Author + information
- Received April 12, 2006
- Revision received June 26, 2006
- Accepted July 23, 2006
- Published online December 19, 2006.
- Eric P. Brass, MD, PhD⁎,⁎ (, )
- Richard Anthony, PhD†,
- Frederick R. Cobb, MD‡,
- Isao Koda, PhD§,
- Jenny Jiao, PhD† and
- William R. Hiatt, MD∥
- ↵⁎Reprint requests and correspondence:
Dr. Eric Brass, Harbor-UCLA Medical Center, Center for Clinical Pharmacology, 1124 West Carson, Torrance, California 90502.
Objectives The current study tested the hypothesis that NM-702 improves treadmill exercise performance in peripheral arterial disease patients with claudication-limited exercise performance.
Background Patients with claudication experience significant disability, owing to their exercise limitation. Therapeutic options to improve exercise performance in these patients are limited. NM-702 is a novel drug that inhibits phosphodiesterase as well as thromboxane A2synthase.
Methods This study was a randomized, multi-center, placebo-controlled, double-blind trial. Patients were randomized to receive 24 weeks of twice-daily treatment with either placebo (intent to treat population, n = 130), 4 mg NM-702 (n = 126), or 8 mg NM-702 (n = 130).
Results After 24 weeks of treatment, 8 mg NM-702 was associated with a statistically significant increased peak walking time on a graded treadmill as compared with placebo (p = 0.004). Peak walking time after 24 weeks was increased by 17.1 ± 49.0% in the placebo arm, 22.1 ± 60.1% in the 4-mg NM-702 arm, and 28.1 ± 50.5% in the 8-mg NM-702 arm. NM-702 at the 8-mg dose for 24 weeks was associated with statistically significant improvements in the treadmill claudication onset time as compared with placebo. In addition, as compared with placebo, NM-702 improved the physical component and physical functioning scores of the Medical Outcomes Study 36-Item Short Form and the walking distance and stair climbing domains of the Walking Impairment Questionnaire. NM-702 was generally well tolerated, but adverse events typical of vasodilators were common.
Conclusions NM-702 used for 24 weeks by patients with claudication was associated with improvements in laboratory- and ambulatory-based exercise performance.
Lower extremity peripheral arterial disease (PAD) is a prevalent manifestation of systemic atherosclerosis. Patients with PAD are at high risk for cardiovascular morbidity and mortality, and strategies to reduce systemic cardiovascular risk are the cornerstone of PAD therapy (1). However, PAD is also associated with substantial functional disability due to impaired walking performance (1,2). Exercise rehabilitation, drug therapy, and selective use of revascularization are options for improving exercise performance in claudicants (1,3). The efficacy of cilostazol, a selective inhibitor of phosphodiesterase type III, in improving walking capacity in PAD patients (4) has led to interest in other phosphodiesterase inhibitors as treatments for claudication.
NM-702 (4-bromo-6-[3-(4-chlorophenyl)propoxy]-5-[3-pyridylmethylamino]-3[2H]-pyridazinone hydrochloride) is an inhibitor of phosphodiesterase (5). NM-702 shows selectivity for phosphodiesterase isoforms III and V on the basis of study of phosphodiesterases extracted from rabbit heart and platelets. NM-702 also inhibits human platelet thromboxane A2synthase at concentrations that inhibit rabbit phosphodiesterase (5). This profile suggests that NM-702 might have efficacy in the treatment of claudication-limited exercise impairment. Phase I and II studies have demonstrated that the drug is well tolerated and might improve treadmill performance in claudicants. The current study was designed to evaluate the potential efficacy of NM-702 at doses of 4 mg or 8 mg twice daily and assess its safety in patients with PAD.
Overview of study design
The study was a 3-armed, randomized, double-blinded, placebo-controlled, multicenter trial conducted at 18 sites in the U.S. (Appendix). After screening evaluations and baseline qualification measurements (see the following text), patients were randomized to twice-daily doses of placebo, 4 mg NM-702, or 8 mg NM-702. Patients remained on study medication for 24 weeks. The protocol was reviewed and approved by appropriate institutional review committees, and all patients provided written informed consent.
A steering committee (E.P.B., R.A., F.R.C., I.K., and W.R.H.) was responsible for design of the study, oversight of its implementation, and analysis of the data. The Committee’s Charter defined its role and independence as well as the sponsor’s commitment to publish the study results in a timely manner.
Patients were recruited from clinics at the study sites and through community-based advertising. Per protocol, all participants were a minimum of 50 years old and had a history of clinically stable claudication for a period of at least 3 months. A diagnosis of PAD was confirmed in each patient on the basis of a resting ankle-brachial index (ABI) of ≤0.90 in the symptomatic leg or for diabetic patients with noncompressible vessels (ABI >1.30), a toe-brachial index (TBI) of ≤0.70. The ABI was calculated with the higher of the 2 systolic blood pressures measured in the upper extremity and the higher of the measured dorsalis pedis or posterior tibial pressures measured in each leg (6). The TBI was measured with standard methodologies (7).
Patients were excluded from study participation if they had ischemic rest pain or ulceration, exercise performance limited by any condition other than claudication, had undergone a revascularization procedure within 12 weeks, suffered a myocardial infarction or stroke within 6 months of potential study drug initiation, had a resting upper extremity blood pressure exceeding 150/100 mm Hg, or were pregnant or lactating. Clinically significant laboratory abnormalities on screening were also exclusionary, including a prothrombin time >2 times the upper limit of normal; a platelet count of <50,000/mm3; a hematocrit >60% or <30%; a hemoglobin A1c>12%; low-density lipoprotein cholesterol >160 mg/dl; a serum creatinine >3.0 mg/dl; or bilirubin, alanine aminotransferase, or aspartate aminotransferase more than 3 times the upper limit of normal.
Patients must have been off claudication pharmacotherapy (including cilostazol or pentoxyfylline) or treatment with other phosphodiesterase inhibitors (for example, sildenafil) for a period of at least 4 weeks before any baseline assessments, and use of these drugs was not permitted until completion of the week-24 efficacy assessments and conclusion of study drug treatment. Patients were permitted to use aspirin or clopidogrel. However, owing to the antiplatelet effects of NM-702, for potential safety reasons participants were not permitted to use combination aspirin plus clopidogrel therapy during the study.
At an initial screening visit, patients underwent a full medical history and physical examination, hemodynamic assessments, electrocardiogram, and familiarization with the treadmill procedures. Participants returned for 3 baseline visits, at each of which a treadmill test to measure peak walking time (PWT) and claudication onset time (COT) was performed. These visits were separated by between 3 and 10 days. At the first of these baseline visits, blood was also obtained for baseline and exclusionary laboratory assessments.
For analysis purposes, a patient’s baseline exercise performance was defined as the median value from the 3 baseline tests for PWT and COT. Patients with a median baseline PWT of ≥90 s and ≤600 s qualified for randomization.
After completion of 3 baseline treadmill tests, patients meeting inclusion/exclusion criteria were randomized to 1 of the 3 treatment arms. At the time of randomization patients also completed the Medical Outcomes Study 36-Item Short Form (SF-36) health-related quality-of-life instrument and the Walking Impairment Questionnaire (WIQ) (8). Patients returned to the study site for visits 6, 12, 18, and 24 weeks after randomization. At each visit study medication was dispensed (except at the 24-week visit), and adverse events and concomitant medications reviewed. At the week 12 and 24 visits, hemodynamic status was assessed, treadmill testing performed, and the SF-36 and WIQ were administered. Active treatment with study medication was discontinued after the 24-week visit, and a final visit for safety purposes occurred 12 months after randomization. Only information on serious adverse events and mortality were collected between the week 24 and month 12 visits.
Claudication-limited exercise performance was quantified with a graded treadmill exercise test. At time 0, the patients stepped onto the treadmill, which was moving at 2 mph at a 0% grade. At 2-min intervals the grade was increased by 2% as per Gardner et al. (9). The patient continued on the treadmill until they could not walk any further, and the time at which this occurred was designated as the PWT. The COT, defined as the time on the treadmill at which claudication symptoms were first reported, was also recorded for each test. All treadmill testing was performed at least 2 h after the last study drug dose. A quality control program was provided to the sites to ensure that the treadmill testing procedures were performed consistently and according to protocol.
The primary end point for this trial was the change in PWT between baseline and 24 weeks. Changes from baseline in treadmill data were analyzed on the basis of log-transformed data [i.e., ln(PWT at 24 weeks) − ln(PWT at baseline), which is the same as the ln ratio]. The use of natural log-transformed data was prospectively defined, owing to the anticipated skewed distribution of the walking time data. Comparisons between treatment arms were done with analysis of covariance (ANCOVA), which included treatment, smoking status, baseline ln(PWT), and investigator site as main effects. Treatment effects on COT were analyzed in a similar manner. Similarly, SF-36 and WIQ data were also analyzed with ANCOVA with adjustment for demographic variables, baseline scores, and smoking and diabetic status. Last observation carried forward was used to replace any missing data.
Changes in limb hemodynamic status between baseline and 12 or 24 weeks were categorized as increased by 15% or more, decreased by 15% or more, or as unchanged (within 15% of baseline). Effects of treatment were assessed with a chi-square test.
The study was designed to allow for independent assessment for the effects of either the 4-mg or the 8-mg dose as compared with placebo, and thus for sample size considerations, each dosed group was independently powered at an alpha = 0.025 level. A sample size of 117 patients/arm would provide 90% power, assuming a 25% increase in PWT in the placebo group, a 65% increase in the treatment group(s), and an 85% coefficient of variation for the absolute PWTs. A sample size of 130 patients/arm was elected to allow for patients who withdrew from the study early.
Statistical inference for the primary efficacy variable (PWT change at 24 weeks) was made on the basis of Hochberg multiple testing procedure for the efficacy of each dose. The requirement for claiming that both doses were efficacious was a p value <0.05 for both the 4-mg NM-702 versus placebo and 8-mg NM-702 versus placebo comparisons. If 1 comparison failed to meet the p < 0.05 criterion, the pair-wise comparison for each dose versus placebo was examined with the more stringent criterion of p < 0.025 required for inferring efficacy at that dose. No adjustments for significance levels were made for other statistical tests performed, and a p < 0.05 was used for statistical inferences.
Demographics and baseline characteristics
A total of 391 patients were randomized. Of this total, 5 were excluded from the intent-to-treat analysis because they never received a dose of study medication. Patients seemed to reflect the general claudication population, because they averaged approximately 67 years of age, they were predominately male, most were either current or past smokers, and approximately 42% were diabetic patients (Table 1).Participants’ ABIs averaged 0.64 on entry, and their baseline performance on the treadmill was consistent with moderate to severe performance limitation due to claudication (Table 1). Six patients, 2 in the 4-mg NM-702 arm and 4 in the 8-mg NM-702 arm, entered the study on the basis of TBI measurements. Patients randomized to the 3 study arms were similar, except that there was a significant difference in the distribution of race amongst the arms (p = 0.003) (Table 1).
Treadmill end points
A total of 11.5% of patients receiving placebo, 17.5% of patients in the 4-mg NM-702 arm, and 22.3% of patients in the 8-mg dosing arm withdrew from the study before completing 24 weeks of treatment and the 24-week treadmill assessment. Of the 15 patients in the placebo group who withdrew, 7 made a decision to withdraw without providing additional information, 5 withdrew owing to an adverse experience, and 1 each were withdrawn owing to a sponsor decision, a physician decision, and after noting failure to meet entry criteria. Similarly, of the withdrawal of 22 patients in the 4-mg NM-702 group, 11 were on the basis of patient decision, 4 owing to an adverse event, 4 were lost to follow-up, 2 were on the basis of physician decision, and 1 was because of death. In the 8-mg NM-702 group, 15 patient withdrawals were on the basis of patient decision, 8 were owing to an adverse event, and 6 patients were lost to follow-up.
On the basis of the protocol’s primary end point analysis, the model-adjusted PWT ratio (least-square mean) for PWT at 24 weeks versus baseline was 1.07 for placebo, 1.12 for 4 mg NM-702, and 1.21 for 8 mg NM-702. The pair-wise comparisons for each dosed arm yielded p = 0.259 for 4 mg NM-702 versus placebo and p = 0.004 for 8 mg NM-702 versus placebo. Thus, the 8-mg NM-702 dose was associated with a statistically significant increase in PWT. The PWT ratio of 1.21 for the 8-mg NM-702 group corresponded to an absolute increase in PWT of 68 s compared with 35 s receiving placebo (Table 2).Expressed as a percent change from baseline, the 8-mg twice daily. NM-702 dose was associated with a 28.1% increase in PWT, as compared with 17.1% receiving placebo. The increase in PWT seemed to be both time and dose dependent (Fig. 1).Importantly, 68% of subjects reported taking their last dose of study medication at least 9 h before the time of the 24-week treadmill test. A linear dose-response relationship was present across the treatment arms (p = 0.004).
Limiting analysis to a per-protocol population that received at least 80% of study drug doses, did not use pentoxyfylline or cilostazol during the treatment period, and completed a week-24 treadmill increased the numeric effect of NM-702 but did not change the statistical conclusions (least-square mean ratio week-24 vs. baseline 1.07 on placebo, 1.19 on 4 mg NM-702, pair-wise p = 0.065 vs. placebo, and 1.27 on 8 mg NM-702, pair-wise p = 0.003 vs. placebo). Approximately 72% of the intent-to-treat cohort qualified for this per-protocol analysis.
The conclusion that 24 weeks of NM-702 treatment improved PWT was supported by analysis of subpopulations stratified by diabetic status, race, or body mass index (data not shown). Smoking status did seem to affect the PWT response (p = 0.007), with current smokers having larger percent increases in PWT in each group, but NM-702 was numerically superior to placebo in each smoking status cohort (data not shown).
Hemodynamic status as assessed by ABI was stable in the majority of patients over the 24 weeks of treatment and was unaffected by NM-702. At week-24, the mean ABI was 0.67 in all treatment arms and did not differ in any group from baseline assessments. With the protocol definition of a 15% change in hemodynamic status as potentially clinically significant, the number of patients showing increases in ABI or TBI was not different between treatment arms (Table 3).
All treatment arms were considered comparable at baseline with respect to SF-36 and WIQ assessments of ambulatory functional status (Table 4).NM-702 treatment for 24 weeks was associated with statistically significant improvements in the SF-36 physical component score, SF-36 assessed physical functioning, WIQ walking distance summary, and the WIQ stair climbing summary (Table 4). In contrast, the SF-36 mental component summary and the WIQ walking speed summary were not affected by treatment.
During the treatment phase of the study (6 months of planned study drug exposure plus 1 month additional observation) there were a total of 3 deaths (1 in the placebo arm and 2 in the 4-mg NM-702 arm). Two of the 3 deaths (1 in the placebo arm and 1 in the 4-mg NM-702 group) occurred within 6 months of randomization but after study drug was discontinued for other reasons. During the additional 5 months of observation (completing 1 year total observation, the last 6 months of which were off drug) there were 5 more deaths (3 in the placebo arm, 1 in the 4-mg NM-702 arm, and 1 in the 8-mg NM-702 arm).
One or more adverse events were experienced by 70%, 77%, and 79% of the subjects in the placebo, 4-mg NM-702, and 8-mg NM-702 treatment arms, respectively. Of these, 86% of the total adverse events were assessed as mild or moderate in intensity by the site investigators. Symptoms anticipated on the basis of NM-702’s mechanism of action were seen at increased frequency in patients receiving NM-702 but were usually mild to moderate in severity and typically were not associated with discontinuation of study medication (Table 5).Symptoms and signs potentially related to tachyarrhythmias were also more common in patients receiving NM-702, and in most cases these were sinus tachycardias. Several cases of atrial flutter or fibrillation were observed in patients taking NM-702, and 4 patients receiving 8 mg twice daily. NM-702 experienced syncope without documentation of the associated cardiac rhythm (Table 5), 2 of which were characterized as serious adverse events (Table 6).Four cases of gout or gout exacerbation were observed in the 8-mg NM-702 arm, with no cases in either of the 2 other arms. No other trends suggesting unanticipated adverse events associated with NM-702 were noted.
Serious adverse events were experienced by 33, 28, and 29 subjects in the placebo, 4-mg NM-702, and 8-mg NM-702 treatment arms, respectively. Given the high prevalence of cardiovascular complications in the PAD population and the mechanism of action of NM-702, particular attention was paid to cardiac serious adverse events (Table 6). There were no trends suggesting that NM-702 was associated with increased frequencies of serious cardiac ischemic events or congestive heart failure.
Because NM-702 affects platelet function, bleeding adverse events were reviewed in detail. Sixty-nine percent of the subjects were taking aspirin during the study, and 16% were taking clopidogrel, and there were no significant differences between treatment groups in the use of these drugs. A total of 5 subjects used coumadin during the study treatment period (1 receiving placebo and 4 receiving 8 mg NM-702). Eleven patients (4 receiving placebo, 1 receiving 4 mg NM-702, and 6 receiving 8 mg NM-702) experienced bleeding or coagulopathy-related adverse events (Table 5). Two of the bleeding adverse events, both in patients taking 8 mg NM-702, were serious. One serious adverse event involved a patient using warfarin who was admitted to the hospital without bleeding but had an international normalized ratio of 11.8. The second serious event involved a patient taking warfarin and aspirin who was hospitalized with a gastrointestinal bleed.
NM-702 use for 24 weeks was associated with statistically significant improvements in treadmill and ambulatory-based performance in patients with claudication. The efficacy of NM-702 in improving treadmill PWT as assessed in the primary efficacy analysis was supported by several pre-specified secondary analyses. Analysis of a per-protocol population enriched for patients who adhered to the medication dosing over 24 weeks demonstrated a numerically larger improvement as compared with the intention to treat population, as would be anticipated if drug exposure was responsible for the effect observed. There was a significant relationship between NM-702 dose and the degree of PWT improvement. NM-702 was also superior to placebo when clinically relevant stratifications of the study population were analyzed. Claudication onset time showed a statistically significant response to NM-702 similar to that observed on analysis of PWT. Most 24-week treadmills were obtained at least 9 h after the last dose of study medication, supporting efficacy of NM-702 throughout the dosing interval. Finally, the laboratory-based improvement in treadmill walking performance was confirmed with the ambulatory-based questionnaires.
The magnitude of improvement seen with NM-702 treatment is likely to be clinically meaningful. The changes in treadmill performance and questionnaire assessments associated with 8 mg NM-702 are very similar to those observed with the use of cilostazol (10), which is approved in the U.S. for the treatment of claudication. Clinical relevance is also supported by the benefit of NM-702 on both instruments used to assess community-based performance in the setting of the patients’ daily activities (Table 4).
Although NM-702 is known to be an inhibitor of phosphodiesterase activities and of thromboxane A2synthase (5), how these effects relate to the improvement in claudication-limited performance is unknown. Cilostazol is also a phosphodiesterase inhibitor (4), suggesting that this activity might be relevant. However, whether the target tissue is the vasculature, muscle, or platelet is unclear. Of note, no changes in large vessel hemodynamic status as assessed by ABI were observed with NM-702 use (Table 3). However, post-exercise ABIs were not assessed, and thus dynamic changes in hemodynamic status can not be excluded.
NM-702 was generally well tolerated, with dropout rates only modestly higher in the 8-mg arm as compared with the placebo arm. Nonetheless, mechanism-based adverse events were common with NM-702 dosing (Table 5). Concerns have been expressed about the cardiovascular risks associated with the use of phosphodiesterase inhibitors in at-risk populations (11,12). Thus, although the number of observed events was small, it was reassuring to note no increase in cardiac ischemic events or heart failure associated with NM-702 (Table 5). Four cases of syncope were associated with 8-mg NM-702 use, 2 of which were characterized as serious adverse events (Table 6). Of note, none of these patients reported using nitrates while taking NM-702. The small number of cases precludes definitive interpretation with respect to either the mechanism of these events or their casual relationship with NM-702 use. Nonetheless, given the theoretical potential for orthostatic hypotension or tachyarrhythmias associated with phosphodiesterase inhibition, attention should be paid to any similar events in future NM-702 trials. Furthermore, although no effect of NM-702 on QT interval was identified in phase I trials, no formal assessment of the drug’s effect on this important parameter was made in the current study.
NM-702 inhibits platelet aggregation, and thus bleeding events might be a risk with the drug’s use. Background therapy with aspirin or clopidogrel was widely used in the study cohort, and bleeding events were rare. Two serious adverse events due to bleeding or coagulopathy occurred in patients taking NM-702. Both of these patients were taking warfarin, raising the possibility of pharmacokinetic and pharmacodynamic drug interactions. Formal assessment of NM-702’s interactions with warfarin and antiplatelet drugs will be required to address these potential interactions.
In conclusion, 8 mg NM-702 twice daily was associated with significant functional improvements in patients with PAD manifesting as claudication. NM-702 was reasonably well tolerated, and no unanticipated safety concerns were established. NM-702 has the potential to provide an additional therapeutic option for patients with claudication if future studies confirm its safety and efficacy.
The authors thank the members of the Data and Safety Monitoring Committee (James Murphy, PhD, Janette Durham, MD, Kathy Hassell, MD, and Mori Krantz, MD), the site investigators, and their staff for their contributions to this study.
For a list of participating study sites and investigators, please see the online version of this article.
This study was supported and sponsored by Nissan Chemical Industries, Ltd., Tokyo, Japan. Drs. Brass, Hiatt, and Cobb were paid consultants to the trial’s sponsor. Drs. Anthony and Jiao are employed by the contract research organization that conducted the study on the sponsor’s behalf. Dr. Koda is employed by the trial’s sponsor. Dr. Cobb died unexpectedly during the preparation of this work.
- Abbreviations and Acronyms
- ankle-brachial index
- analysis of covariance
- claudication onset time
- 4-bromo-6-[3-(4-chlorophenyl)propoxy]-5-(3-pyridylmethylamino)-3(2H)-pyridazinone hydrochloride
- peripheral arterial disease
- peak walking time
- Medical Outcomes Study 36-Item Short Form
- toe-brachial index
- Walking Impairment Questionnaire
- Received April 12, 2006.
- Revision received June 26, 2006.
- Accepted July 23, 2006.
- American College of Cardiology Foundation
- Ikegawa R.,
- Imada T.,
- Tsuruzoe N.,
- Shudo N.,
- Nakamura N.
- Hiatt W.R.,
- Hoag S.,
- Hamman R.F.
- Orchard T.J.,
- Strandness D.E. Jr.
- Hiatt W.R.,
- Hirsch A.T.,
- Regensteiner J.G.,
- Brass E.P.
- Regensteiner J.G.,
- Ware J.E. Jr..,
- McCarthy W.J.,
- et al.