Associated Delivery Systems.” The purpose of this testing is to ensure that the stent platform (i.e., metal stent) functions as intended. The following tests were performed on the Zilver PTX stent and results were submitted for review: o Material Analysis of Nitinol Tubing o Stent Corrosion Resistance o Dimensional Verification o Percent Surface Area o Uniformity of Expansion o Foreshortening o Radial (Hoop) Strength o Stent Integrity o Mechanical Properties o Flex/Kink Evaluation o Crush Resistance o Stress Analysis o Finite Element Analysis (FEA) o Fatigue/Durability Testing o MRI Safety o Radiopacity The review team has completed its review of the stent functional testing and has no outstanding concerns. 4.1.2 Stent Coating Testing For drug-eluting stents, FDA recommends additional testing of the stent coating to ensure that the integrity of the coating will be maintained over the simulated use of the device. The following stent coating tests were performed on the Zilver PTX stent and results were submitted for review: o Coating Thickness and Uniformity o Content Uniformity Along the Stent Length o Microstructure Analysis o In Vitro Drug Release o In Vitro Elution o Potency o Impurities/Degradants o Particulate Analysis The review team has completed its review of the stent functional testing and has no outstanding concerns related to baseline characterization of the product. There are outstanding issues related to the stability testing of the drug coating over time. FDA is working interactively with the sponsor to ensure that these issues are resolved in a timely fashion. 4.1.3 Stent Delivery System Testing FDA recommends delivery system functional testing as outlined in our guidance document “Non-Clinical Tests and Recommended Labeling for Intravascular Stents and Associated Delivery Systems.” The purpose of this testing is to ensure that the stent delivery systems function as intended. The following stent delivery system tests were performed on the Zilver PTX stent and results were submitted for review: o Delivery System Profile o Ease of Access o Deployment Accuracy o Deployment Force o Flex/kink Evaluation o Delivery System Tensile Strength The review team has completed its review of the stent functional testing and has no outstanding concerns. 4.1.4 Animal Studies Cook has conducted six animal studies of the Zilver PTX stent in order to demonstrate the safety of the Zilver PTX Drug-Eluting Stent system. Studies were carried out in a porcine non-atherosclerotic model for durations of 1, 3, and 6 months to determine the safety of the clinical dose of paclitaxel that should be incorporated into the Zilver PTX stent, to evaluate the deliverability and the regional and systemic safety of the Zilver PTX stent, and to determine the short- and long-term pharmacokinetics of the Zilver PTX stent. Animal studies were also utilized to assess the safety of overlapping Zilver PTX stents, and to assess a maximum dose based on a determined drug safety margin. Overall, the animal studies have demonstrated the non-clinical safety of the Zilver PTX stent in the animal model at multiple time points. The Zilver PTX incorporates a drug loading of 3 ug/mm2 and appears to elicit similar biologic responses as the bare metal Zilver stent when implanted in the pig model. Because the animal studies were limited to obtaining sufficient overdose data to support the safety of maximum of two overlapped 80 mm Zilver PTX stents per limb with a maximum stented length of ¬¬¬¬280 mm per patient, the review team believes that conclusions regarding the non-clinical safety of the Zilver PTX stent cannot be drawn for lesion length greater than 140 mm per limb. The stented lengths in the animal studies are consistent with the lesion lengths specified for treatment in the pivotal clinical trial protocol. 4.1.5 Chemistry, Manufacturing, & Controls (CMC) The drug substance, paclitaxel, is purchased from Yunnan Hande Bio-Tech Co., Ltd., of the People’s Republic of China. CMC information describing the manufacture and controls of paclitaxel drug substance is incorporated in the Zilver PTX PMA by authorized reference to Drug Master File (DMF) #11909. The paclitaxel drug substance meets the quality specifications which were developed based on the US Pharmacopoeia (USP) and European Pharmacopoeia (EP). The CMC information for the drug substance was reviewed and found to be acceptable. For the finished product, Cook has provided the details of the manufacturing process, the quantitative composition of the product, and controls for the paclitaxel drug substance used in the manufacture of the Zilver PTX stent, including the analytical test methods and supporting validation data. Cook has also proposed complete finished product specification, including the analytical procedures and supporting validation data. There are outstanding issues related to these specifications. FDA is working interactively with the sponsor to ensure that these issues are resolved in a timely fashion. Cook is collecting stability data which will be submitted to FDA for consideration of a yet-to-be determined shelf life. The shelf life cannot be established until the specifications are finalized and the supporting stability data are reviewed. 4.1.6 Sterilization The Zilver PTX stent and delivery systems are provided sterile. After reviewing the information submitted by the sponsor, the review team has concluded that under the stated exposure conditions, the ethylene oxide cycle will render the Zilver PTX stent and delivery systems sterile at a sterility assurance level of 10-6, or the probability of one survivor in one million products sterilized. Ethylene oxide and ethylene chlorohydrin residual analysis was performed to confirm that residual levels are below stated acceptance criteria. Packaging studies were performed to demonstrate that the current packaging configuration will maintain a sterile barrier to support the forthcoming shelf-life claim. 4.1.7 Biocompatibility Cook has identified all material components of the stent and delivery systems. Cook has provided data from the following biocompatibility evaluations of the finished drug-eluting stent: cytotoxicity, sensitization, intracutaneous reactivity, acute systemic toxicity, in vitro hemolysis, in vivo thrombogenicity, complement activation, plasma recalcification, subchronic toxicity, muscle implantation, material-mediated pyrogenicity, in vivo mouse bone marrow micronucleus assay, bacterial reverse mutation assay, and chromosomal aberration assay. Cook has provided data from the following biocompatibility evaluations of the finished drug-eluting stent and delivery system: cytotoxicity, pyrogenicity, sensitization, intracutaneous reactivity, acute systemic toxicity, direct complement activation, plasma recalcification, in vivo mouse bone marrow micronucleus assay, bacterial reverse mutation assay, chromosomal aberration assay, in vitro hemolysis, and in vivo thrombogenicity. The review team has requested clarifications regarding this testing. FDA is working interactively with the sponsor to ensure that these issues are resolved in a timely fashion. 4.1.8 Manufacturing FDA has reviewed the manufacturing information. Reviews of the reports from the facility inspections are not yet complete. 4.2. Studies of the Drug Substance The active pharmaceutical ingredient in the Zilver PTX stent is paclitaxel provided by Yunnan Hande Bio-Tech Co., Ltd. Paclitaxel, a diterpenoid with a taxane skeleton, is extracted from the bark, branches, or needles of the yew tree, then purified and concentrated by column chromatography, crystallization and recrystallization. The drug substance is a white to off-white crystalline powder and is highly lipophilic, insoluble in water, and freely soluble in: methanol, ethanol, chloroform, ethyl acetate and dimethylsulfoxide. Paclitaxel has a molecular formula of C47H51NO14. Paclitaxel stabilizes the microtubule polymer and inhibits its disassembly, leading to mitotic block with subsequent cell cycle arrest. Paclitaxel (Taxol®) injection is used as an antineoplastic, and is FDA-approved to treat advanced cancer of the ovaries, breast and lung, and Kaposi sarcoma. Paclitaxel delivered locally using devices, including some currently marketed coronary drug-eluting stents, is used as an anti-proliferative agent for the prevention of arterial restenosis. The sponsor has provided a letter from Yunnan Hande Bio-Tech Co., Ltd. authorizing FDA the right to access the DMF for paclitaxel in support of this PMA application. 4.2.1 Safety Pharmacology In the Zilver PTX stent, the systemic levels of drug were not concerning because they were below levels of detection after Day 1. However, the Zilver PTX stent efficiently delivered the drug to the local arterial tissue and regional skeletal muscle such that the concentration of paclitaxel achieved at the stented artery (70µg/g) is about 4 orders of magnitude higher than the concentration of paclitaxel needed for antiproliferative effects (1ng/g). At Day 56, the levels of paclitaxel (160 ng/g) retained in the tissue remained 2 orders of magnitude higher than the half-maximal concentration of paclitaxel for antiproliferative effects. The possible concerns related to the high measured tissue levels of paclitaxel and prolonged retention time were mitigated by the animal and clinical studies, as these data did not reveal any apparent concerns. In the animal studies, there were no pseudoaneurysms, dissections, perforation, systemic illnesses, gait abnormalities, and inflammation and injury were noted to be very low. Please refer to Section 7 for a discussion of the toxicity profile observed in the Zilver PTX clinical studies. Review of these data did not indicate any safety concerns and supported initiation of the human clinical studies of the Zilver PTX stent. 4.2.2 Toxicology Toxicology information on paclitaxel was incorporated by reference to the manufacturer’s DMF. FDA has reviewed this information and there are no remaining concerns. Based on this information, the labeling proposed by the applicant contains the following contraindication: “Women who are pregnant, breastfeeding, or plan to become pregnant in the next 5 years should not receive a Zilver PTX Drug-Eluting Peripheral Stent. It is unknown whether paclitaxel will be excreted in human milk, and there is a potential for adverse reaction in nursing infants from paclitaxel exposure.” 5. ZILVER PTX RANDOMIZED, CONTROLLED TRIAL 5.1. Clinical Evaluation Strategy Clinical data to study the safety and effectiveness of the Zilver PTX stent was obtained from two separate studies: 1) a randomized, controlled trial (RCT) that enrolled 479 subjects from US, Germany and Japan, comparing treatment of de novo or restenotic lesions of the femoropopliteal artery up to 14 cm in length using the Zilver PTX stent and standard of care, percutaneous transluminal angioplasty (PTA); and 2) a global registry of 787 consecutive treated subjects with de novo or restenotic lesions of the femoropopliteal artery with no restriction on lesion length (Figure 4). The combined studies resulted in data from more than 1000 patients with a Zilver PTX stent implanted, providing a statistically powered evaluation of the safety and effectiveness, as well as a larger-scale experience with the study device to allow for surveillance of adverse events that may occur at a low frequency and thus may not be sufficiently evaluable in a randomized trial. Figure 4: Total subject enrollment in the Zilver PTX studies 5.2. Zilver PTX Randomized Controlled Trial (IDE G030251) The Zilver PTX randomized, controlled trial (RCT) was a prospective, controlled, open-label, multi-center study enrolling subjects in the United States, Japan, and Germany with de novo or restenotic lesions of the above-the-knee femoropopliteal artery. Subjects were randomized in a 1:1 ratio to treatment with the Zilver PTX stent (treatment group) or with PTA (control group). Recognizing that balloon angioplasty may not be successful acutely, the trial design mandated provisional stent placement immediately after failure of balloon angioplasty in instances of acute PTA failure. Therefore, subjects with suboptimal (failed) PTA underwent a secondary 1:1 randomization to stenting with either Zilver PTX or bare Zilver stents. The study design is illustrated in Figure 5. Figure 5: Zilver PTX RCT study design The overall study was designed to enroll 480 subjects with symptomatic stenotic or occlusive atherosclerotic lesions, up to 14 cm long, of the above-the-knee femoropopliteal artery. Because at the time of study initiation there was inadequate non-clinical safety information to support clinical treatment using the Zilver PTX stent in lesions greater than 7 cm in length, the first 60 subjects enrolled (Phase 1) were limited to a maximum lesion length of 7 cm, and had additional required imaging for increased evaluation of safety prior to the remaining enrollment. Inclusion/exclusion criteria are identical for the first 60 subjects and subsequent 420 subjects, with the exception of lesion length (limited to 7 cm for the first 60 subjects) and 6-month imaging (IVUS and angiography required for the first 60 subjects). Subjects were randomized to treatment with PTA or treatment with up to two (2) Zilver PTX stents per lesion. One lesion in each limb (total of two lesions per subject) were eligible for treatment. 5.3. Patient Selection Process and Enrollment Criteria The major inclusion and exclusion criteria are summarized below; the full list is in the protocol and the sponsor’s clinical report: Inclusion Criteria The major inclusion criteria for patient entry into the study included the following: • Patient has up to 2 documented stenotic or occluded atherosclerotic lesions (≤ 14 cm long, or ≤ 7 cm for the first 60 subjects enrolled) of the above-the-knee femoropopliteal artery, up to one in each limb, that meet all of the inclusion criteria and none of the exclusion criteria. • Patient has reference vessel diameter of 4 - 9 mm. • Patient has a de novo or restenotic lesion(s) with > 50% stenosis documented angiographically and no prior stent in the target lesion.
• Patient has symptoms of peripheral arterial disease classified as Rutherford Category 2 or greater.
• Patient has a resting ABI < 0.9 or an abnormal exercise ABI if resting ABI is normal. Patient with incompressible arteries (ABI > 1.2) must have a TBI < 0.8. Exclusion Criteria The major exclusion criteria for patient entry into the study included the following: • Patient has significant stenosis (> 50%) or occlusion of inflow tract (proximal ipsilateral, iliofemoral, or aortic lesions) not successfully treated before this procedure (success is measured as < 30% residual stenosis). • Patient has undergone an unsuccessful arterial interventional treatment of the legs (i.e., the treatment resulted in > 30% residual stenosis of a treated lesion) within 30 days prior to the study procedure.
• Patient has experienced complications of an arterial access site in the legs within 30 days prior to the study procedure.
• Patient has any planned surgical or interventional procedure within 30 days after the study procedure.
• Patient has a planned procedure involving arterial interventional treatment of the study leg(s) within the 12-month follow-up period.
• Patient has had previous stenting of the target vessel.
• Patient lacks at least one patent vessel of runoff with < 50% stenosis throughout its course. • Patient has untreated angiographically-evident thrombus in the target lesion. • Patient has a bypass graft with an anastomosis in the target vessel. • Patient has lesions requiring atherectomy (or ablative devices), cutting balloons, cryoplasty balloons, or any other advanced device to facilitate stent delivery. 5.4. Key Study Definitions Adverse Event, Major (MAE) Includes death, clinically-driven target lesion revascularization, ischemia requiring intervention (bypass or amputation of toe, foot or leg) and vascular repair (e.g., dissection requiring surgery). Adverse Event, Early Clinical events that occur within 30 days of the stent procedure. Adverse Event, Late Clinical events that occur 30 or more days after the stent procedure. Ankle-Brachial Index (ABI) The ratio of ankle systolic pressure to brachial systolic pressure. ABI ranges of 0.9 - 1.2 are considered normal, while ABI values < 0.9 indicate peripheral arterial disease and values < 0.4 indicate severe peripheral arterial disease (ischemic pain and ulceration). If the ABI > 1.2, this is likely due to incompressible arteries (commonly observed in association with long-standing diabetes mellitus, extreme old age, or calcinosis) and the toe-brachial index (TBI) should be measured.
ABI measured one minute post-exercise at 3.5 km/hr and 12% grade for five minutes or until symptoms develop. Abnormal exercise ABI is defined as a drop of > 25% from resting ABI.
Classification of Lesion Morphology (TASC):
o Type A femoropopliteal lesions: Single stenoses up to 3 cm in length, not at the origin of the superficial femoral artery (SFA) or the distal popliteal artery.
o Type B femoropopliteal lesions: Single stenoses or occlusions 3-5 cm long, not involving the distal popliteal artery; heavily calcified stenoses up to 3 cm in length; multiple lesions, each less than 3 cm (stenosis or occlusions); single or multiple lesions in the absence of continuous tibial runoff to improve inflow for distal surgical bypass.
o Type C femoropopliteal lesions: Single stenoses or occlusions longer than 5 cm; multiple stenoses or occlusions, each 3-5 cm, with or without heavy calcification.
o Type D femoropopliteal lesions: Complete common femoral artery or SFA occlusions or complete popliteal and proximal trifurcation occlusions.
o Grade A: Small radiolucent area within the lumen of the vessel disappearing with the passage of the contrast material.
o Grade B: Filling defect parallel to the lumen of the vessel disappearing with the passage of contrast material.
o Grade C: Dissection protruding outside the lumen of the vessel persisting after passage of contrast material.
o Grade D: Spiral shaped filling defect with delayed runoff of the contrast material in the distal vessel.
o Grade E: Persistent luminal filling defect with delayed anterograde flow.
o Grade F: Filling defect accompanied by total occlusion.
Freedom from the Clinical Events Committee (CEC)-adjudicated major adverse events of death, target lesion revascularization, target limb ischemia requiring surgical intervention (bypass or amputation of toe, foot, or leg) or surgical repair of the target vessel (e.g., dissection or perforation requiring surgery), and freedom from worsening of the Rutherford classification by 2 classes or to class 5 or 6.
A localized collection of blood, usually clotted, in an organ, space, or tissue, due to a break in the wall of a blood vessel; bleeding into an extravascular space, which could be outside the body. A serious adverse event hematoma or hemorrhage is one requiring transfusion, surgical intervention, or causing death.
Amputation of a leg, either above the knee or below the knee, including amputation of the foot or toes.
Minimum Lumen Diameter (MLD)
Mean minimum lumen diameter (mm) as determined from 2 orthogonal views.
New York Heart Association Classification:
o Class I: The patient has cardiac disease but without resulting limitations of ordinary physical activity. Ordinary physical activity (i.e., walking several blocks or climbing stairs) does not cause undue fatigue, palpitation, dyspnea, or anginal pain. Limiting symptoms may occur with marked exertion.
o Class II: The patient has cardiac disease resulting in slight limitation of ordinary physical activity. Patient is comfortable at rest. Ordinary physical activity such as walking more than two blocks or climbing more than one flight of stairs results in limiting symptoms (e.g., fatigue, palpitation, dyspnea, or anginal pain).
o Class III: The patient has cardiac disease resulting in marked limitation of physical activity. Patient is comfortable at rest. Less than ordinary physical activity (i.e., walking 1-2 blocks or climbing one flight of stairs) causes fatigue, palpitation, dyspnea, or anginal pain.
o Class IV: The patient has dyspnea at rest that increases with any physical activity. Patient has cardiac disease resulting in inability to perform any physical activity without discomfort. Symptoms of cardiac insufficiency or anginal syndrome may be present even at rest. If any physical activity is undertaken, discomfort is increased.
No flow identified within the arterial segment by ultrasound and/or angiogram.
A demonstrably open treated segment (i.e., < 50% diameter stenosis, including the region within ± 5 mm proximal and/or distal to the target lesion) as assessed via duplex ultrasound (PSV < 2) and/or angiography (note: in cases where both imaging modalities are available, the angiography will take precedence). Patency, Primary The proportion of treated segments with uninterrupted (intervention-free) patency (i.e., < 50% diameter stenosis) since the initial procedure. Failure of primary patency occurs at the first occurrence of one of the following: acute PTA failure, loss of patency (i.e., > 50% diameter stenosis, including the region within ± 5 mm proximal and/or distal to the study segment, as assessed via ultrasound (PSV > 2) and/or angiography; in cases where both imaging modalities are available, the angiography will take precedence), re-intervention in the study segment due to > 50% angiographic diameter, stenosis, total occlusion of the treated segment, surgical bypass due to treated segment restenosis (see definition), or amputation of the extremity due to treated segment restenosis (see definition).
Patency, Assisted Primary
The proportion of patients who require an intervention after the initial procedure to assist patency maintenance of the target lesion, without an episode of total occlusion. Failure of primary assisted patency occurs at the time of total occlusion of the original lesion, or when the stented segment is surgically bypassed or amputation of the extremity occurs due to restenosis or occlusion.
The proportion of patients who require an intervention after the initial procedure to assist patency maintenance of the target lesion following an episode of total occlusion. Failure of secondary patency occurs at the time the treated segment is surgically bypassed or amputation of the extremity occurs due to restenosis or occlusion.
Percent Diameter Stenosis (%DS)
Calculated using the formula 100 x (1-(MLD/RVD)).
PTA Failure, Acute
One or more of the following occurring at the time of the procedure: (a) an inadequate angiographic and/or hemodynamic result as defined by a 30% or greater residual stenosis, lesion recoil, or intimal flaps; (b) flow-limiting residual dissections; (c) a 5 mm Hg or greater trans-stenotic pressure gradient; or (d) acute occlusion of the vessel.
PTA Failure, Long-Term
≥ 50% diameter stenosis including the region within ± 5 mm proximal and/or distal to the target lesion within 12 months of the index PTA procedure, confirmed and documented by duplex ultrasound or angiography.
Reference Vessel Diameter, Proximal (RVDprox)
Diameter of normal segment immediately proximal to the treated region.
Reference Vessel Diameter, Distal (RVDDist)
Diameter of normal segment immediately distal to the treated region.
Acute or progressive renal insufficiency leading to the need for dialysis.
A rise in creatinine of more than 30% above the pre-procedure level resulting in a creatinine level > 2.0 mg/dl that does not spontaneously resolve.
Recurrence of ≥ 50% diameter stenosis within ± 5 mm proximal and/or distal to the target lesion as measured by duplex ultrasound (PSV > 2) or angiography (note: in cases where both imaging modalities are available, the angiography will take precedence).
o Class 0: Asymptomatic, no hemodynamically significant occlusive disease.
o Class 1: Mild claudication.
o Class 2: Moderate claudication.
o Class 3: Severe claudication.
o Class 4: Ischemic rest pain.
o Class 5: Minor tissue loss.
o Class 6: Major tissue loss.
o Type 0: No strut fractures.
o Type I: Single strut fracture only.
o Type II: Multiple single strut fractures that can occur at different sites.
o Type III: Multiple strut fractures resulting in complete transection of the stent, without displacement of the stent segments.
o Type IV: Multiple strut fractures resulting in displacement of segments of the stent (includes spiral fractures that could result in stent displacement without complete transaction).
Improvement by at least two Rutherford classes above the pre-treatment clinical level at follow-up.
Vessel with < 30% residual stenosis determined angiographically immediately after the procedure. Target Lesion Revascularization (TLR) A re-intervention performed for ≥ 50% diameter stenosis (confirmed by angiography) within ± 5 mm proximal and/or distal to the target lesion after documentation of recurrent clinical symptoms of PAD following the initial procedure. Toe-Brachial Index (TBI) The ratio of toe systolic pressure to brachial systolic pressure. TBI values > 0.8 are considered normal, while values < 0.8 indicate peripheral arterial disease. Thrombosis, Acute Ultrasound, angiographic or clinically proven complete occlusion of an artery occurring within 24 hours of the procedure. Thrombus, Intraluminal Subtotal occlusion of the artery due to the presence of thrombus. Thrombosis, Late Ultrasound, angiographic or clinically proven complete occlusion of an artery occurring 30 days or more after the procedure. Thrombosis, Subacute Ultrasound, angiographic or clinically proven complete occlusion of an artery occurring after 24 hours, but within 30 days of the procedure. Walking Impairment Questionnaire (WIQ) An independent measure of patient-perceived walking performance for patients with PAD and/or intermittent claudication. This questionnaire estimates walking distance, walking speed, and stair-climbing capacity. Improvement is observed through an increase of walking distance or walking speed. 5.5. Study Endpoints, Hypotheses, and Success Criteria Primary Safety Endpoint The primary safety endpoint was event-free survival (EFS) evaluated at 12-month follow-up, defined as freedom from the Clinical Events Committee (CEC)-adjudicated major adverse events of death, target lesion revascularization, target limb ischemia requiring surgical intervention (bypass or amputation of toe, foot, or leg) or surgical repair of the target vessel (e.g., dissection or perforation requiring surgery), and freedom from worsening of the Rutherford classification by 2 classes or to class 5 or 6. The hypotheses for the primary safety analysis were: H0: EPTX ≤ EPTA - δ HA: EPTX > EPTA – δ;
where EPTX is the event-free survival rate for the Zilver PTX stent group (treatment), EPTA is that for the PTA control group and δ is the non-inferiority margin of 10%.
The Z-statistic is given by , where:
o = Kaplan-Meier estimated event-free survival at 12 months post-procedure for the PTA per-protocol cohort
o = Kaplan-Meier estimated event-free survival at 12 months post-procedure for the Zilver PTX per-protocol cohort
o = minimum difference of practical interest (in this case 10%)
o SE = Standard Error = , where and are the estimates of the variances of the Kaplan-Meier estimates using the Greenwood formula.
The primary safety null hypothesis was to be rejected in favor of the alternative if – > ZαSE – , for α=0.05.
The primary safety endpoint was to be analyzed on the per-protocol population on a per-patient basis.
Primary Effectiveness Endpoint
The primary effectiveness endpoint of the RCT was primary patency evaluated at 12-month follow-up as assessed by duplex ultrasound or by angiography when available. Patency was defined as a demonstrably open treated segment (i.e., < 50% diameter stenosis, including the region within ± 5 mm proximal and/or distal to the target lesion) as assessed via duplex ultrasound (PSV < 2.0) and/or angiography. The hypotheses for the primary effectiveness analysis were: Ho: PPPTA = PPPTX vs. Ha: PPPTA ≠ PPPTX; where PPPTA is the primary patency rate for the PTA arm (control) evaluated at 12-month follow-up and PPPTX is that for the Zilver PTX arm (treatment). The primary effectiveness endpoint was to be analyzed on a per-lesion basis using the Z- statistic from a Generalized Estimating Equation (GEE) model. Because two lesions from the same patient cannot be considered independent, use of the GEE model was proposed to account for the correlations between them. The GEE model was to use the binomial distribution, logit link function, and due to the possibility of bilateral treatment (up to two lesions per patient), the model incorporated an unstructured covariance matrix. This analysis would utilize SAS Proc Genmod specifying the patient as the subject factor in the repeated statement. The Z-statistic is given by , where is the maximum likelihood estimate of the treatment effect and is the corresponding standard error of the estimate. The Zilver PTX stent was considered to yield significantly better results than the control (PTA) group if the null hypothesis was rejected in favor of the alternative for a two-sided P-value < 0.05 and the observed primary patency rate for the Zilver PTX stent was greater than the observed primary patency rate for conventional PTA. The primary effectiveness analysis was to be conducted using the intent-to-treat population on a per-lesion basis. Secondary Endpoints Several secondary endpoints were pre-defined in the protocol. There was no pre-specified hypothesis testing of the secondary endpoint or any pre-planned control of type I error. The secondary endpoints include: o Procedural success o 12-month clinical success based on an improvement of 2 Rutherford categories o 12-month clinical improvement based on an improvement of 1 Rutherford category o 12-month secondary patency o 12-month restenosis rate o 12-month thrombosis rate o 12-month functional status improvement as measured by ABI and the Walking Impairment Questionnaire 5.6. Subject Disposition and Data Analysis Sets The definitions of the intent-to-treat (ITT), per-protocol (PP), and as-treated (AT) analysis populations are listed in Table 2. Table 2: Definition of Intent-To-Treat, Per-Protocol, and As-Treated populations The primary safety analysis was performed on the per-protocol population on a per-patient basis, while the primary effectiveness analysis was performed on the intent-to-treat population on a per-lesion basis. The safety and effectiveness analyses using the other populations were performed as secondary analyses. Enrollment in the RCT began on March 21, 2005 and was completed on August 25, 2008. Four hundred seventy-nine subjects were enrolled at 55 sites within the United States, Germany, and Japan. Of the 479 subjects enrolled, 238 subjects were randomized to the control (PTA) group, 236 subjects were randomized to the treatment (Zilver PTX) group, and five (5) subjects were treated as live cases and were assigned to the Zilver PTX treatment group without undergoing randomization. Of the 479 subjects enrolled, 81.4% (390/479) were enrolled in the United States, 11.1% (53/479) were enrolled in Japan, and 7.5% (36/479) were enrolled in Germany. Follow-up availability is outlined in Table 3, which includes the number of subjects eligible for each follow-up time point and the number of subjects for whom clinical, core laboratory x-ray, and diagnostic core laboratory ultrasound data were available at each follow-up time point on an intent-to-treat basis. The number of subjects eligible for each follow-up time point was determined by subtracting the number of deaths, subjects withdrawn, and subjects lost to follow-up from the total number of subjects eligible for the previous follow-up. The panel will be asked to comment on the adequacy of the long-term follow-up data provided. Table 3: RCT Follow-up availability The five subjects treated as live cases and enrolled in the treatment group were excluded in the intent-to-treat and the per-protocol populations. These subjects were included in the as-treated subject population. The ITT population included the 238 subjects (251 lesions) randomized to the control (PTA) group and 236 subjects (247 lesions) randomized to the treatment (Zilver PTX) group. The PP population included 236 subjects (249 lesions) in the control group and 235 subjects (246 lesions) in the treatment group. The AT population included 236 subjects (249 lesions) in the control group and 242 subjects, including the 5 live cases (253 lesions), in the treatment group. The disposition of randomized subjects who had different assignments between the ITT and AT analysis is provided in Table 4 below. Table 4: Differences between ITT and AT populations (for first and second randomization) In the control group, 125 lesions in 120 subjects experienced acute PTA failure and underwent a second randomization to either placement of a bare Zilver stent(s) (59 subjects, 62 lesions) or a Zilver PTX stent(s) (61 subjects, 63 lesions). 5.7. Subject Demographics and Other Baseline Characteristics Baseline subject demographics are generally comparable between treatment and control groups with respect to age, gender, ethnicity, height, weight, and body mass index (Table 5). Females accounted for 36.1% and 34.4% of control and treatment subjects, respectively. The mean ages for the respective groups were 67.7 and 67.9 years. The treatment group had numerically higher incidence of hypercholesterolemia, previous MI and pulmonary disease, and a statistically significantly higher incidence of hypertension. Baseline angiographic data are generally well-matched, with the exception of the lesion class, degree of calcification, and presence of inflow tract stenosis, all of which are higher in the treatment group. In the control group, 38.1% of subjects had lesions TASC class C or D compared to 48.1% in the treatment group; the difference in the distribution was not statistically significant. All of these clinical predictors favor the control group. Table 5: Demographics and patient characteristics (ITT Population) 5.8. Safety Results 5.8.1 Primary Safety Endpoint The primary safety endpoint analysis was conducted using the per-protocol population on a per-patient basis. At 12 months, there were 37 failures and 15 instances of censoring in the control group, and 22 failures and 16 instances of censoring in the treatment group. The Kaplan-Meier (KM) estimates of the 12-month EFS rates were 90.4% in the Zilver PTX treatment group and 83.9% in the PTA control group (Figure 6). These KM estimates were used to calculate the Z-statistic and the p-value for the non-inferiority hypothesis regarding the 12-month EFS rate. The p-value is < 0.01, indicating that the primary safety endpoint was met with a non-inferiority margin of 10%. Figure 6: Kaplan-Meier estimates for event-free survival to 24 months Secondary analyses of the primary safety endpoint were also performed on the ITT and AT populations, and both yielded p-values of less than 0.01, indicating that non-inferiority of the 12-month EFS rate was demonstrated between the treatment and control groups using all defined analysis populations. Information about each individual adverse event is provided in Table 6. There were 13 subject deaths (4 control, 9 treatment group) during the 12-month follow-up period, but none were adjudicated by the CEC to be related to the study device or to the procedure. The rates for clinically-driven TLR, amputation, and worsening of Rutherford classification by 2 classes or to class 5 or 6 are listed below. Since the assessment of clinically-driven TLR involves subjective clinical judgment in an unblinded trial, the use of this endpoint is potentially subject to bias. Table 6: Individual rates of major adverse events at 12 months Major Adverse Event Control (PTA)1 Treatment (Zilver PTX)1 P-value Difference (95% CI)3 Clinically-driven TLR 16.3% (36/221)2 9.6% (21/219) 0.04 (0.5, 12.9) Worsening of Rutherford classification by 2 classes or to a class 5 or 6 0.9% (2/221)2 0.0% (0/219) 0.49 - Amputation 0.0% (0/221) 0.5% (1/219) 0.49 - 1 Denominator in this table is the number of patients remaining free from MAE at 12 months plus the number that have experienced a MAE prior to 12 months. 2 One patient experienced a worsening Rutherford and a TLR and is included in both categories in this table. 3 Confidence interval is the difference in percentages. 5.8.2 Adverse Events The rates of other adverse events are provided in Table 7. They were balanced between groups, and no patterns that raise concerns about device toxicity are discernable. While there was a trend for higher pulmonary and renal events and a statistically higher rate of gastrointestinal events in the treatment arm, the relevance of these findings to the safety of the Zilver PTX is not apparent. Table 7: All adverse events occurring within 12 months follow-up 5.8.3 Stent Thrombosis The RCT protocol specified a standardized antiplatelet medical regimen for all subjects. Subjects were to receive clopidogrel (or ticlopidine for Japanese subjects) starting at least 24 hours prior to the procedure, or a loading dose of clopidogrel during the procedure, followed by clopidogrel for 60 days post-procedure along with aspirin indefinitely. Nearly all subjects were discharged taking aspirin and clopidogrel, and approximately 90% of subjects continued to take aspirin from 1 month through 2 years. Usage of clopidogrel was lower, with approximately 85% of subjects remaining on this medication at 1 month, 70% at 3 months, and 60% through 2 years. Compliance with this medical regimen was comparable between the treatment and control groups at all time points. The KM estimates for freedom from stent thrombosis for the Zilver PTX treatment arm, provisional Zilver PTX subjects, all subjects who received a Zilver PTX in the RCT, and all provisional bare metal Zilver subjects are presented in Table 8. The observed thrombosis rates were comparable among these subjects groups at both time points. No thromboses were detected in RCT subjects who did not receive a stent. Table 8: Stent thrombosis events Group 12 Months 24 Months KM Estimate Thromboses Remaining KM Estimate Cumulative Thromboses Remaining Zilver PTX treatment group 97.5% 6 219 97.1% 7 198 Provisional Zilver PTX group 100.0% 0 60 100.0% 0 58 Zilver PTX treatment group + provisional Zilver PTX group 98.0% 6 279 97.7% 7 256 Provisional bare Zilver group 96.4% 2 51 96.4% 2 50 5.9. Effectiveness Results 5.9.1 Primary Effectiveness Endpoint Analysis The primary effectiveness endpoint analysis was conducted using the intent-to-treat population on a per-lesion basis. Subjects who completed 12-month or later follow-up or who experienced a loss of primary patency prior to 12-month follow-up were included in this analysis, for a total of 467 of the 498 eligible lesions. The GEE model with only the treatment effect using the unstructured covariance matrix on the ITT population resulted in a p-value less than 0.01 for the treatment effect, indicating rejection of the null hypothesis. Since the difference is in favor of the Zilver PTX treatment group, the Zilver PTX stent is concluded to yield a significantly better 12-month primary patency rate than PTA. Secondary analyses of the primary effectiveness endpoint of the PP and AT populations also yielded p-values less than 0.01. The estimated means and 95% confidence intervals calculated from the ITT, PP and AT populations are listed in Table 9. Table 9: Primary effectiveness endpoint analyses When the GEE model is used with three binary covariates (presence/absence of diabetes, de novo/restenotic lesion, and occluded/stenosed lesion) in addition to the treatment effect, the p value of the treatment effect from the GEE model is still < 0.01 after adjustment for these three covariates. Of the three covariates, the covariate of occluded/stenosed lesion is found to be significant (p-value from the GEE model < 0.01). The worst-case analyses of the primary effectiveness endpoint for the ITT, PP and AT populations (using GEE model with treatment effect only) all have p-values less than 0.01 for the treatment effect, indicating rejection of the null hypothesis. At 12 months, there are 167 patency failures (out of 251) for the control and 41 patency failures (out of 246) for the treatment group, with 11 instances of censoring in the control and 23 instances of censoring in the treatment group. The KM estimated primary patency rates at 12 month are 32.7% for the control group and 82.7% for the treatment group (Figure 7). Figure 7: Kaplan-Meier estimates for primary patency to 24 months In addition to the primary effectiveness analysis, the GEE model (with only treatment effect in) was run for several secondary analyses involving the primary effectiveness endpoint without adjustment for multiplicity. In these analyses, the subjects who underwent secondary randomization (bailout stenting) were not counted as failures. The treatment of initial procedural failure (i.e., stenting for acute PTA failure) was not counted as a revascularization; only subsequent revascularizations were counted towards the calculation of this endpoint. The results are summarized in Table 10 and some of these comparisons are discussed in detail in the following sections. Table 10: Additional analyses of the primary effectiveness endpoint In all of the comparisons, Zilver PTX has either superior or similar primary patency as compared to PTA or the bare metal Zilver stent. The primary effectiveness analysis results are mainly affected by the high rate of acute patency failures (approximately 50%) in the PTA control group. The panel will be asked to comment on the significance of this finding on the interpretation of the effectiveness results. 5.9.2 Primary Patency of Provisional Bare Zilver vs. Provisional Zilver PTX Comparing the primary patency results of subjects who received provisional (i.e. after bailout) stenting using the bare metal Zilver stent and those who received provisional stenting using the Zilver PTX provides some analysis of the effect of the paclitaxel drug coating on primary patency, since the underlying stent platform is nearly identical in both groups. In all analysis groups, the primary patency of the Provisional Zilver PTX group was statistically superior to that of the Provisional Bare Zilver group (see the fourth row of Table 10; p < 0.05 for the ITT, PP, and AT populations). The KM analysis indicates the same conclusion (Figure 8). Figure 8: Kaplan-Meier estimates for primary patency for provisional bare Zilver vs. provisional Zilver PTX (ITT population) 5.9.3 Primary Patency of Optimal PTA + Provisional Bare Metal Zilver vs. Primary Zilver PTX Since current standard of clinical care for patients with symptomatic disease of the femoropopliteal arteries consists of either PTA or bare metal stenting, comparing the primary patency rates of subjects in the Zilver PTX arm vs. those who received either optimal PTA (i.e. did not experience an acute patency failure) or who underwent provisional stenting with a bare metal Zilver stent after failed PTA provides some comparison of the effectiveness of primary stenting using Zilver PTX versus current standard of care. In this analysis, the primary patency in subjects who experienced primary stenting with Zilver PTX was statistically superior to optimal PTA with provisional bare metal stenting in the AT population, as indicated in Figure 9: Figure 9: Kaplan-Meier estimates for primary patency for current standard of care vs. primary Zilver PTX (AT population) To assess whether the observed treatment effect for this non-randomized comparison was solely due to differences in underlying covariates, a covariate analysis utilizing the GEE model was performed. The following variables were included in the GEE model with an unstructured covariance matrix: treatment effect, BMI, gender, total occlusion, lesion length, and smoking status. The treatment effect comparing Optimal PTA + Provisional Bare Zilver vs. Primary Zilver PTX was still significant (p < 0.01) after adjusting for these covariates. 5.9.4 Optimal PTA + Provisional Bare Zilver + Provisional Zilver PTX vs. Primary Zilver PTX As an evaluation of the effectiveness of the Zilver PTX stent when acute PTA failure is not considered a loss of primary patency, the combined primary patency results for optimal PTA, provisional bare Zilver stenting, and provisional Zilver PTX stenting were compared to the results for primary Zilver PTX stenting. As seen in the KM curves in Figure 10, the primary patency of primary treatment with Zilver PTX was statistically superior to the combined primary patency rates of the other three groups (p-value = 0.02). Figure 10: Kaplan-Meier estimates for primary patency for Optimal PTA + Provisional Bare Zilver + Provisional Zilver PTX vs. Primary Zilver PTX (ITT) 5.9.5 Secondary Effectiveness Endpoints The key secondary effectiveness endpoints are provided in Table 11. Rutherford classification, ankle-brachial index (ABI), and symptoms of walking impairment improved significantly in both the PTA control group and the Zilver PTX treatment group compared to baseline measurements. Table 11: Secondary endpoint estimates and 95% confidence intervals PTA Control Group Zilver PTX Treatment Group Procedural success 57.3% (50.8%, 63.5%) 95.0% (91.5%, 97.4%) 12-month clinical success 57.8% (50.7%, 64.6%) 54.9% (47.8%, 61.8) 12-month clinical improvement 77.7% (71.4%, 83.2%) 76.2% (69.8%, 81.9%) 12-month secondary patency 98.5% (95.5%, 99.5%) 98.2% (95.2%, 99.3%) 12-month restenosis rate1 33.5% (26.9%, 39.1%) 23.5% (18.6%, 29.5%) 12-month thrombosis rate n/a 2.0% (0.7%, 3.9%)2 ABI3 0.21 (0.18, 0.25) 0.24 (0.20, 0.28) Speed score3 28.6 (22.1, 35.0) 28.2 (21.9, 34.5) Distance score3 31.4 (25.1, 37.7) 32.8 (26.5, 39.1) Climb score3 22.8 (16.0, 29.6) 19.8 (12.8, 26.8) 1 Estimated based on GEE model for primary patency; the 12-month restenosis rate is the converse of the 12-month primary patency rate when acute PTA failure is not considered a loss of primary patency and lesions are further treated with stenting (bare Zilver or Zilver PTX). 2 Estimate reported for all Zilver PTX subjects (i.e., including provisional Zilver PTX in PTA control group). 3 Estimates reported for the change in value from pre-procedure to 12 months. 5.10. Sex/Gender Analysis The effect of sex/gender on the performance of the Zilver PTX stent was analyzed post-hoc by comparing the primary safety and effectiveness outcomes for the 167/474 (35%) female subjects and 307/474 (65%) male subjects. The event-free survival rates for female were 88.9% and 78.8% for the treatment and control arms, respectively. The rates for male were 90.5% and 84.5% for the treatment and control arms. The difference in outcome as a function of sex/gender for the primary safety endpoint was not statistically significant, suggesting that the event-free survival rate is comparable among male and female subjects within both the control and treatment arms (Figure 11). A covariate analysis utilizing a logistic regression model was performed on the primary safety endpoint. The model included the treatment effect, gender and gender/treatment interaction, and other demographic and clinical covariates. Neither the treatment/gender interaction nor the gender main effect were significant (p = 0.74 and 0.58, respectively). Figure 11: Kaplan-Meier estimates for event-free survival for female and male subjects For the primary effectiveness endpoint, a covariate analysis utilizing the GEE model was performed. The treatment effect was significant (p < 0.01) after adjusting for gender, hypercholesterolemia, smoking status, total occlusion, lesion length, and treatment/gender interaction. The p-value for the treatment-gender interaction was 0.052. This borderline significant interaction suggests that the magnitude of the treatment benefit may potentially be greater for males than for females. Despite the potentially significant interaction, a pronounced treatment effect can be observed for both genders. For females, the primary patency rates were 79.1% and 38.1% for the treatment and control arms, respectively. For males, the rates were 84.5% and 29.7% for the treatment and control. Figure 12: Kaplan-Meier estimates for primary patency for female and male subjects 5.11. Substudies Key substudies in the RCT included analyses of drug pharmacokinetics, angiographic/IVUS follow-up, and stent fractures. 5.11.1 Pharmacokinetic Substudy A subgroup of 60 subjects from the Zilver PTX treatment group was assigned to the pharmacokinetic (PK) sub-study to investigate the pharmacokinetic profile of the eluted paclitaxel in humans. Each subject in the PK substudy was subjected to venous blood draws at 3 of a possible 11 time points, according to a sparse sampling design: post-procedure (time 0), 20 min, 40 min, and 1, 1.5, 2, 3, 4, 6, 8, and 12 hrs. The results demonstrate that minimal paclitaxel was delivered systemically (Cmax < 10 ng/mL), and less than 1 ng/mL remained in the plasma at the 8- and 12-hour time points (Table 12). The results suggest rapid paclitaxel elution and low plasma drug levels. Table 12: Pharmacokinetic (PK) substudy results 5.11.2 Angiographic/IVUS Substudy An angiographic study was conducted on a subset of 80 subjects, 40 each from the treatment and control groups. The lesions in the substudy subjects from the Zilver PTX treatment group were evaluated with IVUS immediately post-stenting and at the 12-month follow-up visit, and all substudy subjects regardless of assigned treatment were evaluated with angiography at the 12-month follow-up. Because angiography or IVUS of femoropopliteal lesions during follow-up has a non-negligible risk of complications and is not standard of care in the absence of symptoms, the protocol did not require invasive imaging follow-up for the entire RCT cohort. In the 55 subjects who underwent both angiographic follow-up in the imaging substudy and duplex ultrasound protocol per the standard RCT protocol at 12 months, there was general agreement between the ultrasound and angiographic patency results, as indicated in Table 12. The results suggest that the protocol-specified non-invasive ultrasound follow-up for the RCT is a valid measure of patency. Table 13: Correlation of 12-month angiographic and ultrasound follow-up In the IVUS imaging subset, there were 2 instances of aneurysm and stent malapposition detected immediately after stent implantation (2/85 or 2.4%). There were no aneurysms or cases of stent malapposition detected via angiography or IVUS at the 6-month or 12-month follow-up visits and none of the findings were associated with clinical sequelae, although a Type I stent fracture (single strut) was detected at 12 months in the location of the aneurysm in one subject. 5.11.3 Stent Integrity High-resolution flat-plate stent x-rays were taken in subjects who received a bare metal Zilver or Zilver PTX stent prior to discharge and at 12 months to study stent integrity. X-rays were also taken at 6 months in the first 60 subjects enrolled. These x-rays were obtained in at least two approximately orthogonal views, including one view with the leg extended straight at the knee and one view with the leg bent 90º at the knee. Stent fractures were detected in 0.0% and 0.9% of stents prior to discharge and at 12 months, respectively (Table 14). Each of the fractures was located in the distal SFA, with two Type I (single strut fracture) and two Type III (multiple strut fractures resulting in complete transection of the stent, without displacement of the stent segments) fractures. Both of the stents with a Type III fracture were found to have been elongated during the implant procedure. This is a known predisposition to stent fracture. As indicated above, one of the Type I fractures occurred at the location of a pre-existing arterial aneurysm. None of the four subjects in whom a stent fracture was detected experienced a loss of primary patency or a TLR at 12 months. Table 14: Stent fracture incidence Imaging Timepoint Fracture Rate, % (n/N stents) Pre-Discharge 0.0% (0/528) 12 Months 0.9 % (4/457) 5.12. Summary of RCT Results The Zilver PTX RCT met its pre-specified primary safety and effectiveness endpoints of event-free survival and primary patency, respectively. The results for the control arm were consistent with those typically reported for PTA procedures in symptomatic femoropopliteal arteries. There were no signals regarding elevated rates of death, stent thrombosis, or stent fracture. However, due to the RCT sample size of 479 subjects, limited conclusions can be drawn regarding the detection of rare adverse events. 6. GLOBAL REGISTRY The Zilver PTX global registry was a prospective, non-randomized, open-label, multi-center registry intended to enroll up to 760 consecutive subjects with de novo or restenotic (including in-stent restenosis) lesions of the above-the-knee femoropopliteal artery at up to 50 centers in Europe, Canada, and Korea. Follow-up was for 2 years and evaluations included ABI, x-rays, and an ultrasound subset (provided duplex ultrasound was considered standard of care at the treating institution) at 6 and 12 months. All subjects enrolled in the study were treated with the Zilver PTX stent. The study was conducted with the intention that safety data from the registry would be combined with the subject data from the RCT to create a multi-study pool of more than 1000 Zilver PTX subjects and thereby more fully establish the rate of potentially rare device- or drug-related adverse events. Patient selection criteria were similar to the RCT to facilitate poolability of the two studies, with the notable exception that the registry had no limit on lesion length and permitted the implantation of up to 4 overlapping Zilver PTX stents. If more than four stents were necessary, bare metal Zilver stents were permitted to be implanted. Additionally, patients with ejection fraction 25 - 30% and NYHA Class III heart failure were eligible. While the global registry incorporated statistical hypotheses and success criteria, these analyses were only intended to support marketing approval outside the United States. The IDE review team did not review the global registry protocol prior to initiation of enrollment, and neither the IDE nor PMA review teams incorporated the study hypotheses or success criteria into their respective reviews of the safety or effectiveness of the Zilver PTX stent. The PMA review team considers this registry data to be informative, but observational in nature. 6.1. Study Endpoints The following endpoints and events were evaluated in the registry cohort: o Event-free survival (EFS) o Procedural success o Deployment success of the Zilver PTX stent o Clinical success based on an increase of 2 Rutherford categories o Clinical improvement based on an increase of I Rutherford category o Primary patency of the Zilver PTX stent o Assisted primary patency of the Zilver PTX stent o Secondary patency of the Zilver PTX stent o Incidence of rare individual adverse events, including stent thrombosis o Functional status as measured by ABI o Walking Impairment Questionnaire o Quality of Life Questionnaire o Stent fracture 6.2. Subject Disposition Between April 7, 2006 and June 18, 2008, 787 subjects were enrolled at 30 sites in Europe (specifically Germany, Italy, Netherlands, Czech Republic, Switzerland, Sweden, Belgium, Spain, Russia, France, and Poland), Canada, and South Korea. The vast majority of subjects (92.6%) were enrolled in Europe, while 2.9% and 4.4% of subjects were enrolled in Canada and South Korea, respectively. The study follow-up schedule included clinical assessment at pre-discharge and at 1, 6, 12, and 24 months, and ultrasound imaging and stent x-rays at pre-discharge and at 6 and 12 months. Telephone contact was scheduled for 3, 9, and 18 months. Clinical follow-up data were available for 740 subjects at 12 months and 500 subjects at 24 months. Approximately 95% of eligible subjects were seen for their 12-month follow-up visit and 90% had stent x-rays and duplex ultrasound completed (Table 15). Table 15: Registry imaging follow-up availability 6.3. Demographics and Other Baseline Characteristics Subject medical history included a high incidence of diabetes (36%), hypercholesterolemia (58%), hypertension (80%), and past or current smoking (80%). A total of 1722 Zilver PTX stents were implanted in 900 lesions during the study procedure, with more than 60% of subjects being treated with at least 2 stents. Lesions treated in the study had a mean length of 100 ± 82 mm, with 38% classified as total occlusions and 24% having been previously treated, including 130 lesions that had been previously stented. Subject characteristics are presented in Table 16. Table 16: Registry subject characteristics The registry subject demographics and comorbidities are similar to those of subjects in the RCT, with some exceptions. Lesion characteristics were more severe in the registry subjects, with longer lesion lengths and the inclusion of subjects with in-stent restenosis. The degree of calcification and the number of runoff vessels favors the registry. 6.4. Results 6.4.1 Event-Free Survival Rate The 6- and 12-month EFS rates for the Zilver PTX stent were 97.4% and 89.0%, respectively (Figure 13). This is consistent with the results from the RCT, in which the EFS rate in the Zilver PTX treatment group was 97.0% at 6 months and 90.4% at 12 months. Months Post-procedure Estimate Standard Error Cumulative Number Failed Cumulative Number Censored Number Remaining 0 100.0% 0% 0 6 781 1 99.2% 0.3% 6 13 768 6 97.4% 0.6% 20 42 725 12 89.0% 1.2% 80 143 564 18 82.3% 1.5% 115 292 380 24 79.3% 1.7% 128 419 240 Figure 13: Kaplan-Meier estimates for event-free survival Major adverse events that occurred in the registry are listed in Table 17. There were 4 CEC-adjudicated deaths through 24 months. All were adjudicated by the CEC as procedure-related, but not device-related. Table 17: Major adverse events through 24 months 6.4.2 Primary Patency Primary patency at 12 months was 83% by Kaplan-Meier estimate (Figure 14). In the RCT, 12-month primary patency of the Zilver PTX treatment group was 83.1%. Months Post-Procedure Estimate Standard Error Number Failed Number Censored Number Remaining 0 99.9% 0.1% 1 0 840 1 99.0% 0.3% 8 2 831 6 96.4% 0.6% 30 18 793 12 83.0% 1.3% 138 50 653 Figure 14: Kaplan-Meier estimates for primary patency 6.4.3 Procedural Success Of the 900 lesions treated, 97.5% (845/867) had < 30% residual stenosis at the end of the procedure and were therefore procedural successes. 6.4.4 Rutherford Classification Rutherford classifications improved significantly (p < 0.01) from pre-procedure through 24 months. Prior to treatment, most subjects were classified as Rutherford class 2 or 3 (88.7%, 673/759). At 12 and 24 months, more than two-thirds of the subjects were categorized as Rutherford class 0 or 1 (72.1%, 489/678 and 67.1%, 284/423 at 12 and 24 months, respectively). Rutherford categories improved by 1 class in 84.1% (551/655) of subjects and by 2 classes in 65.6% (430/655) of subjects. 6.4.5 Functional Status Significant functional status improvements were observed in ankle/brachial systolic pressure index (ABI) (0.64 ± 0.26, n = 700-pre procedure, 0.90 ± 0.24, n = 671 at 12 month follow-up, 0.87 ± 0.21, n = 411 at 24 months), walking distance, and walking speed from pre-procedure to the 6-month, 12-month, and 24-month follow-ups (p < 0.01). Quality of life measures also improved significantly from pre-procedure to the 6-month, 12-month, and 24-month follow-up assessments (p < 0.01). 6.4.6 Stent Integrity Stent fractures were detected in 1.5% of stents (22/1432) at 12 months. Three stent fractures were Type I (single strut fracture), 5 were Type II (multiple strut fractures resulting in complete transection of the stent, without displacement of the stent segments), 4 were Type III (multiple single strut fractures), and 14 were Type IV (multiple strut fractures resulting in displacement of segments of the stent). All stent fractures occurred in separate subjects. Fifteen of the 22 stents with fractures (68%) were associated with quantifiable stent elongation during the implant procedure, including 11 (79%) of the Type IV fractures. 7. SUMMARY OF CLINICAL EVIDENCE The Zilver PTX Drug-Eluting Stent was evaluated in two large clinical studies. The RCT compared treatment of femoropopliteal lesions with the Zilver PTX vs. PTA in lesions up to 140 mm in length. In the RCT, 479 subjects were enrolled, with 306 treated with the Zilver PTX stent. The global registry consecutively enrolled an additional 787 subjects, all of whom were treated with Zilver PTX stents. No limits were placed on lesion length in the registry, and up to 4 overlapping stents were allowed. Combined, the AT population of the two studies consisted of 1092 subjects in whom Zilver PTX stents were placed (Figure 15). Figure 15: Total number of subjects receiving the Zilver PTX stent 7.1. Primary Patency and Event-Free Survival as a Function of Lesion Length The combined primary patency and EFS data from the RCT and the global registry are presented descriptively in Tables 18 - 19 as a function of various lesion lengths. Lesions 7 – 14 cm in length were treated in both studies, while lesions > 14 cm were only treated in the registry. As expected, longer lesions were associated with lower primary patency and EFS rates.
Table 18: Event-free survival as a function of lesion length
Lesion Length Event-Free Survival (# of subjects at risk at time of procedure)
Optimal PTA Bare Zilver Zilver PTX
< 7 cm 80.4% (88) 85.8% (36) 92.7% (543) 7-14 cm 75.3% (26) 86.7% (15) 91.5% (296) > 14 cm 50.0% (4) 100.0% (2) 79.6% (233)
Table 19: Primary patency as a function of lesion length
Lesion Length Primary Patency (# of lesions at risk at time of procedure)
Optimal PTA Bare Zilver Zilver PTX
< 7 cm 70.1% (93) 71.0% (40) 88.7% (616) 7-14 cm 45.0% (27) 70.6% (17) 83.6% (312) > 14 cm1 50.0% (4) 100.0% (2) 68.7% (222)
1 While the RCT was not intended to enroll subjects with lesion lengths > 14 cm, six lesions exceeded this length according to core laboratory measurements
7.2. Stent Integrity
Stent fractures were detected in 4/457 (0.9%) stents in the randomized study and 22/1432 (1.5%) in the registry study, for an overall rate of 1.4% (Table 20). This is consistent with published literature rates. It is important to note that 65% (17/26) stent fractures were in stents that experienced elongation during stent implantation. To minimize the risk of stent fracture due to elongation at implantation, the sponsor proposed that the following wording be included in the Instructions for Use for the Zilver PTX stent:
“NOTE: Care should be taken to hold the hub stationary and to remove any slack in the introducer catheter to ensure the stent is not stretched or compressed lengthwise during deployment (i.e., so that the stent is deployed to its proper length). Full deployment of the stent length will occur when the distal end of the sheath has been retracted past the proximal part of the stent.”
Table 20: Stent fracture rates for both the RCT and registry
The panel will be asked to comment on the significance of the stent fracture results and the adequacy of stent fracture evaluation in the two studies.
7.3. Particulate Effects
The issue of morbidity from drug particulates and from device-related debris would manifest clinically as embolization effects. In the RCT, embolism distal to the study vessel on the day of the procedure or within 30 days post-procedure was reported for one subject receiving a bare metal Zilver stent after acute PTA failure (1.8% or 1/56) and one subject receiving a Zilver PTX stent (0.3% or 1/305). In the registry, embolism was reported for six subjects (0.8%, 6/787), all of whom were successfully treated. Beyond 30 days, embolism distal to the study vessel was reported for three RCT subjects after 184, 717, and 792 days in the RCT, and blue toe syndrome was reported for one RCT subject at both 354 and 471 days. The reported distal embolisms >30 days all occurred in association with reintervention in the study vessel that included stenting or placement of a stent-graft, and the events of blue toe syndrome were not adjudicated as device-related. There were no reported events of distal embolism or blue toe syndrome beyond the day of the procedure in the registry. These rates are within literature-derived values and suggest that the risk of downstream embolization of drug coating is within acceptable levels.
Systemic toxicity due to paclitaxel could manifest as neutropenia, as documented in clinical studies of the chemotherapeutic drug Taxol, in which paclitaxel is the active pharmaceutical ingredient. In these studies, this effect was dose-dependent, with 14% of subjects who received 135 mg/m2 Taxol and 27% of subjects who received 175 mg/m2 Taxol developing neutropenia. Five subjects in the RCT, two with bare metal Zilver stents and three with Zilver PTX stents, and one subject in the Zilver PTX registry study had neutrophil counts indicative of neutropenia (absolute neutrophil count < 1.5 x 1000/mm3). No case of neutropenia was determined to be related to the paclitaxel coating on the Zilver PTX stent or to participation in the study, and the incidence of neutropenia was not increased for subjects receiving a Zilver PTX stent. 7.5. Hypersensitivity Hypersensitivity reactions were commonly reported in clinical trials of the chemotherapeutic drug Taxol, in which paclitaxel is the active pharmaceutical ingredient. These events could also be attributable to the excipients used to manufacture the drug, which are not present in the Zilver PTX stent. There were 33 site-reported cases (28 in the RCT and 5 in the registry) of drug reaction (including contrast reaction) or hypersensitivity allergic reaction in 31 subjects (Table 21). None of the reactions were attributed to the paclitaxel or the stent material. The majority of reactions (21/33) were reported due to the antiplatelet medications ticlopidine, which was allowed under the protocol for Japanese subjects. Table 21: Hypersensitivity reactions 7.6. Stent Thrombosis Rates of stent thrombosis or abrupt stent closure of both the bare Zilver and the Zilver PTX stents at 12 and 24 months were low. In the RCT, 54/56 (96.4%) of subjects receiving the bare Zilver stent after receiving provisional stenting in the PTA control group did not experience stent thrombosis by 12 months, and no subsequent events were observed from 12 to 24 months after implant. Among all subjects who received a Zilver PTX (including those assigned to the Zilver PTX arm of the RCT, treated provisionally with the Zilver PTX in the RCT, and enrolled in the global registry), the KM estimates of freedom from thrombosis was 97.2% and 96.5% at 12 and 24 months, respectively Figure 22. There were no stent thromboses or abrupt stent closures in the RCT subjects who received optimal PTA and therefore did not receive any stent. Figure 16: Kaplan-Meier estimates for freedom from study lesion thrombosis for all Zilver PTX subjects In both studies, the majority of stent thrombosis events occurred while subjects were still on the protocol-recommended anti-platelet regimen (Table 22). Table 22: Stent thrombosis and discontinuation of recommended antiplatelet regimen Study Study Group Total Number of Events Events Occurring while on Protocol-Recommended Antiplatelet Regimen Events Occurring after Discontinuation of Protocol-Recommended Antiplatelet Regimen % of Events Occurring while Still on Protocol-Recommended Antiplatelet Regimen Randomized Control 2 2 0 100% Treatment 8 5 31 63% Single Arm 31 22 92 71% 1 One subject remained on clopidogrel, but not aspirin 2 Three subjects remained on clopidogrel, but not aspirin The results demonstrate comparable thrombosis rates between the uncoated Zilver and coated Zilver PTX stents. In addition, the Zilver PTX thrombosis rates during the first 12 months were higher than during the subsequent 12 months. While the significance of this finding is not clear, the different rates may reflect the elution of the drug and the endothelialization of the implanted stents over time and a subsequent decrease of the risk of these types of events. The panel will be asked to comment on the significance of the stent thrombosis results and to provide recommendations for how information regarding the antiplatelet regimen should be incorporated into the product labeling if approved. 8. FDA PERSPECTIVE REGARDING THE PRE-MARKET DATA To demonstrate the safety and effectiveness of the Zilver PTX Drug-Eluting Stent, the sponsor enrolled subjects in two studies involving a total of 1266 subjects. The randomized, controlled trial (RCT) met its pre-specified primary safety and effectiveness endpoints. Additional sub-analyses comparing various treatment strategies, including provisional Zilver PTX versus provisional bare metal stenting, and Zilver PTX versus standard therapy with optimal PTA or provisional bare metal stenting, also suggested benefit for the Zilver-PTX. The global registry provided additional supporting information regarding the safety of the device and the incidence of rare adverse events, such as stent fracture, thrombosis, aneurysm, and embolization. When evaluating whether the results of these studies support the pre-market safety and effectiveness of the Zilver PTX Drug-Eluting Stent for the proposed indications, the following points should be considered. 8.1. Proposed Indications for Use The sponsor proposes the following indications for use: “The Zilver® PTX® Drug-Eluting Stent is indicated for improving luminal diameter for the treatment of de novo or restenotic symptomatic lesions in vascular disease of the above-the-knee femoropopliteal arteries having reference vessel diameter from 4 mm to 9 mm and total lesion lengths up to 140 mm per limb and 280 mm per patient. To avoid involvement of the common femoral artery, the most proximal stent end should be placed at least 1 cm below the origin of the superficial femoral artery. To avoid involvement of the below-the-knee popliteal artery, the most distal stent end should be placed above the plane of the femoral epicondyles.” One key parameter of the proposed indications is the specification of the total lesion length per subject. Enrollment in the randomized, controlled trial of the Zilver PTX was limited to subjects with up to two lesions of the femoropopliteal arteries, with a maximum individual lesion length of 140 mm. The global registry of the Zilver PTX did not include any limitation on lesion length or number of lesions to be treated, and the mean treated lesion length was 100 ± 82 mm, with 25% of lesions greater than 140 mm in length and 87.2% of subjects receiving treatment of a single lesion. Since data regarding lesions > 140 mm were collected wholly from the global registry, one challenge with interpreting these data is the lack of a prospectively defined comparator, particularly a concurrent control. The data provided in Table 19 above suggest a trend towards lower patency and higher revascularizations with lesion lengths > 140 mm. In addition, the non-clinical animal studies conducted to support IDE approval for the RCT evaluated the safety of implanting Zilver PTX stents overlapped to a length of 140 mm in each of two limbs, using both the clinical drug dose on the stent as well as higher doses (maximum of 3X for this stented length). The FDA review team believes that the indications for use for any device should be supported by available clinical and non-clinical data.
8.2. Stent Thrombosis
Due to the risk of stent thrombosis and other thrombotic events, the RCT protocol specified a standardized dual antiplatelet regimen for all subjects to follow for 60 days post-procedure, followed by aspirin therapy indefinitely. Overall, compliance with the regimen was fairly high, and stent thrombosis rates were 3.4% for all Zilver PTX subjects at 24 months. While the risks and consequences of stent thrombosis have been a topic of clinical interest and research for coronary drug-eluting stenting, they are less well-characterized for peripheral stenting. As a result, there is less established evidence available regarding the clinical significance of the Zilver PTX thrombosis data and the optimal antiplatelet regimen to recommend for Zilver PTX recipients.
8.3. Bailout Stenting
There was a high rate of acute PTA failures in the control arm of the RCT (50.4%). While this may be representative of clinical outcomes during standard PTA procedures in the femoropopliteal arteries in current clinical practice, these acute failures were a main factor in allowing the Zilver PTX to meet its primary effectiveness endpoint of superior primary patency as compared to PTA. The primary patency rates for the two study arms appear to follow similar trends after the procedure.
In addition, among subjects who experienced acute PTA failure and required bailout stenting, primary patency was statistically superior for provisional Zilver PTX stenting as compared to provisional bare metal Zilver stenting. These results suggest a patency benefit for the drug coating of the stent. While this analysis was unpowered and not prospectively defined, the provisional stenting arm assignment of the subjects who failed PTA was random and therefore assignment bias was likely low. These factors should be kept in mind when considering the strength of the conclusions drawn from these analyses.
8.4. Long-Term Follow-Up Availability and Results
While the primary endpoints of the RCT were measured at one year of follow-up, the RCT protocol specified follow-up out to five years and at the time of PMA submission. All subjects were eligible for their two-year follow-up visit and such data were available for approximately 84 % of subjects. The data suggest a sustained primary patency benefit for Zilver PTX as compared to PTA, but given the paucity of high-level data regarding the long-term effects of drug-eluting stents in the femoropopliteal arteries, it is not clear what follow-up duration is adequate to fully assess the safety and durability of Zilver PTX implantation.
8.5. Stent Fracture
High-resolution flat-plate x-rays were taken to detect stent fracture at various time points from discharge to 3 years post-procedure, depending on the study. The overall rate of detected stent fractures across both the RCT and global registry was 1.4%. However, the clinical significance of femoropopliteal stent fractures is still being explored, as is identification of the optimal time points for fracture assessment.
9. POST-APPROVAL STUDY
Note: The inclusion of a Post-Approval Study section in this summary should not be interpreted to mean that FDA has made a decision or is making a recommendation on the approvability of this PMA device. The presence of a post-approval study plan or commitment does not in any way alter the requirements for premarket approval and a recommendation from the Panel on whether the benefits outweigh the risks. The premarket data must reach the threshold for providing a reasonable assurance of safety and effectiveness before the device can be found approvable and any post-approval study could be considered.
The FDA review team has made the recommendation that if the Zilver PTX Drug-Eluting Stent is approved, a post-approval study (PAS) or studies should be required as a condition of approval for this first-of-a-kind device. Through review of the pre-market data, the review team has identified the following areas where collection of post-market data may be appropriate, and recommends that a PAS be conducted to assess these endpoints longer-term (> 1 year):
• Major adverse events as classified in the pre-market study
• Stent thrombosis
• Stent fracture
The sponsor submitted a PAS protocol proposal on June 10, 2011. An overview of the proposed PAS protocol is provided below. The review team’s perspective on the proposal is included in the assessment following the proposal overview. FDA is working interactively with the sponsor to refine the post-approval study. The full PAS protocol submission can be found in Attachment 1 (Tab 9).
9.1. Overview of Sponsor’s Proposed Post-Approval Study
The sponsor has proposed a PAS that consists solely of continued follow-up through 5 years of the RCT subjects who presented with symptomatic stenotic or occlusive atherosclerotic lesions of the above-the-knee femoropopliteal artery. The study will compare the EFS rate in subjects in the Zilver PTX stent group to that of the PTA control subjects.
9.1.1 Study Design Description and Hypothesis
The sponsor does not plan to enroll new subjects for the PAS. The primary endpoint will be tested using the Z-statistic, according to the hypothesis that subjects in the Zilver PTX stent treatment group will have a non-inferior EFS rate at five years compared to subjects in the PTA control group, with a non-inferiority margin of 10%.
The Z-statistic is given by:
o EFSPTA = Kaplan-Meier estimated event-free survival at 60 months post-procedure for the PTA per-protocol cohort
o EFSPTX = Kaplan-Meier estimated event-free survival at 60 months post-procedure for the Zilver PTX per-protocol cohort
o δ = minimum difference of practical interest (in this case 10%)
o SE = standard error , where V(EFSPTX) and V(EFSPTA) are the estimates of the variances of the Kaplan-Meier estimates using the Greenwood formula.
The primary safety null hypothesis will be rejected in favor of the alternative if:
EFSPTX – EFSPTA > Zα SE – δ, for α = 0.05.
Analysis of the primary endpoint will be performed using a Kaplan-Meier analysis on the per-protocol set of subjects, as previously defined in the RCT.
Using Kaplan-Meier estimates of EFS of 77.8% and 73.6% at 60 months for the Zilver PTX stent group and PTA control group respectively, an exact non-inferiority test for two binomial populations, δ =0.10, α = 0.05, and 80% power, the required sample size per treatment group is 113, for a total sample size of 226 subjects. The current rate of subjects lost to follow up or withdrawn from the study is less than 5% percent per year, suggesting that adequate subjects will be available to assess EFS at 5 years.
9.1.2 Data Collection (Endpoints)
The primary endpoint is designed to evaluate event-free survival (EFS) at five years. EFS is defined as in the RCT: freedom from the CEC- adjudicated major adverse events of death, target lesion revascularication, target limb ischemia requiring surgical intervention (bypass or amputation of toe, foot, or leg) or surgical repair of the target vessel (e.g., dissection or perforation requiring surgery), and freedom from worsening of the Rutherford classification by 2 classes or to class 5 or 6.
The secondary endpoints are the rates of stent fracture and late stent thrombosis. These endpoints will be evaluated descriptively and will not be subjected to formal hypothesis testing.
9.1.3 Follow-up Visits and Length of Follow-up
Office visits will be scheduled at 3 , 4, and 5 years for follow up and testing activities including functional status assessment via the Walking Impairment Questionnaire (WIQ) and EQ-5D, bilateral resting ABI (using the baseline protocol) and a Rutherford classification. All stent subjects, and a subset of 50 subjects overall in the PTA group, will be evaluated with duplex ultrasound annually. X-rays will be obtained for all stent subjects at 3 and 5 year clinical visits per the current RCT protocol.
A subject will be considered lost to follow-up only when attempts at contact via telephone, mail, and through the last known contact for the subject has been unsuccessful.
9.2. FDA Perspective on the Proposed Post-Approval Study
The review team believes that the sponsor’s proposal to follow-up the Zilver PTX and control arms of the RCT cohort up to 5 years would provide a sufficient length of long-term follow-up. However, the proposed study design does not provide a sample size adequate to evaluate the incidence of less common events such as stent fracture and late stent thrombosis. The review team believes that the PAS should be statistically powered to analyze these rarer events. While the sponsor calculated a sample size of 113 subjects per arm to assess EFS at 5 years, a total of 900 subjects would be needed to adequately assess the incidence of events such as stent fracture and thrombosis that may occur at a rate of 1 – 2%.
Enrollment of additional subjects from outside the RCT would be necessary to provide this number of subjects, and would also allow for an assessment of the generalizability of the RCT data to the non-RCT patient population. One potential approach to obtaining the necessary sample size would be to re-consent the global registry subjects for longer-term follow-up. However, re-consenting these subjects may be challenging. In addition, one potential risk with this approach would be the introduction of selection bias; that is, the subjects who choose to re-consent for long-term follow-up may not be entirely representative of the global registry cohort as a whole.
The panel will be asked to comment on the design of a PAS should the PMA for the Zilver PTX be approved, particularly with respect to the need to enroll new subjects from outside the RCT and whether re-consenting global registry subjects would provide valid scientific evidence to satisfy the PAS goals.
The data presented in this PMA characterize the safety and effectiveness of the Zilver PTX stent when used to treat patients with symptomatic atherosclerotic stenosis of the femoropopliteal artery. The panel will be asked to fully assess the significance of these results and comment on the risk to benefit ratio of using the Zilver PTX stent to treat these patients.