|
|
||||||||||
|
J Am Coll Cardiol, 2006; 48:15-20, doi:10.1016/j.jacc.2006.02.055 © 2006 by the American College of Cardiology Foundation |
,c
,d
,a
,a
,a
,a
* Henry Ford Heart and Vascular Institute, Detroit, Michigan
University of Alberta, Edmonton, Alberta, Canada
National Institute of Cardiology, Warsaw, Poland
Procter & Gamble Pharmaceuticals, Cincinnati, Ohio
|| Montreal Heart Institute, Montreal, Quebec, Canada
¶ Baylor College of Medicine, Houston, Texas
# Cleveland Clinic Foundation, Cleveland, Ohio.
Manuscript received September 16, 2005; revised manuscript received January 6, 2006, accepted January 9, 2006.
* Reprint requests and correspondence: Dr. W. Douglas Weaver, Henry Ford Heart and Vascular Institute, Henry Ford Hospital (K-14), 2799 West Grand Boulevard, Detroit, Michigan 48202. (Email: wweaver1{at}hfhs.org).
Preliminary findings of this study were presented as Oral Abstract/Late Breaking Sessions of the 54th Annual Scientific Sessions of the American College of Cardiology, Orlando, Florida, March 7, 2005.
| Abstract |
|---|
|
|
|---|
BACKGROUND: PG-116800 is an oral MMP inhibitor with significant antiremodeling effects in animal models of MI and ischemic heart failure.
METHODS: In an international, randomized, double-blind, placebo-controlled study, 253 patients with first ST-segment elevation MI and ejection fraction between 15% and 40% were enrolled 48 ± 24 h after MI and treated with placebo or PG-116800 for 90 days. Major efficacy end points were changes in LV volumes as determined by serial echocardiography, and clinical and safety outcomes were also collected.
RESULTS: In total, 203 patients (80%) completed 90 days of treatment and had evaluable baseline and 90-day echocardiograms. The proportion of patients with anterior MI (78% vs. 81%) and primary percutaneous coronary intervention (90% vs. 91%) along with baseline LV ejection fraction (35.5% vs. 36.8%) did not differ between PG-116800-treated and placebo-treated patients. There was no difference in the change in LV end-diastolic volume index from days 0 to 90 with PG-116800 versus placebo (5.09 ± 1.45 ml/m2 vs. 5.48 ± 1.41 ml/m2, p = 0.42). Changes in LV diastolic volume, LV systolic volume, LV ejection fraction, sphericity index, plus rates of death or reinfarction were not significantly improved with PG-116800. PG-116800 was well tolerated; however, there was increased incidence of arthralgia and joint stiffness without significant increase in overall musculoskeletal adverse events (21% vs. 15%, p = 0.33).
CONCLUSIONS: Matrix metalloproteinase inhibition with PG-116800 failed to reduce LV remodeling or improve clinical outcomes after MI.
| ||||||||
Matrix metalloproteinases (MMPs) are a family of zinc-dependent proteolytic enzymes that promote ECM degradation (6,7). Enhanced activity of MMPs has been directly associated with cardiac ECM degradation and LV remodeling, leading to progressive heart failure in animal models and human patients (610). Both MMP levels and enzyme activity are increased after MI, suggesting their influence on LV remodeling (1116). Matrix metalloproteinase inhibition prevents LV remodeling and improves systolic function in experimental models of MI (1721) and heart failure (2224).
Matrix metalloproteinase inhibitors have been evaluated clinically for a number of indications, including cancer, osteoarthritis, and rheumatoid arthritis. These studies have reported dose- and duration-dependent musculoskeletal side effects (arthritis, tendon nodules, decreased joint range of motion) termed the "musculoskeletal syndrome" that is generally reversible on cessation of the study drug (25,26). The precise etiology and mechanism(s) involved in these musculoskeletal adverse events are unknown; however, broad-spectrum MMP inhibition and particularly MMP-1 (collagenase) inhibition has been hypothesized to be causally related.
PG-116800 is an oral MMP inhibitor of the hydroxamic acid class with high affinity for MMP-2, -3, -8, -9, -13, and -14 and low affinity for MMP-1 and -7, which might produce cardiac antiremodeling benefits without musculoskeletal side effects. We hypothesized that selective MMP inhibition with PG-116800 may attenuate post-MI LV remodeling. Accordingly, we conducted a phase II, international, multicenter, randomized clinical trialthe PREMIER (Prevention of Myocardial Infarction Early Remodeling) trialto determine the effects of PG-116800 on LV remodeling parameters and clinical outcomes.
| Methods |
|---|
|
|
|---|
Eligible patients were ages 18 to 80 years, with ST-segment elevation MI yielding elevated cardiac markers (creatine kinase-MB isoenzyme, troponin I or troponin T >3 times the upper limit of normal) and LV ejection fraction
15% and
40% as measured by echocardiography at the local center 48 ± 24 h after MI. Patients receiving primary percutaneous intervention (PCI), fibrinolysis, or no reperfusion therapy were all eligible. Exclusion criteria were prior MI, documented or suspected history of heart failure or depressed LV ejection fraction, severe renal insufficiency (creatinine >2.5 mg/dl), congenital heart disease, autoimmune or connective tissue disease, chronic substance abuse or psychiatric illness, hepatic impairment/elevated liver chemistries (alananine aminotransferase, aspartate aminotransferase, gamma-glutamyl-transpeptidase, or bilirubin >1.6x the upper limit of normal), uncontrolled hypertension (blood pressure >160/100 mm Hg), or severe blood dyscrasias (platelet count <100,000; hemoglobin <11 g/dl for men or <10 g/dl for women; or absolute neutrophil count <1,000/µl). In addition, patients were excluded if there was persistent cardiogenic shock, refractory pulmonary edema, or hemodynamic instability (blood pressure <90/50 mm Hg). Patients receiving cytochrome P450 inducers (carbamazepine, phenobarbital, or phenytoin) or P450 inhibitors (fluvoxamine, ketoconazole, itraconazole, or fluconazole) were also excluded.
Patients meeting the above criteria provided informed consent, had baseline blood and urine specimens collected, were randomized to either PG-116800 or placebo, and received the study drug at the earliest possible time. Follow-up assessments were scheduled at hospital discharge/day 7 and on days 14, 30, 45, 60, 75, 90, 120, and 180. These visits included detailed inquiry of any musculoskeletal symptoms, assessment of upper extremity range of motion (goniometry), and examination of hand tendon abnormalities. Clinical events were collected and recorded from enrollment through 90 days, whereas vital status and adverse events were collected 180 days after enrollment.
Baseline echocardiograms were performed 48 ± 24 h after acute MI and sent to the Cleveland Clinic Echocardiography Core Laboratory for later LV dimension and ejection fraction quantification. Echocardiograms were obtained serially at days 14, 30, 90, 120, and 180 and forwarded to the Echocardiography Core Laboratory for blinded interpretation and analysis. Site investigators and sonographers received detailed training to obtain optimal standardized echocardiographic images, and perfluorocarbon-based contrast agents were used for LV cavity opacification. At the Echocardiography Core Laboratory, all echocardiograms were reviewed by two teams of a sonographer plus cardiologist to select the optimal image (noncontrast- vs. contrast-enhanced) for endocardial border definition. The echocardiogram with optimal endocardial border definition was selected as the primary data source for LV dimensions, volumes, and ejection fraction. Internal review of the first 100 paired echocardiograms (baseline, day 30, and day 90) was performed and verified no regional variability in echocardiography image quality and <10% interobserver variability in LV dimension measurements.
Study end points.
The primary study end point was the change from baseline at 90 ± 5 days in LV end-diastolic volume index (
LVEDVI) as measured by two-dimensional echocardiography. Secondary end points included LVEDVI change from baseline at 30 days along with LV volumes, ejection fraction, and sphericity change at 30 and 90 days, clinical outcomes, mortality, and safety outcomes. The LV end-diastolic and end-systolic diameters were measured at the cardiac base and at midcavity in the parasternal long axis view, and at the base, midcavity, and lower third of the LV cavity in the apical two- and four-chamber views. The LV long axis was measured in the apical two- and four-chamber views from the apex tip to the anterior/lateral corner of the mitral annulus. End-diastolic volume was derived by multiplying the largest long axis (L) by the largest end-diastolic diameter using the formula: EDV = 3.42 (Diametermax x L) 6.44 (27). The LV ejection fraction was calculated using the multiple diameter method, which uses the average of the LV diameters measured at end-diastole and end-systole from the parasternal long axis and apical views together along with an estimate of fractional long axis shortening (28).
Pertaining to clinical outcomes, severe heart failure was defined as dyspnea, edema, fatigue symptoms accompanied by pulmonary edema on chest radiography, need for intravenous diuretics, or hospitalization. During initial hospitalization, reinfarction was defined as ischemic symptoms plus either new ST-segment elevation or 25% re-elevation of creatine kinase-MB or cardiac troponins from the preceding value. After discharge, reinfarction was determined by ischemic symptoms plus elevated creatine kinase-MB/cardiac troponin greater than the upper limit of normal. Recurrent ischemia was defined as ischemic symptoms not meeting reinfarction criteria that were substantiated by new ECG abnormalities or leading to urgent/unplanned catheterization or revascularization. Recorded musculoskeletal adverse events included the following symptoms or findings: arthralgia, myalgia, joint stiffness, new tendon thickening or nodule, or
30% decrease in shoulder/elbow range of motion at two goniometry measurements.
Statistical analysis.
Comparability of treatment groups for baseline characteristics and clinical outcomes were assessed with t tests and Fisher exact tests. Day 30 and day 90 echocardiographic changes (treatment efficacy) were compared using t tests for the least-square means and confirmed by an analysis of variance model with factors for treatment and stratification. The benefit of PG-116800 in decreasing LVEDVI versus placebo was expected to be
4 ml/m2 with standard deviation 10 ml/m2. Allowing for an attrition rate of 20%, enrollment was targeted at 250 patients to generate two treatment groups of 100 patients sufficient to detect the target LVEDVI (primary end point) benefit of 4 ml/m2 with 80% power and two-sided alpha = 0.05. Statistical analyses were performed using SAS version 8.2 (SAS Institute Inc., Cary, North Carolina). One-sided p values were reported for the primary end point, whereas two-sided p values are reported otherwise.
An independent Data and Monitoring Committee reviewed adverse event reports from study patients and monitored PG-116800 safety data from other clinical trials to safeguard study subjects and provide guidance to the sponsor (Procter & Gamble). The PREMIER trial expert panel functioned as an advisory committee for the trial and had complete access to all echocardiographic and clinical data after unblinding. All expert panel members and coauthors had a substantial role in trial design, data accrual, and data interpretation. Clinical data were collected in an electronic data collection form and entered into database managed by the trial sponsor. All final analyses were conducted by the sponsor in association with the expert panel.
| Results |
|---|
|
|
|---|
Patient baseline and clinical characteristics are shown in Table 1. Study groups were well matched overall without significant differences in demographic variables, prognostic characteristics, or hospital treatments. The study population had mean age of 59.8 ± 10 years, and two-thirds of patients were enrolled in Poland. Most patients (79%) had anterior ST-segment elevation MIs and were treated with primary PCI (91%). Baseline LV ejection fraction (36% vs. 37%) and time interval from MI to study drug administration (54.4 ± 14.0 h vs. 53.9 ± 14.6 h) did not differ in the PG-116800 and placebo groups, respectively. Patients in both treatment groups had high rates of post-MI beta-blocker and angiotensin-converting enzyme inhibitor/angiotensin receptor blocker use exceeding 90%.
|
LVEDVIat 90 days. There was no difference in LVEDVI after 30 or 90 days of treatment with PG-116800 versus placebo. The LV systolic and diastolic volumes, ejection fraction, and sphericity index did not differ in the PG-16800 and placebo treatment groups. Although underpowered, subgroup analysis showed no significant differences in LVEDVI reduction associated with PG-116800 dose (p = 0.27). The 90-day change in LVEDVI did not differ significantly between patients receiving placebo (n = 104), original protocol/high-dose PG-116800 (n = 40), or amended protocol/reduced-dose PG-116800 (n = 59) (5.48 ± 1.41 ml/m2 vs. 7.29 ± 2.28 ml/m2 vs. 3.60 ± 1.88 ml/m2), respectively.
|
|
|
|
| Discussion |
|---|
|
|
|---|
Both MMP gene expression and activity are increased after MI, leading to collagen and elastic fiber degradation, structural remodeling of the extracellular matrix, impaired collagen formation, and myocardial fibrosis (1116). Inhibition of MMP prevents progressive ventricular remodeling in MI and heart failure experimental models (1724). PG-116800 is a selective MMP inhibitor with a high affinity for MMP-2, -3, -8, -9, -11, -13, and -14 and minimal affinity for MMP-1 and -7. PG-116800 and its dehydrated salt form PG-530742 had impressive outcomes in preclinical animal studies, significantly reducing LV volumes along with infarct zone collagen content in a post-MI porcine model (21) and reducing LV volumes plus improving ejection fraction in a microembolization, chronic heart failure canine model (24). Prior studies also suggested that MMP-1 inhibition is associated with musculoskeletal complaints and that MMP-7 positively contributed to wound healing. Thus PG-116800 looked particularly promising because of its MMP selectivity and its expected favorable efficacy-toxicity profile.
Major findings.
In this first therapeutic human study of MMP inhibition after acute MI, we report that 90 days of treatment with PG-116800 initiated 48 h after MI had no beneficial effect on LV remodeling (
LV end-diastolic volume index) or clinical outcomes. PG-116800 was generally well tolerated and produced no significant increase in overall musculoskeletal adverse outcomes. Based on prior studies, 90 days should have provided ample time to observe a change in LVEDVI (29). We did encounter considerable individual variability among the serial LVEDVI measurements; however, the absolute change per group (
5 ml/m2) and variability (14%) were similar to the prestudy estimates, giving us adequate statistical power to observe a change. It is unlikely that a different cardiac imaging modality (i.e., cardiac magnetic resonance imaging) or diagnostic assessment of LV remodeling would have affected our results. Echocardiography was chosen to assess LV remodeling because it was readily available in most centers, permitted efficient screening of potential patients, and had substantiated the benefits of angiotensin-converting enzyme inhibitors and beta-blockers on LV volumes and ejection fractions in prior post-MI studies (3032).
Study limitations. The combined use of contemporary MI therapies may have diminished any potential therapeutic efficacy of MMP inhibition in our study patients. Primary PCI, beta-blockers, angiotensin-converting enzyme inhibitors, and statins were used by >90% of patients in both treatment groups. Because these contemporary therapies along with aldosterone inhibition (33) prevent LV remodeling, it is possible that an additional salutary PG-116800 effect may have been diminished or gone undetected.
Inadequate PG-116800 dosing also may have affected our results. At study onset, we planned to administer 200 mg of PG-116800 twice daily for 90 days. Based on reports of musculoskeletal adverse events in a parallel trial using PG-116800 for osteoarthritis and subsequent concerns raised by Health Canada, we amended our protocol, reduced PG-116800 cumulative dose by 33%, and completed the study with PG-116800treated patients receiving a dosing regimen lacking prior experimental efficacy. It is uncertain whether more potent MMP inhibition might prevent LV remodeling; conversely our results showed no dose response relationship between PG-116800 and LVEDVI change and no diminution of drug efficacy after protocol change and study drug dose reduction.
Conclusions. Contrary to animal and preclinical data, our disappointing results cast doubt on whether MMP inhibition might ameliorate post-MI remodeling. Selective MMP inhibition with PG-116800 is ineffective for treating this condition, yet our current understanding of which MMP(s) contribute most to ventricular remodeling is rudimentary. Future research in post-MI patients is needed to better identify the clinical and biologic determinants of LV remodeling, to understand the timing and sequence of MMP activation and inhibition, and to identify alternative, more-specific MMP targets (34). Additionally, the association between MMP inhibition and musculoskeletal adverse events needs increased study because our results suggest that selective sparing of MMP-1 inhibition is not sufficient to avoid all MMP musculoskeletal side effects.
| Appendix |
|---|
|
|
|---|
| Footnotes |
|---|
a Drs. Brum, Cusmano, Krzeski, and Lyon are employees of Procter & Gamble Pharmaceuticals. ![]()
b Dr. Theroux has equity holdings in Procter & Gamble. ![]()
c Dr. Armstrong has received research grant support and consultancy fees from Procter & Gamble Pharmaceuticals. ![]()
d Drs. Ruzyllo, Quinones, Theroux, and Weaver received consultancy fees for participating on PREMIER Trial Expert Panel. ![]()
| References |
|---|
|
|
|---|
Related Article
This article has been cited by other articles:
![]() |
Y. Ikeda, M. Hoshijima, and K. R. Chien Toward Biologically Targeted Therapy of Calcium Cycling Defects in Heart Failure Physiology, February 1, 2008; 23(1): 6 - 16. [Abstract] [Full Text] [PDF] |
||||
![]() |
F. G. Spinale Myocardial Matrix Remodeling and the Matrix Metalloproteinases: Influence on Cardiac Form and Function Physiol Rev, October 1, 2007; 87(4): 1285 - 1342. [Abstract] [Full Text] [PDF] |
||||
![]() |
K. Uemura, M. Li, T. Tsutsumi, T. Yamazaki, T. Kawada, A. Kamiya, M. Inagaki, K. Sunagawa, and M. Sugimachi Efferent vagal nerve stimulation induces tissue inhibitor of metalloproteinase-1 in myocardial ischemia-reperfusion injury in rabbit Am J Physiol Heart Circ Physiol, October 1, 2007; 293(4): H2254 - H2261. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. R. Dixon, C. L. Grines, and W. W. O'Neill The Year in Interventional Cardiology J. Am. Coll. Cardiol., July 17, 2007; 50(3): 270 - 285. [Full Text] [PDF] |
||||
![]() |
P. Zymek, D.-Y. Nah, M. Bujak, G. Ren, A. Koerting, T. Leucker, P. Huebener, G. Taffet, M. Entman, and N. G. Frangogiannis Interleukin-10 is not a critical regulator of infarct healing and left ventricular remodeling Cardiovasc Res, May 1, 2007; 74(2): 313 - 322. [Abstract] [Full Text] [PDF] |
||||
![]() |
D. Kelly, G. Cockerill, L. L. Ng, M. Thompson, S. Khan, N. J. Samani, and I. B. Squire Plasma matrix metalloproteinase-9 and left ventricular remodelling after acute myocardial infarction in man: a prospective cohort study Eur. Heart J., March 5, 2007; (2007) ehm003v1. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. N. DeMaria, O. Ben-Yehuda, G. K. Feld, G. S. Ginsburg, B. H. Greenberg, W. Y.W. Lew, J. A.C. Lima, A. S. Maisel, J. Narula, D. J. Sahn, et al. Highlights of the Year in JACC 2006 J. Am. Coll. Cardiol., January 30, 2007; 49(4): 509 - 527. [Full Text] [PDF] |
||||
![]() |
J. Lipiecki, N. Durel, and J. Ponsonnaille Which patients with ischaemic heart disease could benefit from cell replacement therapy? Eur. Heart J. Suppl., December 1, 2006; 8(suppl_H): H3 - H7. [Abstract] [Full Text] [PDF] |
||||
![]() |
I. Ernens, D. Rouy, E. Velot, Y. Devaux, and D. R. Wagner Adenosine Inhibits Matrix Metalloproteinase-9 Secretion By Neutrophils: Implication of A2a Receptor and cAMP/PKA/Ca2+ Pathway Circ. Res., September 15, 2006; 99(6): 590 - 597. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | SUBSCRIPTIONS | CURRENT ISSUE | PAST ISSUES | CARDIOSOURCE | SEARCH | HELP | FEEDBACK |