Prostanoids for intermittent claudication
Abstract
Background
Peripheral arterial disease (PAD) is a common cause of morbidity in the general population. While numerous studies have established the efficacy of prostanoids in PAD stages III and IV, the question of the role of prostanoids as an alternative or additive treatment in patients suffering from intermittent claudication (PAD II) has not yet been clearly answered. This is an update of a Cochrane Review first published in 2004.
Objectives
To determine the effects of prostanoids in patients with intermittent claudication (IC) Fontaine stage II.
Search methods
For this update, the Cochrane Peripheral Vascular Diseases Group Trials Search Co‐ordinator (TSC) searched the Specialised Register (last searched January 2013) and CENTRAL (2012, Issue 12). Clinical trials databases were searched for details of ongoing or unpublished studies. In addition, reference lists of relevant articles were checked.
Selection criteria
Randomised clinical trials of prostanoids versus placebo or alternative ('control') treatment in people with intermittent claudication were considered for inclusion.
Data collection and analysis
Two authors independently assessed trial quality and extracted data. Primary outcomes included pain‐free walking distance (PFWD) and maximum walking distance (MWD), presented as mean change in walking distance during the course of the trial (% improvement) and as final walking distance (that is walking distance, in metres, after treatment) for the prostanoid and control groups.
Main results
Eighteen trials with a total of 2773 patients were included (16 in the original review and a further two in this update). As the majority of trials did not report standard deviations for the primary PFWD and MWD outcomes, it was often not possible to test for the statistical significance of any improvements in walking distance between groups. The quality of individual trials was variable and usually unclear due to insufficient reporting information. Comparison between trials was hampered by the use of different treadmill testing protocols, including different walking speeds and gradients. Such limitations in the data and the trial heterogeneity meant it was not possible to meaningfully pool results by meta‐analysis.
Four trials compared prostaglandin E1 (PGE1) with placebo; individual trials showed significant increases in walking distances with administration of PGE1 and in several trials the walking capacity remained increased after termination of treatment. Compared with pentoxifylline, PGE1 was associated with a higher final PFWD and MWD but these results were based on final walking distances rather than changes in walking distance from baseline. When PGE1 was compared with other treatments including laevadosin, naftidrofuryl and L‐arginine, improvements in walking distances over time were observed for both PGE1 and the alternative treatment, but it was not possible from the data available to analyse statistically whether or not one treatment was more effective than the other.
Six studies compared various preparations of prostacyclins (PGI2) with placebo. In one study using three different dosages of iloprost, PFWD and MWD appeared to increase in a dose‐dependent manner; iloprost was associated with headache, pain, nausea and diarrhoea, leading to a higher rate of treatment withdrawal. Of three studies using beraprost sodium, one showed an improvement in PFWD and MWD compared with placebo while two showed no significant benefit. Beraprost sodium was associated with an increased incidence of drug‐related adverse events. Of two studies on taprostene, the results of one in particular must be interpreted with caution due to an imbalance in walking capacity at baseline.
Comprehensive, high quality data on outcomes such as quality of life, ankle brachial index, venous occlusion plethysmography and haemorrheological parameters were lacking.
Authors' conclusions
Whilst results from some individual studies suggested a beneficial effect of PGE1, the quality of these studies and of the overall evidence available is insufficient to determine whether or not patients with intermittent claudication derive clinically meaningful benefit from the administration of prostanoids. Further well‐conducted randomised, double blinded trials with a sufficient number of participants to provide statistical power are required to answer this question.
Plain language summary
Prostanoids for intermittent claudication
Intermittent claudication (IC) is a symptom of lower limb ischaemia that results from peripheral arterial disease (PAD). It is evident as muscle pain (ache, cramp, numbness or sense of fatigue) in the leg muscles that occurs during exercise and is relieved by a short period of rest. Prostaglandin E1 (PGE1) and prostacyclin (PGI2), also known as prostanoids, are vasoactive drugs used in PAD to reduce arterial insufficiency. The aim of this review was to evaluate the effects of prostanoids in patients with IC. We identified 18 randomised studies with a total of 2773 participants, of which four studies compared the effects of PGE1 versus placebo. Overall, there was insufficient high quality evidence to suggest that PGE1 improves walking distances in people with IC. There was also a lack of evidence to determine if PGE1 was more effective than laevadosin, naftidrofuryl or L‐arginine. Evidence on the efficacy of prostacyclin was inconclusive. Results suggest that, compared with PGE1, prostacyclin may be associated with an increased occurrence of side effects including headache, diarrhoea and facial flushing.
Authors' conclusions
Background
Description of the condition
Peripheral arterial disease (PAD) is an important cause of morbidity in the general population, and intermittent claudication (IC) is the primary symptom of lower limb ischaemia. IC is usually caused by haemodynamically significant lesions of the aortoiliac, femoropopliteal or infrapopliteal vessels. IC is a clinical diagnosis given for muscle pain (ache, cramp, numbness or sense of fatigue) in the leg muscles, which occurs during exercise and is relieved by a short period of rest. Functional independence is threatened in people with IC due to limitations on mobility. A survey by The Scottish Vascular Audit Group reported that claudicants had a significantly impaired quality of life in all respects, from general health, pain, vitality and social parameters to mental and emotional wellbeing (Pell 1995). IC may progress to critical limb ischaemia (CLI), causing constant and intractable pain preventing sleep and is often associated with ulceration and gangrene of the foot. Ultimately, IC can lead to limb loss, surgical revascularisation, cardiovascular disease and even death (Dormandy 1999). Recent figures published in the Scottish Health Survey showed that the prevalence of symptomatic disease increases with age, affecting approximately 1.7% of people aged 16 to 54 years up to 7.4% in those aged > 75 years (The Scottish Health Survey 2010). Over a 16‐year period from 1991 to 2007, there were 41,953 new admissions for PAD in Scottish hospitals (Inglis 2012). Results of follow‐up studies of patients with IC have demonstrated that approximately 20% will experience a worsening of symptoms during two to seven years of follow up (Leng 1993). Of those that progress to CLI, approximately 50% of patients die within five years of presenting with symptoms (Davis 2005). The cost of treating CLI in the UK has been estimated at over £200 million per annum (Beard 2000). The five‐year cumulative incidence of amputation is low (about 1%). However, IC is a marker of generalised atherosclerosis and therefore an indicator associated with increased risk of premature death (Dormandy 1999).
Description of the intervention
Most patients with stable IC are managed conservatively (Coffman 1991; Ernst 1993). There are two main objectives of therapy: a) to retard the rate of progression of atherosclerosis, and b) to diminish the arterial insufficiency and thereby to treat the symptoms. Reduction of risk factors for progression of disease is the mainstay of the conservative medical management. Antithrombotic therapy with antiplatelet agents, especially aspirin, can lower the incidence of associated cardiovascular events (ATC 2002).
Since the majority of patients with IC are not at high risk of limb loss, the primary therapeutic goal is to improve exercise performance and the functional status of the patients. Exercise training has been demonstrated to be a very effective therapy to improve claudication‐limiting physical activity in patients with PAD (Leng 2000). However, there are data that suggest that, due to restrictions in training ability or poor compliance, only 30% of all patients with IC benefit from physical training (de la Haye 1992). Therefore, as an alternative or additive treatment, pharmacologic measures with so called vasoactive drugs may be applied.
Compared with exercise training, the place of vasoactive drugs such as prostaglandins, pentoxifylline, naftidrofuryl, cilostazol or buflomedil in the management of patients suffering from IC is less clear. Analysis of numerous publications on the efficacy of these vasoactive substances has shown that in the majority of studies, methodological errors and inaccuracies with regard to patient selection, investigative methods and statistical evaluation were evident (De Backer 2000; Heidrich 1992). However, a few studies have been published indicating a favourable effect of these drugs on the microcirculation or haemorrheological parameters (Hiatt 2001; Van den Brande 1998).
Prostaglandin E1 (PGE1) and prostacyclin (PGI2) have been used for the treatment of PAD for more than two decades. PGE1 is rapidly inactivated during passage through the lungs and therefore has to be given either intra‐arterially or in large intravenous doses (Belch 1997). While intra‐arterial application was recommended initially, intravenous infusions of PGE1 are nowadays preferred. PGI2 is a very potent but chemically unstable and short‐lived compound and thus has limited clinical use. Therefore more stable analogues such as iloprost, beraprost, taprostene or ciprostene, with similar pharmacodynamic profiles, have been developed.
How the intervention might work
The direct vasodilatory properties and subsequent changes in blood flow and antiplatelet activity are generally regarded as the most important effects of PGE1 and other prostanoids (Matsuo 1998; Scheffler 1991). However, many other pharmacological effects have been described that seem to be responsible for clinical efficacy (Grant 1992). The demonstrated effects of prostanoids include inhibition of neutrophilic activation and release of toxic oxygen radicals from activated leukocytes (white blood cells); reduction of adhesion of monocytes and macrophages; antiproliferative activity on vascular smooth muscle cells; reduction of extracellular matrix formation; an inhibitory action on platelet function; increase in fibrinolytic activity; influence on lipid metabolism; reduction of vascular cholesterol content; improvement of oxygen and glucose metabolism; and an influence on rheological parameters by increased erythrocyte (red blood cell) deformability.
Why it is important to do this review
After early enthusiasm for preliminary clinical findings, a more critical opinion concerning the clinical value of PGE1 and PGI2 has developed. There is now convincing evidence that PGE1 and iloprost, administered by intermittent daily intravenous infusions, are effective in patients with CLI (Verstraete 1994), and they are therefore recommended in such patients (Dormandy 2000). While the use of prostanoids in the management of severe symptomatic peripheral ischaemia (a condition for which no alternative drug therapy exists) can be recommended, it is still unknown whether patients with IC might benefit from this treatment. This is an update of a Cochrane Review first published in 2004.
Objectives
To determine the efficacy of prostanoids (PGE1, PGI2 and analogues such as iloprost, beraprost, taprostene, ciprostene) in improving walking distance and other objective and subjective outcomes in patients with IC Fontaine stage II.
Methods
Criteria for considering studies for this review
Types of studies
Randomised clinical trials that consider the effects of prostanoids for the treatment of patients with IC Fontaine stage II. Controlled prospective trials with parallel groups formed without randomisation were excluded.
Types of participants
Patients with IC (Fontaine stage II). There were no restrictions for age or gender.
Types of interventions
Treatment with PGE1, PGI2 and analogues administered by any route, over any time period, compared with placebo or any alternative ('control') treatment.
Types of outcome measures
Primary outcomes
-
Pain‐free walking distance (PFWD) or the initial claudication distance (ICD), which is the distance walked on a treadmill before the onset of pain
-
Maximum walking distance (MWD) or the absolute claudication distance (ACD), which is the maximum or absolute distance walked on a treadmill
-
Drug side effects
For trials which reported walking time, we converted this to walking distance in metres using the reported walking gradient and speed. We planned to convert data obtained using graded treadmill tests into data comparable with the data obtained using a constant load treadmill test where this was possible.
Secondary outcomes
-
Quality of life
-
Ankle brachial index (ABI), also known as ankle brachial pressure index
-
Venous‐occlusion plethysmography
-
Haemorrheological parameters
-
Necessity of vascular surgery
-
Occurrence of amputation
-
Mortality
-
Cardiovascular events
Search methods for identification of studies
We did not apply any language restrictions on publications, or any restrictions regarding publications status.
Electronic searches
For this update, the Cochrane Peripheral Vascular Diseases Group Trials Search Co‐ordinator (TSC) searched the Specialised Register (last searched January 2013) and the Cochrane Central Register of Controlled Trials (CENTRAL) (2012, Issue 12), part of The Cochrane Library (www.thecochranelibrary.com). See Appendix 1 for details of the search strategy used to search CENTRAL. The Specialised Register is maintained by the TSC and is constructed from weekly electronic searches of MEDLINE, EMBASE, CINAHL, and AMED, and through handsearching relevant journals. The full list of the databases, journals and conference proceedings which have been searched, as well as the search strategies used, are described in the Specialised Register section of the Cochrane Peripheral Vascular Diseases Group module in The Cochrane Library (www.thecochranelibrary.com).
Ongoing studies and unpublished studies
The following trials databases were searched (February 2013) by AA using the terms 'intermittent claudication' for details of ongoing and unpublished trials:
-
World Health Organization International Clinical Trials Registry Platform (http://apps.who.int/trialsearch/);
-
ClinicalTrials.gov (http://clinicaltrials.gov/);
-
Current Controlled Trials (http://www.controlled‐trials.com/).
Searching other resources
We searched the reference lists of articles retrieved by electronic searches for additional citations. We contacted trialists for further information in cases where there were missing data or doubts about whether to include trials in the review.
Data collection and analysis
Selection of studies
For this update, two review authors (AA and LR) used the selection criteria to identify trials for inclusion. Disagreements were resolved by discussion. We obtained full versions of articles that potentially met the inclusion criteria based on the title or abstract and assessed these independently against the inclusion criteria. The reason for each study's exclusion is presented in the 'Characteristics of excluded studies' table.
Data extraction and management
Both review authors (AA and LR) independently extracted the data for the new studies considered using the standard checklist developed by the Cochrane Peripheral Vascular Diseases Group.
The following outcome data were sought (if assessed in the individual studies): PFWD and MWD on a standard treadmill; drug side effects; ABI; venous‐occlusion plethysmography; haemorrheological parameters; necessity of vascular surgery; occurrence of amputation; death; cardiovascular events. Authors of publications were contacted for additional details. Data were entered into Review Manager (RevMan), the systematic review software from The Cochrane Collaboration (RevMan 2011).
Assessment of risk of bias in included studies
Two review authors (AA, LR) independently assessed the risk of bias using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). The risk of bias tool provides a strict protocol to assess allocation (selection bias), blinding (performance bias and detection bias), incomplete outcome data (attrition bias), selective reporting (reporting bias) and other potential sources of bias. For each of the six domains, we assessed the risk of bias as 'low risk', 'high risk' or 'unclear risk' of bias, with unclear risk of bias indicating that there was insufficient information available to permit an assessment of either low or high risk of bias. Disagreements were resolved by discussion between the two review authors.
Measures of treatment effect
We planned to pool the data on the PFWD and the MWD from each trial to arrive at an overall estimate of the effectiveness of the pharmacological interventions. We aimed to calculate the percentage change in the walking distance during the course of the trial and then, where possible, calculate the mean difference in the prostanoid group compared with the control group. For continuous outcomes such as PFWD and MWD, statistical analysis was presented as mean difference (MD) with 95% confidence interval (CI). For dichotomous outcomes, we used odds ratios (OR) with 95% CI. However, few studies presented the standard deviations (SD) of the walking distances and therefore it was seldom possible to analyse the mean difference in change in walking distance from baseline to follow up between the treatment and control groups. Therefore the final walking distance, that is the walking distance at the end of treatment (in metres), was also presented and pooled, where possible. There are obvious limitations in presenting differences in final walking distances between treatment and control groups, the main one being that this does not take into account any differences in walking distances between the two groups at baseline, that is before treatment. Furthermore, the relationship between treadmill speed and walking distance is not linear but is also influenced by the gradient of the treadmill. Individual studies used a range of speeds and gradients for treadmill testing; and while these measures were converted into final walking distances in metres, it could be argued that the walking distance on a gradient of 7 ° would be different to the walking distance on a gradient of 10 °. Converting these results into a standard measure of final walking distance in metres is likely to result in significant heterogeneity when pooling the studies. Nevertheless, it was decided to use all available data in order to make the analysis as complete as possible.
Unit of analysis issues
The unit of randomisation was the individual participant in all included studies. All of the trials involved repeat observations on participants at different points in time, ranging from one week to one year, and thus were prone to unit of analysis errors. Data on PFWD, MWD and all other outcomes were presented for all times reported in the studies and analysed, where possible.
Dealing with missing data
Where necessary, we contacted the authors of included trials to clarify data and obtain missing data.
Assessment of heterogeneity
All analyses were conducted on an intention‐to‐treat basis. When the individual trials did not use intention‐to‐treat analyses, the analyses in this review were on the basis of the data (absolute numbers) provided in the included trial report. The statistical appropriateness of combining the trials was based on tests of heterogeneity, which consider whether differences in treatment effect over individual trials are consistent with natural variation around a constant effect. We assessed trial heterogeneity using the Chi2 test and I2 statistic, which describes the percentage of the variability in effect estimates that is due to heterogeneity rather than sampling error (chance). Where heterogeneity was identified (P < 0.1, or I2 > 50%), we investigated the reason for heterogeneity. If no apparent reason was found, we conducted a random‐effects model meta‐analysis to incorporate heterogeneity among trials. In the absence of heterogeneity we used a fixed‐effect model.
Assessment of reporting biases
We planned to assess reporting biases by using funnel plots if more than 10 studies were included in the meta‐analysis.
Data synthesis
Where two or more studies with low methodological and statistical heterogeneity were included, we performed a meta‐analysis. Results for changes in walking distances were stated in percentage values. For final walking distances, the results were stated in metres. Numeric values were provided as weighted mean differences (WMD) with the 95% confidence intervals (CI) in parenthesis. A P value of < 0.05 was considered as significant for analysis of potential differences between the treatment groups.
Where heterogeneity was identified (P < 0.1, or I2 > 50%), we investigated the reason for heterogeneity. If no apparent reason was found, we conducted a random‐effects model meta‐analysis to incorporate the heterogeneity among trials. In the absence of heterogeneity we used a fixed‐effect model.
Subgroup analysis and investigation of heterogeneity
We anticipated that the trials would not be homogeneous. Therefore we planned to do a subgroup analysis of the included trials according to variables such as duration, dose and route of administration.
Results
Description of studies
See: Characteristics of included studies; Characteristics of excluded studies; Characteristics of studies awaiting classification
Results of the search
For this update 20 studies were retrieved from the search of the Specialised Register. Five additional studies were found from the search of CENTRAL which were not in the Specialised Register. All 25 studies were screened by title and abstract. Two studies (Creager 2008; Milio 2006) were deemed relevant and included in the update. From the updated search, 14 studies were excluded (Acciavatti 2001; Anon 2011; Barradas 1989; Bieron 1993; Diehm 1990; Esato 1995a; Esato 1995b; Fitscha 1985; Goya 2003; Ishitobi 1991; Linhart 1998; Okadome‐Kenchiro 1992; Sakaguchi‐Shukichi 1990; Wang 2009). Three studies were deemed not relevant as the treatment or outcome did not meet the scope of the review. One study (Valerio 2012) is ongoing and awaiting publication. One study (Nakagawa 1998) is awaiting translation before it can be determined if it is relevant for inclusion in the review. Four studies (Blume 1986; Creutzig 1988; Diehm 1989; Lièvre 2000) were duplicate publications of studies already included in the original review.
Included studies
Of the 18 included studies (16 in the original review and two from the update), 13 (Belch 1997; Blume 1986; Böger 1998; Creager 2008; Creutzig 1988; Diehm 1989; Diehm 1997; Hepp 1996; Lièvre 1996; Lièvre 2000; Mangiafico 2000; Mohler 2003; Scheffler 1994) excluded all participants with changes in walking distances exceeding more than 20% to 25% from baseline during a wash‐out phase of about one to four weeks in which no study drugs were administered. The most frequent, additional exclusion criteria for trial entry were: pregnancy, congestive heart failure, decompensated renal failure, haemodynamically relevant aortic and pelvic arterial occlusion, respiratory disorders, myocardial infarction within the previous six months, insulin‐dependent diabetes mellitus, indication for revascularisation procedures, severe uncontrolled hypertension (diastolic blood pressure > 120 mmHg), orthopaedic and neurological diseases that impaired walking performance, thrombocytosis (increase in the numbers of circulating platelets) with count > 400,000/ml, and second and third degree atrioventricular block.
Four studies compared PGE1 with placebo (Belch 1997; Blume 1986; Diehm 1989; Mangiafico 2000), four studies compared PGE1 with pentoxifylline (Hepp 1996; Luk'Janov 1995; Milio 2006; Scheffler 1994), one study compared PGE1 with laevadosin (Creutzig 1988), one study compared PGE1 with naftidrofuryl (Diehm 1989), one study compared PGE1 with L‐arginine (Böger 1998), six studies compared PGI2 with placebo (Creager 2008; Lièvre 1996; Lièvre 2000; Mohler 2003; Virgolini 1989; Virgolini 1990), one study compared PGI2 with pentoxifylline (Creager 2008) and one study compared PGI2 with hydroxy‐ethyl starch (Müller‐Bühl 1987).
Sixteen studies measured walking distances in metres (Belch 1997; Blume 1986; Böger 1998; Creager 2008; Creutzig 1988; Diehm 1989; Diehm 1997; Hepp 1996; Lièvre 1996; Lièvre 2000; Luk'Janov 1995; Mangiafico 2000; Milio 2006; Mohler 2003; Müller‐Bühl 1987; Scheffler 1994). Of the 16 studies, 15 used a constant load test with gradients ranging between 3.2 ° (Lièvre 1996), 5 ° (Blume 1986; Creutzig 1988; Hepp 1996; Luk'Janov 1995; Mangiafico 2000; Scheffler 1994), 7 ° (Milio 2006), 10 ° (Belch 1997; Diehm 1989; Lièvre 2000; Mohler 2003; Müller‐Bühl 1987) and 12 ° (Böger 1998; Diehm 1997) and speeds ranging between 2 km/h (Belch 1997), 3 km/h (Blume 1986; Böger 1998; Creutzig 1988; Diehm 1997; Hepp 1996; Lièvre 2000; Luk'Janov 1995; Mangiafico 2000; Milio 2006; Mohler 2003; Müller‐Bühl 1987; Scheffler 1994), 3.2 km/h (Lièvre 1996) and 3.5 km/h (Diehm 1989). One study used a graded test with a starting incline of 0 ° which was increased by 2 ° every 2 minutes, with a walking speed of 3.2 km/h (Creager 2008). Two studies measured walking distances in seconds (Virgolini 1989; Virgolini 1990) on a slope of 7.5 º and at a speed of 1.5 mph. The walking times in these studies were converted into walking metres using the gradient and walking speed reported by the study authors.
Three studies (Blume 1986; Milio 2006; Müller‐Bühl 1987) provided no information about the intake of antiplatelet or vasoactive drugs during the trial. In the studies by Belch (Belch 1997), Creager (Creager 2008), Mohler (Mohler 2003) and Mangiafico (Mangiafico 2000) the participants continued a previously indicated therapy with antiplatelet drugs. The participants in the remaining 11 studies were not treated by any vasoactive or antiplatelet drug during the study. Only eight studies (Böger 1998; Creutzig 1988; Diehm 1989; Hepp 1996; Lièvre 1996; Mangiafico 2000; Mohler 2003; Scheffler 1994) provided detailed information about concomitant vascular exercise. In the study by Scheffler 1994 all three groups received intensive physical training in addition to treatment with either PGE1 or pentoxifylline.
Prostanoids were administered by intra‐arterial or intravenous infusion, or orally.
For further details of the characteristics of the included studies see Characteristics of included studies.
Excluded studies
In total, 20 studies (six from the original review and 14 from the update) were excluded because they did not meet the inclusion criteria. See Characteristics of excluded studies for details of the reasons for exclusion. In brief, seven trials were not randomised controlled trials (Barradas 1989; Bieron 1993; Diehm 1990; Ishitobi 1991; Rudofsky 1987; Rudofsky 1988; Sakaguchi‐Shukichi 1990). Two of those trials (Rudofsky 1987; Rudofsky 1988) were included in the original review but we excluded them for this update as we believe that they were not randomised controlled trials. Four studies (Acciavatti 2001; Anon 2011; Linhart 1998; Okadome‐Kenchiro 1992) had no placebo or control group, two studies (Hay 1987; Waller 1986) administered different drug dosages to patients limiting comparability, one study (Ylitalo 1990) had no subgroup analysis despite three of the participants having rest pain, one study (Fitscha 1985) had incomplete data, while a further study (Wilkinson 1988) did not present any data. Two studies were excluded as all patients had CLI (Esato 1995a; Esato 1995b) while a further study (Goya 2003) was excluded as patients had arteriosclerotic changes in the carotid artery. Finally, one study (Wang 2009) was excluded as not all patients had IC and the outcomes measured were not relevant for the review.
Risk of bias in included studies
The quality of the included trials was variable. See Figure 1 and Figure 2 for a graphical presentation of the risk of bias. Insufficient information was the main reason for the 'unclear' rating for most trials.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Allocation
For random sequence generation, only four of the 18 included studies (Belch 1997; Lièvre 1996; Lièvre 2000; Milio 2006) adequately described the methods used, with random number lists used in two studies (Belch 1997; Milio 2006) and random computer generators used in the other two (Lièvre 1996; Lièvre 2000). The remaining 14 studies (Blume 1986; Böger 1998; Creager 2008; Creutzig 1988; Diehm 1989; Diehm 1997; Hepp 1996; Luk'Janov 1995; Mangiafico 2000; Mohler 2003; Müller‐Bühl 1987; Scheffler 1994; Virgolini 1989; Virgolini 1990) did not state the method used to generate the allocation sequence and thus there was insufficient information to permit a judgment of low or high risk of selection bias. In terms of concealing the allocation sequence, only one study (Milio 2006) used sealed envelopes and was therefore at low risk of selection bias. One study (Belch 1997) was deemed to be at high risk due to the fact that it used a random number list, which, according to the Cochrane criteria for assessing risk of bias (Higgins 2011), can introduce selection bias. The remaining 16 studies (Blume 1986; Böger 1998; Creager 2008; Creutzig 1988; Diehm 1989; Diehm 1997; Hepp 1996; Lièvre 1996; Lièvre 2000; Luk'Janov 1995; Mangiafico 2000; Mohler 2003; Müller‐Bühl 1987; Scheffler 1994; Virgolini 1989; Virgolini 1990) did not provide adequate descriptions of methods used to conceal allocation and therefore a judgment on the risk of selection bias could not be made.
Blinding
Although some studies (Belch 1997; Creager 2008; Creutzig 1988; Diehm 1989; Diehm 1997; Hepp 1996; Lièvre 1996; Lièvre 2000; Luk'Janov 1995; Mangiafico 2000; Mohler 2003; Virgolini 1989; Virgolini 1990) were reported as 'double blind', authors did not provide any more information on how blinding of study participants and personnel was maintained throughout the study. As it could not be guaranteed that blinding was ensured and unbroken, these studies were deemed to be at unclear risk of performance bias. Five studies (Blume 1986; Böger 1998; Milio 2006; Müller‐Bühl 1987; Scheffler 1994) were deemed to be at high risk of performance bias. In Blume 1986, PGE1 was administered as an intra‐arterial infusion over 90 minutes while the placebo was an injection, and therefore blinding of the study participants or personnel was impossible. Böger 1998 blinded two of the three treatment groups. Milio 2006 and Müller‐Bühl 1987 both conducted single blinded trials with only the participants blinded to the treatment. Finally, in the study by Scheffler 1994, one of the 'treatment' groups received no treatment. None of the 18 included studies provided any information on the blinding of outcome assessors and therefore judgements on the risk of detection bias could not be made.
Incomplete outcome data
Seven studies (Belch 1997; Böger 1998; Diehm 1989; Lièvre 2000; Mangiafico 2000; Müller‐Bühl 1987; Scheffler 1994) accounted for all data and were deemed to be at low risk of attrition bias. Six studies (Blume 1986; Creager 2008; Creutzig 1988; Diehm 1997; Hepp 1996; Virgolini 1990) were deemed to be at high risk due to large losses to follow up with inadequate explanations. For the remaining five studies (Lièvre 1996; Luk'Janov 1995; Milio 2006; Mohler 2003; Virgolini 1989) there was insufficient information to permit a judgement on risk of attrition bias.
Selective reporting
Seven studies (Böger 1998; Diehm 1989; Diehm 1997; Hepp 1996; Lièvre 1996; Lièvre 2000; Müller‐Bühl 1987) reported data on all pre‐specified outcomes and were therefore at low risk of reporting bias. Seven studies (Belch 1997; Blume 1986; Creager 2008; Creutzig 1988; Mangiafico 2000; Scheffler 1994; Virgolini 1990) did not present data on all pre‐specified outcomes and were deemed to be at high risk of reporting bias. The remaining four studies (Luk'Janov 1995; Milio 2006; Mohler 2003; Virgolini 1989) did not provide sufficient information to permit a judgement on the level of reporting bias.
Other potential sources of bias
Only one study (Müller‐Bühl 1987) appeared to be free from other sources of bias. The remaining studies were either at high risk of bias (Belch 1997; Creager 2008) because of unclear study design and missing inclusion and exclusion criteria (Belch 1997) and sponsorship from a pharmaceutical company (Creager 2008) or at unclear risk of bias (Blume 1986; Böger 1998; Creutzig 1988; Diehm 1989; Diehm 1997; Hepp 1996; Lièvre 1996; Lièvre 2000; Luk'Janov 1995; Mangiafico 2000; Milio 2006; Mohler 2003; Scheffler 1994; Virgolini 1989; Virgolini 1990) because of insufficient information to permit judgement of low or high risk of bias.
Effects of interventions
PGE1 versus placebo (Comparison 1)
Walking distances
Three studies (Blume 1986; Diehm 1997; Mangiafico 2000), with a total of 300 participants (152 PGE1 group and 148 placebo group), analysed the effects of PGE1 versus placebo. Only one study (Blume 1986) with 50 participants (25 PGE1 group and 25 placebo group) investigated the effects of intra‐arterial (ia) infusions on PFWD and MWD. After three weeks of treatment, the mean PFWD improved by 109% (59 m) in PGE1 patients compared to 35% (19 m) in placebo patients, while the mean MWD improved by 74% (72 m) and 18% (17 m) in the two treatment groups respectively. Based on final walking distances, the PFWD improved by 40 m (95% CI 13.42 to 66.58 m, P = 0.003) and MWD by 56 m (95% CI 20.45 to 91.55 m, P = 0.002) in patients treated with intra‐arterial PGE1 compared to a placebo.
The two remaining studies (Diehm 1997; Mangiafico 2000) measured the effects of intravenous (iv) infusions. Both studies measured the PFWD and MWD after four weeks of treatment and individual results showed significant improvements in patients treated with PGE1. Diehm 1997 examined 208 patients (106 PGE1 and 102 placebo) and estimated that after four weeks of treatment PGE1 improved the PFWD by 75% compared to 43% with the placebo, while MWD improved by 65% with PGE1 compared to 42% with the placebo. Standard deviations around the mean differences were not presented and therefore significance levels could not be assessed. Based on final walking distances and compared to a placebo, PGE1 improved PFWD and MWD by 17.2 m (95% CI 16.68 to 17.72 m) and 21.1 m (95% CI 20.62 to 21.58 m), respectively. The study by Mangiafico 2000 was much smaller with only 42 patients (21 in each group) and, after four weeks of treatment, the PFWD improved by 87.5% compared to 4% in the placebo group, and MWD improved by 90% compared to 1% in the placebo group. Significance levels could not be assessed as standard deviations around the differences were not presented in this publication. Based on final walking distances, PGE1 was associated with an increase in PFWD of 51 m (95% CI 35.12 to 66.88 m) while the MWD was 112 m (95% CI 80.67 to 143.33 m) compared to placebo. However, results based on final walking distances must be interpreted with caution as they do not take into account the change from baseline. In this study (Mangiafico 2000), baseline walking distances were not comparable between the treatment groups: the PFWD was 9 m less and the MWD was 12 m less in PGE1 patients compared to the placebo controls. In this study, reporting final walking distances alone led to an under‐reporting on the improvement in PFWD and MWD with PGE1. The two studies could not be combined in a meta‐analysis as they were deemed too heterogeneous (I2 = 97%).
The study by Diehm 1997 also administered treatment for eight weeks and measured walking distances at this time point. After an initial treatment period of four weeks with PGE1 or placebo administered five days a week, an interval treatment period of four weeks was applied where treatment was reduced to bi‐weekly. After this four‐week interval, and eight weeks after treatment first commenced, the mean PFWD in PGE1 patients improved by 100% from baseline compared to 60% in the placebo group. Using final walking distance to estimate improvement, the PFWD was increased by 22.30 m (95% CI 21.74 to 22.86 m) with PGE1 (P < 0.001). At the same time point, MWD improved by 88% compared to 61% in the PGE1 and placebo groups respectively. Compared with placebo, PGE1 led to an improvement of 25.8 m (95% CI 25.27 to 26.33 m) in MWD (P < 0.001). The study by Mangiafico 2000 also measured walking distance at eight weeks although treatment was administered for only four weeks. This study was not included in a meta‐analysis as the study authors did not report the PFWD and MWD at eight weeks in the placebo group. Attempts were made to retrieve this information from the author but we were unsuccessful. Mean PFWD increased by 57% (41 m) while the mean MWD improved by 64% (89 m) from baseline in the PGE1 group. Data on walking distances at eight weeks in the placebo group were not presented.
A further study (Belch 1997) analysed the PGE1 prodrug AS‐013 (59 participants) versus placebo (21 participants) but it could not be included in our analysis because the data were reported in medians rather than means. Those who received PGE1 were split into three groups according to dosing schedule: 2 µg five days a week (n = 19), 5 µg two days a week (n = 18) and 5 µg five days a week (n = 22). The authors only reported walking distances for the highest dose group. At four weeks, there was no significant difference in the median improvement of PFWD between the PGE1 group (20 m, interquartile range (IQR) 33 m) and the placebo group (9 m, IQR 18 m) (P > 0.05). However, the study authors reported a significant difference between the PGE1 group (28 m, IQR 81 m) and the placebo group (4 m, IQR 20 m) (P < 0.05) in the improvement of the MWD. At eight weeks, the median PFWD increased by 20.9 m (IQR 43 m) in PGE1 patients but did not improve in placebo patients (0 m, IQR 29 m) (P < 0.01) while the median MWD improved by 35 m (IQR 68 m) but decreased by 11.2 m (IQR 35 m) in the PGE1 and placebo groups respectively (P < 0.01). However, PFWD and MWD at baseline were not comparable between the placebo and PGE1 patients and thus the results should be interpreted with caution.
| Study | Dose (µg) | Duration (weeks) | PGE | PLC | PGE0 | SD | PGEE | SD | %age | SD% | PLC0 | SD | PLCE | SD | %age | SD% | DIFF (%) |
| 2, 5 d/wk | 4 | 19 | 21 | 44.8 | 16.0 | not stated | |||||||||||
| 5, 2 d/wk | 18 | 50.7 | 33.5 | not stated | |||||||||||||
| 5, 5 d/wk | 22 | 45.0 | 33.0 | 64.5 | 43 | 17 | 26 | ||||||||||
| 2, 5 d/wk | 8 a | 19 | 21 | 44.8 | 16.0 | not stated | |||||||||||
| 5, 2 d/wk | 18 | 50.7 | 33.5 | not stated | |||||||||||||
| 5, 5 d/wk | 22 | 45.0 | 33.0 | 65.9 | 46 | 0 | 46 | ||||||||||
| 10‐20 | 3 | 25 | 25 | 54 | 13 | 113 | 54 | 109 | 54 | 16 | 73 | 41 | 35 | 74 | |||
| 60 | 4 | 106 | 102 | 64.3 | 1.6 | 112.7 | 1.8 | 75 | 66.6 | 1.6 | 95.5 | 2.0 | 43 | 32 | |||
| 8 b | 128.9 | 2.0 | 100 | 106.6 | 2.1 | 60 | 40 | ||||||||||
| 60 | 4 | 21 | 21 | 72 | 16 | 135 | 33 | 88 | 81 | 17 | 84 | 17 | 4 | 84 | |||
| 8 a | 113 | 26 | 57 | not stated | not stated |
PGE = prostaglandin treatment group sample size
PLC = placebo group sample size
PGE0 = PGE baseline walking distance
SD = standard deviation
PGEE = PGE end walking distance
%AGE = percentage improvement of walking distance
SD% = standard deviation of percentage improvement of walking distance
PLC0 = placebo baseline walking distance
PLCE = placebo end walking distance
DIFF = difference in percentage of improvement of PGE and placebo
* Study reported PFWD as a median with IQR.
a = treatment was administered for 4 weeks, treadmill tests conducted at 4 weeks and also at 8 weeks (after a 4‐week period of no treatment).
b = treatment was administered for 5 days a week for 4 weeks and reduced to 2 days a week for a further 4 weeks, treadmill tests conducted at both 4 and 8 weeks.
| Study | Dose (µg) | Duration (weeks) | PGE | PLC | PGE0 | SD | PGEE | SD | %age | SD% | PLC0 | SD | PLCE | SD | %age | SD% | DIFF (%) |
| 2, 5 d/wk | 4 | 19 | 21 | 60.0 | 25.6 | not stated | |||||||||||
| 5, 2 d/wk | 18 | 68.6 | 48.8 | not stated | |||||||||||||
| 5, 5 d/wk | 22 | 67.1 | 68.8 | 95.1 | 42 | 75.6 | 93.0 | 80.1 | 6 | 36 | |||||||
| 2, 5 d/wk | 8 a | 19 | 21 | 60.0 | 25.6 | not stated | |||||||||||
| 5, 2 d/wk | 18 | 68.6 | 48.8 | not stated | |||||||||||||
| 5, 5 d/wk | 22 | 67.1 | 68.8 | 102.1 | 52 | 75.6 | 93.0 | 64.4 | ‐15 | 67 | |||||||
| 10‐20 | 3 | 25 | 25 | 98 | 45 | 170 | 75 | 74 | 97 | 28 | 114 | 51 | 18 | 56 | |||
| 60 | 4 | 106 | 102 | 98.8 | 1.5 | 163 | 1.8 | 65 | 99.8 | 1.4 | 141.9 | 1.7 | 42 | 23 | |||
| 8 b | 186.3 | 2.0 | 88 | 160.5 | 1.9 | 61 | 27 | ||||||||||
| 60 | 4 | 21 | 21 | 140 | 30 | 266 | 62 | 90 | 152 | 38 | 154 | 39 | 1 | 89 | |||
| 8a | 229 | 55 | 64 | not stated | not stated |
PGE = prostaglandin treatment group sample size
PLC = placebo group sample size
PGE0 = PGE baseline walking distance
SD = standard deviation
PGEE = PGE end walking distance
%AGE = percentage improvement of walking distance
SD% = standard deviation of percentage improvement of walking distance
PLC0 = placebo baseline walking distance
PLCE = placebo end walking distance
DIFF = difference in percentage of improvement of PGE and placebo
* Study reported MWD as a median with IQR.
a = treatment was administered for 4 weeks, treadmill tests conducted at 4 weeks and also at 8 week (after a 4‐week period of no treatment).
b = treatment was administered for 5 days a week for 4 weeks and reduced to 2 days a week for a further 4 weeks, treadmill tests conducted at both 4 and 8 weeks.
Quality of life
Quality of life was measured in two studies. Mangiafico 2000 used two methods, the walking impairment questionnaire (WIQ) which specifically assesses changes in walking capacity in response to treatment of IC and the RAND 36‐Item health Survey, which is a generic tool designed to measure health‐related quality of life. After four weeks of treatment, analysis of the WIQ demonstrated significant improvements in the walking impairment, distance, speed and stair climbing scores (all P < 0.001) while the RAND survey showed improvements in physical function (P < 0.001) and bodily pain scores (P < 0.01). At the end of the eight‐week treatment‐free follow up, all WIQ and RAND scores were still increased compared with baseline (P < 0.01). Belch 1997 used an 8‐point questionnaire to determine quality of life and concluded that, overall, patients given a placebo deteriorated by a score of 64 points while those administered the PGE1 prodrug AS‐013 improved by a score of 22 to 73 points, with the most marked difference being observed in physical functioning and leisure activities.
Ankle brachial index (ABI)
ABI was an outcome in four studies (Belch 1997; Blume 1986; Diehm 1997; Mangiafico 2000). No study presented data but all study authors stated that there were no significant differences in ABI between the PGE1 and placebo groups throughout the course of the studies.
Venous occlusion plethysmography
Blume 1986 performed venous occlusion plethysmography and, although data were not reported, the study authors stated that there were no significant changes during or after therapy.
Haemorrheological parameters
The study by Blume 1986 reported that the haemorrheological parameters including haematocrit, erthrocyte aggregation and deformability) were normal at the start of the study and did not change significantly in either the PGE1 or placebo group.
Adverse events
Adverse events were reported in four studies (Belch 1997; Blume 1986; Diehm 1997; Mangiafico 2000). In one study (Blume 1986), treatment was stopped due to side effects (pain, swelling, flushing) in 1/25 placebo patients and 4/25 PGE1 patients. In another trial (Belch 1997), 2/59 PGE1 patients experienced severe treatment‐related events (atrial fibrillation and hypotension). Flu‐like symptoms occurred in 7/80 patients (5/59 PGE1, 2/21 placebo), dyspepsia in 3/59 PGE1 patients and mild injection site reactions occurred in 6/59 PGE1 patients (Belch 1997). Two studies (Diehm 1997; Mangiafico 2000) reported no serious drug‐related adverse events. The rate of less severe events (skin reddening, pain at infusion site, dizziness and nausea) was 12.8% in the 106 PGE1 patients and 7.7% in the 102 placebo patients respectively (Diehm 1997) while skin reddening occurred in 9.5% of PGE1 patients (Mangiafico 2000). It was not possible to pool the adverse events (AEs) data into a meta‐analysis because they differed in severity. Blume 1986 only reported serious AEs that resulted in treatment discontinuation, while Diehm 1997 and Mangiafico 2000 did not present the number of patients.
PGE1 versus pentoxifylline (Comparison 2)
Walking distances
Four studies (Hepp 1996; Luk'Janov 1995; Milio 2006; Scheffler 1994) analysed iv PGE1 versus iv pentoxifylline. Although the mean difference in walking distance could be calculated for each study, the SDs for the mean differences were not reported and therefore the studies could not be pooled in a meta‐analysis. Hepp 1996 conducted a study on 195 patients (97 PGE1, 98 pentoxifylline). Results showed that administration of PGE1 and pentoxifylline significantly increased the PFWD by 218% (83 m to 264 m) and 124% (84 m to 188 m), respectively, and the MWD by 164% (130 m to 343 m) and 146% (131 m to 322 m), respectively. Luk'Janov 1995 measured the PFWD only and reported a significant improvement in PFWD of 119% (89 m to 195 m) in the PGE1 group (n = 42) and 91% (78 m to 149 m) in the pentoxifylline group (n = 40). It was not clear at what time point this was measured and this could not be clarified by the study authors. Meanwhile Milio 2006 showed that, after four weeks of treatment, PFWD increased with both PGE1 (396%) and pentoxifylline (113%), while MWD improved by 260% in 63 PGE1 patients and 118% in 60 pentoxifylline‐treated patients. Based on final walking distances and compared to pentoxifylline, PGE1 was associated with an improvement in PFWD of 212 m (95% CI 130 to 294 m) while MWD increased by 192 m (95% CI 95 to 289 m). The study by Scheffler 1994 comprised 45 patients, 15 received PGE1, 15 pentoxifylline, and 15 underwent physical exercise alone. Four weeks after treatment, the mean PFWD had increased by 119% in the exercise group (72 m to 158 m), 105% in the pentoxifylline group (75 m to 154 m) and 604% in the PGE1 group (81 m to 570 m). Compared to pentoxifylline, PGE1 was associated with an increase in PFWD of 416 m (95% CI 27.7 to 804.3 m). Mean MWD also increased by 99% (131 m to 261 m), 119% (160 m to 350 m) and 371% (158 m to 744 m ) in the physical exercise, pentoxifylline and PGE1 groups respectively. Final MWD was improved by 393 m (95% CI ‐32.5 to 818.5 m) in favour of PGE1. Forty‐four patients were followed up for one year and a reduction in PFWD was observed, with the extent of deterioration differing by treatment. In both the exercise and pentoxifylline groups, the symptomatic walking distance remained at just 30% above pretreatment distance while the PGE1 patients still showed a 149% increase in PFWD.
Two studies (Milio 2006; Scheffler 1994) that reported SDs for final walking distances were combined in a meta‐analysis as low heterogeneity was determined (I2 = 2% and 0% for PFWD and MWD respectively). Combining these two studies (77 PGE1 and 75 pentoxifylline patients) and based on final walking distances, PGE1 was associated with a significant improvement in PFWD (221 m, 95% CI 141 to 300 m, P < 0.001) and MWD (202 m, 95% CI 107 to 297 m, P < 0.001).
| Study | Dose PGE (µg) | Dose PTX (mg) | Duration (weeks) | PGE1 | PTX | PGE0 | SD | PGEE | SD | %age | SD% | PTX0 | SD | PTXE | SD | %age | SD% | DIFF (%) |
| 80 | 400 | 4 | 97 | 98 | 83 | 264 | 218 | 84 | 188 | 124 | 94 | |||||||
| 40 | 400 | 4 | 42 | 40 | 89 | 195 | 119 | 78 | 149 | 91 | 45 | |||||||
| 60 | 200 | 4 | 63 | 60 | 78 | 36 | 387 | 274 | 396 | 82 | 41 | 175 | 180 | 113 | 260 | |||
| 80 | 200 | 4 | 14 | 15 | 81 | 23 | 570 | 727 | 604 | 75 | 41 | 154 | 150 | 105 | 499 |
PGE = prostaglandin treatment group sample size
PTX= pentoxifylline group sample size
PGE0 = PGE baseline walking distance
SD = standard deviation
PGEE = PGE end walking distance
%AGE = percentage improvement of walking distance
SD% = standard deviation of percentage improvement of walking distance
PTX0 = pentoxifylline baseline walking distance
PTXE = pentoxifylline end walking distance
DIFF = difference in percentage of improvement of PGE and pentoxifylline
| Study | Dose PGE (µg) | Dose PTX (mg) | Duration (weeks) | PGE1 | PTX | PGE0 | SD | PGEE | SD | %age | SD% | PTX0 | SD | PTXE | SD | %age | SD% | DIFF (%) |
| 80 | 400 | 4 | 97 | 98 | 130 | 343 | 164 | 131 | 322 | 146 | 18 | |||||||
| 40 | 400 | 4 | 42 | 40 | ‐ | ‐ | ‐ | ‐ | ‐ | ‐ | ‐ | ‐ | ‐ | ‐ | ‐ | ‐ | ‐ | |
| 60 | 200 | 4 | 63 | 60 | 143 | 55 | 515 | 285 | 260 | 148 | 69 | 323 | 264 | 118 | 158 | |||
| 80 | 200 | 4 | 14 | 15 | 158 | 95 | 744 | 697 | 371 | 160 | 133 | 351 | 432 | 119 | 252 |
PGE = prostaglandin treatment group sample size
PTX= pentoxifylline group sample size
PGE0 = PGE baseline walking distance
SD = standard deviation
PGEE = PGE end walking distance
%AGE = percentage improvement of walking distance
SD% = standard deviation of percentage improvement of walking distance
PTX0 = pentoxifylline baseline walking distance
PTXE = pentoxifylline end walking distance
DIFF = difference in percentage of improvement of PGE and pentoxifylline
Ankle brachial index (ABI)
ABI was measured in two studies (Hepp 1996; Luk'Janov 1995). In one study (Hepp 1996), the ABI increased significantly after four weeks of treatment with both PGE1 (0.47 (SD 0.21) to 0.53 (SD 0.21); mean difference (MD) 0.06, +12.8%, P < 0.001) and pentoxifylline (0.53 (SD 0.21) to 0.60 (SD 0.20); MD 0.07, +13.2%, P < 0.001). ABI measurements at final follow up showed a very slight significant improvement in favour of pentoxifylline (MD ‐0.07, 95% CI ‐0.13 to ‐0.01). Luk'Janov 1995 reported that at 12 months the ankle arm pressure index increased from 0.66 (SD 0.17) to 0.74 (SD 0.5) and from 0.66 (SD 0.26) to 0.69 (SD 0.26) in the PGE1 and pentoxifylline groups respectively.
Venous occlusion plethysmography
One study (Milio 2006) measured plethysmographic parameters. At the end of treatment maximal post‐ischaemic flow (PF) was significantly higher in the PGE1 group (16.21) than in the pentoxifylline group (13.47) while minimal vascular resistance (MVR) was significantly lower in the PGE1 patients compared to pentoxifylline‐treated patients (16.93 versus 18.24 respectively). There were no significant differences in muscular flow at rest (MR) nor basal vascular resistance (BVR) between the two groups. Scheffler 1994 reported that plasma viscosity and erthrocyte aggregation decreased in the PGE1 group but not to a statistically significant level (P = NS).
Adverse events
Three studies (Hepp 1996; Milio 2006; Scheffler 1994) measured adverse events. In the study by Hepp 1996, the rate of adverse events four weeks after treatment was similar between PGE1 (5%) and pentoxifylline (7%) patients (OR 0.71, 95% CI 0.22 to 2.31), and remained so at 12‐months follow up (OR 0.32, 95% CI 0.06 to 1.64). Of the adverse events which occurred during the year after treatment, there were two in the PGE1 group (one embolism and one stroke) and six in the pentoxifylline group (two amputations, two bypass operations, one fracture, and one arterial occlusion). Scheffler 1994 reported no serious side effects but reported that side effects including flushing (n = 6), reddening of the vein (n = 4) and diarrhoea (n = 1) occurred. It was not possible to determine which treatment group these side effects occurred in and attempts to clarify this query with the author were unsuccessful. Milio 2006 reported no side effects, while Luk'Janov 1995 provided no information about side effects during treatment with prostanoids.
PGE1 versus laevadosin (Comparison 3)
Walking distances
Creutzig 1988 compared the effects of PGE1 ia (20 participants) with laevadosin ia (20 participants). After three weeks of treatment, the mean PFWD increased significantly in both treatment groups: from 60 m to 198 m (230%) in PGE1 patients and from 69 m to 170 m (146%) in the laevadosin group. Improvements in mean MWD were 239% (90 m to 305 m) and 156% (107 m to 274m) in the PGE1 and laevadosin groups respectively. Significance levels of these results could not be assessed because standard deviations (SD) were not stated in this publication. Follow up was completed at 4, 12 and 36 weeks after treatment discontinuation. With the exception of MWD in PGE1 patients, the PFWD in both treatment groups and MWD in laevadosin patients remained elevated at all time points.
Adverse events
Adverse events occurred in 2/20 PGE1 patients (one nausea and one sweating) and 3/20 laevadosin patients (one nausea, one vomiting, and one thoracic oppression).
PGE1 versus naftidrofuryl (Comparison 4)
Walking distances
One study (Diehm 1989) analysed the effects of PGE1 iv (24 participants) versus naftidrofuryl iv (24 participants). Standard deviations of walking distances were not stated and therefore significance levels could not be assessed: the PFWD increased in the PGE1 group from 136 m to 270 m (98.5%) and in the naftidrofuryl group from 117 m to 230 m (96.6%) after three weeks of treatment. Three months after treatment was initiated, the mean PFWD in the PGE1 group increased to 306 m while at the same follow‐up point the mean PFWD fell to 210 m in the naftidrofuryl group.
Ankle brachial index (ABI)
The study by Diehm 1989 showed no significant difference between the groups in the ABI of the right leg (P = 0.1) and the ABI of the left leg (P = 0.9).
Adverse events
The rate of adverse events (vein reddening, pain, headache, change in blood pressure, diarrhoea, dizziness, nausea, discomfort and paraesthesia) was 20.8% in the PGE1 group compared to 91.6% in the naftidrofuryl group (OR 0.02, 95% CI 0 to 0.14) but treatment was never discontinued.
PGE1 versus L‐arginine (Comparison 5)
Walking distances
Böger 1998 investigated the effects of PGE1 versus L‐arginine in 26 participants (13 PGE1 and 13 L‐arginine). After three weeks of treatment, the mean PFWD improved in the L‐arginine group from 52 m to 147 m (185%) and in the PGE1 group from 52 m to 128 m (147%). Corresponding results for the MWD were an improvement from 93 m to 216 m (132%) in the L‐arginine group and from 93 m to 199 m (114%) in the PGE1 group. We could not include the data in our statistical program because SDs of the changes in walking distances were not reported but the study authors stated that there was no significant differences between treatments.
Quality of life
This study also measured claudication‐associated pain using a 10‐point scale to estimate intensity. Results showed that both treatments were associated with a significant improvement in pain, with the improvement most pronounced in the L‐arginine group (P < 0.05).
Ankle brachial index (ABI)
Böger 1998 detected no difference in the ABI between patients treated with PGE1 and L‐arginine.
PGI2 versus placebo (Comparison 6)
Walking distances
Six studies (Creager 2008; Lièvre 1996; Lièvre 2000; Mohler 2003; Virgolini 1989; Virgolini 1990) measured the efficacy of prostacyclins against a placebo: one (Creager 2008) compared iloprost, three studies (Lièvre 1996; Lièvre 2000; Mohler 2003) compared beraprost, and the remaining two studied taprostene (Virgolini 1989; Virgolini 1990). Creager 2008 conducted a study in which 260 patients received three different dosages of iloprost: 50 µg twice‐daily (n = 87), 100 µg twice‐daily (n = 86), and 150 µg twice‐daily (n = 87). In contrast, 84 patients were given a placebo. Data on PFWD and MWD were presented as mean percentage changes from baseline to follow up. At six months, PFWD improved by 3.3% in the placebo group compared to 7.7%, 8.8% and 11.2% in the iloprost 50 µg, 100 µg and 150 µg groups respectively, while MWD improved by 3.2%, 7.1%, 13.7% and 25.7% in the placebo and three iloprost groups respectively. As the mean walking distances and SDs after treatment were not reported, it was not possible to perform statistical analyses on the data. Contact was made with the author to obtain the raw data but attempts were unsuccessful.
Three studies (Lièvre 1996; Lièvre 2000; Mohler 2003) analysed the effects of beraprost (BPS) versus placebo. All three studies administered a 120 µg dose of BPS; Lièvre 1996 administered treatment for 12 weeks, Lièvre 2000 administered treatment for six months, while the study by Mohler 2003 administered treatment for one year. Walking distances at 12 weeks were used in order to make the results comparable but data from other time points were also reported. The studies by Lièvre 2000 and Mohler 2003 could not be included in a meta‐analysis as the study authors only provided log‐transformed values without SDs. Lièvre 1996 reported baseline PWFD and MWD and presented results as percentage change from baseline to follow up. As the SDs for the percentage change were presented, it was possible to conduct statistical analysis on this study. Results showed that while BPS was associated with an improvement in PFWD, it was not statistically significant (MD 56%, 95% CI ‐13% to 126%). The same was true for MWD, which improved by 99% and 61% in the two groups respectively (MD 38%, 95% CI ‐23% to 99%). These results were supported by Mohler 2003, who reported that BPS was not more effective than placebo in the improvement of walking distances at 12, 18 and 24 weeks (all P > 0.05). The study by Lièvre 1996 also reported change in walking distance in metres but as they did not present SDs for the final distances it was not possible to include this study in a meta‐analysis based on final walking distance.
The study by Lièvre 1996 also administered doses of 60 µg and 180 µg of BPS daily. BPS administered at a dose of 60 µg improved PFWD by 129% (SD 210%) compared with placebo (58% (SD 107%)) but the improvement was not statistically significant (MD 76%, 95% CI ‐0.4% to 142%). The MWD increased by 142% (SD 318%) and 61% (SD 92%) in the BPS 60 µg and placebo groups respectively, but again this improvement was not statistically significant (MD 81%, 95% CI ‐19% to 181%). At a higher dose of 180 µg, PFWD was improved by 51% (SD 114%) compared to 58% (SD 107%) with placebo (MD ‐7%, 95% CI ‐56% to 41%) while MWD was improved by 69% (SD 133%) and 61% (SD 92%) in the BPS 180 µg and placebo groups respectively (MD 8%, 95% CI ‐42% to 58%).
In the two studies by Virgolini 1989 and Virgolini 1990, PFWD and MWD were reported in seconds. For the purpose of this review, the walking times were converted into metres. The study by Virgolini 1989 evaluated taprostene versus placebo in 30 participants (15 taprostene and 15 placebo). Analysis of the data suggested that taprostene was not as effective as a placebo in improving PFWD (MD ‐9 m, 95% CI ‐12 to ‐6 m) but it is important to note that the baseline walking distance was higher in the placebo group. Taprostene was associated with an increase in MWD of 11 m (95% CI 0 to 22 m) but again the baseline MWD was higher in the placebo group, thus suggesting that the effect of taprostene was greater than as reported in the analysis. In the study by Virgolini 1990, 54 patients were randomised to PGI2 and 54 were given a placebo. The baseline PFWD and MWD were comparable between the two treatment groups. Although walking distances appeared to improved with taprostene, this was not to a statistically significant degree (PFWD MD 6 m, 95% CI ‐9 to 22 m; MWD MD 18 m, 95% CI ‐11 m to 48 m).
| Study | Dose (µg) | Duration (weeks) | PGI2 | PLC | PGI20 | SD | PGI2E | SD | %age | SD% | PLC0 | SD | PLCE | SD | %age | SD% | DIFF (%) |
| 100 | 26 | 87 | 84 | 105 | 81 | 7.7 | 120 | 88 | 3.3 | 4.4 | |||||||
| 200 | 86 | 124 | 96 | 8.8 | 5.5 | ||||||||||||
| 300 | 87 | 129 | 88 | 11.2 | 7.9 | ||||||||||||
| 60 | 12 | 42 | 41 | 132 | 72 | 254 | 129 | 210 | 131 | 73 | 190 | 58 | 107 | 71 | |||
| 120 | 42 | 142 | 69 | 270 | 114 | 204 | 56 | ||||||||||
| 180 | 39 | 137 | 74 | 186 | 51 | 114 | ‐7 | ||||||||||
| 120 | 26 | 209 | 213 | 130 | 65 | 280 | 115 | 133 | 71 | 245 | 84 | 31 | |||||
| 120 | 52 | 385 | 377 | 85 | 101 | 19 | 90 | 104 | 15 | 4 | |||||||
| 25 ng/kg/min | 1 | 15 | 15 | 54.07 | 4.35 | 22 | 63.66 | 4.41 | 4 | 18 | |||||||
| 6 ng/kg/min | 1 | 54 | 54 | 67.12 | 25.41 | 16 | 60.75 | 54.94 | 12 | 4 |
PGI2 = prostaglandin treatment group sample size
PLC = placebo group sample size
PGI20 = PGI2 baseline walking distance
SD = standard deviation
PGI2E = PGI2 end walking distance
%AGE = percentage improvement of walking distance
SD% = standard deviation of percentage improvement of walking distance
PLC0 = placebo baseline walking distance
PLCE = placebo end walking distance
DIFF = difference in percentage of improvement of PGI2 and placebo
| Study | Dose (µg) | Duration (weeks) | PGI2 | PLC | PGI20 | SD | PGI2E | SD | %age | SD% | PLC0 | SD | PLCE | SD | %age | SD% | DIFF (%) |
| 100 | 26 | 87 | 84 | 105 | 81 | 7.1 | 120 | 88 | 3.2 | 3.9 | |||||||
| 200 | 86 | 124 | 96 | 13.7 | 10.5 | ||||||||||||
| 300 | 87 | 129 | 88 | 25.7 | 22.5 | ||||||||||||
| 60 | 12 | 42 | 41 | 203 | 123 | 384 | 142 | 318 | 206 | 145 | 291 | 61 | 92 | 81 | |||
| 120 | 42 | 224 | 94 | 428 | 99 | 182 | 38 | ||||||||||
| 180 | 39 | 207 | 112 | 357 | 69 | 133 | 3 | ||||||||||
| 120 | 26 | 209 | 213 | 275 | 229 | 467 | 70 | 271 | 240 | 378 | 39 | 31 | |||||
| 120 | 52 | 385 | 377 | 164 | 191.4 | 16.7 | 170 | 195.7 | 15 | 1.7 | |||||||
| 25 ng/kg/min | 1 | 15 | 15 | 224.87 | 21.28 | 14 | 213.4 | 6.28 | 2 | 12 | |||||||
| 6 ng/kg/min | 1 | 54 | 54 | 177.56 | 62.89 | 16 | 158.86 | 93.27 | 7 | 9 |
PGI2 = prostaglandin treatment group sample size
PLC = placebo group sample size
PGI20 = PGI2 baseline walking distance
SD = standard deviation
PGI2E = PGI2 end walking distance
%AGE = percentage improvement of walking distance
SD% = standard deviation of percentage improvement of walking distance
PLC0 = placebo baseline walking distance
PLCE = placebo end walking distance
DIFF = difference in percentage of improvement of PGI2 and placebo
Quality of life
The study by Lièvre 2000 demonstrated a marginally significant improvement in quality of life in favour of BPS treatment (P = 0.049), with specific items such as 'going out', 'general condition', 'relationships with people' and 'concerns about health' showing increased satisfaction. Mohler et al (Mohler 2003) observed no significant differences in quality of life parameters between oral prostacyclin and placebo.
Ankle brachial index (ABI)
Lièvre 2000 specified ABI as an outcome. Data were not presented but the study authors reported that there was no difference in ABI during the exercise tests between the two treatment groups.
Adverse events
Four of the six studies (Creager 2008; Lièvre 1996; Lièvre 2000; Mohler 2003) measured adverse events associated with PGI2. In the study by Creager 2008 one patient in the placebo group died but there were no major amputations. The rate of revascularisation was 5.7%, 5.8% and 1.1% in the iloprost 50 µg, 100 µg and 150 µg groups respectively, compared to 6% in the placebo group. Headache was reported in 44%, 64% and 67% of the iloprost 50 µg, 100 µg and 150 µg groups compared to 16% of placebo‐treated patients. There was no significant difference in the rate of flushing between the two treatments (5% iloprost versus 4% placebo). Pain in the extremity, jaw pain, nausea and diarrhoea occurred more frequently in the iloprost group compared to placebo. Cardiovasular events occurred at a similar rate between the two groups. The rate of adverse events leading to discontinuation of the study medications was two to three‐fold higher in the iloprost groups than with placebo (31%, 57% and 53% for 50 µg, 100 µg and 150 µg iloprost groups respectively, compared to 14% in the placebo group). Mohler 2003 reported five cases of myocardial infarction and four cardiovascular deaths in the placebo group and one cardiovascular death in the BPS group. No limb amputation occurred but four BPS and eight placebo patients underwent limb revascularisation. In the study by Lièvre 1996, the rate of adverse events did not differ significantly by treatment group: 10 in the placebo group, 17 in the BPS 60 µg group, 13 in the BPS 120 µg group, and 19 in the BPS 180 µg group (P = 0.113). The most common adverse events were gastrointestinal disorders, headache, skin complaints and flushes. Lièvre 2000 reported that fewer critical cardiovascular events (that is death and myocardial infarction) occurred in the BPS group (OR 0.28, 95% CI 0.1 to 0.78). Arterial thrombosis of the leg occurred in 8 BPS and 14 placebo patients (OR 0.57, 95% CI 0.23 to 1.38). Drug‐related adverse events occurred at a significantly higher rate in the patients treated with BPS (OR 2.66, 95% CI 1.40 to 5.03). Side effects led to discontinuation of the study drug in 18 BPS and 31 placebo patients respectively.
PGI2 versus pentoxifylline (Comparison 7)
Walking distances
Creager 2008 also compared 270 iloprost patients with 86 patients who received pentoxifylline. At six months, PFWD increased by 24%, 28.9% and 31.2% for the groups treated with twice‐daily 50 µg, 100 µg and 150 µg iloprost respectively, compared to 34.3% in patients treated with pentoxifylline. Furthermore, the MWD increased by 7.7%, 8.8% and 11.2% for the three iloprost dosage groups respectively, compared to 13.9% with pentoxifylline. However, the study authors did not report SDs so it was impossible to determine if the improvement in PFWD and MWD observed with pentoxifylline was statistically significant.
Adverse events
One death occurred in the pentoxifylline group (OR 0.11, 95% CI 0 to 2.71). Revascularisation rates were similar between the two treatment groups, both when all iloprost doses were combined (OR 1.86, 95% CI 0.48 to 5.09) and when the three doses were compared individually (50 µg twice‐daily OR 2.56, 95% CI 0.48 to 13.57; 100 µg twice‐daily OR 2.59, 95% CI 0.49 to 13.74; 150 µg twice‐daily OR 0.49, 95% CI 0.04 to 5.49). No major amputations occurred in this study. Headache was reported in 44% of patients receiving iloprost 50 µg twice‐daily, in 64% of those receiving iloprost 100 µg twice‐daily and in 67% of those receiving 150 µg twice‐daily, compared to 19% of patients treated with pentoxifylline. Additionally, while the frequency of flushing was comparable between the pentoxifylline (2%) and iloprost 50 µg twice‐daily (5%) groups, it increased in a dose‐respondent manner to 23% in patients receiving 100 µg twice‐daily and 31% in patients receiving 150 µg twice‐daily iloprost. Furthermore, diarrhoea occurred more often in iloprost‐treated patients (13%) than pentoxifylline‐treated patients (6%). On the other hand, mild dyspepsia occurred at twice the frequency in patients receiving pentoxifylline (13%) compared to those receiving 50 to 100 µg iloprost twice‐daily (6%). The occurrence of adverse events led to discontinuation of treatment in 31%, 57% and 53% of the twice‐daily 50 µg, 100 µg and 150 µg iloprost groups respectively, compared to 15% of the pentoxifylline group.
PGI2 versus hydroxy‐ethyl starch (HES) (Comparison 8)
Walking distances
One study (Müller‐Bühl 1987) compared iv iloprost (11 participants) versus iv hydroxyl‐ethyl starch (HES) (12 participants). At the end of a two‐week treatment period, PFWD improved by 36.5 m (59%) in PGI2 patients compared to 17.9 m (31%) in HES patients while the corresponding improvements in MWD were 50.2 m (49%) and 44.6 m (45%) in the two treatment groups respectively. The improvements in final PFWD (22 m, 95% CI ‐8.9 m to 53.3 m) and final MWD (7.4 m, 95% CI ‐49.8 m to 64.6 m) were not statistically significant.
Ankle brachial Index (ABI)
The Doppler brachiocrural index did not change significantly in either treatment group. At baseline the ABI was 0.63 (SD 0.38) in the iloprost group and rose to 0.77 (SD 0.55) after 10 days of treatment; while in the HES group, the baseline ABI was 0.55 (SD 0.25) and rose to 0.71 (SD 0.38) after 10 days. The ABI remained higher than baseline at both two and six weeks after treatment in both groups.
Venous occlusion plethysmography
A slight increase in calf blood flow was shown by venous occlusion plethysmography at rest and after tourniquet ischaemia, but this was not significant.
Haemorrheological parameters
Study authors (Müller‐Bühl 1987) reported that the haematocrit level declined to the lowest mean value of 40.6% at the end of HES treatment but returned to normal levels during the six‐weeks follow up. The change in haematocrit level in the iloprost group was less pronounced, with a lowest mean value of 42.3%. In both groups, neither plasma viscosity nor erythrocyte aggregation changed significantly. Furthermore, plasma thrombin time and the thromboelastogram were in the normal range before and after both iloprost and HES treatment.
Adverse events
Flushing occurred in 82% of iloprost patients at the beginning of infusion while mild headache and nausea also occurred. However, none of these events led to the discontinuation of iloprost treatment.
Discussion
We reviewed 18 randomised controlled trials studying the clinical effectiveness of prostanoids in patients with intermittent claudication (IC).
Summary of main results
Results of some individual studies included in this review suggested that PGE1 improved walking capacity in patients with IC. Five studies reported that walking capacity even continued to increase after termination of treatment (Belch 1997; Creutzig 1988; Hepp 1996; Mangiafico 2000; Scheffler 1994). Two studies (Diehm 1989; Scheffler 1994) suggested that PGE1 is beneficial as an addition to physical training and leads to a greater improvement in walking distances than with exercise alone. However, the quality of the studies is variable and the individual results could not be strengthened by pooled analysis. Therefore, the overall evidence is insufficient to determine whether or not patients with IC derive clinically meaningful benefit from administration of PGE1. Data concerning the effects of beraprost sodium are conflicting with one study reporting an increase in walking distances while two studies demonstrated it had no significant effect. Results from individual trials also suggested that walking distances improved in a dose‐dependent manner with iloprost, a prostacyclin analogue (PGI2). However iloprost was also associated with an increased incidence of drug‐related adverse events, particularly at higher doses. Pentoxifylline may be a useful alternative as it was associated with an improvement in walking distances, but it did not produce fewer drug‐related side effects than PGI2.
Overall completeness and applicability of evidence
The search identified studies of the majority of well‐known treatments for IC, including prostaglandin, pentoxifylline and PGI2. However, studies on cilostazol were not identified in the completed searches.
Three studies provided no information about participants' characteristics (especially risk factors for the development of atherosclerosis, for example smoking habits, hypertension, dyslipidaemia or diabetes mellitus) and therefore a potential heterogeneity between the study groups that may influence the final results could not be rated (Diehm 1989; Luk'Janov 1995; Müller‐Bühl 1987). The additional effect of vascular training could not be included in our analysis because only eight studies clearly mentioned this important part of therapy, and described a daily, controlled or supervised training for all randomised groups (Böger 1998; Creutzig 1988; Diehm 1989; Hepp 1996; Lièvre 1996; Mangiafico 2000; Mohler 2003; Scheffler 1994). The remaining studies provided no information concerning the start or continuation of a training program. It is possible that the lack of this important information could have influenced our final results. Three studies did not mention whether other vasoactive or antiplatelet drugs were continued during the study (Blume 1986; Milio 2006; Müller‐Bühl 1987). The possible administration of these drugs could have lead to false positive results and should therefore should be considered in the interpretation of this publication.
The major problem of our review was the limited comparability of the included studies, as our analyses revealed a significant heterogeneity between the studies. Additionally, different study designs have to be considered in the interpretation of the included studies used; the duration of administration and therefore the total dosages of the study drugs differed significantly between the studies. In addition, different walking treadmills were used to determine walking distances, with gradients ranging from 5 ° to 12 ° and walking speeds ranging from 2 km/h to 3.5 km/h. Due to the difference in treadmill testing, results were not always comparable. For example, it would be expected that a patient walking at a slower speed or on a lower gradient would be able to walk further than a patient walking at a faster speed or a higher gradient. Therefore, final walking distance is not always a true indication of the effectiveness of treatment but is also influenced by the speed and gradient of the treadmill. Ideally all studies would use a standardised protocol to measure walking distances. However this was not the case. To be as complete as possible, final walking distances were converted regardless of the treadmill test but the review authors did so knowing that this would introduce heterogeneity between individual studies. Seventeen of the 18 included studies used a constant load test and one used a graded test (Creager 2008). Authors of this study did not present final walking distances but rather a percentage change from baseline. Without the final walking distance and standard deviations, it was not possible to convert the results of the graded test to make them comparable with a constant load test. However, as this study was not included in a meta‐analysis, it had no effect on the overall result. Most of the studies enrolled only a small number of participants and therefore their statistical power appears low, generating a considerable beta‐error. Additional statistical problems occurred in the analysis of 15 studies (Belch 1997; Blume 1986; Böger 1998; Creager 2008; Creutzig 1988; Diehm 1989; Diehm 1997; Hepp 1996; Lièvre 2000; Luk'Janov 1995; Mangiafico 2000; Milio 2006; Mohler 2003; Müller‐Bühl 1987; Scheffler 1994) because the standard deviations of the results were missing and therefore significance levels could not be assessed.
Quality of the evidence
A restriction of this review was that the majority of studies did not provide enough information to permit a judgement on the risk of selection, performance, attrition, detection and reporting biases. Attempts were made to contact authors for more information but only one author (Milio 2006) responded with the required information. The major problem was the lack of information about methods of randomisation and blinding, and the number of withdrawals and dropouts. In addition, several methodological errors of the published data were evident. In five studies baseline values of walking distances were determined by a single measurement (Luk'Janov 1995; Müller‐Bühl 1987; Virgolini 1989; Virgolini 1990). This could have generated false results because the remaining studies observed variations of walking distances in several participants of more than 20% between two measurements (performed in the pre‐treatment period) and excluded these patients from participation in the study (Belch 1997; Blume 1986; Böger 1998; Creager 2008; Creutzig 1988; Diehm 1989; Diehm 1997; Hepp 1996; Lièvre 1996; Lièvre 2000; Mangiafico 2000; Mohler 2003; Scheffler 1994).
Potential biases in the review process
In this systematic review we identified all the randomised controlled trials (RCTs) that compared prostanoids to a placebo or control treatment. Open or cohort studies were not included because prostanoids are a well studied drug over a long period of time, and there is a significant number of RCTs available. We believe our search for RCTs has been inclusive and it is unlikely that our standardised methods of study selection and data extraction could have introduced bias. Due to the heterogeneity of the included studies, lack of standard deviations reported in the publications and the variable presentation of the outcomes by the trialists, we did not pool all the available data. This is likely to have resulted in the conclusion of this review being less robust.
Two studies (Virgolini 1989; Virgolini 1990) presented walking distances in seconds. In order to make these results comparable with other studies, we converted the walking times in seconds to walking distances in metres using the gradient and speed reported by the authors. While this may have introduced some bias into the review, we considered that not including the data would have introduced even more bias.
Due to the lack of data presented in studies, it was not possible to account for background exercise and treadmill testing. Variations in exercise levels prior to participating in the study may have introduced bias into the results. However as authors of the studies did not clarify this, there was no way to assess the level of bias.
In most cases we presented final walking distances instead of the percentage change in walking distance because of a lack of standard deviations reported in the original publications. There are obvious limitations in presenting final walking distances, the main one being that the final value does not take into account the change from baseline. Furthermore, variations in treadmill testing meant that walking distances were not always comparable between individual studies, and this is likely to result in significant heterogeneity when pooling the studies. Nevertheless it was decided to use all available data in order to make the analysis as complete as possible.
Agreements and disagreements with other studies or reviews
To date, no systematic review has examined the effect of prostanoids on intermittent claudication (IC). Three systematic reviews (Ernst 1994; Frampton 1995; Salhiyyah 2012) have measured the effectiveness of pentoxifylline on improving walking distance in IC. A greater improvement in PFWD and TWD has been shown for pentoxifylline‐treated patients compared with placebo. However, the studies were poor quality and the large degree of heterogeneity between them resulted in the reviews concluding that the overall clinical effect of pentoxifylline was unclear. One review (Ruffolo 2010) measured the effectiveness of prostanoids in treating critical limb ischaemia (CLI) and, while there were some positive results regarding rest pain relief, ulcer healing and amputations, the reviewers determined that there was no conclusive evidence on the long‐term effectiveness and safety of different prostanoids in patients with CLI.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Comparison 1 PGE1 versus placebo, Outcome 1 Mean change in pain‐free walking distance (%).
Comparison 1 PGE1 versus placebo, Outcome 2 Final pain‐free walking distance (meters).
Comparison 1 PGE1 versus placebo, Outcome 3 Mean change in maximal walking distance (%).
Comparison 1 PGE1 versus placebo, Outcome 4 Final maximal walking distance (meters).
Comparison 2 PGE1 versus pentoxifylline (Ptx), Outcome 1 Mean change in pain‐free walking distance (%).
Comparison 2 PGE1 versus pentoxifylline (Ptx), Outcome 2 Final pain‐free walking distance (meters).
Comparison 2 PGE1 versus pentoxifylline (Ptx), Outcome 3 Mean change in maximal walking distance (%).
Comparison 2 PGE1 versus pentoxifylline (Ptx), Outcome 4 Final maximal walking distance (meters).
Comparison 2 PGE1 versus pentoxifylline (Ptx), Outcome 5 Ankle brachial index.
Comparison 2 PGE1 versus pentoxifylline (Ptx), Outcome 6 Adverse events.
Comparison 3 PGE1 versus laevadosin (energy rich phosphates), Outcome 1 Mean change in pain‐free walking distance (%).
Comparison 3 PGE1 versus laevadosin (energy rich phosphates), Outcome 2 Final pain‐free walking distance (meters).
Comparison 3 PGE1 versus laevadosin (energy rich phosphates), Outcome 3 Mean change in maximal walking distance (%).
Comparison 3 PGE1 versus laevadosin (energy rich phosphates), Outcome 4 Final maximal walking distance (meters).
Comparison 4 PGE1 versus naftidrofuryl, Outcome 1 Mean change in pain‐free walking distance (%).
Comparison 4 PGE1 versus naftidrofuryl, Outcome 2 Final pain‐free walking distance (meters).
Comparison 4 PGE1 versus naftidrofuryl, Outcome 3 Ankle brachial index left side.
Comparison 4 PGE1 versus naftidrofuryl, Outcome 4 Ankle brachial index right side.
Comparison 4 PGE1 versus naftidrofuryl, Outcome 5 Adverse events.
Comparison 5 PGE1 versus L‐arginine, Outcome 1 Mean change in pain‐free walking distance (%).
Comparison 5 PGE1 versus L‐arginine, Outcome 2 Final pain‐free walking distance (meters).
Comparison 5 PGE1 versus L‐arginine, Outcome 3 Mean change in maximal walking distance (%).
Comparison 5 PGE1 versus L‐arginine, Outcome 4 Final maximal walking distance (meters).
Comparison 5 PGE1 versus L‐arginine, Outcome 5 Ankle brachial index.
Comparison 6 PGI2 versus placebo, Outcome 1 Mean change in pain‐free walking distance (%).
Comparison 6 PGI2 versus placebo, Outcome 2 Final pain‐free walking distance (meters).
Comparison 6 PGI2 versus placebo, Outcome 3 Mean change in maximum walking distance (%).
Comparison 6 PGI2 versus placebo, Outcome 4 Final maximal walking distance (meters).
Comparison 6 PGI2 versus placebo, Outcome 5 Adverse events.
Comparison 7 PGI2 versus pentoxifylline (Ptx), Outcome 1 Death.
Comparison 7 PGI2 versus pentoxifylline (Ptx), Outcome 2 Revascularisation.
Comparison 8 Iloprost versus hydroxy‐ethyl starch (HES), Outcome 1 Mean change in pain‐free walking distance (%).
Comparison 8 Iloprost versus hydroxy‐ethyl starch (HES), Outcome 2 Final pain‐free walking distance (meters).
Comparison 8 Iloprost versus hydroxy‐ethyl starch (HES), Outcome 3 Mean change in maximal walking distance (%).
Comparison 8 Iloprost versus hydroxy‐ethyl starch (HES), Outcome 4 Final maximal walking distance (meters).
Comparison 8 Iloprost versus hydroxy‐ethyl starch (HES), Outcome 5 Ankle brachial index.
Comparison 8 Iloprost versus hydroxy‐ethyl starch (HES), Outcome 6 Venous‐occlusion plethysmography ‐ at rest.
Comparison 8 Iloprost versus hydroxy‐ethyl starch (HES), Outcome 7 Venous‐occlusion plethysmography ‐ reactive hyperaemia.
| Study | Dose (µg) | Duration (weeks) | PGE | PLC | PGE0 | SD | PGEE | SD | %age | SD% | PLC0 | SD | PLCE | SD | %age | SD% | DIFF (%) |
| 2, 5 d/wk | 4 | 19 | 21 | 44.8 | 16.0 | not stated | |||||||||||
| 5, 2 d/wk | 18 | 50.7 | 33.5 | not stated | |||||||||||||
| 5, 5 d/wk | 22 | 45.0 | 33.0 | 64.5 | 43 | 17 | 26 | ||||||||||
| 2, 5 d/wk | 8 a | 19 | 21 | 44.8 | 16.0 | not stated | |||||||||||
| 5, 2 d/wk | 18 | 50.7 | 33.5 | not stated | |||||||||||||
| 5, 5 d/wk | 22 | 45.0 | 33.0 | 65.9 | 46 | 0 | 46 | ||||||||||
| 10‐20 | 3 | 25 | 25 | 54 | 13 | 113 | 54 | 109 | 54 | 16 | 73 | 41 | 35 | 74 | |||
| 60 | 4 | 106 | 102 | 64.3 | 1.6 | 112.7 | 1.8 | 75 | 66.6 | 1.6 | 95.5 | 2.0 | 43 | 32 | |||
| 8 b | 128.9 | 2.0 | 100 | 106.6 | 2.1 | 60 | 40 | ||||||||||
| 60 | 4 | 21 | 21 | 72 | 16 | 135 | 33 | 88 | 81 | 17 | 84 | 17 | 4 | 84 | |||
| 8 a | 113 | 26 | 57 | not stated | not stated | ||||||||||||
| PGE = prostaglandin treatment group sample size * Study reported PFWD as a median with IQR. a = treatment was administered for 4 weeks, treadmill tests conducted at 4 weeks and also at 8 weeks (after a 4‐week period of no treatment). b = treatment was administered for 5 days a week for 4 weeks and reduced to 2 days a week for a further 4 weeks, treadmill tests conducted at both 4 and 8 weeks. | |||||||||||||||||
| Study | Dose (µg) | Duration (weeks) | PGE | PLC | PGE0 | SD | PGEE | SD | %age | SD% | PLC0 | SD | PLCE | SD | %age | SD% | DIFF (%) |
| 2, 5 d/wk | 4 | 19 | 21 | 60.0 | 25.6 | not stated | |||||||||||
| 5, 2 d/wk | 18 | 68.6 | 48.8 | not stated | |||||||||||||
| 5, 5 d/wk | 22 | 67.1 | 68.8 | 95.1 | 42 | 75.6 | 93.0 | 80.1 | 6 | 36 | |||||||
| 2, 5 d/wk | 8 a | 19 | 21 | 60.0 | 25.6 | not stated | |||||||||||
| 5, 2 d/wk | 18 | 68.6 | 48.8 | not stated | |||||||||||||
| 5, 5 d/wk | 22 | 67.1 | 68.8 | 102.1 | 52 | 75.6 | 93.0 | 64.4 | ‐15 | 67 | |||||||
| 10‐20 | 3 | 25 | 25 | 98 | 45 | 170 | 75 | 74 | 97 | 28 | 114 | 51 | 18 | 56 | |||
| 60 | 4 | 106 | 102 | 98.8 | 1.5 | 163 | 1.8 | 65 | 99.8 | 1.4 | 141.9 | 1.7 | 42 | 23 | |||
| 8 b | 186.3 | 2.0 | 88 | 160.5 | 1.9 | 61 | 27 | ||||||||||
| 60 | 4 | 21 | 21 | 140 | 30 | 266 | 62 | 90 | 152 | 38 | 154 | 39 | 1 | 89 | |||
| 8a | 229 | 55 | 64 | not stated | not stated | ||||||||||||
| PGE = prostaglandin treatment group sample size * Study reported MWD as a median with IQR. a = treatment was administered for 4 weeks, treadmill tests conducted at 4 weeks and also at 8 week (after a 4‐week period of no treatment). b = treatment was administered for 5 days a week for 4 weeks and reduced to 2 days a week for a further 4 weeks, treadmill tests conducted at both 4 and 8 weeks. | |||||||||||||||||
| Study | Dose PGE (µg) | Dose PTX (mg) | Duration (weeks) | PGE1 | PTX | PGE0 | SD | PGEE | SD | %age | SD% | PTX0 | SD | PTXE | SD | %age | SD% | DIFF (%) |
| 80 | 400 | 4 | 97 | 98 | 83 | 264 | 218 | 84 | 188 | 124 | 94 | |||||||
| 40 | 400 | 4 | 42 | 40 | 89 | 195 | 119 | 78 | 149 | 91 | 45 | |||||||
| 60 | 200 | 4 | 63 | 60 | 78 | 36 | 387 | 274 | 396 | 82 | 41 | 175 | 180 | 113 | 260 | |||
| 80 | 200 | 4 | 14 | 15 | 81 | 23 | 570 | 727 | 604 | 75 | 41 | 154 | 150 | 105 | 499 | |||
| PGE = prostaglandin treatment group sample size | ||||||||||||||||||
| Study | Dose PGE (µg) | Dose PTX (mg) | Duration (weeks) | PGE1 | PTX | PGE0 | SD | PGEE | SD | %age | SD% | PTX0 | SD | PTXE | SD | %age | SD% | DIFF (%) |
| 80 | 400 | 4 | 97 | 98 | 130 | 343 | 164 | 131 | 322 | 146 | 18 | |||||||
| 40 | 400 | 4 | 42 | 40 | ‐ | ‐ | ‐ | ‐ | ‐ | ‐ | ‐ | ‐ | ‐ | ‐ | ‐ | ‐ | ‐ | |
| 60 | 200 | 4 | 63 | 60 | 143 | 55 | 515 | 285 | 260 | 148 | 69 | 323 | 264 | 118 | 158 | |||
| 80 | 200 | 4 | 14 | 15 | 158 | 95 | 744 | 697 | 371 | 160 | 133 | 351 | 432 | 119 | 252 | |||
| PGE = prostaglandin treatment group sample size | ||||||||||||||||||
| Study | Dose (µg) | Duration (weeks) | PGI2 | PLC | PGI20 | SD | PGI2E | SD | %age | SD% | PLC0 | SD | PLCE | SD | %age | SD% | DIFF (%) |
| 100 | 26 | 87 | 84 | 105 | 81 | 7.7 | 120 | 88 | 3.3 | 4.4 | |||||||
| 200 | 86 | 124 | 96 | 8.8 | 5.5 | ||||||||||||
| 300 | 87 | 129 | 88 | 11.2 | 7.9 | ||||||||||||
| 60 | 12 | 42 | 41 | 132 | 72 | 254 | 129 | 210 | 131 | 73 | 190 | 58 | 107 | 71 | |||
| 120 | 42 | 142 | 69 | 270 | 114 | 204 | 56 | ||||||||||
| 180 | 39 | 137 | 74 | 186 | 51 | 114 | ‐7 | ||||||||||
| 120 | 26 | 209 | 213 | 130 | 65 | 280 | 115 | 133 | 71 | 245 | 84 | 31 | |||||
| 120 | 52 | 385 | 377 | 85 | 101 | 19 | 90 | 104 | 15 | 4 | |||||||
| 25 ng/kg/min | 1 | 15 | 15 | 54.07 | 4.35 | 22 | 63.66 | 4.41 | 4 | 18 | |||||||
| 6 ng/kg/min | 1 | 54 | 54 | 67.12 | 25.41 | 16 | 60.75 | 54.94 | 12 | 4 | |||||||
| PGI2 = prostaglandin treatment group sample size | |||||||||||||||||
| Study | Dose (µg) | Duration (weeks) | PGI2 | PLC | PGI20 | SD | PGI2E | SD | %age | SD% | PLC0 | SD | PLCE | SD | %age | SD% | DIFF (%) |
| 100 | 26 | 87 | 84 | 105 | 81 | 7.1 | 120 | 88 | 3.2 | 3.9 | |||||||
| 200 | 86 | 124 | 96 | 13.7 | 10.5 | ||||||||||||
| 300 | 87 | 129 | 88 | 25.7 | 22.5 | ||||||||||||
| 60 | 12 | 42 | 41 | 203 | 123 | 384 | 142 | 318 | 206 | 145 | 291 | 61 | 92 | 81 | |||
| 120 | 42 | 224 | 94 | 428 | 99 | 182 | 38 | ||||||||||
| 180 | 39 | 207 | 112 | 357 | 69 | 133 | 3 | ||||||||||
| 120 | 26 | 209 | 213 | 275 | 229 | 467 | 70 | 271 | 240 | 378 | 39 | 31 | |||||
| 120 | 52 | 385 | 377 | 164 | 191.4 | 16.7 | 170 | 195.7 | 15 | 1.7 | |||||||
| 25 ng/kg/min | 1 | 15 | 15 | 224.87 | 21.28 | 14 | 213.4 | 6.28 | 2 | 12 | |||||||
| 6 ng/kg/min | 1 | 54 | 54 | 177.56 | 62.89 | 16 | 158.86 | 93.27 | 7 | 9 | |||||||
| PGI2 = prostaglandin treatment group sample size | |||||||||||||||||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Mean change in pain‐free walking distance (%) Show forest plot | 3 | 508 | Std. Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 1.1 Intra‐arterial, 3 weeks | 1 | 50 | Std. Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 1.2 Intravenous, 4 weeks | 2 | 250 | Std. Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 1.3 Intravenous, 8 weeks | 1 | 208 | Std. Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 2 Final pain‐free walking distance (meters) Show forest plot | 3 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 2.1 Intra‐arterial, 3 weeks | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.2 Intravenous, 4 weeks | 2 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.3 Intravenous, 8 weeks | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3 Mean change in maximal walking distance (%) Show forest plot | 3 | 508 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 3.1 Intra‐arterial, 3 weeks | 1 | 50 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 3.2 Intravenous, 4 weeks | 2 | 250 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 3.3 Intravenous, 8 weeks | 1 | 208 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 4 Final maximal walking distance (meters) Show forest plot | 3 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 4.1 Intra‐arterial, 3 weeks | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4.2 Intravenous, 4 weeks | 2 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4.3 Intravenous, 8 weeks | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Mean change in pain‐free walking distance (%) Show forest plot | 4 | 429 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 2 Final pain‐free walking distance (meters) Show forest plot | 4 | 429 | Mean Difference (IV, Fixed, 95% CI) | 220.62 [140.80, 300.44] |
| 3 Mean change in maximal walking distance (%) Show forest plot | 3 | 348 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 4 Final maximal walking distance (meters) Show forest plot | 3 | 347 | Mean Difference (IV, Fixed, 95% CI) | 201.93 [107.33, 296.54] |
| 5 Ankle brachial index Show forest plot | 2 | 240 | Mean Difference (IV, Fixed, 95% CI) | ‐0.06 [‐0.12, 0.00] |
| 6 Adverse events Show forest plot | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 6.1 4 weeks | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 6.2 12 months | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Mean change in pain‐free walking distance (%) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 2 Final pain‐free walking distance (meters) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 3 Mean change in maximal walking distance (%) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 4 Final maximal walking distance (meters) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Mean change in pain‐free walking distance (%) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 2 Final pain‐free walking distance (meters) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 3 Ankle brachial index left side Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 4 Ankle brachial index right side Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 5 Adverse events Show forest plot | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Mean change in pain‐free walking distance (%) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 2 Final pain‐free walking distance (meters) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 3 Mean change in maximal walking distance (%) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 4 Final maximal walking distance (meters) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 5 Ankle brachial index Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Mean change in pain‐free walking distance (%) Show forest plot | 6 | 1575 | Mean Difference (IV, Fixed, 95% CI) | 56.00 [‐13.85, 125.85] |
| 2 Final pain‐free walking distance (meters) Show forest plot | 5 | 1402 | Mean Difference (IV, Random, 95% CI) | ‐3.66 [‐18.77, 11.46] |
| 3 Mean change in maximum walking distance (%) Show forest plot | 6 | 1575 | Mean Difference (IV, Fixed, 95% CI) | 38.0 [‐23.83, 99.83] |
| 4 Final maximal walking distance (meters) Show forest plot | 5 | 1328 | Mean Difference (IV, Fixed, 95% CI) | 12.36 [1.84, 22.87] |
| 5 Adverse events Show forest plot | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 5.1 Critical cardiovascular events | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 5.2 Arterial thrombosis | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 5.3 Drug‐related adverse events | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Death Show forest plot | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 2 Revascularisation Show forest plot | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 2.1 50 µg iloprost | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.2 100 µg iloprost | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.3 150 µg iloprost | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Mean change in pain‐free walking distance (%) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 2 Final pain‐free walking distance (meters) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 3 Mean change in maximal walking distance (%) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 4 Final maximal walking distance (meters) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 5 Ankle brachial index Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 6 Venous‐occlusion plethysmography ‐ at rest Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 7 Venous‐occlusion plethysmography ‐ reactive hyperaemia Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
