Oral Beraprost Sodium, a Prostaglandin I2 Analogue, for Intermittent Claudication
A Double-Blind, Randomized, Multicenter Controlled Trial
Background—Beraprost sodium (BPS) is a new stable, orally active prostaglandin I2 analogue with antiplatelet and vasodilating properties. We report the results of a phase III clinical trial of BPS in patients with intermittent claudication.
Methods and Results—Patients (n=549) with a pain-free walking distance of between 50 and 300 m were entered into a 4-week single-blind placebo run-in phase. Patients whose pain-free walking distance had changed by <25% were then randomized to receive either BPS (40 μg TID, n=209) or placebo (n=213) in a double-blind manner for 6 months. Pain-free and maximum walking distances were measured on the occasion of treadmill exercise tests performed at baseline and 1.5, 3, 4.5, and 6 months after randomization. Success was defined as an improvement of >50% in pain-free walking distance at month 6 and in ≥1 earlier treadmill exercise test in the absence of critical cardiovascular events. Success was observed more frequently in the BPS group (43.5%) than in the placebo group (33.3%, P=0.036). Pain-free walking distances increased by 81.5% and 52.5%, respectively, in the BPS and placebo groups (P=0.001) and maximum walking distances by 60.1% and 35.0%, respectively (P=0.004). The incidence of critical cardiovascular events was 4.8% in the BPS group and 8.9% in the placebo group.
Conclusions—These results show that BPS is an effective symptomatic treatment of patients with intermittent claudication. The beneficial effects of BPS on critical cardiovascular events should be confirmed in appropriate clinical trials.
Intermittent claudication is a common presentation of lower-limb arterial disease, with a prevalence of 1% to 7% in men 50 to 75 years old.1 2 Medical therapy for patients with intermittent claudication has 2 goals: improvement of symptoms and prevention of cardiovascular events. Many agents, including vasodilators, have been tried for the relief of symptoms, although their effectiveness is still subject to debate. The antiplatelet and vasodilating properties of epoprostenol (prostaglandin [PG] I2) make it a theoretically interesting agent in the treatment of peripheral vascular disease.3 Although promising results were reported for its use in the treatment of intermittent claudication,4 its very short half-life and high chemical instability precluded its use as a therapeutic agent and led to the development of more stable synthetic analogues. Iloprost, to date the most extensively studied of these analogues, has been shown to provide pain relief and to accelerate ulcer healing in patients with stage III or IV critical leg ischemia5 and has been licensed in some countries for the treatment of thromboangiitis obliterans. The need for an intravenous infusion of iloprost, although acceptable for the treatment of critical ischemia, makes it unacceptable for the long-term treatment of intermittent claudication.
Beraprost sodium (BPS) is a new stable, orally active PGI2 analogue with antiplatelet and vasodilating properties.6 7 8 It has linear kinetic characteristics at doses ranging from 20 to 60 μg TID, with a half-life for the active isomer of ≈0.6 hours after repeated doses in healthy volunteers.7 We assessed the effects of BPS on walking distances in patients with intermittent claudication in the Beraprost et Claudication Intermittente (BERCI) dose-effect study.9 In this 12-week study, BPS increased the pain-free walking distance, compared with placebo, at doses of 20 and 40 μg TID but not at 60 μg. Because 40 μg TID appeared to offer the best combination of both the antiplatelet and the symptomatic effects of BPS, this dose regimen was chosen for the phase III trial that we report in this article. This trial (BERCI-2) was designed to demonstrate the 6-month efficacy of BPS on walking distances and cardiovascular events in patients with intermittent claudication.
BERCI-2 was a randomized, double-blind, multicenter, placebo-controlled trial conducted in France and Italy. After a 4-week single-blind placebo run-in phase, eligible patients were randomly assigned through a computer network (Minitel) to receive placebo or BPS 40 μg TID for 6 months in a double-blind manner. Randomization was stratified by clinical center in blocks of 2 treatments. The study was conducted in accordance with good clinical practices, the current version of the Declaration of Helsinki, and local regulations. The protocol was approved by an ethical committee, and patients were required to give informed written consent before being included.
The main study outcome was success, defined as an improvement of >50% in the treadmill exercise pain-free walking distance compared with that at baseline (day of randomization, D0), at the 6-month test, and in ≥1 of the earlier tests in the absence of critical cardiovascular events.
Critical cardiovascular events were defined as death of cardiovascular origin (fully confirmed or sudden death), myocardial infarction, unstable angina, coronary angioplasty, coronary artery bypass graft surgery, stroke, transient ischemic attack, critical leg ischemia (ischemia with trophic signs necessitating urgent medical or surgical measures to avoid amputation), subacute critical ischemia (continuous rest pain for ≥2 weeks requiring analgesics), peripheral angioplasty, peripheral bypass graft surgery, and amputation at any level.
Changes in pain-free and maximum walking distances, subjective walking distance, and the incidence of critical events were secondary outcomes. In addition, changes in ankle-brachial indices (ABIs) were analyzed, although they were not considered a secondary outcome by the protocol.
Quality of life was assessed as a tertiary outcome by use of the Quality of Life Subjective Profile questionnaire,10 which includes 17 items exploring functional, relational, and interior life and evaluates the patient’s satisfaction related to changes in each item. This was completed at baseline and month 6.
Clinical examination data and vital signs were recorded at D0 and months 1.5, 3, and 6. Treadmill exercise tests were performed at entry into the run-in phase (D −28), at inclusion (D0), and 1.5, 3, 4.5, and 6 months after randomization (at D0). Pain-free and maximum walking distances and ABIs before and immediately after exercise were measured on these occasions. Treadmill exercise tests were performed at 3 km/h and 10% gradient by the same operator and at the same time of the day for a given patient throughout the trial. In the absence of onset of claudication, the test was stopped after 24 minutes (1200 m). Critical cardiovascular events were looked for during the 6-month period after randomization even in patients who had withdrawn from the study prematurely. Every potential critical cardiovascular event was fully documented and evaluated blindly by 3 experienced cardiologists. Standard biological analyses were performed at D0 and month 6. Patients performed a subjective assessment of treatment usefulness at 6 months. The patients completed the quality-of-life questionnaires in the absence of the investigator and mailed them directly to the coordinating center.
Male or female patients 35 to 75 years old with intermittent claudication due to noninflammatory arterial disease of the lower limbs of >6 months’ duration who had been fully investigated by arteriography or Doppler scan and were stable during the previous 3 months were considered for inclusion. Those for whom invasive interventions were not indicated, with a pain-free walking distance of 50 to 300 m at the D −28 treadmill exercise test; who had no ischemic rest pain, ulceration, or gangrene; and who were not on antiplatelet therapy were eligible for entry into the run-in phase of the study. Patients were required to have no other disease that affected walking or was likely to interfere with follow-up, no history of myocardial infarction or stroke in the previous 3 months, no unstable or severe angina pectoris, and no insulin-dependent diabetes mellitus. Patients were given appropriate dietary recommendations and were counseled to stop smoking.
At the end of the placebo run-in phase (D0), patients were eligible for randomization only if their pain-free walking distance had changed by <25% between the 2 run-in treadmill exercise tests performed at D −28 and D0. Finally, anticoagulants and vasodilators were not allowed during the trial.
Assuming a 50% success rate in the BPS group and 35% in the placebo group, with α=5% (1-sided) and β=10%, the sample size required was 368. We decided to evaluate 400 patients.
All randomized patients who had taken ≥1 tablet of the study treatment were considered for analysis. Although the sample size was calculated with a 1-sided approach, all the tests performed were 2-sided. Except where specified, results are given as mean±SD.
The 2 treatment groups were compared for the proportion of responders by Fisher’s exact test. The incidence of the first cardiovascular critical event was analyzed by the log-rank test.
Because walking distances and their change from baseline at any measurement time were not distributed normally, they were analyzed as log-transformed values [log (value/baseline), where log=natural logarithm]. Consequently, the results are presented as geometric means. For the sake of clarity, however, walking distances at baseline are described as mean±SD. Because transformed values were not normally distributed either, they were analyzed with Wilcoxon’s test.
Because a number of patients did not complete the trial for different reasons, missing values were estimated by use of the conservative rule of last observation carried forward.
The limits of the CIs for the percentage of change from baseline were obtained by the inverse transformation of the limits of the CI calculated on the transformed data.
Satisfaction related to changes in quality of life was subjected to a principal-component analysis. The first factor of this analysis was linear combination of satisfaction related to changes in the 17 items of the questionnaire and was used as an evaluation of the global satisfaction related to changes in quality of life. The 2 groups were compared for global satisfaction related to changes in quality of life by Student’s t test. Satisfaction related to change in each item was analyzed by Wilcoxon’s rank sum test.
Between November 1993 and September 1995, 549 patients were entered into the 4-week single-blind run-in phase (Figure 1⇓). Of these, 424 patients were randomized to receive either BPS or placebo in the double-blind 6-month active treatment phase. The main reason for nonrandomization was a change in pain-free walking distance of >25% during the run-in phase (83 patients). Other reasons for noninclusion were biological abnormality (10 patients), withdrawal of consent (10 patients), an aggravation of symptoms (5 patients), and the occurrence of an adverse event (6 patients). In addition, 2 randomized patients withdrew before receiving the study drug, leaving 422 patients for analysis. The remaining treatment groups were well balanced, with 209 in the BPS group and 213 in the placebo group. Patients were mainly men, with a mean age of 62±9 years, and few of them had a history of myocardial infarction, stroke, or angina pectoris (Table 1⇓). At entry into the double-blind period, the BPS and placebo groups were similar, especially with regard to the duration of intermittent claudication, previous treatments, walking distances, and ABIs. There were significant differences only with regard to the number of patients following lifestyle modification prescriptions (57.9% in the BPS group versus 47.4% in the placebo group, P=0.032), the number of patients with presence of the left popliteal or anterior tibial pulse (higher in the BPS group, P<0.004), and the level of physical activity (higher in the BPS group, P=0.030).
The mean duration of intermittent claudication at entry into the run-in phase of the study was 5 years 8 months, ranging from 5 months to 28 years. As shown in Table 1⇑, a majority of patients had had previous medical treatment (73.7%) and/or surgical treatment (27.0%) for peripheral arterial disease.
The most frequent concomitant medications at inclusion were antihypertensive and antianginal drugs (45.9%; calcium channel blockers, ACE inhibitors, nitrates, or α-adrenergic receptor–blocking agents), lipid-lowering drugs (32.7%), oral antidiabetics (15.6%), and diuretics (14.5%).
Of the 422 patients who received study medication in the double-blind phase of the study, a total of 74 patients (17.5%) discontinued treatment prematurely (Figure 1⇑). The main reason for withdrawal was an “adverse event,” which included both nonserious and serious adverse events and critical cardiovascular events (Table 2⇓). Six of the patients who had prematurely discontinued study treatment nevertheless subsequently performed the month 6 treadmill exercise test.
A total of 162 patients were reported to be responders with regard to the main study outcome: success, defined as an improvement of >50% in pain-free walking distance at the 6-month and ≥1 earlier treadmill exercise test, in the absence of critical cardiovascular events. There was a significant difference between the treatment groups, the BPS group having a larger number of responders, 91 (43.5%), than the placebo group, 71 (33.3%), P=0.036.
Throughout the 6-month treatment period, the BPS group showed longer mean pain-free walking distances and mean absolute walking distances than the placebo group (Figure 2⇓). There was a significant increase in the percentage change in pain-free walking distance at 6 months for patients in the BPS group compared with the placebo group (geometric mean increase, 81.5% for the BPS group and 52.5% for the placebo group, P=0.001). Figure 3a⇓ compares the change from baseline pain-free walking distance between the 2 groups. Likewise, the percentage change in absolute walking distance was significantly different in favor of the BPS group, as shown in Figure 3b⇓, with a geometric mean increase of 60.1% for the BPS group and 35.0% for the placebo group (P=0.004). The difference between the BPS and the placebo groups was 36 m for the increase in pain-free walking distance and 70 m for the increase in absolute walking distance at 6 months compared with baseline.
An improvement of >50% in pain-free walking distance at the 6-month treadmill exercise test was achieved in 56.5% of patients in the BPS group and in 42.2% in the placebo group (P=0.0047).
ABI measured before treadmill exercise remained stable in both groups. There was no difference between the 2 groups in the loss in ABI during the treadmill exercise test. No difference between the groups was found regarding compliance with lifestyle modification prescriptions and tobacco consumption.
When results at month 6 for pain-free and maximum walking distances were analyzed with adjustment for a set of background variables, which included smoking, hypertension, diabetes, and dyslipidemia, no interaction was found.
Incidence of Critical Cardiovascular Events
Critical cardiovascular events were assessed at 6 months in all patients but 2, 1 of whom had already presented with a critical event during follow-up. Nearly twice as many patients reported critical cardiovascular events in the placebo group (19 of 213, 8.9%) as in the BPS group (10 of 209, 4.8%), although this difference was not statistically significant (P=0.092). The most common critical cardiovascular event in both treatment groups was arterial thrombosis of the leg (BPS, 8 events; placebo, 14 events).
Changes in Quality of Life
There was a marginally significant difference in favor of the BPS group for global satisfaction related to changes in quality of life (P=0.049). There were significant differences in favor of the BPS group for satisfaction related to changes regarding the items “going out” (P=0.017), “general condition” (P=0.046), “relationships with people” (P=0.034), and “concerns about health” (P=0.025). In addition, the treatment was deemed useful at 6 months by 92% of patients in the BPS group and 83% of patients in the placebo group (P=0.023).
No patient receiving BPS died, whereas 2 patients receiving placebo died suddenly. Adverse events (including critical cardiovascular events) considered by the investigators to be drug-related were observed in 35 patients (16.7%) in the BPS group and in 15 patients (7.0%) in the placebo group. In patients receiving BPS, the most common drug-related adverse events were headache (n=13) and vasodilatation (n=11), compared with headache (n=4) and arterial thrombosis of the leg (n=3) in the placebo group. Forty-nine patients experienced serious or nonserious adverse events (including cardiovascular critical events) that led to the permanent discontinuation of the study drug, 31 in the placebo group and 18 in the BPS group. There were no differences between the groups at any time regarding blood pressure and heart rate or in the standard biological analyses performed at D0 and month 6.
The BERCI-2 study showed that BPS, 40 μg TID, over a period of 6 months increases both the pain-free and the maximum walking distances, with respect to placebo, in patients with intermittent claudication. Treatment success, defined as an increase in pain-free walking distance of >50% at month 6 and ≥1 other treadmill exercise test with no critical event during follow-up, was significantly more frequent (by 30.6%) in the BPS group than in the placebo group. Successes were due to the conjunction of higher increases in pain-free walking distance and a lower risk of critical event in the BPS group, which is clinically meaningful. It was necessary to treat only 10 patients to gain 1 success.
Two reviews of trials of drug treatment of intermittent claudication have detected methodological failures in many of these trials.11 12 The design of the present trial responded to the criticisms of the former studies. Notwithstanding the quality of the design, missing treadmill exercise values may have impaired the final quality of the results. Because it was impossible to avoid having missing exercise test values in a 6-month trial, we decided to replace the missing values in a conservative way (see Methods section) so that the analysis could be performed according to the intention-to-treat principle. In addition, the only patient lost to follow-up with no known critical event was in the placebo group. This patient was counted as having had no critical event during follow-up, and it can therefore be considered that the assessment of critical events was done under the “maximum-bias hypothesis.”
Large increases in walking distances, both pain-free and maximum, were observed in the placebo group. This is a common feature of most of the trials that used treadmill exercise tests to assess intermittent claudication and can be explained by the effects of exercise itself or habituation to the treadmill test.13 14 A regression-to-the-mean phenomenon cannot be excluded either, although it should have been partly avoided by excluding from randomization patients with great changes in pain-free walking distance during the run-in part of the study.15 The mean differences between the BPS and placebo groups for the increases in pain-free (36 m) and in absolute (70 m) walking distances were in the range of what was observed in clinical trials of pentoxifylline and cilostazol.16 17 18
In a meta-analysis of 12 study arms from 11 trials, the difference between the pentoxifylline and placebo groups was 29 m for the increase in pain-free walking distance and 48 m for the increase in absolute walking distance.16 In a 12-week pilot study, the difference between cilostazol and placebo was 39 m and 106 m for the increases in pain-free and absolute walking distances, respectively.17 Compared with BERCI-2, the patients in this study had shorter baseline pain-free (78 m) and absolute (169 m) walking distances, but the treadmill exercise was performed at a 12.5% gradient instead of 10%. In addition, there was an unexpected decrease in walking distances at week 12 in the placebo group. In another 16-week study of cilostazol, the difference between the active and placebo groups was 28 m for the increase in pain-free walking distance and 62 m for the increase in absolute walking distance.18 In this latter study, however, a variable-grade, constant-speed treadmill exercise protocol was used, which makes the results difficult to compare with those of BERCI-2.
The absence of any effect of the treatment on the ABI before treadmill exercise and on the loss in this index during exercise could be anticipated from the results of the BERCI dose-effect study9 and are in accordance with the results of most clinical trials in intermittent claudication.11 19 20 In 1 study of cilostazol, however, the increase in walking distance was accompanied by a small but significant increase in the ABI compared with placebo.18
BPS had no significant effect on blood pressure and heart rate. This posed, once more, the problem of the mechanism of action of BPS in particular and prostanoids in general in the treatment of intermittent claudication. The trend toward fewer critical cardiovascular events observed in the BPS group suggests that BPS might exert antithrombotic effects in the clinical setting. The sample size and the duration of the trial were not, however, planned to allow the detection of an effect of the treatment on critical events with enough power.
The results of the quality-of-life assessment show that patients felt better on BPS than on placebo, the improvement in walking distance being paralleled by an improvement in relational and interior life. It is not surprising that no significant improvement in individual functional items was shown, because our patients suffered no major impact of their illness on everyday activities.
Among prostanoids, prostaglandin E1 (PGE1) was recently evaluated in intermittent claudication.21 22 23 The largest published study21 evaluated 213 patients with a maximum walking distance of 50 to 200 m at treadmill exercise. PGE1 was infused 5 days a week for 4 weeks. After 4 weeks, pain-free walking distances increased by 101% in the PGE1 group, compared with 60% in the placebo group, a relative increase comparable to that observed at 6 months in BERCI-2. In a smaller placebo-controlled study, a PGE1 prodrug induced short-term, dose-related increases in both pain-free and maximal walking distances.22 In a small, uncontrolled study,23 a 7-day infusion of PGE1 increased resting calf and skin blood flows compared with baseline but left the ABI unchanged. It seems, therefore, that oral BPS and intravenous PGE1 have comparable effects on walking distances. The oral route is a great advantage of BPS in intermittent claudication, a condition for which long-term treatment is necessary.
Although the exact mechanism of action of BPS remains to be elucidated, the present study shows that its ability to increase both pain-free and maximum walking distances makes it an effective drug in the treatment of intermittent claudication. Although not significant and observed in patients who did not receive the usual antiplatelet therapy, the 46% decrease in cardiovascular events observed in the BPS group is an interesting result, which should be confirmed in an appropriate clinical trial.
Coordinating Center: S. Azoulay, S. Morand. Data Handling Center: J.P. Boissel, C. Fernandez, M. Lievre, C. Mercier, I. Njem. Data Handling Center: Service de Pharmacologie Clinique, Lyon, France. Steering Committee: S. Azoulay, J.P. Boissel, J.N. Fiessinger, J.P. Girre, and M. Lievre.
Investigators: In France: Drs Leseche (Clichy), Azorin, Minveille (Bobigny), Barroussel (Annemasse), Dany (Limoges), Dumeril (Lyon), Eisenman (Strasbourg), Gallot-Lavallée (Saint Etienne), Hamel-Desnos (Caen), Hachulla (Lille), Page (Grenoble), Henry (Essey les Nancy), Janbon (Montpellier), Malgrange (Reims), Ponsonnaille (Clermont Ferrand), Sandreau (Lorient), Schmidt (Nancy), Testart, Pasquis (Rouen), Chabanier (Bayonne), Bizé (Cambo les Bains), Boutboul (Marseille), Breton, Breda (Béziers), Frédéric (Tours), Clerson (Wattrelos), Copere, Polverelli (Roanne), Cozic, Ferry (Nantes), El Sawy (St Martin d’Hères), Feuillette (Vallet), Fieux, Dufour (Paris), Gorlier (La Seyne/Mer), Dagorn (Toulon), Grégoire (Tallard), Tafani (Gap), Huin, Peyretou (Cenon), De la Chevasnerie (Perigny), Lafont, Alessandrini (Marseille), Lejay (Vieux Conde), Dalle (Conde/Escaut), Merijen, Diss (Mallemort), Ferrière, Musso (Eaunes), Speiser (Muret), Pradeau (Limoges), Carne (Noth), De Sainte Lorette, Dyan (Paris), Sicard (St Romain/Cher), Lefèvre (St Aignan/Cher), Villarroya, Deleporte (Wattrelos), Hardy (Argenteuil), Attia (Clichy), Guingal, Cordon, Kayanakis (Bayonne), Longhi (Pau), Sollier (Argeles/Mer), Lapeyre (Albi), Constans, Mouysset (Agen), Garrigues (Saint Orens), Dandine (Montauban), Olivier (Rouen), Elbèze (St Laurent du Var), Degeilh (Toulouse). In Italy: Profs Dipperri (Siena), Agrifoglio (Milan), Belcaro (Pescara), Catalano (Milan), Cernigliaro (Novara), Pedone (Forli), Pessina (Padova), Fogari (Pavia).
This trial was financed by an unrestricted grant from Laboratories Hoechst Houde (France). We would like to thank Alison Foote for her editorial assistance in the preparation of the manuscript.
This trial was financed by an unrestricted grant from Laboratories Hoechst Houde (France). Drs Lievre and Boissel received an institutional grant from Laboratories Hoechst Houde for data management and analysis of the BERCI-2 study. In addition, they are members of the steering committee of the DIABHYCAR study, for which their institution has received a grant from Laboratories Hoechst Houde. Drs Morand and Besse are employees of Laboratories Hoechst Houde.
- Received December 15, 1999.
- Revision received January 18, 2000.
- Accepted February 21, 2000.
- Copyright © 2000 by American Heart Association
Fowkes FGR, Housley E, Cawood EHH, et al. Edinburgh Artery Study. prevalence of asymptomatic and symptomatic peripheral arterial disease in the general population. Int J Epidemiol. 1991;20:384–391.
Kato R, Uji Y, Matsumoto K. Phase I study of beraprost sodium (TRK-100), an epoprostenol derivative: repeated oral administration for 10 days. Jpn J Clin Pharmacol Ther. 1989;20:529–539.
Radaack K, Wyderski RJ. Conservative management of intermittent claudication. Ann Intern Med. 1990;113:135–146.
Dahllöf AG, Holm J, Scherstén T, et al. Peripheral arterial insufficiency: effects of physical training on walking tolerance, calf blood flow, and blood flow resistance. Scand J Rehab Med. 1976;8:19–26.
Hiatt WR, Regensteiner JG, Hargarten ME, et al. Benefit of exercise conditioning for patients with peripheral arterial disease. Circulation. 1990;81:602–609.
Davis CE. The effect of regression to the mean in epidemiologic and clinical studies. Am J Epidemiol. 1976;104:493–498.
Hood SC, Moher D, Barber GG. Management of intermittent claudication with pentoxifylline: meta-analysis of randomized controlled trials. Can Med Assoc J. 1996;155:1053–1059.
Dawson DL, Cutler BS, Meissner MH, et al. Cilostazol has beneficial effects in treatment of intermittent claudication. Circulation. 1998;98:678–686.
Coffmann JD, Mannick JA. Failure of vasodilator drugs in arteriosclerosis obliterans. Ann Intern Med. 1972;76:35–39.
Bagger JP, Helligsoe P, Randsbaek F, et al. Effect of verapamil in intermittent claudication: a randomised, double-blind, placebo-controlled, cross-over study after individual dose-response assessment. Circulation. 1997;95:411–414.
Belch JJF, Bell PRF, Creissen D, et al. Randomised, double-blind, placebo-controlled study evaluating the efficacy and safety of AS-013, a prostaglandin E1 prodrug, in patients with intermittent claudication. Circulation. 1997;95:2298–2302.