the dominant mechanism of restenosis in patientsundergoing stent implantation. The proliferation andmigration of vascular smooth muscle cells (VSMCs) areinduced by cytokines such as transforming growth fac-tor β and platelet derived growth factor activated at thesite of PTCR. These medial smooth muscle cells arebelieved to migrate into the intima and result in neointi-mal hyperplasia with cellular proliferation and extracel-lular matrix production. Agents that affect the prolifera-tion and migration of VSMCs may be effective inpreventing restenosis.

Tranilast, N-(3’4’-demethoxycinnamoyl)-anthranilicacid (N-5), inhibits release or production of chemicalmediators, cytokines, and active oxygen transport (byvirtue of its membrane stabilizing action) by variousinflammatory cells and macrophages. It also interfereswith the proliferation and migration of VSMC-inducedby platelet-derived growth factor and transforminggrowth factor β1. The anti-inflammatory effect of trani-last was demonstrated by inhibition of cyclooxygenase-2

Percutaneous transluminal coronary revascularization(PTCR) is widely used for the treatment of ischemicheart disease, but restenosis requiring repeat proce-dures remains problematic, occurring in approximately30% to 40% of patients treated with conventional angio-plasty and 15% to 30% of patients treated with stentimplantation.1-4 The process of restenosis is multifactor-ial and includes elastic recoil, negative remodeling (atleast in nonstented segments), excess matrix formation,and neointimal hyperplasia. Neointimal proliferation is

From the aMayo Clinic, Rochester; bStanford University; cCooper Hospital UniversityMedical Center, Camden; dHospital Cardiologique, Lille; eCleveland Clinic Founda-tion; fJefferson University Hospital, Philadelphia; gHeartcentre, Rotterdam; hUniver-sity of Texas Medical School and Texas Heart Institute, Houston; and iSmithKlineBeecham, Collegeville.Supported by grants from SmithKline Beecham Pharmaceuticals.Submitted July 1, 1999; accepted September 19, 1999.Reprint requests: David R. Holmes Jr, MD, Division of Internal Medicine and Car-diovascular Diseases, Mayo Clinic, 200 First St SW, Rochester, MN 55905.Copyright © 2000 by Mosby, Inc.0002-8703/2000/$12.00 + 0 4/1/103546

The PRESTO (Prevention of Restenosis with Tranilastand its Outcomes) protocol: A double-blind,placebo-controlled trialDavid Holmes, MD,a Peter Fitzgerald, MD,b Sheldon Goldberg, MD,c J-M LaBlanche, MD,d A. Michael Lincoff, MD,e

Michael Savage, MD,f P.W. Serruys, MD,g James Willerson, MD,h Jeffrey R. Granett, MD,i Robert Chan, MD,i NeilH. Shusterman, MD,i and Marcia Poland, MAi Rochester, Minn; Stanford, Calif; Camden NJ; Lille, France; Cleve-land, Ohio; Philadelphia, Pa; Rotterdam, The Netherlands; Houston, Texas; and Collegeville, Pa

Background Tranilast is a unique drug in clinical development for the prevention of restenosis after percutaneoustransluminal coronary revascularization (PTCR). Tranilast interferes with proliferation and migration of vascular medialsmooth muscle cells induced by platelet-derived growth factor and transforming growth factor β1. Collagen synthesis in vas-cular medial smooth muscle cells is inhibited by tranilast, which also inhibits the release or production of cyclooxygenase-2and restores cytokine-induced nitric oxide production. These mechanisms may contribute to the reduction of angiographicrestenosis after coronary intervention previously reported in clinical studies.

Methods The primary objective of this multicenter study of 11,500 patients is to compare the composite clinical eventrate of death, myocardial infarction, or the need for ischemia-driven target vessel revascularization of tranilast (300 and450 mg twice daily) for 1 or 3 months with that of placebo in patients undergoing PTCR with or without stenting for single ormultiple vessels over a 9-month period. The lesions can be de novo or restenotic. All revascularization procedures and theuse of glycoprotein IIb/IIIa agents are permitted. The inclusion criteria are meant to allow an “all comer” approach for gen-eralization of results to the broadest possible PTCR population. A subset population (n = 2000) will undergo 9-month follow-up angiography, 1000 of which will also undergo intravascular ultrasound (n = 1000). This study is the first tranilast trial tobe conducted in a Western population to confirm the improved angiographic findings reported in Japanese patients and todetermine if the clinical sequelae of restenosis are also reduced.

Conclusion This multicenter study is the largest restenosis trial planned to date. It will test whether tranilast, a drug withmultiple actions aimed at affecting proliferation and migration of vascular smooth muscle cells, can reduce clinical, angio-graphic, and intravascular ultrasound assessments of restenosis. (Am Heart J 2000;139:23-31.)

expression induced by interleukin-1β stimulated fibro-blasts.5 In addition, tranilast restores cytokine-inducednitric oxide production.6

In vivo studies of tranilast using the balloon angio-plasty model in rabbits,7-8 rats,9 and the photochemi-cally induced injury model in spontaneously hyperten-sive rats and mice10 have shown a significant reductionin neointimal formation. Tranilast was also found toinhibit neointimal formation and adventitial remodelingin the pig model of balloon restenosis and to inhibitneointima in stents.11

Two double-blind, placebo-controlled trials12-13 con-ducted in Japan with quantitative coronary angiographyhave shown tranilast 600 mg/day to be significantlymore effective than placebo in reducing angiographicrestenosis (relative reductions 58% and 51%; P < .001).In these trials, tranilast was started on the day after theintervention. In stented patients treated with tranilast600 mg/day, restenosis rates were also improved com-pared with stent implantation alone (P < .05).14 The rel-ative reduction in the rate of restenosis with tranilastplus stenting was 42% when compared with stent alone.

Although the first double-blind, placebo-controlledtrial was not powered to detect a difference in clinicalevents, 65 placebo-treated patients and 59 tranilast-treated patients were followed up for 1 year (KisseiPharmaceutical Co, personal communication, 1998).The frequency of adverse cardiovascular events was15% in patients treated with tranilast and 23% forpatients treated with placebo (a relative reduction of34%, not statistically significant). In patients undergoingdirectional coronary atherectomy,15 the relative reduc-tion in the 1-year event rate was 60% tranilast withcoronary atherectomy versus coronary atherectomyalone (12.5% vs 31.6%; P < .03).

Based on the results obtained in these trials, tranilastis hypothesized to be effective in reducing the fre-quency of cardiovascular events and angiographicrestenosis in patients undergoing PTCR, but the effec-tive dose in Western populations is not known. ThePRESTO study is designed to determine the safety andefficacy of tranilast in a Western population.

Study objectivesThe primary objective is to compare the composite

clinical event rate (death, myocardial infarction [MI], orneed for ischemia-driven target vessel revascularization)of tranilast (300 mg and 450 mg twice daily) for 1 or 3months of treatment with that of placebo over an obser-vation period of 9 months in patients undergoing PTCR.

There are 5 secondary objectives, the first 2 of whichare subset populations. (1) To compare angiographicparameters at 9 months after the procedure. The princi-pal comparison will be minimal luminal diameter; sec-ondary comparisons are late loss; percent of patientsand lesions with ≥50% narrowing of at least 1 treated

vessel and those with ≥50% loss of initial treatmentgain; and the ratio of late loss to initial gain. (2) To com-pare the morphologic and morphometric findings fromintravascular ultrasound (IVUS) of the target and refer-ence segment(s) in stented and nonstented patients at 9months after PTCR. The principal comparison will beneointimal plaque volume within the stented segment(s)at follow-up; secondary comparisons will include meanluminal area; vessel area; plaque area; and plaque composi-tion. This two-dimensional analysis will be performed inboth stented and nonstented patients. (3) To assess safety.(4) To compare the cost effectiveness (ie, estimation ofthe incremental cost per clinical end point averted). (5)To evaluate the population pharmacokinetics.

Patient selectionInclusion criteria

Men and women more than 18 years of age undergo-ing elective or emergency PTCR for single or multiplevessels are included. All noninvestigational proceduresare permitted (eg, directional coronary atherectomy,rotational atherectomy, laser), as are all types and sizesof lesions (eg, de novo or restenotic, native coronaryartery or vein graft lesions) and all noninvestigationalstents. Women must be postmenopausal, surgicallysterilized, or using a double-barrier method of birthcontrol. Concomitant treatment with glycoproteinIIb/IIIa agents, ticlopidine, or clopidogrel is permitted.

Exclusion criteriaThe exclusion criteria can be divided into 3 cate-

gories: (1) procedural, (2) concurrent illnesses, or (3)potential for medication interactions. The latter 2 cate-gories are based on the reported adverse effects oftranilast or contrast dye. Tranilast has been reportedto increase serum levels of liver enzymes16 and todecrease uric acid to below the lower limit of normal.In addition, tranilast inhibits the liver cytochrome P450CYP2C9 isozyme and has been reported to interferewith hemostasis when coadministered with warfarin.In patients receiving metformin, treatment will be with-held after PTCR until radiocontrast dye has been com-pletely excreted.

Procedural exclusions include (1) non-Q-wave MIwithin 48 hours before PTCR; (2) preplanned, stagedPTCR, (3) PTCR in the 3 months before study entry ona vessel that is not being revascularized at time of theentry procedure, and (4) radiotherapy or investigationalstents in any vessel within 9 months before study entry.

Exclusions for concomitant conditions include (1)chronic liver disease and/or liver enzyme levels ≥1.5times the upper limit of normal; (2) history of uric acidstones, gout, or the need for uric acid surveillance; (3)serum creatinine level of ≥1.8 mg/dL within 2 weeks ofstudy entry; (4) hemoglobin <10 g/dL within 2 weeksof study entry; (5) alcohol and/or drug abuse in the past

American Heart JournalJanuary 2000Holmes et al24

6 months; (6) severe concomitant illness that wouldpreclude the patient from completing a 9-month trial;(7) a history of nephrotoxicity from radiocontrast dye;and (8) pregnancy or breast feeding.

Exclusions because of concomitant medications includethe use of (1) warfarin or dicoumarol, (2) phenytoin, (3)investigational drugs within 30 days or 5 half-lives of theirlast dose before initiating study drug (whichever islonger), (4) patients with a history of exposure to trani-last, and (5) patients who cannot tolerate hydration with150 mL/h during the first 4 hours after PTCR.

Study designThis is a double-blind, randomized (1:1:1:1:1 ratio),

placebo-controlled, 9-month trial (Figure 1). Patientswho have a successful PTCR, defined as at least 1 lesiondilated to <50% residual stenosis without the complica-tion of death, MI, or need for CABG before the firstdose of study medication, are eligible for the study. Ifabrupt closure occurs a second PTCR is permissiblebefore dosing. Randomization is being carried out withan interactive voice randomization system. The investi-gational site will telephone the randomization systemto obtain the study medication number to assign toeach patient and will also register whether the patientis in the angiographic and/or IVUS subsets. This study isbeing conducted in accordance with Good ClinicalPractice, the Declaration of Helsinki, and US 21 Code ofFederal Regulations part 50, Protection of Human Sub-

jects, and part 56, Institutional Review Boards. The pro-tocol and informed consent have been approved byeach investigator’s institutional review board before ini-tiation of the study.

Patients will receive study medication twice daily inas close to 12-hour intervals as possible. The first doseof study medication will be administered 4 to 8 hoursafter completion of the successful index PTCA as longas the patient has voided at least once during that time.Every effort will be made to give the patient 2 doses onthe day of the procedure. Tranilast doses are 300 mgand 450 mg twice daily for 3 months versus the samedoses for 1 month followed by 2 months of placebo.Medication is blinded by overencapsulating each tabletof tranilast or placebo.

All visits are required for all patients who are alivebecause more than one clinical event can occur in the9-month follow-up. The visit intervals are shown inFigure 2. The patient is to be seen by the investigatoror subinvestigator at the end of 1 and 3 months of treat-ment and again at 9 months. All other study visits maybe performed by the investigator’s designate for labora-tory test blood draws, which are sent to a central labo-ratory, and by study personnel telephoning the patientfor updated clinical history, including assessment ofclinical outcome. Routine laboratory testing andinterim medical history is required at each visit duringthe 3-month treatment period and once in the 6-monthfollow-up period.

American Heart JournalVolume 139, Number 1, Part 1 Holmes et al 25

Figure 1

Study design. CABG, Coronary artery bypass grafting; TVR, target vessel revascularization; bid, twice daily.

The primary efficacy variable is the composite clinicalend point (Figure 3). This primary parameter will beanalyzed for subset populations based on demographicvariables, target vessel variables, and procedural vari-ables (Figure 4). Other secondary efficacy parametersinclude (1) individual components of death (all causes),MI, or all target vessel revascularizations within 9months; (2) cardiovascular death within 9 months; (4)need for any revascularization (ie, either for signs andsymptoms of recurrent ischemia and/or angiographicfindings); (5) quantitative coronary angiographic (QCA)results; and (6) IVUS results.

QCA Core laboratories

Angiograms performed in the United States andCanada will be sent for QCA to Thomas Jefferson Uni-versity Hospital, Coronary Angiography Laboratory;those performed in the rest of the world will be sent toLille CORE Laboratory in France. Cross-validation ofpreselected angiograms before the initiation of PRESTOwas performed to ensure that the QCA readings werenot different in the 2 laboratories. Both laboratoriesperform QCA with the Cardiovascular MeasurementSystem (Medis Medical Imaging Systems, Ridgefield,Conn).17,18

MethodsTo ensure that 400 patients per treatment group with

both a baseline and follow-up angiogram are enrolled,an additional 25% of patients will be asked to give con-sent for a 9-month angiogram. Acceptable angiogramsmust be on 35-mm cine x-ray film or CD-ROM. Eachpatient is to receive 100 to 200 mcg of intracoronarynitroglycerine before all required films. Molsidomine isalso permitted, but 2 minutes must elapse before film-ing. Before the scheduled 9-month angiogram, investi-gators must document whether there are clinical signsor symptoms of ischemia warranting revascularization.If an angiogram is required at <4 months and the base-line target vessel is revascularized, a repeat angiogramwill not be required; if the target vessel is not revascu-larized, a repeat 9-month angiogram is required.

Quantitative and qualitative angiography assessmentswill be performed by the core laboratories. Qualitativeassessment for all patients will be performed by the inves-tigators as well. For the nonangiographic patient subset,investigators will estimate the percent of residual stenosisat baseline for all revascularized vessels and identify alldiseased vessels with <30% stenosis. In addition, thesenonangiographic sites will send baseline and target vesselrevascularization angiograms to the Core Laboratories tovalidate the target lesion and vessel revascularization.Qualitative assessment includes the following: identifyingthe site of restenosis; specific segment; American Collegeof Cardiology/American Heart Association score and eachfeature component of that score; Thrombolysis in Myocar-dial Infarction flow grade; presence of thrombus; and the

American Heart JournalJanuary 2000Holmes et al26

Figure 2

Follow-up chart for all patients who have received at least 1 dose of medication (ie, intention-to-treat population).ECG, electrocardiogram.

dissection grade. In addition, the stent number, type,aggregate length, and diameter will be recorded.

Baseline QCA of a 20% sample of nonangiographic subset

To assess whether the 2000 angiographic patient sub-set is representative of the whole population, the corelaboratories will perform QCA on a random sample(20%) of 9500 patients who are not scheduled for a 9-month repeat angiogram.

Quantitative coronary ultrasoundmethodCore laboratories

IVUS studies performed in the United States and Canadawill be sent to Stanford University IVUS Laboratory, andstudies performed in the rest of the world will be sent toCardiolysis Laboratory in The Netherlands. Cross valida-tion of quantitative coronary ultrasound results before theinitiation of PRESTO was undertaken to ensure that QCAreadings were not different in the 2 core laboratories.

MethodsOne hundred patients per treatment group who par-

ticipate in the 9-month angiography substudy will alsohave an IVUS performed. Motorized pullback is pre-ferred during the baseline IVUS but is mandatory at thefollow-up IVUS study. Interventionalists will place thetip of the IVUS catheter at least 10 mm distal to eachintended target lesion; pullback (0.5 mm/s) is initiatedfrom the distal reference segment through each targetlesion and includes at least 10 mm of proximal refer-ence segment. Start/stop times are recorded. The IVUSaudio dialogue consists of documenting the catheterposition, pullback speed, start time, and pullback timeas well as coronary segment and the end of the pull-back time. All pullback sequences will be performedwith the calibration marker off. A recording of the cali-bration will be performed before the case and thezoom setting should not change throughout the pull-back acquisition. When performing the follow-up IVUSin a stented vessel, the tip of the catheter is placed 10mm beyond the distal stent border; motorized pullbackis at a speed of 0.5 mm/s from distal reference segmentthrough the stent to at least 10 mm proximal to thestent edge.

The core laboratories will evaluate each IVUS slice byquantitative and qualitative parameters. The quantitativemeasurements include volumetric measurements of thelumen, vessel, and plaque computed according to Simp-son’s rule (Volume = Sum of cross-section of plaqueareas within the stent) or as a volume index in non-stented segments. Plaque vessel and stent volumes arecomputed by integrating area measurements over thestent length. Area-length plots will be tracked for each

patient at follow-up to study the distribution of plaqueas a function of stent and reference segment length. Thequalitative analysis includes the following in both thelesion and the reference segment: dominant plaque typein each lesion and proximal to the lesion (fibrofatty,fibrous, and calcific); calcium (deep and superficial tothe intima) will be recorded as the maximum arc.

Population pharmacokineticsApproximately 1000 patients (200 per treatment

group) worldwide will have blood drawn for populationpharmacokinetics beginning 1 week after the start ofdouble-blind medication and concluding after 1 monthof treatment. Four samples will be taken over a 12-hourperiod from each patient. The individual plasma concen-tration time data will be pooled to perform populationpharmacokinetics. The influence of various covariates(such as race, age, weight, smoking status, renal func-tion, and sex) on the pharmacokinetic variables will beexamined. An exploratory analysis to examine the rela-tion between the pharmacokinetic variables and efficacyand/or adverse experiences may also be performed.

American Heart JournalVolume 139, Number 1, Part 1 Holmes et al 27

Figure 3

Primary efficacy variable: composite clinical events. ULN,Upper limit of normal; CABG, coronary artery bypass grafting;TVR, target vessel revascularization; ECG, electrocardiogram.

Other assessments In the angiographic subset of patients, C-reactive pro-

tein will be measured before the initial dose of studymedication, 2 days later, and after 1 and 3 months oftreatment. This subset will also have a lipid profile (totalcholesterol, calculated LDL cholesterol, HDL cholesterolby the ultracentrifugation method, and triglycerides)before the first dose and after 1 month of treatment.

Data analysisSample size justification

Determination of the sample size is based on theassumption that the 9-month placebo composite clini-cal event rate is 18% and a 30% relative reductionamong any of the 4 tranilast groups is clinically desir-able. The 18% was derived from trials of both stentedand nonstented patients considered at high and lowrisk for an event.1,2,19 It is expected that approximately70% to 80% of patients will receive at least 1 stent andthe population will contain a high number of at-riskpatients and high-risk lesions. Randomizing 2300patients to each treatment group will provide 90%power to detect a 30% relative reduction (ie, a reduc-tion from 18% to 12.6%) among any or all tranilastgroups with 2-sided log-rank tests at familywise α of.00125. The modified Bonferroni procedure, as out-

lined in Hochberg,20 will be used to adjust for multipletreatment comparisons.

The angiographic patient subset sample size is basedon minimal lumen narrowing. The minimum clinicallyrelevant treatment difference is 0.2 mm and an SD of0.7 mm is assumed. An angiographic subset of 400patients per arm will have a 93% power to detect thedesired treatment difference at familywise α of .05.

The IVUS patient subset sample size is based on aclinically relevant reduction in plaque volume betweentranilast and placebo of 20% in the stented vessel andan expected mean (±SD) plaque volume (untreated) of41.91 ± 17.67. This mean was based on data from Stan-ford University Core Laboratory and confirmed by Car-diolysis data. The study has the power to detect a 20%difference with 80% power at the .05 level of signifi-cance. Nonstented vessels will also be analyzed.

Methods of analysisThe primary efficacy variable will be analyzed by eval-

uating the time to the first occurrence of any of the 3components of the composite. The 9-month cumulativerisk will be estimated for each treatment group by theKaplan-Meier method. Significance levels for pairwisecomparisons will be derived from stratified log-ranktests. Cox’s proportional hazards models (with center

American Heart JournalJanuary 2000Holmes et al28

Figure 4

Hazard ratios (95% confidence intervals) will be calculated for the primary efficacy parameter for each of these subsets in patients orlesions, as appropriate.

and treatment group in the model) will be used to calcu-late hazard ratios and associated 95% confidence inter-vals for each pairwise comparison. In addition, multi-variate logistic regression will be used to control for thepotential effect of the variables listed in Figure 4.

Secondary efficacy analyses of the individual compo-nents of the primary end point will be analyzed by χ2

or Fisher exact test and odds ratios with 95% confi-dence intervals. Cardiovascular death and the need forany revascularization (ie, both for signs and symptomsof ischemia and/or angiographic findings) (CABG orrepeat PTCR of the target vessel) will be subjected to asimilar analysis.

Planned interim analysis by theStatistical Data Analysis Center ofthe Data Safety Monitoring Board

Formal interim analyses will be performed after 33%,66%, and 100% of the planned number of patients havebeen randomized. Interim analyses of the primary endpoint (for the possibility of early termination from evi-dence of treatment harm) and all-cause mortality (forthe possibility of early termination from evidence oftreatment benefit) will use log-rank tests without strati-fication. Early termination for treatment benefit will beconsidered if the statistical significance is at familywiseα of .00001 (type I error of 0.0012 at the final analysis)and for treatment harm at an α .018 (type I error of .05at the final analysis).

Health economic assessmentsMethods

The pharmacoeconomic analysis will examine differ-ences between treatment groups in (1) intended effectof treatment (avoidance of composite clinical events),and (2) unintended effects (hospitalizations, physicianvisits, and prescribed medications for adverse events).The primary economic evaluation is that of cost-effec-tiveness (estimation of the incremental cost of tranilasttherapy compared with placebo per end point eventaverted including deaths and event-free survival).

Health economic statistical analysisHealthcare use will be compared by univariate analy-

sis of the following: (1) number of adverse eventsrequiring hospitalization or medical treatment; (2) totalnumber of hospital admissions for treatment of clinicalevents; (3) duration of hospital stay for treatment ofeither clinical end point events or adverse events; (4)number and type of revascularization procedures per-formed; and (5) total admissions to rehabilitation andskilled nursing facilities. The costs of treating each com-ponent of the composite clinical event will be esti-mated with cost prediction models based on data onhospital resource use and cost. Because some variation

in costs may be related to baseline characteristics ofstudy subjects, multivariate analysis will be performedfor all cost estimates. Covariates will be included inthese analyses if they differ significantly between treat-ment groups or if they are deemed a priori to be impor-tant determinants of cost (such as patient age andcomorbidity as measured by the Charlston index21).

Study organizationSteering Committee

The role of the Steering Committee is to direct andprovide academic leadership with respect to the overallconduct of the PRESTO trial, including interacting withthe Data Safety Monitoring Board and Clinical EventCommittee. The Steering Committee will review publi-cations and consider requests for substudies related tothe PRESTO protocol. A complete list of members is inthe Appendix.

Data Safety Monitor BoardAn independent Data Safety Monitoring Board will

review safety data on a continual basis and make recom-mendations to the sponsor and the Steering Committeeregarding the future conduct of the study. The eventscomprising the composite clinical end point will not beconsidered serious adverse events requiring expeditedreporting to regulatory authorities. A Statistics and DataAnalysis Center will make the following data available tothe Data Safety Monitoring Board by treatment group ina double-blind manner (ie, labeling groups A through Ewhere designation is not in ascending order and theorder is randomly rotated after each formal review):demography, laboratory results, deaths, serious adverseexperiences (including the primary clinical end point),withdrawals because of adverse experiences, concomi-tant medications, and study drug compliance. A com-plete list of membership is in the Appendix.

Clinical Events CommitteeAn independent Clinical Events Committee chaired

by Kenneth W. Mahaffey, MD, from Duke ClinicalResearch Institute will review all events that constitutethe composite clinical end point. In addition, the Clini-cal Events Committee will confirm the cause of deathas cardiovascular or noncardiovascular as well as con-firm that no serious adverse experience reported to thesponsor is a clinical event. When there is a disagree-ment between the committee and the site principalinvestigator, the Clinical Events Committee will discussthe reason for disagreement with the principal investi-gator before rendering a final decision. If agreementcannot be reached, the Clinical Events Committee deci-sion will be considered final. In the event that a deathor other serious adverse experience was considered bythe investigator to constitute a clinical event and thecommittee determines that it is not, the committee will

American Heart JournalVolume 139, Number 1, Part 1 Holmes et al 29

notify the sponsor. The sponsor will notify the DataSafety Monitoring Board Statistics and Data AnalysisCenter and the investigator. This event will be desig-nated a serious adverse experience.

DiscussionRestenosis remains a major problem for interven-

tional cardiology despite the technical advances thathave occurred in the field, such as development andwidespread use of intracoronary stenting. It is a particu-lar problem for treatment of more complex lesions,such as diffuse disease or small vessels and patientswith multivessel diseases. When restenosis occurs, itoften requires subsequent evaluation and treatmentnegating some of the early economic benefits of PTCRcompared with coronary artery bypass surgery. Pharma-cologic approaches to prevention have in general beeneither unsuccessful or have been awkward to deliver(eg, probucol,22 which although effective, required 1month of pretreatment).

The PRESTO trial is focused on testing the safety andefficacy of tranilast. This trial has several advantages:(1) the drug has salutory effects on pathophysiologicmechanisms of restenosis; (2) the drug has been foundeffective in the prevention of angiographic restenosis insmall randomized trials; (3) the drug can be startedafter a successful PTCR; and (4) the patient and lesioncriteria are very inclusive, thus representing the “realworld” of interventional cardiology. As the largest trialfor prevention of restenosis in interventional cardiologywith the largest core laboratory IVUS and angiographicsubstudies, it will expand the knowledge base in inter-ventional cardiology and should help in attempts tooptimize the care of patients undergoing PTCR.

We acknowledge the statistical support fromSmithKline Beecham: Nicole Blackman, MS, statistics;Rezaul K. Khandker, PhD, and Michael J. Lacey, MS,health economics; Robert Z. Harris, PhD, pharmacoki-netics; and Patricia L. Adams, MS, and Trish McBride,BSRN, for their operational support of the protocol.

References1. Serruys PW, deJaegere M, Kiemaeneig F, et al for the Benestent

Study Group. A comparison of balloon-expandable-stent implanta-tion with balloon angioplasty in patients with coronary artery dis-ease. New Engl J Med 1994;331:489-95.

2. Fischman DL, Leon MB, Baim DS, et al for STRESS (Stent RestenosisStudy Investigators). A randomized comparison of coronary-stentplacement and balloon angioplasty in the treatment of coronaryartery disease. N Engl J Med 1994;331:496-501.

3. Nobuyoshi M, Kimura T, Nosaka H, et al. Restenosis after success-ful coronary angioplasty: serial angiographic follow-up of 229patients. J Am Coll Cardiol 1988;12:616-23.

4. Savage MP, Fischman DL, Rake R, et al. Efficacy of coronary stentingvs balloon angioplasty in small coronary arteries. J Am Coll Cardiol1998;31:307-11.

5. Inoue H, Oshima H, Kono H, et al. Suppressive effects of tranilast onthe expression of inducible cyclooxygenase (COX2) in interleukin-1β-stimulated fibroblasts. Biochem Pharmacol 1997;53:1941-4.

6. Hishikawa K, Nakaki T, Hirahashi J, Marumo T, Saruta T. Tranilastrestores cytokine-induced nitric oxide production against platelet-derived growth factor in vascular smooth muscle cells. J Cardiovas-cular Pharmacol 1996;28:200-7.

7. Fukuyama J, Ichikawa K, Miyazawa K, et al. Tranilast supprerssesintimal hyperplasia in the balloon injury model and cuff treatmentmodel in rabbits. Jpn J Pharmacol 1996;70:321-7.

8. Fukuyama J, Ichikawa K, Hamano S, Shibato N. Tranilast suppressesthe vascular intimal hyperplasia after balloon injury in rabbits fed ahigh cholesterol diet. Eur J Pharmacol 1996;318:327-32.

9. Ichikawa K, Fukuyama J, Miyazawa K, et al. Suppressive effect ofan anti-allergic drug, tranilast, on the vascular intimal thicking byballoon catheter. Pharmacometrics 1995;50:539-45.

10. Kikuchi S, Umemura K, Kondo K, Nasashima M. Tranilast suppressesintimal hyperplasia after photochemically induced endothelial injuryin the rat. Eur J Pharmacol 1996;295:221-7.

11. Ishiwata S, Chronos NAF, Lin G, et al. Tranilast inhibits neointimalformation in balloon-injured and stented pig coronary arteries anddecreases negative remodelling after injury [abstract]. J Am CollCardiol 1998;31:144A.

12. Tamai H, for the Treat Group. Inhibitory effect of tranilast onrestenosis after percutaneous transluminal coronary angioplasty(PTCA). Results of a late phase II clinical placebo controlled doubleblind study. J Clin Exp Medicine 1995;175:577-94.

13. Tamai H, et al. Clinical evaluation of tranilast in the prevention ofrestenosis after percutaneous transluminal coronary angioplasty(PTCA). A phase III multicenter randomized double blind placebo-controlled trial. J Clin Therap Med 1996;12:65-85.

14. Hsu Y-S, Tamai H, et al. Efficacy of of tranilast on restenosis aftercoronary stenting [abstract]. Circulation 1996;94 (suppl):I-620.

15. Kosuga K, Tamai H Uedo K, et al. Effectiveness of tranilast onrestenosis after directional coronary atherectomy. Am Heart J1997;134:712-8.

16. Package insert of Rizaben (tranilast). Toyko: Kissei Pharmaceutical; 1998.17. The EPISTENT Investigators. Randomised placebo-controlled and

balloon-angioplasty—controlled trial to assess safety of coronarystenting with use of platelet glycoprotein IIb/IIIa blockade. Lancet1998;352:87-92.

18. Hausleiter J, et al. Comparison of different quantitative coronaryanalysis systems: ARTREK, CAAS, and CMS. Cathet CardiovascDiagn 1996;37:14-22.

19. Reiber JHC, Serruys PW, editors. Progress in quantitative coronaryarteriography. Dordrecht: Kluwer Academic; 1994.

20. Hockberg Y. A sharper Bonferroni procedure for multiple tests ofsignificance. Biometrika 1988;75:800-2.

21. Charlson ME, Pompei P, Ales KL, et al. A new method of classifyingcomorbidity in longitudinal studies: development and validation. JChron Dis 1987;40:373-83.

22. Tardif J-C, Cote G, Lesperance J. Probucol and multivitamins in theprevention of restenosis after coronary angioplasty. New Engl JMed 1997;337:365-72.

AppendixSteering Committee members (excluding the authorswho are also Publication Subcommittee members)

Barry R. Davis, MD, PhD, University of Texas Schoolof Public Health, Houston; James J. Ferguson, MD, Texas

American Heart JournalJanuary 2000Holmes et al30

Heart Institute, Houston; Lars Grip, MD, GothenburgUniversity, Sahlgrens Hospital, Gothenburg, Sweden;Spencer B. King III, MD, Emory University Hospital,Atlanta, Ga; Jeffrey Popma, MD, Brigham and Women’sHospital, Boston; Leonard M. Shapiro, MD, PapworthHospital, Cambridge, United Kingdom; James E. Tcheng,MD, Duke University, Durham, NC; Paul Yock, MD,Stanford University Medical Center, Calif; A.M. Zeiher,PhD, Med Klinik IV Kardiologie, Frankfurt, Germany

Data Safety Monitoring BoardCardiologists: Eric Topol, MD, Committee Chair,

Cleveland Clinic, Ohio; M. Bertrand, MD, HopitalCardiologique, Lille, France

Hepatologist: Willis C. Maddrey, MD, University ofTexas Medical Center at Dallas

Nephrologist: J. David Wallin, MD, Louisiana StateUniversity, New Orleans

Hematologist: A. Koneti Rao, MD, Temple UniversitySchool of Medicine, Philadelphia, Pa

Statistician: Frank E. Harrell Jr, PhD, University ofVirginia School of Medicine, Charlottesville

Statistics and Data Analysis Center: Shelly Sapp, MS,Cleveland Clinic, Ohio

American Heart JournalVolume 139, Number 1, Part 1 Holmes et al 31

AVAILABILITY OF JOURNAL BACK ISSUES

As a service to our subscribers, copies of back issues of the American Heart Journal for the preceding 5 years are maintainedand are available for purchase from Mosby until inventory is depleted. The following quantity discounts are available: 25% offon quantities of 12 to 23, and one third off on quantities of 24 or more. Please write to Mosby, Inc., Subscription Services,11830 Westline Industrial Drive, St. Louis, MO 63146-3318, or call (800)453-4351 or (314)453-4351 for information onavailability and prices of particular issues. If unavailable from the publisher, photocopies of complete issues may be purchasedfrom Bell & Howell Information and Learning, 300 N. Zeeb Rd., Ann Arbor, MI 48106, (734)761-4100 or (800) 521-0600.