Intervention Review

You have free access to this content

Treatment of hypertension in peripheral arterial disease

  1. Deirdre A Lane,
  2. Gregory YH Lip*

Editorial Group: Cochrane Vascular Group

Published Online: 4 DEC 2013

Assessed as up-to-date: 19 MAR 2013

DOI: 10.1002/14651858.CD003075.pub3


How to Cite

Lane DA, Lip GYH. Treatment of hypertension in peripheral arterial disease. Cochrane Database of Systematic Reviews 2013, Issue 12. Art. No.: CD003075. DOI: 10.1002/14651858.CD003075.pub3.

Author Information

  1. University of Birmingham Centre for Cardiovascular Sciences, City Hospital, Birmingham, UK

*Gregory YH Lip, University of Birmingham Centre for Cardiovascular Sciences, City Hospital, Dudley Road, Birmingham, B18 7QH, UK. g.y.h.lip@bham.ac.uk.

Publication History

  1. Publication Status: New search for studies and content updated (no change to conclusions)
  2. Published Online: 4 DEC 2013

SEARCH

 
Characteristics of included studies [ordered by study ID]
DAPHNE 2002

MethodsDAPHNE study: single-blind placebo medication for first six weeks, followed by double-blind randomisation
Method of randomisation not mentioned
Dropouts: 12 (29.3%) from the doxazosin group; 12 (30.8%) from the hydrochlorothiazide (HCTZ) group
Both drugs could be doubled in dose at two-weekly intervals until the goal diastolic blood pressure of 90 mmHg was achieved


ParticipantsLocation: outpatient clinics at two hospitals in the Netherlands
Inclusion criteria: peripheral arterial disease (PAD) patients, 80 men aged 45-70 years
Doxazosin group: 41 (8 dropouts, n = 29)
HCTZ group: 39 (12 dropouts, n = 27)


InterventionsDoxazosin 1 mg, 2 mg, 4 mg, 8 mg, 16 mg once daily (od) or HCTZ 12.5 mg, 25 mg, 50 mg, 100 mg. The dose of either doxazosin or HCTZ was increased at two-weekly intervals until the target blood pressure was achieved

Duration: three years


OutcomesArterial intima-media thickness (IMT) of 20 longitudinal arterial wall segments of the carotid and femoral arteries


NotesFor the outcome arterial (IMT) the data presented are the change from baseline to three-years follow-up, for the average of 20 mean far and near walls of carotid and femoral arteries combined for the two treatment groups. However, data are also presented for carotid and femoral arteries separately, and combined arteries for each of the following: average of maximum far and near walls, average of far walls, and average of near walls. There was a significant change in IMT for each of the groups of arteries examined, for both the HCTZ and doxazosin groups. Data for the change in arterial IMT from baseline to three-years follow-up are only available for 27 patients receiving HCTZ and 29 patients receiving doxazosin.

For the outcome peripheral vascular surgery, the type of surgery was not specified. Five patients in each treatment group (HCTZ n = 31; doxazosin n = 33) received peripheral vascular surgery for progressive PAD.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNo description given, only that patients were randomised

Allocation concealment (selection bias)Unclear riskNo description given, only that patients were randomised

Blinding (performance bias and detection bias)
Participant
Unclear riskTrial described as double-blind but no description given

Blinding (performance bias and detection bias)
Physician
Unclear riskTrial described as double-blind but no description given

Blinding (performance bias and detection bias)
Outcome assessor
Unclear riskNo information given regarding the blinding of the outcome assessor

Incomplete outcome data (attrition bias)
All outcomes
High riskSimilar number of dropouts in both groups but follow-up data only available for 69.2% and 70.7% in the HCTZ and doxazosin groups, respectively

Selective reporting (reporting bias)Unclear riskThe authors reported all the data for the measures they stated in the methods section of the paper

Other biasUnclear riskMeasurement bias: measurements of the IMT from the carotid and femoral arteries were "quality controlled by repeated measurement procedure", but intra- and inter-observer variability was not reported in this paper.

Diehm 2011

MethodsMulti-centre, prospective, randomised, double-blind, active-controlled trial


ParticipantsLocation: multi-centre, all sites in Germany, participants recruited April 2006 to December 2008
Inclusion criteria: patients aged ≥ 40 years with essential hypertension and confirmed PAD patients at Fontaine Stage II. Hypertension criteria checked twice. First at screening (visit 1), hypertension was defined as systolic blood pressure (SBP) 140 - 179 mmHg and diastolic blood pressure (DBP) 90 -109 mmHg with or without treatment with anti-hypertensive drugs. Second check was at baseline (visit 3), SBP > 130 mmHg and DBP > 85 mmHg. Other inclusion criteria: aspirin 100 mg and/or clopidogrel 75 mg or phenprocoumon (stable for ≥ 3 months) stable background cardiovascular medication; treadmill variability in absolute claudication distance (ACD) ≤ 25% between screening and baseline (week 4); ACD between 100 m and 300 m at baseline.

Although 177 participants were randomised only results report intention-to-treat (ITT) analysis on n = 163. 163 participants, mean (SD) age 66.3 (9.2), all Caucasian and 125 (76.7%) men
Nebivolol group: 84 (19 dropouts, n = 65 in per protocol analyses). Mean (SD) age 66.9 (9.8) years; 68 (81.0%) men
HCTZ group: 79 (17 dropouts, n = 62 in per protocol analyses). Mean (SD) age 65.5 (8.5) years; 57 (72.1%) men


InterventionsNebivolol 5 mg od or HCTZ 25 mg od

Duration: 6 months


OutcomesPrimary outcome: % change in IC distance (distance walked in metres until onset of pain) between baseline and week 28

Secondary outcomes: ACD (pain-free metres plus distance walked with pain) between baseline and week 28; quality of life; adverse events and serious adverse events (which included cardiovascular outcomes). Although not listed as secondary endpoints the paper also reports on ABI and all-cause mortality which are of interest to this Cochrane review


NotesAlthough 177 participants were randomised, results only report ITT analysis on n = 163. No detail given on the 14 patients randomised who are not included in the ITT analyses. Both treatment groups were well-balanced in terms of baseline differences. The only significant baseline difference was in concomitant use of amlodipine (25.3% versus 10.3% in the nebivolol vs HCTZ group, respectively). PAD defined as history of typical IC for ≥ 6 months; actual proven PAD by objective means such as haemodynamics and non-invasive imaging or angiography; history of previous lower extremity vascular intervention; ABI of the worse leg < 0.90.

All data reported in this Cochrane review are for ITT analyses (as is the convention for RCTs).


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskRandomised 1:1 using a computer-generated randomisation list

Allocation concealment (selection bias)Low riskRandomised 1:1 using a computer-generated randomisation list

Blinding (performance bias and detection bias)
Participant
Low riskTrial described as double-blind and the nebivolol and HCTZ tablets were manufactured to be identical in size and appearance

Blinding (performance bias and detection bias)
Physician
Low riskDouble-blind trial, tablets identical in appearance and allocation done by computer-generated randomisation list

Blinding (performance bias and detection bias)
Outcome assessor
Unclear riskNo detail given on the blinding of the outcome assessors

Incomplete outcome data (attrition bias)
All outcomes
Unclear risk177 participants were randomised yet ITT analyses were only conducted on n = 163. Per protocol analyses were conducted on n = 127 (n = 65 on nebivolol and n = 62 on HCTZ); therefore 36 participants discontinued study medication (n = 19 on nebivolol and n = 17 on HCTZ)

Selective reporting (reporting bias)Unclear riskData were reported on all the outcomes mentioned in the Methods of the study. Additional outcomes are reported in the results section. Data were not available for 14 randomised participants with no explanation given for this missing data

Other biasUnclear riskOf the 266 participants screened only 177 (66.5%) were randomised

HOPE 2004

MethodsMutli-centre, randomised, double-blind, placebo-controlled, two-by-two factorial trial


ParticipantsLocation: 267 centres in 19 countries in North and South America and Europe

HOPE participants were aged ≥ 55 years with existing cardiovascular disease or diabetes mellitus and an additional coronary risk factor (smoking, hypertension, hypercholesterolaemia/low HDL/microalbuminuria) but no heart failure or left ventricular dysfunction

Number of patients: 1725 PAD patients
Mean (SD) age: 66.7 (6.8) years; 1247 (72.3%) male
PAD was defined as either IC with ABI < 0.9, or previous vascular intervention or limb amputation for PAD

Number of PAD patients randomised to ramipril or placebo not stated in HOPE 2004 paper. Unpublished data provided by the authors

Ramipril: 836 (48.5%)

Placebo: 889 (51.5%)


InterventionsSingle-blind run-in period for 7-10 days with ramipril 2.5 mg daily. If tolerated, randomised to ramipril 10 mg od (starting dose of 2.5 mg for 7 days, 5 mg for three weeks, then 10 mg) or matching placebo

Duration: 4.5 years


OutcomesPrimary outcomes: cardiovascular death, non-fatal myocardial infarction, or non-fatal stroke
Secondary endpoints: all primary endpoints individually or all-cause mortality, heart failure hospitalisations, or diabetic complications


NotesHeart Outcomes Prevention Evaluation (HOPE)

We wish to thank Professor Östergren and colleagues for providing additional data on the PAD patients included in the HOPE trial.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskRandomised using a central telephone service

Allocation concealment (selection bias)Low riskRandomised using a central telephone service

Blinding (performance bias and detection bias)
Participant
Low riskDouble-blind trial. 'Randomised using a central telephone service and patients given either ramipril or matching placebo'

Blinding (performance bias and detection bias)
Physician
Low riskDouble-blind trial. 'Randomised using a central telephone service and patients given either ramipril or matching placebo'

Blinding (performance bias and detection bias)
Outcome assessor
Low riskEndpoint classification committee of two masked clinicians reviewed clinical records of all cardiovascular events reported by recruiting centres to determine whether they met endpoint criteria

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskPermanent discontinuation of medication was similar in the ramipril (28.9%) and placebo group (27.3%)

Selective reporting (reporting bias)Low riskNo apparent problems

Other biasLow riskNo apparent problems

INVEST 2003

MethodsPost hoc analysis of the INternational VErapamil-SR/Trandolapril STudy (INVEST). Multicentre, international, randomised, open-label, blinded endpoint study


ParticipantsLocation: 862 centres in 14 countries; patients enrolled September 1997 to February 2003
Inclusion criteria: Patients aged ≥ 50 years with hypertension and clinically stable coronary artery disease [CAD; previous MI, abnormal coronary angiogram (> 50% narrowing of ≥ 1 major coronary artery), abnormalities on 2 different types of stress test or angina pectoris) and a documented history of PAD on the baseline INVEST form. 2699/22,576 (12.0%) participants enrolled in the INVEST study had hypertension, stable CAD, and PAD at baseline

In the subgroup analysis of patients with concomitant PAD at baseline the following were randomised to receive:

Calcium antagonist-based strategy: 1345 participants, mean (SD not reported) age 68.6 years; 48.2% men; 49.3% white

Beta-adrenoreceptor blocker-based strategy: 1354 participants, mean (SD not reported) age 68.8 years; 47.6% men; 46.2% white. 2.5% participants in total were lost to follow-up (no further detail available on these participants)


InterventionsCalcium antagonist-based strategy (verapamil SR ± trandolapril) or a beta-adrenoreceptor blocker-based strategy (atenolol ± HCTZ). The first step was either verapamil SR 240 mg daily or atenolol 50 mg daily. Doses could be adjusted downwards if necessary by physician. Other anti-hypertensive drugs were added in stepped fashion to achieve blood pressure control

Step 2: verapamil SR/trandolapril 240/2 (4) mg daily or atenolol 50 mg daily plus HCTZ 25 mg daily

Step 3: verapamil SR/trandolapril 180/2 (4) mg twice daily or atenolol 50 mg twice daily plus HCTZ 25 mg twice daily

Step 4: verapamil SR/trandolapril 180/2 (4) mg twice daily plus HCTZ 25 mg daily or atenolol 50 mg twice daily plus HCTZ 25 mg twice daily plus trandolapril 2 (4) mg daily

Step 5: any other non-study anti-hypertensive drug (except beta-adrenoreceptor blockers for calcium antagonist-based strategy patients and calcium antagonists for beta-adrenoreceptor blocker strategy patients) could be added in if needed

Duration of treatment: 24 months


OutcomesPrimary outcome: composite endpoint of the first occurrence of all-cause death, non-fatal MI or stroke by ITT analysis

Other outcomes: all-cause mortality, cardiovascular mortality, total MI (fatal plus non-fatal), and total stroke (fatal plus non-fatal)

Two additional composite outcomes: an event in the primary outcome or poor/fair QoL at the final visit or an event in the primary outcome or first occurrence of a vascular procedure (carotid endarterectomy/stent, amputation, percutaneous peripheral vascular intervention, or aortic aneurysm resection/repair/stent) during follow-up


NotesDefinition of PAD for patients included in this post hoc analysis was a documented history of PAD on the baseline INVEST form. No detailed information regarding the diagnosis of PAD was available.

There are no significant baseline differences between the treatment groups for those patients with concomitant PAD but patients with PAD were significantly older and had more co-morbidities at baseline than those without PAD.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskRandomised using an electronic system available via the Internet

Allocation concealment (selection bias)Unclear riskRandomised using an electronic system available via the Internet and investigator informed electronically of treatment allocation. However treatment was open label

Blinding (performance bias and detection bias)
Participant
High riskOpen-label trial

Blinding (performance bias and detection bias)
Physician
High riskOpen-label trial. Physicians could down titrate initial dose of verapamil or HCTZ monotherapy or up-titrate in a stepped fashion as detailed in the Intervention section. The final step allowed the physician to add in any non-study anti-hypertensive (except beta-adrenoreceptor blockers for calcium antagonist-based strategy patients and calcium antagonists for beta-adrenoreceptor blocker strategy patients) at their discretion to control blood pressure

Blinding (performance bias and detection bias)
Outcome assessor
Low riskOutcomes adjudicated by a blinded events committee

Incomplete outcome data (attrition bias)
All outcomes
Unclear risk2.5% participants with concomitant PAD were lost to follow-up (no further detail available on these participants)

Selective reporting (reporting bias)High riskData analysed on the ITT population. In the INVEST protocol and main results papers the individual components of the composite endpoints are given by treatment strategy but only two composite endpoints are reported in the subgroup analysis in PAD patients

Other biasHigh riskPost hoc analysis and therefore the patients were not randomised to treatment on the basis of their PAD diagnosis. Definition of PAD for patients included in this post hoc analysis was a documented history of PAD on the baseline INVEST form. No detailed information regarding the diagnosis of PAD was available

NORMA 2011

MethodsNebivolol or Metoprolol in Arterial Occlusive Disease (NORMA) Trial (ISRCTN06278310), prospective, randomised, double-blind, single-centre trial


ParticipantsLocation: single-centre in Germany
Inclusion criteria: Patients with stable IC (Fontaine Stage II) for ≥ 6 months and an ABI of < 0.90 were recruited. All patients had stage I arterial hypertension (SBP 140 - 159 mmHg and DBP 90 - 99 mmHg) or a previous diagnosis of stage I arterial hypertension currently under treatment. Men and post-menopausal women were recruited. Previous treatment with nebivolol or carvedilol was not permitted

128 participants randomised to receive nebivolol (n = 65) or metoprolol (n = 63) but endpoint analysis (and demographics) only available on n = 109 (85.2%)

Nebivolol group: 52 (13 dropouts); mean (SD) age 66.7 (8.3) years; 45 (86.5%) men

Metoprolol group: 57 (6 dropouts); mean (SD) age 65.9 (7.9) years; 41 (71.9%) men


InterventionsNebivolol 5 mg od or metoprolol 95 mg od.

Duration: 48 weeks of treatment


OutcomesOutcomes not specified as primary and secondary. However, the study was powered to detect a difference in flow-mediated dilatation (FMD) between the groups.

Outcomes were the change in ABI, IC distance (metres until onset of pain), ACD (distance beyond which exercise could not be continued because of claudication pain), QoL, and endothelial-dependent FMD between baseline and 48 weeks of treatment

Although not listed as secondary endpoints the paper also reports on all-cause mortality and revascularisation (number requiring a revascularisation procedure) which are of interest to this Cochrane review.


NotesOnly post-menopausal women were recruited (to exclude the effect of female hormones on endothelial function). No significant differences in any baseline characteristics between the two treatment groups; borderline significant difference (P = 0.06) in the number of men in each group (with more men in the nebivolol group).

In the data analysis section the study authors state that analyses were performed for two populations: safety population (those who received ≥ 1 dose of double-blind medication) and an endpoint population (all patients for whom the endpoint variables were available). However, only the endpoint analyses are reported.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNo description of how the treatment allocation was generated other than that the patients were randomised

Allocation concealment (selection bias)Unclear riskNo description of how the treatment allocation was generated other than that the patients were randomised

Blinding (performance bias and detection bias)
Participant
Unclear riskTrial described as double-blind but no description given

Blinding (performance bias and detection bias)
Physician
Unclear riskTrial described as double-blind but no description given

Blinding (performance bias and detection bias)
Outcome assessor
Unclear riskClinical monitoring, data management, and statistical analysis were performed by a company but unclear it the study authors were involved in the outcome assessments

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskAuthors estimated a 20% dropout rate after screening. Endpoint data were not available on 19 randomised patients. 13 participants randomised to nebivolol dropped out. Six participants randomised to metoprolol dropped out

Selective reporting (reporting bias)High riskThe study authors reported on all of the outcomes stated in the methods but although 128 participants were randomised, endpoint analyses were only available on n = 109. In the data analysis section the study authors state that analyses were performed for two populations: safety population (those who received ≥ 1 dose of double-blind medication) and an endpoint population (all patients for whom the endpoint variables were available). However, only the endpoint analyses are reported. Authors state that the study had 80% power to detect a 2.0% change in FMD between the treatment groups, based on their planned sample size of n = 51 in each group

Other biasUnclear riskOnly post-menopausal women were recruited (to exclude the effect of female hormones on endothelial function). No significant differences in any baseline characteristics between the two treatment groups; borderline significant difference (P = 0.06) in the number of men in each group (with more men in the nebivolol group)

Overlack 1994

MethodsTwo-centre, randomised, double-blind, placebo-controlled trial

Dropout rate: 3.2% in treatment group; 1.7% in placebo group


ParticipantsLocation: two centres supervising five general practices, Germany

Inclusion criteria: patients with newly diagnosed or pre-treated essential hypertension, defined as sitting DBP of 95 -104 mmHg and one of nine concomitant diseases or therapies: hyperlipidaemia, type II diabetes mellitus, ischaemic heart disease, cardiac arrhythmias, PAD, nephropathy with proteinuria, chronic obstructive pulmonary disease (COPD), or degenerative joint disease treated with non-steroidal anti-inflammatory drugs (NSAIDS). Subgroup with PAD patients, defined as a history of IC ≥ 6 months, a pain-free walking distance of 80 to 200 m (Fontaine Stage IIb), and angiographic or ultrasound evidence of iliac or femoral occlusion

Number of patients: 490 (only 54 with PAD)

Males: 263; females: 227

Age: 40-75 years

Perindopril group: 253 (26 with PAD)
Placebo group: 237 (28 with PAD)


InterventionsPerindopril 4 mg od or placebo

Duration: six weeks


OutcomesPAD patients: Doppler ankle pressures, pain-free and MWD, SBP, DBP and heart rate


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNo description of how the treatment allocation was generated other than that the patients were randomised

Allocation concealment (selection bias)Unclear riskNo description of how treatment allocation was concealed other than that the patients were randomised

Blinding (performance bias and detection bias)
Participant
Unclear riskNo description of how the participant was blinded to treatment allocation other than that the trial was double-blind

Blinding (performance bias and detection bias)
Physician
Unclear riskNo description of how the physician was blinded to treatment allocation other than that the trial was double-blind

Blinding (performance bias and detection bias)
Outcome assessor
Unclear riskBlood pressure readings were measured using an automated device which printed out the values and dated them. These printouts were included with the follow-up records. All data analyses were conducted centrally in Munich, Germany using the DataEase system. There was no other description of blinding to the other outcome measures including Doppler index, and claudication and MWD

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskThere was overall dropout rate of 3.2% and 1.7% in the treatment and placebo groups, respectively, but the actual number of PAD patients with follow-up data was not clear

Selective reporting (reporting bias)High riskThe study authors do not report all the data collected for all the various subgroups, just the most disease-specific or relevant outcomes

Other biasLow riskAll data analyses were conducted centrally in Munich using the DataEase system

Schweizer 1998

MethodsRandomised, placebo-controlled trial. Method of randomisation not mentioned. Dropouts: one from the treatment group; two from the placebo group


ParticipantsLocation: Germany

Inclusion criteria: patients with Fontaine Stage IIb (walking distance < 200 m on treadmill at 4 km/h at a 10° gradient) PAD and concomitant chronic stable angina pectoris and mild hypertension plus one other risk factor from undergoing percutaneous transluminal coronary angioplasty (PTCA). PAD was diagnosed based on arterial angiography and colour-coded duplex ultrasound, with occlusions (diameter ≤ 5 cm) or subtotal stenoses in the distal superficial femoral artery present for > 6 months

Number of patients: 98

Verapamil group: 49 (1 dropout = 48)
Placebo group: 49 (2 dropouts = 47)


InterventionsVerapamil 240 mg twice daily, or placebo

Duration: six months


OutcomesAssessments before, immediately after, and at six weeks and six months from baseline of: the degree of stenosis measured by angiography; superficial femoral artery IMT measured by colour duplex ultrasound; SBP was measured in the arm and posterior tibial artery by means of continuous wave Doppler ultrasound in the PTCA-treated leg - the ankle/brachial SBP ratio was then calculated; distance to claudication; SBP, DBP and ventricular septal thickness


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNo description other than being random

Allocation concealment (selection bias)Unclear riskNo description of allocation concealment other than in a randomised, double-blind manner

Blinding (performance bias and detection bias)
Participant
Unclear riskNo description of blinding of the participant except for the study being double-blind

Blinding (performance bias and detection bias)
Physician
Unclear riskNo description of blinding of the physician except for the study being double-blind

Blinding (performance bias and detection bias)
Outcome assessor
Unclear riskThe colour-coded duplex scans were reviewed by two independent experts unaware of which group the participant was allocated to

Incomplete outcome data (attrition bias)
All outcomes
Low riskOne verapamil patient and two placebo patients were placed on other medication during the trial and their data were not included in the analyses

Selective reporting (reporting bias)High riskDistance to claudication after treatment not reported

Other biasUnclear riskMeasurement of the layer thickness: in separate experiments with three observers and four repeated measures of layer thickness for 36 patients, intra-observer variability was 3.1% and inter-observer variability was 5.9%

Zankl 2010

MethodsPilot study, single-centre, prospective, single-blind


ParticipantsLocation: single centre at an academic hospital in Germany, recruited April 2004 to February 2006
Inclusion criteria: PAD patients at Fontaine Stage IIa or higher with mild to moderate arterial hypertension, 36 participants (26 men), aged 18-80 years
Telmisartan group: 18 (0 dropouts). Median (95% CI) age 57 (53-64) years
Placebo group: 18 (0 dropouts). Median (95% CI) age 63 (56-75) years


InterventionsTelmisartan 40 mg or 80 mg od or placebo. Initial dose of telmisartan was 40 mg od. After 4 weeks of treatment, study medication uptitrated to 80 mg daily if SBP was ≥ 150 mmHg at this visit

Duration: 12 months


OutcomesPrimary endpoint: AWD

Secondary endpoints: FMD, carotid IMT, ABI, and disease-related QoL


Notesn = 40 randomised to treatment but only 36 completed the study and were analysed (attrition bias). The four who dropped were out due to premature discontinuation of the study medication, due to non-adherence. No detail is given on which drug they were randomised to. IMT was assessed by more than one sonographer but they were trained using a standardised protocol (not specified in paper) and used a common and extensively validated imaging protocol. A minimum of three measurements at different sites on each common carotid artery were obtained and IMT values were averaged. Power calculation suggests that n = 18 in each arm had 80% power to detect a 1.5% difference in MWD.

Study supported by Bayer.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNo description given, only that patients were randomised

Allocation concealment (selection bias)Unclear riskNo description given, only that patients were randomised

Blinding (performance bias and detection bias)
Participant
Unclear riskSingle-blind but no description given on blinding

Blinding (performance bias and detection bias)
Physician
Unclear riskPhysician was unaware of treatment allocation but no description of blinding given

Blinding (performance bias and detection bias)
Outcome assessor
Unclear riskNo detail is given on the blinding of the outcome assessor

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskn = 40 randomised to treatment but only 36 completed the study and were analysed. The four dropped out due to premature discontinuation of the study medication due to non-adherence. No detail is given on which drug they were randomised to

Selective reporting (reporting bias)Unclear riskAll outcomes mentioned in the methods are included in the results. Data not available for the four (10%) participants who dropped out

Other biasUnclear riskBaseline differences between the two groups despite randomisation (not reported) but ANCOVA was performed which found no confounding effects of differences in baseline values on treatment. Very small sample size

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

ABCD 2003The data for PAD patients were not available. All patients had concomitant diabetes mellitus

Ahimastos 2006Patients were not hypertensive at baseline

ASCOT 2005The data for PAD patients were not available

Bernardi 1988The treatment period for each arm in this trial was two weeks only and the minimum for inclusion is one month

Bogaert 1983This study fits the inclusion criteria except that the data were presented as a line graph in the results. The author has been successfully contacted but the relevant data are no longer available

Bostrom 1986Poorly designed study with only seven patients and no clear outcome measure. Attempts to contact the authors have been unsuccessful

Branchereau 1995The active drug, ifenprodil, is not an anti-hypertensive drug although it blocks alpha1-adrenoreceptors

Brown 1998This study is of treatment titration and not of outcome

Casiglia 1997Walking distance data were presented as a line graph and attempts to contact the authors have been unsuccessful

Coto 1989The active drug in this double-blind, placebo-controlled trial was an antiplatelet drug, not anti-hypertensive medication

Coto 1991This study was not a randomised controlled trial

Diehm 1993This study is a randomised controlled trial but the results are confounded by the potential addition of a nitrate to lower blood pressure further

Domschky 1977The active drug in this study was not an anti-hypertensive medication

Gastmann 1987Only published as an abstract but the abstract was not available

Hiatt 1985The treatment period for this trial was two weeks only and the minimum for inclusion is one month

Ingram 1982This study was not a randomised controlled trial

Jageneau 1977The active drug in this study was not an anti-hypertensive medication

Kalus 1995The treatment in this trial comprised a single dose only, and the minimum for inclusion is one month

Klieber 1986The treatment period for this trial was only 25 days and the minimum for inclusion is one month

Larsen 1969This trial did not compare two anti-hypertensive therapies or one therapy against placebo

Laurent 1994Hypertensive patients were actively excluded from this study

Leeman 1995This paper was concerned with abdominal aortic aneurysms and not peripheral occlusive disease

Lepantalo 1983Although this trial measured calf blood flow under treatment with beta-blockers, the patient cohort did not have PAD

Lepantalo 1984The treatment period for this trial was three weeks only and the minimum for inclusion is one month

Lepantalo 1985aThe treatment period for this trial was 10 days only and the minimum for inclusion is one month

Lepantalo 1985bThis is a retrospective case control study and not a randomised controlled study

Liakishev 2008This is a comment on the results of the ONTARGET trial not an original data paper

Lievre 1996The drug in this study, beraprost, is not an accepted anti-hypertensive medication

LIFE 2002The data for PAD patients were not available

Liu 1997The treatment in this trial comprised only one dose and the minimum for inclusion is one month

Mann 1998This paper describes a study design only and contains no results

Natali 1989The treatment period was 10 days only in this study and the minimum for inclusion is one month

Nelson 1978The active drug in this study was not an anti-hypertensive medication

Novo 1985This was not a randomised, controlled trial of anti-hypertensive therapy in peripheral vascular disease

Novo 1986Only one patient in this study was symptomatic of peripheral vascular disease. Attempts to contact the authors have been unsuccessful

Novo 1996Cross-over trial of captopril against ticlopidine and placebo, never tested against either drug though it was included in a sequence of treatments but at the same time in both treatment groups

ONTARGET 2008The data for PAD patients were not available

OPERA 2002PAD patients not included in this study

Panzner 1992Patients with peripheral vascular disease were not studied

POISE 2008Wrong disease group. This study included patients with atherosclerotic disease undergoing non-cardiac surgery

Reichert 1975The treatment period was two weeks only and not the minimum of one month

Roberts 1987Cross-over trial, with the data from both or all phases presented lumped together, and not analysed separately

Roberts 1992Cross-over trial with the data from both or all phases presented lumped together, and not separately

Rouffy 1989Hypertensive patients accounted for 40% of the cohort. The data were not presented for the hypertensive patients alone and thus the study could not be included. Attempts to contact the authors have been unsuccessful

Schweizer 1996Patients were not hypertensive at the outset of this study

Siniscalchi 1993The outcome measures in this study did not include any specific for PVD, i.e. walking distance

Smith 1982The treatment period for this trial was two weeks only and the minimum for inclusion is one month

Solomon 1991This study is of adequate design but the standard deviation was missing from the published data and the authors have failed to respond to requests for further data

Spence 1993Patients were not hypertensive at the outset of this study

Staessen 1978The active drug in this study was not an anti-hypertensive medication

Stumpe 1993This paper included patients with PAD but does not use measurements of PAD as outcome measures

Stumpe 1995This paper does not contain patients with PVD as a main or subgroup

Sutton-Tyrell 1995This is a good study of mortality with anti-hypertensive therapy but the therapies were mixed and the paper did not specify which patient was on which dose etc. Also, carotid stenosis and PAD were presented together in the results. Attempts to contact the authors have failed

Svendsen 1986Cross-over trial, with the data from both or all phases presented lumped together, and not separately

Takeda 2010This study did not include PAD patients

VALUE 2006The data for PAD patients were not available

Van de Ven 1994This study was not a randomised controlled trial

Weibull 1992This study did not include PAD patients as a study or subgroup

Winterfeld 1984This study was not a randomised controlled trial

 
Characteristics of studies awaiting assessment [ordered by study ID]
Ahimastos 2013

MethodsRandomised, double-blind, placebo-controlled trial

ParticipantsLocation: three hospitals in Australia (Melbourne, Townsville and Brisbane), participants recruited between 10 May 2008 and 23 August 2011
Inclusion criteria: ABI < 0.90 at rest in at least one leg; history of IC (unilateral or bilateral) and stable for the previous 6 months; and a stable medication regimen for at least 6 months. Hypertension was NOT one of the inclusion criteria although half (n = 106) the patients were hypertensive at baseline. Patients with resting brachial blood pressure of > 160/100 mmHg were excluded

Ramipril group: 106 (7 drop-outs). Mean (SD) age 65.5 (5.3) years; 87 (82.1%) men
Placebo group: 106 (5 drop-outs). Mean (SD) age 65.5 (7.1) years; 90 (84.9%) men

InterventionsRamipril 10 mg od or matching placebo

Duration: 24 weeks

OutcomesPrimary outcomes: pain-free walking time (time to onset of claudication pain) and maximum walking time

Secondary outcomes: ABI; stenosis severity (duplex ultrasound of lower limb arteries); patient-reported symptoms and functional status (assessed by the Walking Impairment Questionnaire); and health-related QoL

NotesRandomisation process reported as 'tamper-proof', with participants randomised into blocks of 10 to receive ramipril or matching placebo. All investigators, analysts and participants were blinded to drug assignment and baseline data when they performed follow-up measurements. No cross-over. Not all participants were hypertensive at baseline.

 
Comparison 1. ACE inhibitors versus placebo

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Cardiovascular events1Odds Ratio (M-H, Fixed, 95% CI)Totals not selected

 2 Claudication distance1Mean Difference (IV, Fixed, 95% CI)Totals not selected

 3 Maximum walking distance1Mean Difference (IV, Fixed, 95% CI)Totals not selected

 4 Ankle brachial pressure index1Mean Difference (IV, Fixed, 95% CI)Totals not selected

 
Comparison 2. Calcium antagonists versus placebo

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Degree of diameter stenosis1Mean Difference (IV, Fixed, 95% CI)Totals not selected

 2 Ankle brachial pressure index1Mean Difference (IV, Fixed, 95% CI)Totals not selected

 3 Arterial intima-media thickness1Mean Difference (IV, Fixed, 95% CI)Totals not selected

 
Comparison 3. Thiazide diuretics versus alpha-adrenoreceptor blocking drugs

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Arterial intima-media thickness (IMT)1Mean Difference (IV, Fixed, 95% CI)Totals not selected

 
Comparison 4. Angiotensin-II receptor antagonist versus placebo

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Maximum walking distance at 12 months (m)Other dataNo numeric data

 2 Intima-media thickness at 12 months (cm)Other dataNo numeric data

 3 Ankle-brachial pressure indexOther dataNo numeric data

 
Analysis 4.1 Comparison 4 Angiotensin-II receptor antagonist versus placebo, Outcome 1 Maximum walking distance at 12 months (m).
Maximum walking distance at 12 months (m)

StudyTelmisartan (n = 18)Placebo (n = 18)

Zankl 2010Median (95% CI)Median (95% CI)

Zankl 2010191 (157 - 226)132 (103 - 192)

 
Analysis 4.2 Comparison 4 Angiotensin-II receptor antagonist versus placebo, Outcome 2 Intima-media thickness at 12 months (cm).
Intima-media thickness at 12 months (cm)

StudyTelmisartan (n = 18)Placebo (n = 18)

Zankl 2010Median (95% CI)Median (95% CI)

Zankl 20100.08 (0.07 - 0.09)0.09 (0.08 - 1.00)

 
Analysis 4.3 Comparison 4 Angiotensin-II receptor antagonist versus placebo, Outcome 3 Ankle-brachial pressure index.
Ankle-brachial pressure index

StudyTelmisartan (n = 18)Placebo (n = 18)

Zankl 2010Median (95% CI)Median (95% CI)

Zankl 20100.60 (0.60 - 0.77)0.52 (0.48 - 0.67)

 
Comparison 5. Beta-adrenoreceptor blockers versus thiazide diuretics

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Change in intermittent claudication distance1Mean Difference (IV, Fixed, 95% CI)Totals not selected

 2 Absolute claudication distance1Mean Difference (IV, Fixed, 95% CI)Totals not selected

 3 Change in ankle brachial pressure index1Mean Difference (IV, Fixed, 95% CI)Totals not selected

 4 All-cause mortality1Odds Ratio (M-H, Fixed, 95% CI)Totals not selected

 5 Non-fatal cardiovascular events1Odds Ratio (M-H, Fixed, 95% CI)Totals not selected

 
Comparison 6. Beta-adrenoreceptor blockers versus beta-adrenoreceptor blockers

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 All-cause mortality1Odds Ratio (M-H, Fixed, 95% CI)Totals not selected

 2 Ankle brachial pressure index1Mean Difference (IV, Fixed, 95% CI)Totals not selected

 3 Intermittent claudication distance1Mean Difference (IV, Fixed, 95% CI)Totals not selected

 4 Absolute claudication distance1Mean Difference (IV, Fixed, 95% CI)Totals not selected

 5 Need for revascularisation1Odds Ratio (M-H, Fixed, 95% CI)Totals not selected

 
Comparison 7. Calcium antagonist-based strategy versus beta-adrenoreceptor blocker-based strategy

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Composite endpoint of death, non-fatal MI, or non-fatal stroke1Risk Ratio (M-H, Fixed, 95% CI)Totals not selected

 2 Composite endpoint of death, non-fatal MI or stroke, and revascularisation1Risk Ratio (M-H, Fixed, 95% CI)Totals not selected