Omega-3 fatty acids for intermittent claudication

  • Conclusions changed
  • Review
  • Intervention

Authors


Abstract

Background

Omega-3 fatty acids have been used in the treatment and prevention of coronary artery disease although current evidence suggests they may be of limited benefit. Peripheral arterial disease and coronary artery disease share a similar pathogenesis so omega-3 fatty acids may have a similar effect on both conditions. This is an update of a review first published in 2004 and updated in 2007.

Objectives

To determine the clinical and haematological effects of omega-3 supplementation in people with intermittent claudication.

Search methods

For this update the Cochrane Peripheral Vascular Diseases Group Trials Search Co-ordinator searched the Specialised Register (last searched September 2012) and the Cochrane Central Register of Controlled Trials (CENTRAL) (2012, Issue 9).

Selection criteria

Randomised controlled trials of omega-3 fatty acids versus placebo or non-omega-3 fatty acids in people with intermittent claudication.

Data collection and analysis

One review author identified potential trials. Two review authors independently assessed trial quality and extracted data. We contacted study authors for additional information if necessary.

Main results

Nine studies were included representing 425 participants. All studies compared omega-3 fatty acid supplementation with placebo lasting from four weeks to two years. Three studies with long treatment periods administered additional substances, making any observed effects impossible to attribute to omega-3 fatty acids and were excluded from the statistical analyses. One study did not express any mean values and, therefore, could not be included in statistical analyses.

No significant differences between intervention and control groups were observed in pain-free walking distance (mean difference (MD) 11.62 m, 95% confidence interval (CI) -67.74 to 90.98), maximal walking distance (MD 16.99 m, 95% CI -72.14 to 106.11), ankle brachial pressure index (MD -0.02, 95% CI -0.09 to 0.05), total cholesterol levels (MD 0.27 mmol/L, 95% CI -0.48 to 1.01), high-density lipoprotein cholesterol levels (MD 0.00 mmol/L, 95% CI -0.16 to 0.15), low-density lipoprotein cholesterol levels (MD 0.44 mmol/L, 95% CI -0.31 to 1.19), triglyceride levels (MD -0.39 mmol/L, 95% CI -1.10 to 0.33), systolic blood pressure (MD 5.00 mmHg, 95% CI -11.59 to 21.59) or plasma viscosity (MD 0.03 mPa/s, 95% CI -0.02 to 0.08).

There was some limited evidence that blood but not plasma viscosity levels decreased with treatment and gastrointestinal side effects such as nausea, diarrhoea and flatulence were observed in two studies.

Authors' conclusions

Omega-3 fatty acids appear to have little haematological benefit in people with intermittent claudication and there is no evidence of consistently improved clinical outcomes (quality of life, walking distance, ankle brachial pressure index or angiographic findings). Supplementation may also cause adverse effects such as nausea, diarrhoea and flatulence. Further research is needed to evaluate fully short- and long-term effects of omega-3 fatty acids on the most clinically relevant outcomes in people with intermittent claudication before they can be recommended for routine use.

摘要

背景

使用歐美茄3脂肪酸(Omega3 fatty acids)治療間歇性跛行(Intermittent claudication)症狀

目前已知歐美茄3脂肪酸可以有效的用在治療和預防冠狀動脈疾病上。另外歐美茄3脂肪酸有可能也可以用在周邊動脈疾病上,因為這兩種疾病之症狀具有高度相似性。

目標

確認間歇性跛行病患使用歐美茄3脂肪酸對於臨床和血液學的影響。

搜尋策略

藉由Cochrane Peripheral Vascular Diseases Group trials register (最後一次搜尋為2007年8月)及Cochrane Library的Cochrane Central Register of Controlled Trials (CENTRAL)(最後一次搜尋為 Issue 3, 2007年)找出相關試驗。此外,我們還搜尋藥廠、生產omega3食品的廠商,以及omega3專門的營養機構的網站的文獻。

選擇標準

在間歇性跛行病患,將歐美茄3脂肪酸,和安慰劑或非歐美茄3脂肪酸,對於其效果進行比較的隨機性對照試驗。

資料收集與分析

有1個作者確認可能符合的試驗,有2個作者獨立評估試驗品質和萃取數據,必要時會連絡研究作者以取得更多資訊。

主要結論

有6個包括313名受試參與者的研究被納入,所有的研究都比較歐美茄3脂肪酸補充品和安慰劑,持續至少4星期至兩年。其中有兩個長期治療的研究,因為有合併使用其他的藥物,因此無法確認為歐美茄3脂肪酸所產生的效果,也導致這些試驗被排除在統計分析之外。在介入治療組和對照組間並沒有觀察到下列指標上有顯著差異:踝肱壓力指數(AnkleBrachial Pressure Index ABPI)的數值(加權平均差(WMD)值為−0.02,95%CI值介於−0.09至0.05間)、收縮壓(WMD值為5.00毫米汞柱,95%CI值介於−11.59至21.59之間)、血漿黏度(WMD值為0.03毫帕/秒,95%CI值為−0.02至0.08之間)、無痛行走距離(WMD值為7.46公尺,95%CI值介於−25.47至40.39之間)或最大行走距離(maximal walking distance,MWD)(WMD值為0.27公尺,95%CI值介於−39.59至40.13之間)。但是血液的黏稠度有降低。有2個研究發現有腸胃之副作用。歐美茄3脂肪酸補充品會增加低密度脂蛋白膽固醇(LDL cholesterol) (WMD值為0.80毫莫耳/公升,95% CI值介於0.34至1.26間)和總膽固醇(WMD值為0.64毫莫耳/公升,95% CI值介於0.08至1.20間)的含量。

作者結論

對於間歇性跛行之病患,歐美茄3脂肪酸有血液方面之功效,雖然其效果有限。但是在本篇回顧中並沒有發現任何證據可以一致性的改善臨床上的預後(生活品質、無痛行走距離、最大行走距離、踝肱壓力指數和造影結果)。補充品可能會引發諸如增加總膽固醇和低密度脂蛋白膽固醇含量之類的不良事件。在此一領域應該要進行更多研究,以評估對於更多臨床相關預後的短期和長期效應。

翻譯人

此翻譯計畫由臺灣國家衛生研究院(National Health Research Institutes, Taiwan)統籌。

總結

使用歐美茄3脂肪酸治療間歇性跛行。 間歇性跛行症狀是一種在運動或移動的期間,血液沒有將足夠的氧氣供給至肌肉,造成小腿肌肉有緊繃的感覺,最後會造成病患產生行動遲緩或是無法行動的狀況。這種症狀好發於下肢血液的血管出現窄化,導致有慢性下肢動脈疾病之患者。通常會建議具有輕度下肢動脈疾病的患者要停止吸菸、保持運動且要服用低劑量的 aspirin來預防心臟病發作或中風。目前並沒有被廣泛接受可以用來治療跛行的藥物。歐美茄3脂肪酸是一種出現在魚油、雞蛋、水果和蔬菜中的多元不飽和脂肪酸,是一種必需的營養成分但是身體內無法自行產生。本回顧納入了6個包含有313名病患的研究,是將歐美茄3脂肪酸補充物和其他脂肪酸進行比較,在這些研究的基礎下發現,歐美茄3脂肪酸補充品並沒會改善行走距離或踝肱壓力指數,在服用補充品七周至兩年的過程中會降低血液的黏稠度。 由結果看來使用歐美茄3脂肪酸似乎可以對血液產生一些好處,但是幾乎沒有證據可以證明歐美茄3脂肪酸對臨床預後的改善。令人注意的還有低密度脂蛋白膽固醇和總膽固醇含量會在服用歐美茄3脂肪酸補充品的四週至四個月中出現些微的增加現象。在思及使用歐美茄3脂肪酸的好處時,一定也要考慮到引發血液中膽固醇含量增加的可能。

摘要

ω−3脂肪酸治疗间歇性跛行

研究背景

目前已知ω−3脂肪酸可以有效地治疗和预防冠状动脉疾病。人们认为ω−3脂肪酸也有可能用于治疗周围动脉疾病,因为这两种疾病的发病机制具有高度相似性。

研究目的

评估ω−3脂肪酸补充剂对间歇性跛行患者的临床和血液学的影响。

检索策略

检索了Cochrane Peripheral Vascular Diseases Group trials register(2007年8月)和Cochrane Central Register of Controlled Trials (CENTRAL)(2007年第3期)中收录的相关试验,以及制药公司、富含ω−3脂肪酸食品的生产公司、研究ω−3脂肪酸的营养机构的网站中的相关文献。

标准/纳入排除标准

比较使用ω−3脂肪酸与安慰剂或非ω−3脂肪酸对间歇性跛行患者疗效的随机对照试验。

数据收集与分析

由一名评价者检索文献,两名评价者独立对文献质量进行评价并提取数据,必要时联系文献作者以获取更多信息。

主要结果

总共纳入6个研究,总计313名受试者参与,研究比较了ω−3脂肪酸补充剂和安慰剂的效果,持续时间为四星期至两年。两项研究被排除,其原因是试验中合并使用了其他药物,导致无法确认ω−3脂肪酸的作用效果。

在干预治疗组和对照组间并没有观察到下列指标在统计学上的显著差异: 踝肱压力指数(ABPI) (加权平均差WMD: 0.02,95%CI: −0.09~0.05)、收缩压(WMD: 5.00 mmHg,95%CI: 11.59~21.59)、血浆粘度(WMD: 0.03 mPa/s ,95%: CI −0.02~0.08)、无痛行走距离(WMD: 7.46m,95%CI: 25.47~40.39)或最大行走距离(MWD)(WMD: 0.27m,95%CI: 39.59~40.13)。但是血液的粘稠度有降低。有2个研究观察到ω−3脂肪酸补充剂的胃肠道副作用。ω−3脂肪酸补充剂会增加低密度脂蛋白 (WMD: 0.80mmol/L,95%CI: 0.34~1.26)和总胆固醇(WMD: 0.64 mmol/L ,95%CI: 0.08~1.20)的含量。

作者结论

对于间歇性跛行的患者,ω−3脂肪酸有改善血液学指标的功效,但这种效果非常有限。本研究中没有发现任何证据可以支持其具有改善临床预后的效果(生活品质、无痛行走距离、最大行走距离、踝肱压力指数和造影结果等)。补充剂可能会引发诸如增加总胆固醇和低密度脂蛋白胆固醇含量之类的不良事件。需要进行更多研究,以评估ω−3脂肪酸对更多临床相关结局指标的长期或短期影响。

アブストラクト

間欠性跛行に対するオメガ-3脂肪酸

背景

オメガ-3脂肪酸は冠動脈疾患の治療および予防に用いられているが、最新のエビデンスはその有益性が限定的であることを示唆している。末梢動脈疾患と冠動脈疾患の発症機序には共通点があるため、オメガ-3脂肪酸が両疾患の病態に同程度の効果をもたらす可能性がある。本レビューは2004年発表、2007年に更新されたレビューの再更新である。

目的

間欠性跛行患者に対するオメガ-3脂肪酸補充の臨床的効果および血液学的効果を検証すること。

検索戦略

本更新版では、Cochrane Peripheral Vascular Diseases Group Trials Search Co-ordinatorがSpecialised Register(最終検索日2012年9月)およびCochrane Central Register of Controlled Trials (CENTRAL)(2012年第9版)を検索した。

選択基準

間欠性跛行患者を対象に、オメガ-3脂肪酸をプラセボまたはオメガ-3脂肪酸以外の脂肪酸と比較したランダム化比較試験。

データ収集と分析

1名のレビュー著者が対象候補となる試験を同定した。2名のレビュー著者が独立して試験の質を評価し、データを抽出した。必要に応じて研究著者らに追加情報を問い合わせた。

主な結果

参加者425例を対象とした9件の研究を対象とした。すべての研究において、オメガ-3脂肪酸補充とプラセボ投与を4週間から2年間比較した。3件の長期投与試験ではオメガ-3脂肪酸以外の物質も投与しており、観察された効果がオメガ-3脂肪酸に起因するとは断定できなかったため、統計解析から除外した。1件の研究では平均値が全く示されていなかったため、統計解析に組み入れることができなかった。

無痛歩行距離(平均差[MD]11.62 m, 95%信頼区間[CI]-67.74〜90.98)、最大歩行距離(MD 16.99 m、95%CI -72.14〜106.11)、足関節上腕血圧比(MD -0.02、95%CI -0.09〜0.05)、総コレステロール値(MD 0.27 mmol/L、95%CI -0.48〜1.01)、高密度リポ蛋白コレステロール濃度(MD 0.00 mmol/L、95%CI -0.16〜0.15)、低密度リポ蛋白コレステロール濃度(MD 0.44 mmol/L、95%CI -0.31〜1.19)、トリグリセリド濃度(MD -0.39 mmol/L、95%CI -1.10〜0.33)、収縮期血圧(MD 5.00 mmHg、95%CI -11.59〜21.59)および血漿粘度(MD 0.03 mPa/s、95%CI -0.02〜0.08)について、介入群と対照群との間に有意差は認められなかった。

2件の研究では、オメガ-3脂肪酸投与によって血漿粘度ではなく血液粘度が低下し、悪心、下痢および鼓腸など消化管の副作用が発現するという限られたエビデンスが得られた。

著者の結論

間欠性跛行患者に対するオメガ-3脂肪酸の血液学的有益性はほとんど認められず、臨床アウトカム(QOL、歩行距離、足関節上腕血圧比または血管造影所見)が一貫して改善するというエビデンスは得られていない。また、オメガ-3脂肪酸補充は悪心、下痢、鼓腸などの有害作用を引き起こす可能性がある。オメガ-3脂肪酸のルーチン使用を推奨する前に、間欠性跛行患者において臨床的に最も関連性のあるアウトカムに対するオメガ-3脂肪酸の短期および長期効果を完全に評価するための研究がさらに必要である。

訳注

《実施組織》厚生労働省「「統合医療」に係る情報発信等推進事業」(eJIM:http://www.ejim.ncgg.go.jp/)[2015.12.29]《注意》この日本語訳は、臨床医、疫学研究者などによる翻訳のチェックを受けて公開していますが、訳語の間違いなどお気づきの点がございましたら、eJIM事務局までご連絡ください。なお、2013年6月からコクラン・ライブラリーのNew review, Updated reviewとも日単位で更新されています。eJIMでは最新版の日本語訳を掲載するよう努めておりますが、タイム・ラグが生じている場合もあります。ご利用に際しては、最新版(英語版)の内容をご確認ください。

Plain language summary

Omega-3 fatty acid supplementation for intermittent claudication

Intermittent claudication is a pain in the calf due to a lack of blood needed to supply those muscles with oxygen during exercise or movement, ultimately resulting in the person to slow or stop movement. It is the most common presenting symptom for people with long-standing lower limb arterial disease resulting from narrowing of the arteries that supply the lower limbs with blood. This narrowing most commonly occurs through the process of atherosclerosis in which an artery wall thickens as a result of the accumulation of fatty materials such as cholesterol and triglycerides. People with mild lower limb arterial disease are advised to stop smoking, exercise, and take an antiplatelet agent to prevent heart attack or stroke. Medication to improve walking distance is only of limited value.

Omega-3 fatty acids are a type of fatty acid found in fish oils, eggs, fruits and vegetables. They are essential nutrients, as the body cannot make omega-3 fatty acids by itself. It is thought that omega-3 fatty acids may be beneficial in heart disease through their effect on several different biological mechanisms. As heart disease and intermittent claudication have similar disease processes it is possible that omega-3 fatty acids will have similar effects on both diseases and could potentially increase pain-free walking distance, quality of life and other measures of benefit in those suffering from intermittent claudication.

The review included nine randomised controlled trials with 425 participants, comparing omega-3 fatty acid supplementation with other fatty acids. On the basis of these studies, omega-3 fatty acid supplementation did not improve walking distance, blood pressure in the leg or any other measure of clinical benefit. There was some limited evidence to suggest that omega-3 fatty acid supplementation may reduce blood viscosity (the resistance of blood to flow), which when high could potentially contribute to intermittent claudication. There was no evidence to suggest that omega-3 fatty acid supplementation reduced plasma (the liquid component of blood) viscosity or improved the levels of different types of cholesterol or any other components of blood tested. Side effects such as nausea, diarrhoea and flatulence were observed in two studies.

概要

使用ω−3脂肪酸治疗间歇性跛行

间歇性跛行症状是一种在运动或行走的时候,因血液没有将足够的氧气供给肌肉,而造成的小腿肌肉的紧绷感觉,最终会造成患者产生行动迟缓或无法行动的结果。这种症状好发于因下肢血管出现狭窄而导致慢性下肢动脉疾病的患者。通常建议具有轻度下肢动脉疾病的患者要戒烟、多锻炼且要服用低剂量阿司匹林来预防心脏病的发作或中风。目前还没有可以广泛应用于治疗此类跛行的药物。

ω−3脂肪酸是一种富含于鱼油、鸡蛋、水果和蔬菜中的多不饱和脂肪酸,是一种机体必需的但却无法通过自身合成的营养成分。

本综述纳入了6项研究,总计313名患者。这些研究比较了ω−3脂肪酸补充剂和其他脂肪酸对治疗间歇性跛行的效果。根据纳入的研究,我们发现ω−3脂肪酸补充剂在改善行走距离或踝肱压力指数上没有统计学差异,但坚持服用补充剂七周至两年后会降低血液的粘稠度。从研究结果可以看出,使用ω−3脂肪酸可能对血液会产生一定好处,但是几乎没有证据可以证明ω−3脂肪酸可以改善临床预后。需要我们注意的还有,在持续四周至四个月服用ω−3脂肪酸补充品时,低密度脂蛋白胆固醇和总胆固醇含量会出现微量增加。在谈及使用ω−3脂肪酸的好处时,也一定要考虑到该类脂肪酸有引发血液中胆固醇含量增加的可能。

翻译注解

本摘要由重庆医科大学中国循证卫生保健协作网(China Effective Health Care Network)翻译。

Translated by: China Effective Health Care Network

平易な要約

間欠性跛行に対するオメガ-3脂肪酸補充

間欠性跛行では、運動や動作の際に筋肉に必要な酸素を供給する血液が不足するため、ふくらはぎに痛みを生じ、最終的に動作が緩慢になったり停止したりする。間欠性跛行は下肢動脈疾患に長期間罹患している患者に最も多い主症状で、下肢に血液を供給する動脈の狭窄が原因である。動脈狭窄はアテローム性動脈硬化の過程で発生することが最も多く、コレステロールやトリグリセリドといった脂質の蓄積により動脈壁の厚みが増す。軽度の下肢動脈疾患患者は、心臓発作や脳卒中を予防するために、禁煙、運動および抗血小板薬の服用を指示される。歩行距離の改善に薬物治療では限定的な効果にとどまる。

オメガ-3脂肪酸は魚油、卵、果物および野菜に含有される脂肪酸の一種である。オメガ-3脂肪酸は、必須栄養素であり、体内でオメガ-3脂肪酸を合成することはできない。オメガ-3脂肪酸は、複数の異なる生物学的機序を介して、心疾患に有益性をもたらすと考えられている。心疾患と間欠性跛行の発病過程は類似しているため、オメガ-3脂肪酸が両方の疾患に同程度の効果をもたらし、間欠性跛行患者の無痛歩行距離、QOLおよび有益性に関するその他の指標を改善させる可能性がある。

このレビューでは、オメガ-3脂肪酸補充を他の脂肪酸補充と比較した、参加者425例を対象とした9件のランダム化比較試験を選択した。これらの研究によると、オメガ-3脂肪酸を補充しても歩行距離、脚の血圧、その他臨床的に有益な指標は改善しなかった。オメガ-3脂肪酸補充が血液粘度(血流に対する抵抗性)を低下させることを示唆する限られたエビデンスが得られた。なお、血液粘度が高いと間欠性跛行の一因となる可能性がある。オメガ-3脂肪酸補充によって血漿(血液の液体成分)粘度が低下する、または種々の型のコレステロール濃度や他の血液成分の検査値が改善することを示唆するエビデンスは得られなかった。吐き気、下痢、鼓腸などの副作用が2件の研究で認められた。

訳注

《実施組織》厚生労働省「「統合医療」に係る情報発信等推進事業」(eJIM:http://www.ejim.ncgg.go.jp/)[2015.12.29]《注意》この日本語訳は、臨床医、疫学研究者などによる翻訳のチェックを受けて公開していますが、訳語の間違いなどお気づきの点がございましたら、eJIM事務局までご連絡ください。なお、2013年6月からコクラン・ライブラリーのNew review, Updated reviewとも日単位で更新されています。eJIMでは最新版の日本語訳を掲載するよう努めておりますが、タイム・ラグが生じている場合もあります。ご利用に際しては、最新版(英語版)の内容をご確認ください。

Background

Description of the condition

Intermittent claudication is the most common presenting symptom among people with chronic lower limb arterial disease (Ouriel 2001). Ischaemia of the calf muscles causes exercise-induced lower leg discomfort that classically resolves with rest. In severe progressive disease, pain also occurs at rest and is usually localised in the foot (Ouriel 2001).

The condition becomes more common with advancing age. The prevalence has been estimated at 1% to 2.5% in people aged 50 to 60 years, rising to 5% to 9% in people over 65 years of age (Meijer 1998; Stoffers 1991). Asymptomatic peripheral arterial disease (often defined pragmatically as an ankle brachial pressure index (ABI) less than 0.90 in either leg) is even more common, affecting 17% of men and 20% of women in the over-55 year age group (Meijer 1998). Symptomatic and asymptomatic disease are strongly associated with future cardiovascular (CV) events (Criqui 1992).

Management of intermittent claudication includes the targeting of risk factors such as smoking, diabetes, hypertension and hypercholesterolaemia (Regensteiner 2002). Treatment of symptoms may include various vasoactive drugs, exercise regimens, and more complex interventions such as bypass surgery and angioplasty (Ouriel 2001).

Description of the intervention

The omega-3 fatty acid family consists of eicosapentaenoic acid (EPA), docosapentaenoic acid (DPA) and alpha-linolenic acid (LNA). EPA and DPA are the active forms of the omega-3 family and are found in certain fish oils, for example salmon and mackerel. Green leafy vegetables and soybean oil contain LNA, which, when consumed, is converted to EPA and then docosahexaenoic acid (DHA).

Previous Cochrane reviews have investigated the potential benefit of omega-3 fatty acids in cystic fibrosis (Oliver 2011), type II diabetes (Hartweg 2008), dysmenorrhoea (Proctor 2001), schizophrenia (Irving 2006), pregnancy (Makrides 2006), and CV disease (Hooper 2004), with generally inconclusive results.

How the intervention might work

Since the 1990's much interest has surrounded the use of omega-3 fatty acids (usually in the form of fish oils) to target the origin and development of coronary artery disease. Through effects on lipoprotein metabolism, haemostatic function, inhibition of smooth muscle cell proliferation and anti-arrhythmic actions, omega-3 fatty acids have been found both to prevent the development of and to reduce the mortality from coronary artery disease (Abeywardena 2001). The UK's National Institute for Health and Clinical Excellence (NICE) guidelines on secondary prevention in people following a myocardial infarction (MI) have found omega-3 fatty acids to be a cost-effective therapy (Cooper 2007), although evidence would now suggest that in people with CV disease, omega-3 fatty acid supplements do not reduce adverse CV outcomes (Halim 2012).

At a molecular level omega-3 fatty acids are thought to achieve their effects through several different molecular pathways, including alteration of physical and chemical properties of cellular membranes, direct interaction with and modulation of membrane channels and proteins, regulation of gene expression via nuclear receptors and transcription factors, changes in eicosanoid profiles and conversion of omega-3 fatty acids to bioactive metabolites (Mozaffarian 2011).

Why it is important to do this review

Atherosclerosis is a systemic disorder with shared risk factors which are independent of the specific arterial system involved. For this reason, omega-3 fatty acids may also prove beneficial in peripheral arterial disease. This review has sought to identify and evaluate all studies involving the use of omega-3 fatty acids in people with intermittent claudication to determine the effectiveness of omega-3 fatty acids based on available evidence. Areas requiring further investigation have also been identified.

Objectives

To determine the effects of omega-3 fatty acid supplementation on clinical, haematological and biochemical parameters in people with intermittent claudication. This was assessed by the effect of omega-3 fatty acid supplementation on pain-free walking distance (PFWD), maximal walking distance (MWD), quality of life scores, ankle brachial pressure index (ABI), angiographic changes, amputation rate, revascularisation procedures, mortality, vascular events, plasma lipid levels, systolic blood pressure (SBP), diastolic blood pressure (DBP) and blood/plasma viscosity.

Methods

Criteria for considering studies for this review

Types of studies

All randomised controlled trials of omega-3 fatty acid supplementation versus placebo or alternative therapy, published in any language, with any method of randomisation were considered to be eligible.

Types of participants

We included trials involving people with chronic (greater than six months) intermittent claudication secondary to atherosclerotic disease. We excluded trials that included participants with acute lower limb ischaemia or who had undergone lower limb vascular surgery or angioplasty within the previous three months because the authors felt that the effects of surgery might make it difficult to attribute any changes in condition to omega-3 fatty acid supplementation.

Types of interventions

Trials involving any form, dose and duration of omega-3 fatty acid supplementation were eligible. Supplementation could involve fatty acid capsules or powders, or dietary manipulation (for example, increased consumption of oily fish).

Types of outcome measures

Primary outcomes
  • Quality of life scores.

  • Walking distance (pain-free and maximal).

  • Ankle brachial pressure index.

  • Angiographic findings.

Secondary outcomes
  • Lipid levels (total cholesterol, low-density lipoprotein (LDL) cholesterol, very-low-density lipoprotein (VLDL) cholesterol, high-density lipoprotein (HDL) cholesterol, triglycerides).

  • Blood pressure (systolic and diastolic).

  • Progression to amputation.

  • Revascularisation procedures.

  • Non-fatal myocardial infarction (MI), non-fatal cerebrovascular accident (CVA).

  • All-cause and vascular mortality.

  • Blood and plasma viscosity.

  • Adverse effects of therapy.

Search methods for identification of studies

There were no restrictions for language.

Electronic searches

For this update the Cochrane Peripheral Vascular Diseases Group Trials Search Co-ordinator (TSC) searched the Specialised Register (last searched September 2012) and the Cochrane Central Register of Controlled Trials (CENTRAL) 2012, Issue 9, 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, AMED, and through handsearching relevant journals. The full list of the databases, journals and conference proceedings that 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).

Searching other resources

For the original review, the review authors sought additional trials through cross-referencing and personal communication with colleagues. In addition, they searched literature from websites of Roche Pharmaceutical (Roche), and Seven Seas Pharmaceuticals (Seven Seas), manufacturers of omega-3 rich foods (Columbus Eggs), and websites of nutritional organisations dedicated to omega-3 fatty acids (Fish Foundation; Omega-3) for relevant references.

Data collection and analysis

Selection of studies

Two review authors (TS or AC and either WH or JP) obtained and reviewed abstracts following the identification of trials for possible inclusion. If the criteria for type of study, participants and intervention were met, we obtained the full-text paper. We contacted trial authors if the published articles did not provide sufficient information to determine whether the study should be included.

Data extraction and management

TS or AC reviewed each study and extracted the following information: method of allocation; degree of blinding; exclusions post-randomisation; losses to follow-up; country where the study was undertaken; number, age and sex of participants; inclusion criteria; exclusion criteria; treatment; control; duration; outcomes; and additional quality items (described below). WH, JP or AW (of the Cochrane PVD Review Group) independently extracted data.

Several authors provided additional data. Leng provided mean and standard deviation (SD) values for ABI and walking distance (Leng 1998). Mori provided mean and SD values for cholesterol and triglyceride levels, along with details of the randomisation process and participant characteristics (Mori 1992). Carrero provided walking distance values and information on the study's method of double-blinding (Carrero 2005). Conway provided details on randomisation/concealment, exclusion criteria, inclusion criteria and numerical values for each outcome (Conway 2005).

Data from the following studies were not available.

  • Berrettini 1996: data on the seven participants in the study with peripheral arterial disease.

  • Carrero 2006: further information regarding allocation concealment and blinding process.

  • Ishikawa 2010: data on the potential subset of participants included in study with intermittent claudication.

  • Mori 1992: information on the severity of peripheral arterial disease among participants.

  • Ramirez-Tortosa 1999: information on the actual method of participant allocation and whether the control group received dietary, exercise and smoking cessation advice.

  • Schiano 2008: mean values for all outcomes. No information regarding randomisation process and allocation concealment.

  • Stricker 2008: no information regarding concealment of allocation and mean values for some outcomes.

  • Woodcock 1984: further information regarding participant characteristics and withdrawals/losses to follow-up.

Assessment of risk of bias in included studies

We assessed risk of bias in each included study according to the guidance of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011, Chapter 8.5-6). We assessed the following domains of trial quality:

  • random sequence generation (selection bias);

  • allocation concealment (selection bias);

  • blinding of participants and personnel (performance bias);

  • blinding of outcome assessment (selection bias);

  • incomplete outcome data (attrition bias);

  • selective reporting (reporting bias);

  • other biases.

We gave trials a quality rating of 'low risk', 'unclear risk' or 'high risk' of bias for each of these domains in line with guidance from the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). A description of the reasons for our ratings are given in the 'Risk of bias' tables which were completed for each study.

Measures of treatment effect

We performed statistical analyses according to the statistical guidelines for review authors in the Cochrane Peripheral Vascular Diseases Group using Review Manager 5 (RevMan 2011) software.

Since the data were continuous, we extracted mean and SD values for each outcome and analysed using the mean difference (MD), with 95% confidence intervals (CI), as the summary statistic. If appropriate, we pooled results for the purpose of meta-analysis. We excluded three studies from the analysis process as they administered additional active substances to their intervention groups, which made it impossible to attribute any observed effects to omega-3 fatty acids alone (Carrero 2005; Carrero 2006; Leng 1998). One study could not be included in the analysis process as they did not express mean values (preferring median (25th; 75th percentile) values) for any outcomes; this made it impossible to combine the results of this study with the other included studies in a meta-analysis (Schiano 2008).

Unfortunately, no study presented SDs for the differences between pre- and post-intervention values. This meant that the MD analysis was carried out on the difference between treatment and control values in the post-trial period, ignoring the baseline values. Since all the studies were randomised, baseline values should have been balanced between groups. Some papers presented standard errors (SE) rather than SDs. These figures were, therefore, converted to SDs with the formula SD = √N x SE.

Dealing with missing data

When there were issues with missing data, we contacted the original investigators to request the relevant missing data. When the requested data were not received, only available data were analysed. We analysed all data on an intention-to-treat (ITT) basis.

Assessment of heterogeneity

We evaluated heterogeneity in the studies by examining the characteristics and quality of included studies, along with the I2 statistic for heterogeneity. We used both fixed-effect and random-effects models of analysis to investigate the presence of heterogeneity further.

Assessment of reporting biases

We assessed reporting bias by scrutinising the study results using the 'Risk of bias' tables in Review Manager 5 (RevMan 2011). If we had identified a sufficient number of trials, a funnel plot would have been constructed to help detect bias.

Data synthesis

Four of the studies measured outcomes on more than one occasion (Carrero 2005; Carrero 2006; Leng 1998; Mori 1992). Therefore, a decision had to be made on which values to consider. Mori 1992 took measurements at two, four, and ten weeks. As the ten week measurements were taken six weeks after cessation of therapy (by which time the effects of omega-3 fatty acid supplementation might have subsided), the four-week values were used in the analysis. Leng 1998 measured outcomes at six, 12, 18 and 24 months. The 24-month data were used because it was felt that the maximal effect of the intervention therapy would be observed at this time. Carrero 2005 took measurements at three, six, nine and 12 months. The 12-month values were used to allow the maximal effects of therapy to be analysed. Carrero 2006 measured outcomes at six and 12 months. The 12-month values were again used for the same reasons as mentioned above. A separate meta-analysis for short- and long-term outcomes was considered but not undertaken because only one study looked at long-term outcomes and this was already excluded from the meta-analysis (Leng 1998).

Subgroup analysis and investigation of heterogeneity

Insufficient data existed to perform any subgroup analyses (e.g. based on severity of disease or presence of diabetes).

Sensitivity analysis

We considered performing sensitivity analysis; however, the small number of studies included in the meta-analyses made this impractical.

Results

Description of studies

Results of the search

For this update, following screening of studies from the Specialised Register and CENTRAL, an additional 14 studies were identified for possible inclusion.

Included studies

Summary details of included studies are given in the Characteristics of included studies table.

We included three additional studies in this update (Carrero 2006; Schiano 2008; Stricker 2008), making a total of nine included studies (Carrero 2005; Carrero 2006; Conway 2005; Gans 1990; Leng 1998; Mori 1992; Schiano 2008; Stricker 2008; Woodcock 1984). The nine studies involved 425 participants and were based in six different countries: the UK (Conway 2005; Leng 1998; Woodcock 1984), Australia (Mori 1992), Italy (Schiano 2008), the Netherlands (Gans 1990), Switzerland (Stricker 2008), and Spain (Carrero 2005; Carrero 2006). Seven of the studies were small, involving between 19 and 60 participants. The largest study involved 120 participants (Leng 1998).

Types of participants

All studies required participants to have symptoms of intermittent claudication. In addition, five studies required confirmation of disease by Doppler pressure indices (Carrero 2005; Carrero 2006; Gans 1990; Leng 1998; Woodcock 1984). Two studies required confirmation of disease by Doppler pressure indices and the presence of stenosis or occlusion of a leg artery on duplex scanning or angiogram (Schiano 2008; Stricker 2008). Mori 1992 required angiographic confirmation of peripheral arterial disease. Gans 1990 also used standardised treadmill testing to ensure that participants had stable disease.

Six studies included both males and females (Conway 2005; Gans 1990; Leng 1998; Schiano 2008; Stricker 2008; Woodcock 1984), whereas Carrero 2005, Carrero 2006 and Mori 1992 included only male participants.

Studies differed in terms of exclusion criteria.

Woodcock 1984 described no exclusion criteria.

Gans 1990 excluded participants who had unstable angina or an MI within the past three months; any illness with rapid progression; rest pain; gangrene; DBP greater than 100 mm Hg; poorly controlled diabetes; vasculitis; high or low platelet count; haematocrit greater than 55%; people with fish allergy; and those on lipid lowering or anti-platelet therapy.

Mori 1992 excluded people with renal impairment or generalised inflammatory disorders.

Conway 2005 excluded people with critical ischaemia; those with haemorrhagic disorders; those on warfarin therapy; those with diabetes and those with aspirin-sensitive asthma.

Leng 1998 excluded people with critical ischaemia; previous or impending arterial surgery or angioplasty (within the previous three months); unstable angina or MI (within the previous three months); severe intercurrent illness including severe liver disorders and malignancy; epilepsy; pregnancy; and treatment with anti-coagulants, other oils, lithium or phenothiazines.

Carrero 2005 and Carrero 2006 excluded people who were eligible for revascularisation surgery; those with endocrine or metabolic disturbances; those with a cardiac history; and those who resided outside of the Granada region of Spain.

Schiano 2008 excluded people with rest pain; trophic lesions in affected limb; acute coronary syndrome or CV event in previous six months; significant renal, hepatic or inflammatory disease; history of hypersensitivity; other contraindications to n-3 polyunsaturated fatty acids (PUFAs).

Stricker 2008 excluded participants with acute, intermittent illness; thromboangiitis obliterans; renal insufficiency; acute stroke or MI in previous two months; current anticoagulation medicine; liver cirrhosis; presence of a malignant tumour.

Interventions and placebos

All studies except Schiano 2008 compared omega-3 fatty acid supplementation with placebo. Five studies used omega-3 fatty acid supplementation in the form of fish oil capsules (Conway 2005; Gans 1990; Mori 1992; Schiano 2008; Woodcock 1984). Gans 1990 used 1.8 g EPA and 1.2 g DHA per day, Mori 1992 used 2.8 g EPA and 1.8 g DHA per day, Woodcock 1984 used 1.8 g EPA per day and Conway 2005 used 1.7 g EPA and 1.15 g DHA per day. Schiano 2008 differed slightly in that participants were given 1 g omega-3 fatty acid capsules twice daily. The capsules used in the study contained at least 85% of EPA and DHA with a mean ratio of EPA/DHA of 0.9:1.5. Placebo preparations included omega-6 fatty acids (Gans 1990 used 3 g linoleic acid/corn oil); mono-unsaturated fatty acids (Mori 1992 used 11.2 g oleic acid/olive oil); or a combination of both (Woodcock 1984 used linoleic acid/corn oil and oleic acid/olive oil). Conway 2005 used 10 g mixed oil capsules, the composition of which resembled the mean UK dietary fatty acid consumption. Unfortunately, its constituents were not specified by the author. Schiano 2008 did not use a placebo preparation and instead the control group remained on pre-enrolment therapy.

Leng 1998 differed in that the intervention group received capsules containing a combination of omega-3 and omega-6 fatty acids (1.12 g gamma-linolenic acid (GLA)/evening primrose oil daily, increasing to 1.68 g after two weeks and 180 mg EPA, increasing to 270 mg). The control group received a placebo preparation containing 2.5 g sunflower oil/linoleic acid).

Carrero 2005 used dairy drinks containing 0.13 g DHA and 0.2 g EPA, as well as 5.12 g oleic acid (a mono-unsaturated omega-9 fatty acid) plus folic acid, and vitamins A, B6, D and E. The control preparation consisted of regular semi-skimmed milk with added vitamins A and D.

Carrero 2006 had two treatment groups. Both received a dairy product containing EPA, DHA, oleic acid, folic acid, and vitamins A, D, E and B6 while the second treatment group also received 20 mg/day of simvastatin. There were also two control groups, which both received 500 mL/day of semi-skimmed milk as a placebo preparation with the addition of 20 mg/day of simvastatin for the second control group.

Stricker 2008 used two tablespoons (35 mL) a day of canola oil (2.24 g of LNA, C18 3n-3 omega-3 fatty acid). The control group received two tablespoons a day of sunflower oil (16.24 g of linolenic acid, C18 2n-6 omega-6 fatty acid).

Studies differed in the changes that they made to participants' existing medication and lifestyles. Mori 1992 asked participants to maintain normal dietary habits and activity levels. However, participants were asked to stop taking aspirin and other non-steroidal anti-inflammatory drugs in the 14 days prior to the study. Gans 1990 asked participants not to change their smoking, dietary or exercise habits during the study period. No changes were made to their CV medication. Similarly, in Woodcock 1984 no changes to prescribed drugs or smoking status were made. In Schiano 2008, pre-enrolment therapy was not modified and concomitant CV treatments were not discontinued. Stricker 2008 instructed participants not to change their habitual alimentation. Conway 2005 referred all participants to a nurse-led exercise programme.

In Leng 1998, participants were advised to increase physical activity levels and stop smoking, but no changes were made to pre-existing medication. Carrero 2005 and Carrero 2006 provided all participants with dietary counselling (advised to increase fruit and vegetable consumption and reduce processed food intake), and advice to exercise daily and to stop smoking. All participants were also given an anti-platelet agent (Triflusal) and a haemorrheological agent (pentoxifylline).

Length of studies

Duration of therapy varied substantially between studies. The longest study lasted two years (Leng 1998), and physiological measurements were recorded every six months. Carrero 2005 took place over 12 months and measurements were recorded at baseline, three, six, nine and 12 months. Carrero 2006 also lasted 12 months with measurements taken at baseline, six months and 12 months. Conway 2005 lasted 16 weeks and measurements were taken at baseline and study completion only. Similarly, Gans 1990 lasted four months and measurements were taken only at baseline and study completion. Schiano 2008 lasted three months with measurements taken at baseline and completion. Stricker 2008 lasted eight weeks with measurements taken at the beginning and end of the trial. Woodcock 1984 lasted seven weeks and again, measurements were only recorded at baseline and completion. The shortest duration of therapy was four weeks (Mori 1992); however, measurements were taken at baseline, two, four, and ten weeks (i.e. ten weeks = six weeks following cessation of therapy).

Outcomes

No studies included angiographic change as an outcome. Only one study included quality of life (Conway 2005). One study looked at fatal/non-fatal CV events, all-cause mortality, progression to amputation and vascular surgery/bypass (Leng 1998). Two studies reported data on blood pressure and side effects of treatment (Gans 1990; Leng 1998). Three studies evaluated blood/plasma viscosity (Gans 1990, Leng 1998; Woodcock 1984). Seven studies included triglycerides as an outcome (Carrero 2005; Carrero 2006; Gans 1990; Mori 1992; Schiano 2008; Stricker 2008; Woodcock 1984). Eight studies evaluated effects on cholesterol levels (Carrero 2005; Carrero 2006; Gans 1990; Leng 1998; Mori 1992; Schiano 2008; Stricker 2008; Woodcock 1984). Seven studies included ABI (Carrero 2005; Carrero 2006; Conway 2005; Gans 1990; Leng 1998; Schiano 2008; Woodcock 1984), and six studies included walking distance walking distance as outcomes (Carrero 2005; Carrero 2006; Conway 2005; Gans 1990; Leng 1998; Woodcock 1984).

Excluded studies

For this update, an additional four studies were excluded (Ishikawa 2010; MacKay 2012; Moller 1998; Olsson 1984), making a total of six excluded studies (see Characteristics of excluded studies table).

Berrettini 1996: this was a randomised, controlled, double-blind trial involving 39 participants with vascular disease (peripheral, cerebral or coronary). The intervention group were given EPA plus DHA daily for 16 weeks. We excluded this study because it was not possible to identify which participants had peripheral arterial disease. We requested data for this subset of participants from the investigator but received no reply.

Ishikawa 2010: this was a prospective, randomised, open-label, blinded endpoint evaluation method trial involving 223 people. These people were a subset of the large JELIS (Japan EPA Lipid Intervention Study) trial, which investigated the effects of EPA on CV events in 18,645 Japanese participants with hypercholesterolaemia. The intervention group were given EPA plus a statin, while the control group received a statin alone for a maximum follow-up period of five years. We excluded this study as it was impossible to ascertain whether participants included in the trial suffered from intermittent claudication. We contacted the authors to seek this information but received no reply.

MacKay 2012: this was a randomised cross-over double-blind study involving 150 people with intermittent claudication. Participants received OMACOR fish oil (850 to 882 mg EPA and DHA) or a placebo for six weeks with the aim of determining if omega-3 fatty acid supplementation in addition to best medical therapy can reduce the increased platelet and endothelial activity that is present in people with intermittent claudication. There were no relevant outcomes for this review.

Moller 1998: this was a double-blind, placebo-controlled trial involving 31 people with stable intermittent claudication. Participants received either 5 g omega-3 fatty acids or placebo (5 g corn oil) daily for 12 months. No significant effects were observed on clinical endpoints in people receiving omega-3 fatty acids compared to those receiving placebo. We excluded this study as it was only available in abstract form and it was not clear if participants were randomly assigned to groups or not. We contacted the authors for further details, but we did not receive any replies.

Olsson 1984: this was a Swedish double-blinded, placebo-controlled trial investigating the effect of omega-3 fatty acids on intermittent claudication. Participants received a daily dosage of 2.6 g linoleic acid plus 300 mg gamma-linolenic acid or a placebo for six months. We excluded this trial as, after translation, it was deemed to be a controlled clinical trial and not a randomised controlled trial.

Ramirez-Tortosa 1999: this was a longitudinal intervention study involving 37 men with peripheral arterial disease (24 in intervention group, 13 in control group) plus 20 in a healthy reference group. The intervention group received three months of oleic acid/olive oil, followed by a three-month wash-out period, then three months of oleic acid and 16 g fish oil. The participants also received advice regarding smoking cessation, diet and exercise. We excluded this study because participants were not randomly allocated into treatment and control groups. The study was also unblinded.

Ongoing Studies

For this update, we classified one study as ongoing (Leyva 2011) (see Characteristics of ongoing studies table).

Leyva 2011: the FLAX-PAD randomised, double-blind controlled trial involves 110 people aged over 40 years old with peripheral arterial disease. Participants receive 30 g of milled flaxseed or placebo per day with the aim of determining the effect of dietary flaxseed on improving symptoms of CV disease. Primary endpoints are incidence of MI and stroke. Secondary measures include: requirement for surgical interventions, exercise and cardiopulmonary performance, cardiac arrhythmias, serum lipid profile, arterial sufficiency, blood pressure, inflammatory profile, platelet function, changes in drug dosage levels and nutrigenomic and biomarker profiles in the blood.

Risk of bias in included studies

Details of the 'Risk of bias' are presented in Figure 1 and Figure 2 and in the Characteristics of included studies table.

Figure 1.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Figure 2.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

We judged only two studies to have a low risk of selection bias (random sequence generation and allocation concealment combined) (Conway 2005; Leng 1998).

We considered seven studies to have 'low' risk of bias for random sequence generation (Carrero 2005; Carrero 2006; Conway 2005; Leng 1998; Mori 1992; Stricker 2008; Woodcock 1984). Two studies provided no description of the randomisation process and we deemed them to have an 'unclear' risk of bias (Gans 1990; Schiano 2008).
We judged three studies to have 'low' risk of bias for allocation concealment (Conway 2005; Gans 1990; Leng 1998). The remaining six studies did not describe how group allocation was concealed and we, therefore, judged them to have an 'unclear' risk of bias (Carrero 2005; Carrero 2006; Mori 1992; Schiano 2008; Stricker 2008; Woodcock 1984).

Blinding

We judged six studies to have 'low' risk of bias from blinding of participants and personnel (performance bias) (Carrero 2005; Leng 1998; Mori 1992; Schiano 2008; Stricker 2008; Woodcock 1984). Three studies reported a 'blind' trial method but did not provide details of the blinding process and we deemed them to have an 'unclear' risk performance bias (Carrero 2006; Conway 2005; Gans 1990).

We considered four studies to have 'low' risk of bias from blinding of outcome assessors (Carrero 2005; Mori 1992; Schiano 2008; Stricker 2008). The five other trials did not report whether outcome assessors were blinded and we judged them to have 'unclear risk' of detection bias (Carrero 2006; Conway 2005; Gans 1990; Leng 1998; Woodcock 1984).

Incomplete outcome data

We judged five trials to have 'low' risk of attrition bias due to incomplete outcome data (Carrero 2006; Conway 2005; Gans 1990; Mori 1992; Schiano 2008). We considered three studies to have an unclear risk of attrition bias with Stricker 2008 lacking an explicit statement to confirm there were no exclusions or withdrawals (Leng 1998; Stricker 2008; Woodcock 1984). Leng 1998 had some unexplained missing outcome data. Woodcock 1984 did not state number of people randomised or describe any withdrawals or exclusions. We judged Carrero 2005 to have 'high' risk of attrition bias due to incomplete outcome data as exclusions were not balanced between study groups.

Selective reporting

We judged all studies except Woodcock 1984 to have 'low' risk of bias from selective reporting as they all reported the prespecified outcomes in the prespecified way. We judged Woodcock 1984 to have 'high' risk of bias as data was missing for most outcomes.

Other potential sources of bias

Carrero 2006 reported a large difference between claudication distance across groups between baseline (pre-study) and T0 (start of study). The trial sample size was also very small with poor matching for age. Mori 1992 provided no data on severity of the variable peripheral arterial disease. We identified no other potential sources of bias for the remaining studies (Carrero 2005; Conway 2005; Gans 1990; Leng 1998; Schiano 2008; Stricker 2008; Woodcock 1984).

Effects of interventions

The studies were clinically heterogeneous in terms of dose, duration of therapy and placebo used. We used both fixed-effect and random-effects models in every analysis; where statistical heterogeneity was detected (by the Chi2 test) we used the random-effects model and MD is presented.

Three studies showed particular clinical heterogeneity. Leng 1998 took place over a longer time period and involved a greater number of participants than the other studies. In addition, it evaluated the effect of both omega-3 and omega-6 fatty acid supplementation. This made it difficult to attribute any particular effects to omega-3 fatty acids, since omega-6 fatty acids have also been described as having lipid-lowering and anti-hypertensive effects (Greene 1997). It used linoleic acid as a placebo, even though this is a precursor compound of GLA. The authors acknowledge in the paper that the linoleic acid given to the control group may have been (at least partly) metabolised into GLA. This study also had the highest number of withdrawals and dropouts. Carrero 2005 and Carrero 2006 also differed from the other studies in that the intervention preparation, as well as containing omega-3 fatty acids, contained oleic acid and certain vitamins (not included in the placebo preparation). In addition, the omega-3 fatty acid dose was very small. Participants in both intervention and control groups were also given antiplatelet therapy and a haemorrheologic agent. Due to the clinical heterogeneity of these studies, we excluded them from the meta-analysis. A secondary analysis involving only these three studies was considered but not performed as it would be impossible to attribute any observed effects to omega-3 fatty acids.

Quality of life

One study included quality of life (Conway 2005). Difference in quality of life using the self reported Short Form - 36 item (SF-36) questionnaire for individual participants between entry and completion at 16 weeks through t-test comparison showed no statistically significant improvement in any of the eight quality-of-life parameters for the intervention group (P value > 0.05). Numerical values were not given in the paper, neither were results for the control group.

Pain-free walking distance

Six studies reported PFWD (Carrero 2005; Carrero 2006; Conway 2005; Gans 1990; Leng 1998; Woodcock 1984).

Gans 1990 found that the mean PFWD increased (after four months) by 18% in the intervention group and by 41% in the control group (no P values quoted in paper). Woodcock 1984 found that after seven weeks of treatment, three participants in the intervention group noted an increase of 46 m in claudication distance, and one person no longer had rest pain. In the control group, one had an increase in claudication distance of 73 m and one person no longer had rest pain. Unfortunately, no data were provided to clarify and further analyse these findings.

Conway 2005 detected a mean improvement in PFWD in the intervention group of 2.87 m, compared to a mean deterioration of 17.3 m in the control group. Despite the trend, the difference was not statistically significant (P value > 0.05, paired t-test).

Overall, there was no statistically significant difference in post-intervention PFWD between intervention and control groups when the results from Gans 1990 and Conway 2005 were combined (MD 11.62 m, 95% CI -67.74 to 90.98). Significant statistical heterogeneity was detected (I2 = 83%), therefore we used a random-effects analysis model.

Of the remaining three studies, once again, only Carrero 2005 and Carrero 2006 detected any statistically significant effect. In Carrero 2005, PFWD increased by up to 3.5 times in the intervention group (by 279.33 m) compared to only 43.5 m in the control group (P value < 0.05). In Carrero 2006, PFWD increased by more than three times in the two treatment groups (S: enriched dairy product and SS: enriched dairy product + simvastatin). In the S group, PFWD increased by 251 m compared to the control group (C) where PFWD increased only by 28 m (P value < 0.05) and, in the SS group, PFWD increased by 317 m compared to 54 m in the control group that also received simvastatin (CS) (P value < 0.05). Leng 1998 found that the PFWD rose after 24 months by 81% in the intervention group and by 26% in the control group. Large SDs prevented these results from being statistically significant.

Maximum walking distance

Three studies included MWD (Conway 2005; Gans 1990; Leng 1998).

Maximum walking distance did not change significantly in any of the studies. Gans 1990 observed a non-significant increase of 18% in the intervention group versus 5% in the control group. In Conway 2005, a mean improvement of 7.64 m was found in the intervention group and a mean deterioration of 26.64 m found in the control group. Despite the magnitude of the difference it was not statistically significant.

Overall, the mean post-trial MWDs did not differ significantly between treatment and control groups (MD 16.99 m, 95% CI -72.14 to 106.11). Significant statistical heterogeneity was detected (I2 = 74%), therefore the random-effects analysis model is presented.

In the remaining study (Leng 1998), a non-significant increase in MWD of 27% was observed in the treatment group but no increase occurred in the control group. Large SDs were present and these results were found to be non-significant.

Ankle brachial pressure index

ABI measures the decrease in blood pressure in the arteries supplying the legs and can indicate blockages in these arteries; it is a ratio of the blood pressure in the arm compared to the legs, therefore a value near 1 indicates that pressures are approximately equivalent. Seven studies included ABI (Carrero 2005; Carrero 2006; Conway 2005; Gans 1990; Leng 1998; Schiano 2008; Woodcock 1984). Gans 1990 found that the mean ABI rose non-significantly by 4% after four months in the intervention group and by 8% in the control group (no P value quoted in the paper). Woodcock 1984 found that ABI did not change in either group after seven weeks, although no data were given. Conway 2005 found that the mean improvement in ABI in the intervention group was 0.03 compared with 0.06 in the control group (not statistically significant).

When results from Conway 2005 and Gans 1990 were combined, the mean post-trial ABI values did not differ significantly between treatment and control groups (MD -0.02, 95% CI -0.09 to 0.05). Fixed- and random-effects models gave identical results.

Of the studies excluded from the meta-analysis, Carrero 2005 and Carrero 2006 detected significant changes between pre- and post-intervention ABI. In Carrero 2005, ABI increased from 0.46 to 0.52 in the intervention group. This is reported as being statistically significant, although the P value was not quoted. There was no statistically significant change in the control group.

In Carrero 2006, ABI increased from 0.47 to 0.54 (P value < 0.05) in the treatment group receiving additional simvastatin. There was no significant change in any of the other groups.

Leng 1998 found that after 24 months, the mean ABI had risen non-significantly in both groups (by 6% in the intervention group and 9% in the control group, no P value quoted).

Schiano 2008 found no significant changes between pre- and post-intervention ABI. In the control group, the median ABI was 0.67 (interquartile range 0.55 to 0.77) at baseline and median 0.68 (interquartile range 0.55 to 0.78) at the end of the study. In the treatment group, ABI remained unmodified (median 0.66, interquartile range 0.53 to 0.69) after omega-3 fatty acid administration.

Angiographic findings

Angiographic findings were not reported in any study.

Total cholesterol

Eight studies reported total cholesterol (Carrero 2005; Carrero 2006; Gans 1990; Leng 1998; Mori 1992; Schiano 2008; Stricker 2008; Woodcock 1984). After four weeks, Mori 1992 found a significant increase of 4% in total cholesterol levels in participants receiving omega-3 fatty acid supplementation (P value < 0.01). Total cholesterol fell by 3% in the control group (no P value quoted in paper). In contrast, Stricker 2008 found a significant decrease in total cholesterol from 4.73 to 4.42 mmol/L (P value = 0.017) in the group receiving omega-3 fatty acid supplementation. Total cholesterol did not change significantly in the control group. No significant changes in total cholesterol levels were found in Gans 1990 or Woodcock 1984.

When the results from Gans 1990, Mori 1992, Stricker 2008 and Woodcock 1984 were pooled, mean post-trial total cholesterol levels did not differ significantly between intervention and control groups (random-effects model; MD 0.27 mmol/L; 95% CI -0.48 to 1.01). Significant statistical heterogeneity was detected (I2 = 64%), therefore the random-effects analysis model is presented.

Of the four studies excluded from the meta-analysis, Carrero 2005 detected a decrease in total cholesterol in the intervention group at nine and 12 months (from 5.43 to 5.15 mmol/L at 12 months, P value < 0.05). No change occurred in the control group. Carrero 2006 found a significant increase in total cholesterol in the control group that did not receive the enriched dairy product or simvastatin (C) (from 5.48 to 5.19 mmol/L at 12 months). There was a decreasing trend in the other groups but no statistically significant changes. No significant change in total cholesterol was detected in Leng 1998 or Schiano 2008.

High-density lipoprotein cholesterol

Eight studies evaluated HDL cholesterol (Carrero 2005; Carrero 2006; Gans 1990; Leng 1998; Mori 1992; Schiano 2008; Stricker 2008; Woodcock 1984). Following four months of treatment, Gans 1990 detected a statistically significant increase in HDL cholesterol in the intervention group (an increase of 17%, P value < 0.05). In the control group, mean HDL cholesterol rose non-significantly by 3% (no P value provided in paper). Mori 1992, Stricker 2008 and Woodcock 1984 found no significant changes in HDL cholesterol.

The meta-analysis (involving Gans 1990; Mori 1992; Stricker 2008; and Woodcock 1984) showed no difference between treatment and control group in mean HDL cholesterol levels at the end of the trial (0.00 mmol/L, 95% CI -0.16 to 0.15). Fixed-effect and random-effects analyses did not differ.

Of the remaining four studies, Carrero 2005 described a fall in HDL cholesterol in both intervention and control groups (from 1.27 mmol/L to 1.17 mmol/L, P value < 0.001 for intervention and from 1.30 mmol/L to 1.22 mmol/L, P value < 0.05 for control). No changes in HDL cholesterol levels were found in Carrero 2006, Leng 1998 or Schiano 2008.

When HDL subtypes were considered, Mori 1992 found that after four weeks of treatment, the fish oil group had a 17% increase in HDL-2C subfraction (P value < 0.05). Leng 1998 found that after 24 months of treatment, those in the fish oil group who had completed the study had a 7.8% increase in HDL-3C levels, as compared to the control group in which there was only a 5.8% increase (P value < 0.01).

Low-density lipoprotein cholesterol

Seven studies evaluated LDL cholesterol (Carrero 2005; Carrero 2006; Gans 1990; Leng 1998; Mori 1992; Schiano 2008; Stricker 2008). Mori 1992 found a 17% increase in LDL cholesterol level in the intervention group (P value < 0.01) but a 7% reduction in the control group (P value < 0.01). In contrast, Stricker 2008 found that LDL cholesterol levels decreased in the group receiving omega-3 fatty acids from 2.74 mmol/L to 2.42 mmol/L (P value = 0.007) after eight weeks. No statistically significant difference between pre- and post-treatment LDL cholesterol was found by Gans 1990.

When the results from Gans 1990, Mori 1992 and Stricker 2008 were combined, there was no significant difference found between mean post-trial LDL cholesterol levels in the treatment group when compared with the control group (MD 0.44 mmol/L, 95% CI -0.31 to 1.19). Significant statistical heterogeneity was detected (I2 = 75%), therefore the random-effects analysis model is presented.

Of the four remaining studies, only Carrero 2006 detected any significant changes in LDL cholesterol levels. The control group that did not receive the enriched dairy product or simvastatin (C) saw a rise in LDL cholesterol (from 2.71 to 2.98 mmol/L at 12 months), while there was a downward trend but no significant changes in the other groups. Carrero 2005, Leng 1998 and Schiano 2008 detected no significant changes in LDL cholesterol levels.

Very low-density lipoprotein cholesterol

Only two studies evaluated VLDL cholesterol (Gans 1990; Leng 1998). No significant change in VLDL levels was observed in the Gans 1990 study. Leng 1998 found that at six months, those who went on to complete the study had significantly higher VLDL cholesterol levels as compared to the control group (1.2 mmol/L (SD 0.97) compared with 1.06 mmol/L (SD 0.96), P value < 0.05).

Since Leng 1998 was excluded from the meta-analysis, the mean difference (MD) was only calculated for Gans 1990. This showed no difference between mean post-trial VLDL cholesterol levels in intervention and control groups (MD 0.08 mmol/L, 95% CI -0.47 to 0.63).

Triglycerides

Seven studies evaluated triglycerides (Carrero 2005; Carrero 2006; Gans 1990; Mori 1992; Schiano 2008; Stricker 2008; Woodcock 1984). Gans 1990 found that after four months, mean triglyceride levels fell by 32% (P value < 0.01) in the treatment group but remained unchanged in the control group. Mori 1992 found that after four weeks, there was a significant drop of 26% in triglyceride levels in the intervention group (P value not quoted in paper), whereas levels did not change significantly in the control group. Woodcock 1984 found that, after seven weeks, mean triglyceride levels fell by 27% in the intervention group but increased slightly in the control group (P value not quoted). Stricker 2008 found no significant changes in triglyceride levels. Unfortunately, this study could not be included in the meta-analysis as, like Schiano 2008, only median values (25th and 75th percentile) were expressed for this outcome.

Despite the three studies having positive results, when the results were pooled no significant difference was found between post-intervention treatment and control groups (MD -0.39 mmol/L, 95% CI -1.10 to 0.33). Statistical heterogeneity was present in this analysis (I2 = 72%), therefore the random-effects model MD is presented. The lack of a significant difference between intervention and control groups may have been due to the fact that the analysis did not take into account the higher baseline triglyceride levels in the intervention group compared to the control group in the Gans 1990 study.

Of the studies not included in the meta-analysis, only Schiano 2008 found any significant differences in triglyceride levels. There was a decrease in triglyceride levels from 171 mmol/L to 108 mmol/L (P value = 0.03) after three months in the treatment group. There was also a significant difference between the control and intervention group in relative changes from baseline (P value = 0.03). There was no significant change from baseline in the control group. Carrero 2005 and Carrero 2006 found no change in triglyceride levels in either the intervention or control group.

Blood pressure

Only two studies evaluated blood pressure (Gans 1990; Leng 1998). In Gans 1990, DBP fell significantly by 13% in the intervention group (P value < 0.05) and by 14% in the control group (P value < 0.05). SBP fell significantly in the control group (by 7%, P value < 0.05) but no significant change was observed in the intervention group. Mean post-trial DBP was lower in the intervention group (MD -2.00, 95% CI -3.74 to -0.26) but this was due to a higher mean baseline DBP in the control group. Mean SBP at the end of the trial showed no significant difference between treatment and control groups (MD 5.00, 95% CI -11.59 to 21.59).

Leng 1998 (which was excluded from the meta-analysis) detected a statistically significant reduction in SBP of 5% (P value < 0.05) in the intervention group versus an increase of 0.25% in the control group. There were no significant changes in DBP.

Mortality and vascular events

Only one study reported mortality and vascular events (Leng 1998). There was no significant difference between treatment and control groups in the incidence of all-cause mortality with three deaths (5%) in each group (Peto odds ratio (OR) 1.0, 95% CI 0.2 to 5.0), and CV mortality with two deaths (3%) in each group (Peto OR 1, 95% CI 0.1 to 7.3). The intervention group had a lower incidence of non-fatal coronary events with six (10%) events in the intervention group versus nine (15%) events in the control group (Peto OR 0.64, 95% CI 0.2 to 1.9), but this was not statistically significant (P value > 0.05). The control group had a lower incidence of non-fatal stroke/transient ischaemic attack with one (1.7%) event in the intervention group versus three (5%) events in the control group (Peto OR 0.64, 95% CI 0.2 to 1.9), and lower limb angioplasty/bypass surgery with one (1.7%) in the intervention group versus three (5%) in the control group (Peto OR 2.8, 95% CI 0.4 to 20.3), but neither difference was statistically significant (P value > 0.05). Similarly, progression to critical limb ischaemia/amputation did not differ significantly between groups with one (1.7%) in the intervention group versus zero in the control group (Peto OR 0.1, 95% CI 0 to 6.8, no P value quoted in paper).

Blood viscosity

This laboratory test can be performed at varying shear rates. Three studies reported blood viscosity (Gans 1990; Leng 1998; Woodcock 1984). In the Woodcock study, a statistically significant reduction was found at shear rates of both 230/s and 23/s in the intervention group (9.13 mPa/s (1.38) to 7.88 mPa/s (0.85), P value < 0.05 at shear rate of 23/s) (Woodcock 1984). No change was found in the control group. The Gans 1990 study found a statistically significant reduction in blood viscosity at low shear rates (0.945/s and 0.081/s) in the intervention group only (52.9 mPa/s (4.2) to 50.9 mPa/s (3.9), P value < 0.01 at shear rate of 0.945/s). The Leng study found no significant change in blood viscosity in either group (described as high shear rates but exact rate not given) (Leng 1998).

Since this haematological test was carried out at different shear rates, the results were not pooled for analysis. However, the fact that two studies detected statistically significant falls in blood viscosity makes it likely that a meta-analysis would have indicated a significant effect.

Plasma viscosity

Three studies evaluated plasma viscosity but none detected any significant change in plasma viscosity in intervention or control groups (Gans 1990; Leng 1998; Woodcock 1984). When the results from Woodcock 1984 and Gans 1990 were combined, there was no significant difference in post-intervention plasma viscosity between groups (MD 0.03 mPa/s, 95% CI -0.02 to 0.08). Fixed-effect and random-effects models did not differ.

Adverse effects of therapy

The Gans 1990 study stated that neither group experienced significant side effects from fatty acid supplementation. Likewise, Schiano 2008 stated that no adverse effects were seen. Conway 2005 reported that four people withdrew from the study because they were unable to tolerate therapy due to either nausea or a difficulty in swallowing the capsules (three people from the intervention group and one in the placebo group). Carrero 2005, Carrero 2006, Mori 1992, Stricker 2008 and Woodcock 1984 did not refer to side effects of therapy. In the Leng 1998 study, the incidence of gastrointestinal upset was significantly higher in the intervention group (50% versus 38%, P value < 0.05). However, it should be noted that the intervention group also received an omega-6 fatty acid.

Discussion

Summary of main results

The review included nine studies representing 425 participants, examining the effects of omega-3 fatty acid supplementation. On the basis of these studies, omega-3 fatty acid supplementation did not improve ABI or walking distance. It would also appear to have no effect on lipid levels (total cholesterol, LDL cholesterol, VLDL cholesterol, HDL cholesterol, triglycerides), blood pressure, mortality or vascular events. Blood viscosity was reduced with seven weeks to two years of supplementation. There appears to be a surrogate benefit but there is little evidence for improved clinical outcomes. Omega-3 fatty acid supplementation may also be associated with gastrointestinal side effects such as nausea, diarrhoea and flatulence.

Overall completeness and applicability of evidence

This Cochrane review involved 425 participants including both men and women. Certain patient groups (e.g. those with a diagnosis of diabetes) were excluded from the majority of studies; thus, findings from this review may not necessarily be applicable to these groups of patients.

Trials involving any form, dose and duration of omega-3 fatty acid supplementation were eligible. In the majority of trials, omega-3 fatty acids supplementation came in the form of capsules. No studies included in this review used dietary manipulation (e.g. increasing oily fish consumption) as a means of supplementation.

The small sample sizes used in three of the trials may also have resulted in a failure to detect significant clinical effects. Although two studies included a power calculation (Conway 2005; Leng 1998), Leng 1998 was only powered to detect a significant change in cholesterol levels and Conway 2005 was only powered to detect a 20% change in PFWD.

It is interesting to note that none of the studies used an inactive placebo. This may have masked certain effects of omega-3 fatty acids. Both oleic acid (a monounsaturated fatty acid) and linoleic acid (an omega-6 fatty acid) have been shown to have therapeutic effects on lipids and blood pressure (Sirtori 1992). A previous study involving omega-3 fatty acids (Belluzzi 1996) found that Miglyol (capric and caprillic acid) was a suitable inactive placebo.

The short time-span of the included studies may also have contributed to the lack of observed effects on clinical outcomes. With the exception of Carrero 2005, Carrero 2006 and Leng 1998 (which were all excluded from the meta-analysis), the trial periods ranged from four weeks to four months. Mori 1992 followed up the participants for six weeks after treatment but only four weeks of therapy was given. Woodcock 1984 administered only seven weeks of therapy and Schiano 2008 eight weeks. Stricker 2008 used three months of therapy, while Conway 2005 and Gans 1990 used four months. No long-term follow-up of participants was performed.

Insufficient doses of omega-3 fatty acids may have been used in at least two of the studies. While Leng 1998 administered treatment for a substantial period of time (two years), only a small daily dose of EPA was used (270 mg). Similarly, Carrero 2005 and Carrero 2006 used only 130 mg DHA and 200 mg EPA per day. Previously published studies have shown that a combined intake of 1.5 g to 4 g of EPA and DHA per day is required in order to observe anti-thrombotic, triglyceride-lowering, blood and plasma viscosity-lowering and endothelium-modifying effects. The three remaining studies (those included in the meta-analysis) do appear to have used therapeutic doses of omega-3 fatty acids.

As discussed earlier, two (Gans 1990; Woodcock 1984) of the three studies (Gans 1990; Leng 1998; Woodcock 1984) investigating blood viscosity found statistically significant reductions in blood viscosity in the intervention group. Blood viscosity can be measured at varying shear rates. Similar to blood pressure, blood viscosity can vary greatly between systole (high shear rate) and diastole (low shear rate). It is, therefore, important to test blood viscosity at both high and low shear rates; however, only Woodcock 1984 did so. Despite only testing at a low shear rate, Gans 1990 detected a significant change in blood viscosity between groups. Leng 1998, however, only tested blood viscosity at high shear rates and did not detect a significant difference between groups. If blood viscosity had also been measured at a low rate, it is possible that different results may have been obtained. High blood viscosity is associated with intermittent claudication (Lowe 1993), although it remains unclear whether lowering blood viscosity is clinically beneficial.

As only four studies discussed adverse effects of therapy, it is difficult to determine whether the dose or formulation of omega-3 fatty acid is related to adverse effects (Conway 2005; Gans 1990; Leng 1998; Schiano 2008). The fact that two of the studies reported that no participants reported side effects suggest that omega-3 fatty acids are generally well tolerated (Gans 1990; Schiano 2008). Leng 1998 reported increased gastrointestinal side effects despite a small daily dose of EPA suggesting that side effects are unlikely to be dose related. The intervention group in this study also included omega-6 fatty acids making it impossible to attribute the side effects specifically to omega-3 fatty acids. The severity and burden of these side effects suffered is unclear. Conway 2005 and Gans 1990 used very similar daily doses of omega-3 fatty acids. Three participants from the intervention group compared to one participant in the control group withdrew from Conway 2005 due to nausea or an inability to take the medication, whereas no adverse effects were reported in Gans 1990. One potential explanation for the difference between the two studies is that in Gans 1990, the dose was divided in to six different capsules whereas in Conway 2005 the full daily dose was contained in one capsule. It is, therefore, possible that the way the omega-3 fatty acids are administered may contribute to the occurrence of adverse effects. None of the studies that used dairy drinks to administer the omega-3 fatty acids discussed side effects of therapy so it is, therefore, impossible to compare with giving omega-3 fatty acids in capsules.

To monitor compliance among participants, Leng 1998, Mori 1992 and Schiano 2008 counted the number of returned capsules. Similarly, Stricker 2008 assessed compliance by the return of empty bottles. Leng 1998 also directly asked participants about compliance. In addition, five of the studies (Carrero 2005; Carrero 2006; Gans 1990; Mori 1992; Woodcock 1984) measured either plasma or platelet phospholipid fatty acid composition among participants and found increases in omega-3 fatty acid concentrations in the intervention groups. As well as confirming compliance, these findings demonstrated that the omega-3 supplements had been absorbed from the gastrointestinal tract. Leng 1998 found compliance to be high among those who completed the study, with 95% taking at least 1.12 g GLA and 180 mg EPA daily and 74% taking the full dose of 1.68 g GLA and 270 mg EPA. Similarly, Mori 1992 found compliance to be 99% for the fish oil group and 98% for the control group. Schiano 2008 found compliance with omega-3 fatty acid treatment to be satisfactory (> 94%). Carrero 2005 and Carrero 2006 describes compliance as "good" (assessed by regular telephone calls to participants and collection of empty dairy containers). Conway 2005, Gans 1990 and Woodcock 1984 did not report compliance figures.

The primary outcome of quality of life was only investigated by one of the studies included in this review. No studies provided data on angiographic findings (another of our primary outcomes). Clinical outcomes were generally less commonly reported than biochemical or haematological outcomes, with blood pressure and side effects only evaluated by a minority of studies (Gans 1990; Leng 1998; Schiano 2008), and only one study publishing data on mortality, morbidity and progression to amputation (Leng 1998). Economic consequences of therapy were not assessed in any of the studies and may have been useful.

Quality of the evidence

Nine studies were included in the review involving 425 participants from six countries. The quality assessment process for the systematic review of omega-3 for intermittent claudication found little evidence of major bias; however, there is some uncertainty around key elements in particular threats to the internal validity arising from possible poor blinding and allocation concealment. There was a small amount of evidence to suggest that the results were biased by random sequence generation, incomplete outcome data and selective reporting. More details can be found in the 'Risk of bias' tables.

In addition, three trials involved the administration of additional substances to participants making it impossible to attribute any effects of treatment specifically to omega-3 fatty acids and were, therefore, excluded from the meta-analysis (Carrero 2005; Carrero 2006; Leng 1998). This meant that the actual number of participants and studies included in the meta-analysis for each outcome was actually small with 205 participants involved in the remaining six studies.

There was also significant clinical and statistical heterogeneity discovered between the included studies suggesting that caution should be applied when interpreting the results of this review.

Potential biases in the review process

In order to minimise any potential bias in the review process every effort was made to conduct this review to the standards recommended in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

The studies were not homogeneous in terms of sample size, dose of therapy, duration of therapy and timing of physiological measurements. Therefore, the appropriateness of pooling individual results had to be assessed. Despite the three most clinically heterogeneous studies being excluded from the meta-analysis process, the Chi2 test statistic still showed significant statistical heterogeneity in the LDL, triglycerides and PFWD analyses. For this reason, caution should be applied when interpreting these results.

The analysis in this review was carried out on post-intervention values only, using the MD between treatment and control groups. This was done because none of the papers provided sufficient data to compare the MD between pre- and post-intervention values between groups. This meant that baseline values were not acknowledged by the analysis. Since the studies were all randomised, the baseline values should have been relatively equal. However, for certain outcomes this was not the case. For example, the baseline triglyceride levels were higher in the omega-3 group than in the control group in Gans 1990. Even though the difference between pre- and post-trial values was greatest for the intervention group, the control group had a lower mean triglyceride level at the end of the study. This led to the MD showing no significant difference between the two, despite the fact that every study detected a fall in triglyceride levels. A further example of this problem was found in the DBP analysis. The baseline DBP levels in Gans 1990 were higher in the control group compared to the treatment group. This led to the MD erroneously indicating that mean post-trial DBP levels were lower in the intervention group, despite a similar reduction in DBP being observed in both intervention and control groups.

Agreements and disagreements with other studies or reviews

The findings of this review disagree slightly to the findings of the original version of this review (Sommerfield 2004). The original review found that individuals with intermittent claudication may lead to a slight increase in total and LDL cholesterol levels. These findings were not replicated in this updated version of the review, which found no impact of omega-3 fatty acids on these measures when evidence of three new studies were included. The two versions of the review are, however, in agreement for all other outcomes.

Authors' conclusions

Implications for practice

On the basis of these studies, omega-3 fatty acid supplementation (in the form of eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA)) should not be recommended for routine use in intermittent claudication unless further evidence emerges in their favour. Omega-3 fatty acid supplementation may reduce blood viscosity but there is no clear evidence of clinical benefit and gastrointestinal side effects may occur.

The fact that all the studies used omega-3 fatty acid capsules or dairy drinks means that no conclusions can be drawn about the benefits of dietary manipulation of omega-3 fatty acid intake, such as increased consumption of oily fish.

Implications for research

The review has shown that more research is required in this area. In particular, an adequately powered study is needed to compare a therapeutic dose of an omega-3 fatty acid with an inactive placebo. Further research comparing different formulations and doses of omega-3 fatty acids may be useful. Important outcomes that require adequate evaluation include walking distance, ankle brachial pressure index (ABI), quality of life, vascular events (fatal and non-fatal) and adverse effects. Long-term follow-up of participants is essential.

From a public health perspective, it would be useful to consider clinical effectiveness as well as the economic aspects of omega-3 supplementation such as those presented by the UK National Institute for Health and Clinical Excellence (NICE) following myocardial infarction (Cooper 2007). Mortality and morbidity outcomes, such as progression to surgery, would be particularly relevant in such an analysis. A study examining the effects of dietary manipulation would also be useful, as people may prefer to consume omega-3 fatty acids in a more natural form, rather than via capsules.

Acknowledgements

We would like to thank Dr Tasmin Sommerfield (TS) for her contribution to earlier versions of this review. We would like to thank the Cochrane Peripheral Vascular Diseases Group and Mr Andrew Wawrzyniak for assisting with the data extraction. We would like to thank Dr Gillian Leng, Dr Trevor Mori, Dr Juan J Carrero and Mr K Conway for providing additional data for their studies.

Data and analyses

Download statistical data

Comparison 1. Omega-3 fatty acids (FAs) versus placebo
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Walking distance2 Mean Difference (IV, Random, 95% CI)Subtotals only
1.1 Pain-free walking distance264Mean Difference (IV, Random, 95% CI)11.62 [-67.74, 90.98]
1.2 Maximum walking distance281Mean Difference (IV, Random, 95% CI)16.99 [-72.14, 106.11]
2 Ankle brachial pressure index282Mean Difference (IV, Fixed, 95% CI)-0.02 [-0.09, 0.05]
2.1 Post-intervention282Mean Difference (IV, Fixed, 95% CI)-0.02 [-0.09, 0.05]
3 Cholesterol4 Mean Difference (IV, Random, 95% CI)Subtotals only
3.1 High-density lipoprotein cholesterol4122Mean Difference (IV, Random, 95% CI)-0.00 [-0.16, 0.15]
3.2 Low-density lipoprotein cholesterol3102Mean Difference (IV, Random, 95% CI)0.44 [-0.31, 1.19]
3.3 Total cholesterol4122Mean Difference (IV, Random, 95% CI)0.27 [-0.48, 1.01]
3.4 Very low-density lipoprotein cholesterol133Mean Difference (IV, Random, 95% CI)0.08 [-0.47, 0.63]
4 Triglycerides381Mean Difference (IV, Random, 95% CI)-0.39 [-1.10, 0.33]
4.1 Post-intervention381Mean Difference (IV, Random, 95% CI)-0.39 [-1.10, 0.33]
5 Blood pressure (BP)1 Mean Difference (IV, Fixed, 95% CI)Subtotals only
5.1 Systolic BP132Mean Difference (IV, Fixed, 95% CI)5.0 [-11.59, 21.59]
5.2 Diastolic BP132Mean Difference (IV, Fixed, 95% CI)-2.0 [-3.74, -0.26]
6 Plasma viscosity251Mean Difference (IV, Fixed, 95% CI)0.03 [-0.02, 0.08]
6.1 Post-intervention251Mean Difference (IV, Fixed, 95% CI)0.03 [-0.02, 0.08]
Analysis 1.1.

Comparison 1 Omega-3 fatty acids (FAs) versus placebo, Outcome 1 Walking distance.

Analysis 1.2.

Comparison 1 Omega-3 fatty acids (FAs) versus placebo, Outcome 2 Ankle brachial pressure index.

Analysis 1.3.

Comparison 1 Omega-3 fatty acids (FAs) versus placebo, Outcome 3 Cholesterol.

Analysis 1.4.

Comparison 1 Omega-3 fatty acids (FAs) versus placebo, Outcome 4 Triglycerides.

Analysis 1.5.

Comparison 1 Omega-3 fatty acids (FAs) versus placebo, Outcome 5 Blood pressure (BP).

Analysis 1.6.

Comparison 1 Omega-3 fatty acids (FAs) versus placebo, Outcome 6 Plasma viscosity.

Appendices

Appendix 1. CENTRAL search strategy

#1MeSH descriptor Arteriosclerosis, this term only893
#2MeSH descriptor Arteriolosclerosis, this term only0
#3MeSH descriptor Arteriosclerosis Obliterans, this term only70
#4MeSH descriptor Atherosclerosis, this term only373
#5MeSH descriptor Arterial Occlusive Diseases, this term only750
#6MeSH descriptor Intermittent Claudication, this term only701
#7MeSH descriptor Peripheral Vascular Diseases, this term only543
#8(atherosclero* or arteriosclero* or PVD or PAOD or PAD)16478
#9(arter* or vascular or vein* or veno* or peripher*) near (occlus* or steno* or obstuct* or lesio* or block*)7225
#10(peripheral near3 dis*)3134
#11(claudic* or hinken*)1417
#12(#1 OR #2 OR #3 OR #4 OR #5 OR #6 OR #7 OR #8 OR #9 OR #10 OR #11)23389
#13MeSH descriptor Fish Oils explode all trees1850
#14MeSH descriptor Fatty Acids, Omega-3 explode all trees1607
#15fatty near3 acid7654
#16omega near3 acid1529
#17*eicosapentanoic or docosahexanoic or docosapentanoic or alpha-linolenic352
#18*eicosapentaen* OR icosapentaenoic OR docosahexaeno*1376
#19fish near3 oil*1408
#20cod near3 oil*68
#21PUFA or EPA or E-EPA or DHA or DPA or ALA2166
#22MeSH descriptor Linseed Oil explode all trees46
#23linseed88
#24MeSH descriptor Flax, this term only54
#25flax* near3 oil66
#26Walnut near3 Oil12
#27Hemp near3 Oil4
#28Rapeseed near3 Oil89
#29Canola near3 Oil59
#30Soy* near3 Oil328
#31Alga* near3 Oil20
#32mackerel or sardine or trout or herring or salmon935
#33(#13 OR #14 OR #15 OR #16 OR #17 OR #18 OR #19 OR #20 OR #21 OR #22 OR #23 OR #24 OR #25 OR #26 OR #27 OR #28 OR #29 OR #30 OR #31 OR #32)10234
#34(#12 AND #33)661

What's new

Last assessed as up-to-date: 30 October 2012.

DateEventDescription
6 November 2012New citation required and conclusions have changedNew author (AC) joined review team. One author (TS) stepped down from author team. Review updated with three new included studies four excluded studies and one ongoing study. Risk of bias tables added. New sections added in line with review format changes. Change in conclusion for effect of omega-3 fatty acids: different findings from previous review versions for serum total cholesterol and LDL cholesterol levels.
6 November 2012New search has been performedSearches re-run. Review updated with three new included studies, four excluded studies and one ongoing study.

History

Protocol first published: Issue 4, 2002
Review first published: Issue 3, 2004

DateEventDescription
6 August 2008AmendedConverted to new review format.
21 August 2007New citation required and conclusions have changedSubstantive amendment.Two new included studies; adverse effects added to conclusions. Dr Jackie Price added as new co-author. Search strategy and dates updated. Copy edits performed and text updated to reflect new trials
12 August 2004AmendedSynopsis added, date of searches amended. Minor copy edits.

Contributions of authors

Tasmin Sommerfield and Andrew Campbell identified possible trials, considered them for inclusion, assessed trial quality, contacted study investigators if data were required that were not present in the published articles and extracted data. William Hiatt and Jackie Price considered trials for inclusion, assessed trial quality and checked data extraction. All authors contributed to the writing of the review.

Declarations of interest

WH has declared that he is in receipt of a current research grant from AstraZeneca to design and oversee a clinical trial with ticagrelor in PAD. This support does not directly compete with this review. 

Sources of support

Internal sources

  • No sources of support supplied

External sources

  • Chief Scientist Office, Scottish Government Health Directorates, The Scottish Government, UK.

    The PVD Group editorial base is supported by the Chief Scientist Office.

  • National Institute for Health Research (NIHR), UK.

    The PVD Group editorial base is supported by a programme grant from the NIHR.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Carrero 2005

MethodsStudy design: randomised controlled trial
Method of randomisation: table of random numbers
Concealment of allocation: not stated
Blinded: claims to be "blinded", unclear whether single or double
Power calculation: no
Number of participants randomised: 60
Number of participants analysed: 56
Number of exclusions post randomisation: 0
Number of withdrawals and reasons: 4 from control group (change of residence, depression, statin prescription)
Intention-to-treat analysis: no
Source of funding: supported in part by a Ph.D. educational grant from the University of Granada. Milk products, reagents and kits were supplied by Puleva Food S.L., Granda, Spain. One of the study authors is an employee of Puleva Biotech
ParticipantsCountry: Spain
Number of participants: 60 (30 in each group)
Age (mean ± SD): intervention group 62.4 ± 1.6 years, control group 65.6 ± 1.7 years
Inclusion criteria: all males diagnosed with PAD (ABI < 0.70) and presenting with intermittent claudication (Fontaine Stage IIb, claudication distance < 200 m)
Exclusion criteria: eligible for revascularisation surgery, endocrine or metabolic disturbances, cardiac history, resident outside of the Granada region of Spain
InterventionsTreatment: 500 mL/day of skimmed milk containing 200 mg EPA, 130 mg DHA plus oleic acid, folic acid and vitamins A, B6, D, E
Control: 500 mL/day semi-skimmed milk plus vitamins A and D
Duration: 12 months
Outcomes1. ABI
2. PFWD
3. Triglycerides
4. Cholesterol (total, LDL, HDL)
Notes-
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskA table of random numbers was used
Allocation concealment (selection bias)Unclear riskNo description of allocation concealment
Blinding of participants and personnel (performance bias)
All outcomes
Low riskFrom correspondence: "dairy drinks in white tetrapak containers, of the same appearance with the key A or B, performed by the supplier"
Blinding of outcome assessment (detection bias)
All outcomes
Low riskFrom correspondence: "The teams dealing with the analysis of samples only received numbered samples. The medical team collecting data did not have indication of the type of drink assigned. Only the team dealing with the production of the dairy drinks would know the meaning of keys A and B. The meaning of keys A or B was disclosed once all the data was obtained"
Incomplete outcome data (attrition bias)
All outcomes
High risk

Details provided of all missing outcome data; however, the withdrawals and exclusion led to an imbalance between the study groups

107 eligible candidates originally identified. 34 then excluded (11 living outside Granada, 12 previous MI, 6 statins prescription before time of inclusion, 5 refused to participate). A further 13 were excluded due to budget restrictions leaving 60 to be randomised

4 were excluded from the control group post randomisation (2 due to change of residence, 1 due to depression and 1 due to statin prescription). 0 were excluded post randomisation from the supplemented group. This meant that 30 were analysed from the supplemented group but only 26 were analysed from the controlled group

Selective reporting (reporting bias)Low riskAll prespecified outcomes were reported in the prespecified way
Other biasLow riskNo other obvious sources of bias

Carrero 2006

MethodsStudy design: randomised controlled trial
Method of randomisation: table of random numbers
Concealment of allocation: unclear
Blinded: at least single blinded "products packaged in Tetra Bricks to perform blind study"
Power calculation: no
Number of participants randomised: 40
Number of participants analysed: 40
Number of exclusions post randomisation: not stated (implied 0)
Number of withdrawals and reasons: 0
Intention-to-treat analysis: yes
Source of Funding: study supported by a Ph.D. educational grant from the University of Granada and by Puleva Biotech SA, Granada Spain
ParticipantsCountry: Spain
Settling/location: hospital
Number of centres: 1
Number of participants: 40 (10 in each group)
Age (mean): 66:66:65:63 years per group
Sex: men
Inclusion criteria: participants diagnosed with PAD (ABI < 0.7) and intermittent claudication (Fontaine stage IIb, claudication distance < 200 m)
Exclusion criteria: eligibility for revascularisation surgery; endocrine or metabolic disturbances (such as hypothyroidism or obesity, BMI > 30 kg/m2; history of cardiac episodes (such as angina pectoris) or previous acute MI; living outside of Granada
Interventions

Treatment: one treatment group received 500 mL/day of a fortified dairy product containing fish oil, oleic acid, folic acid, and other vitamins. The other treatment group consumed the same
fortified product plus 20 mg/day of simvastatin

Control: one control group received 500 mL/day of semi skimmed milk, while the other control group received semi skimmed milk plus simvastatin

Outcomes

1. PFWD

2. ABI

3. Cholesterol (total, HDL, LDL)

4. Triglycerides

NotesGroups very small and poorly matched for age at baseline
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskRandom table of numbers
Allocation concealment (selection bias)Unclear riskAllocation concealment not described
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskStudy states that "Products were packaged in white 500ml Tetra Bricks to perform a blind study." Unclear whether investigators were blinded or not
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskUnclear whether assessors were blinded or not. Detection bias is unlikely to have affected the objective outcomes (ABI, total cholesterol, HDL, LDL, triglycerides) but may have had an effect on PFWD, which is slightly more subjective
Incomplete outcome data (attrition bias)
All outcomes
Low riskNumber of exclusions post randomisation not stated but presumed none. All participants completed the study so no withdrawals
Selective reporting (reporting bias)Low riskPrespecified outcomes were reported in the prespecified way
Other biasUnclear risk

Claudication distance for baseline (when subjects were recruited) and T0 (the start of the study) were very different with no obvious explanation for this

Very small groups poorly matched for age at baseline

Conway 2005

Methods

Study design: randomised controlled trial

Method of randomisation: randomised by computer software

Concealment of allocation: sealed opaque envelopes

Blinded: double blind

Power calculation: yes

Number of participants randomised: 50

Number of participants analysed: 40

Number of exclusions post randomisation: 0

Number of withdrawals and reasons: 10 patients failed to complete the study (6 from treatment group and 4 from control group). There were 6 withdrawals (myocardial infarction, nausea or inability to take medication), 2 lost to follow-up and 2 could not perform treadmill test

Intention-to-treat analysis: yes

Source of funding: not stated

Participants

Country: UK

Number of participants: 50
Age (mean (range)): 66.1 years (49-82 years)

Sex: 35 women; 15 men
Inclusion criteria: participants with intermittent claudication who had been referred to a nurse-led exercise programme
Exclusion criteria: critical ischaemia, warfarin therapy, haemorrhagic disorders, DM, aspirin-sensitive asthma

Interventions

Treatment: 10 g fish oils per day (capsules containing 1.7 g EPA, 1.15 g DHA)

Control: capsules containing 10 g mixed oils
Duration: 16 weeks

Outcomes1. QoL
2. ABI
3. PFWD
4. MWD
Notes-
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskFrom correspondence: "Twenty five patients were randomised by computer software (Microsoft® Excel 2000 Number Generator) to each arm of the trial"
Allocation concealment (selection bias)Low riskFrom correspondence: "Following randomisation, allocations were placed in order, in sealed opaque envelopes in the pharmacy"
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskClaims to be "double-blinded" and in correspondence author states that: "Trial packs containing either medication or placebo were selected sequentially by the pharmacist in accordance with the randomisation code." However, no further information is provided and it is unclear whether there was a visible difference between the medication or placebo packs
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskUnclear whether assessors were blinded or not. Detection bias is unlikely to have had an effect on the QoL or ABI outcomes; however, may have had an impact on assessment of PFWD and MWD
Incomplete outcome data (attrition bias)
All outcomes
Low risk10 participants did not complete the study. 2 participants had an MI (1 from the treatment group and 1 from the control group), 4 withdrew due to nausea or an inability to swallow the capsules (3 from the intervention group, 1 from the control group), 2 were lost to follow-up and 2 were unable to complete treadmill retesting. Power calculation and intention-to-treat analysis was carried out
Selective reporting (reporting bias)Low riskThe prespecified outcomes in the study protocol were reported in the prespecified way
Other biasLow riskNo other obvious source of bias

Gans 1990

Methods

Study design: randomised controlled trial

Method of randomisation: not stated

Concealment of allocation: sealed envelopes

Blinded: double-blind

Power calculation: no

Number of participants randomised: 37

Number of participants analysed: 32

Number of exclusions post randomisation: 3

Number of withdrawals and reasons: 2 dropouts (spinal fracture, ischaemia in non-claudicated leg)

Intention-to-treat analysis: no

Source of funding: not stated

Participants

Country: The Netherlands

Setting: hospital

Number of participants: 32

Age (mean ± SD): fish oil group 65.5 ± 10.4; corn oil group 66.7 ± 5.3

Sex: men 22; women 10

Inclusion criteria: symptoms of intermittent claudication due to atherosclerotic disease (Fontaine IIa and IIb). Stable for 1 year according to ABI and claudication distance on standardised testing

Exclusion criteria: unstable angina or MI in past 3 months; any illness with rapid evolution; rest pain; gangrene; DBP > 100 mmHg; poorly controlled DM; (HbA1c > 12%); vasculitis, thromboangiitis obliterans; platelet count > 500,000x109/L or < 90,000x109/L; haematocrit > 55%; fish allergy; lipid-lowering or platelet active drugs

Interventions

Treatment: 6 capsules fish oil per day (daily dose of 1.8 g EPA + 1.2 g DHA)

Control: placebo 6 capsules corn oil per day (3 g linoleic acid per day)

Duration: 4 months

Outcomes1. BP
2. ABI
3. PFWD and MWD
4. Blood and plasma viscosity
5. Cholesterol (total, HDL, LDL)
6. Triglycerides
Notes-
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskStudy states that "Patients were randomly assigned"; however no further details provided
Allocation concealment (selection bias)Low riskAllocation carried out by "numbered sealed envelope system"
Blinding of participants and personnel (performance bias)
All outcomes
Unclear risk

Claims to be "double-blind" study but no information given regarding method of blinding

Poor blinding is unlikely to have had an effect for the majority of outcomes which were objective, i.e. BP, ABI, blood and plasma viscosity, cholesterol, triglycerides, but may have had an impact on the more subjective outcomes, i.e. PFWD and MWD 

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskUnclear whether assessors were blinded or not. Detection bias unlikely to be an issue for the objective outcomes listed but may potentially have had an effect on the more subjective outcomes
Incomplete outcome data (attrition bias)
All outcomes
Low riskData provided on all missing outcome data. 3 participants (2 from the intervention group and 1 from the control group) were excluded from the study post-randomisation due to concurrent medical conditions (2 were diabetic and 1 required eye surgery). 2 participants in the fish oil group later dropped out; 1 sustained a spinal fracture and 1 developed unstable ischaemia in his non-claudicated leg. 16/18 (89%) of the intervention group and 16/19 (84%) of the control group, therefore, completed the study
Selective reporting (reporting bias)Low riskPrespecified outcomes were reported in the prespecified way
Other biasLow riskNo other obvious sources of bias

Leng 1998

Methods

Study design: randomised controlled trial

Method of randomisation: blocks of 4 using minimisation method - taking into account age, sex, smoking and aspirin treatment

Concealment of allocation: code number allocated by telephone

Blinded: double-blind

Power calculation: yes

Number of participants randomised: 120

Number of participants analysed: 75

Number of exclusions post randomisation: 28

Number of withdrawals and reasons: 17 who "no longer wished to continue"

Intention-to-treat analysis: yes

Source of funding: not stated

Participants

Country: UK

Number of participants: 120; 60 in each group

Age (mean (SE)): treatment group 65 years (0.94); placebo group 66.48 years (0.90)

Sex: men 84; women 36

Inclusion criteria: intermittent claudication on the Edinburgh Claudication Questionnaire, ABI < 0.9 in at least 1 limb, stable > 6 months

Exclusion criteria: clinical evidence of critical ischaemia (rest pain, ulcer, gangrene); previous or impending arterial surgery or angioplasty (within 3 months); unstable angina or MI within previous 3 months; severe intercurrent illnesses; concurrent treatment with anticoagulants, other oils, lithium or phenothiazines; and pregnancy

Interventions

Treatment: 2 capsules evening primrose oil and fish oil bd for 2 weeks then 3 capsules bd for 2 years (280 mg GLA and 45 mg EPA per day)

Control: identical placebo capsules containing 500 mg sunflower oil (linoleic acid)

Outcomes1. Fatal and non-fatal CV events
2. Cholesterol (total, HDL and LDL)
3. Blood and plasma viscosity
4. BP
5. ABI
6. PFWD
Notes-
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskPerformed in blocks of 4 using a minimisation method taking into account age, sex, smoking habits and regular aspirin consumption
Allocation concealment (selection bias)Low riskCode number allocated by telephone
Blinding of participants and personnel (performance bias)
All outcomes
Low riskTreatment and placebo groups were given identical capsules
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskUnclear whether assessors were blinded or not. Unlikely to have affected most outcomes as they were objective measurements analysed in laboratories, however, may have affected PFWD, which entails a little more subjectivity
Incomplete outcome data (attrition bias)
All outcomes
Unclear risk

Most missing outcome data were documented (45 in total). 6 of the intervention group and 11 of the control group chose not to complete the study. In addition, 15 of the intervention group and 13 of the control group were withdrawn by the trial organisers due to either fatal or non-fatal adverse events

However, some additional missing data for some outcomes were not fully reported. For cholesterol and lipoprotein levels, only data from 37 of the 39 people that completed the trial in the treatment group were included in the results. For blood viscosity, data from 37 out of 39 people from the treatment group and 33 out of 36 from the placebo group were included in the results. For plasma viscosity, 37 out of 39 people from the treatment group and 35 out of 37 from the placebo group were included in the results. No reasons for this missing data were provided.

Questionnaires were sent to participants and their general practitioners 2 years after recruitment to identify CV events in those who withdrew from the trial

Selective reporting (reporting bias)Low riskAll prespecified outcomes were reported in the prespecified way
Other biasLow riskNo other obvious sources of bias

Mori 1992

Methods

Study design: randomised controlled trial

Method of randomisation: table of random numbers. Controls matched for age, BMI, creatinine, glucose, cholesterol and triglycerides

Concealment of allocation: not stated

Blinded: double-blind

Power calculation: no

Number of participants randomised: 32

Number of participants analysed: 29

Number of exclusions post randomisation: 0

Number of withdrawals and reasons: 3 withdrawals prior to commencement of the study "for personal reasons"

Intention-to-treat analysis: no

Source of funding: supported by the National Heart Foundation of Australia

Participants

Country: Australia

Number of participants: 32

Age: range 47-71 years; fish oil group 61.9 ± 1.2 ; olive oil group 61.9 ± 1.7 (Mean +/- SEM)

Sex: men

Inclusion criteria: symptomatic and angiographically demonstrated PAD

Exclusion criteria: renal impairment, generalised inflammatory disorders

Interventions

Treatment: 5 capsules of fish oil 3 times per day with meals (2.8 g EPA, 1.8 g DHA per day)

Control: 5 capsules of olive oil 3 times per day (11.2 g oleic acid per day)

Duration: 4 weeks

Outcomes1. Cholesterol
2. Triglycerides
3. Platelet count
NotesNo data on severity of PAD. Some participants underwent peripheral vascular surgery but at least 3 months before start of study. 1 IDDM in intervention group, 1 NIDDM in control group
Controls matched for age and BMI
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskTable of random numbers
Allocation concealment (selection bias)Unclear riskMethod of allocation concealment not stated
Blinding of participants and personnel (performance bias)
All outcomes
Low riskClaims to be a "double-blind" study. Although there are no further details about the blinding process, poor blinding is unlikely to have impacted on the outcomes
Blinding of outcome assessment (detection bias)
All outcomes
Low riskUnclear whether assessors were blinded or not however all outcomes are objective measurements so unlikely to have had an impact
Incomplete outcome data (attrition bias)
All outcomes
Low risk3 people withdrew post randomisation but before commencement of the study (1 from fish oil group; 2 from olive oil group). All 29 participants who commenced the study completed the 10-week trial
Selective reporting (reporting bias)Low riskAll prespecified outcomes were reported in the prespecified way
Other biasUnclear riskNo data on severity of PAD

Schiano 2008

Methods

Study design: randomised controlled trial

Method of randomisation: not stated

Concealment of allocation: not stated

Blinded: single-blind (blind observer)

Power calculation: no

Number of participants randomised: 32

Number of participants analysed: 32

Number of exclusions post randomisation: not stated (implied 0)

Number of withdrawals and reasons: 0

Intention-to-treat analysis: yes

Source of funding: not stated

Participants

Country: Italy

Setting/location: vascular laboratory

Number of centres: 1

Number of participants: 32 (16 in each group)

Age: (median) 66 years for both groups

Sex: 29 men, 3 women

Inclusion criteria: participants referred to vascular laboratory with history of intermittent claudication and resting ABI < 0.9 plus presence of ≥ 1 stenoses > 50% in ≥ 1 artery on duplex scanning

Exclusion criteria: rest pain; trophic lesions in affected limb; acute coronary syndrome or CV event in previous 6 months; significant renal, hepatic or inflammatory disease; history of hypersensitivity; other contraindications to n-3 PUFAs

Interventions

Treatment: n-3 PUFAs 1 g bd* + "usual" (pre-enrolment) therapy

Control/placebo: "usual" pre-enrolment therapy

Duration: 3 months

*capsules containing at least 85% EPA and DHA as ethyl esters in mean ratio of EPA/DHA of 0.9:1.5

Outcomes

1. ABI

2. Cholesterol (total, HDL, LDL)

3. Triglycerides

Notes-
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNo description of randomisation process
Allocation concealment (selection bias)Unclear riskNo description of allocation concealment
Blinding of participants and personnel (performance bias)
All outcomes
Low riskPersonnel assessing outcome stated to be "blinded". Participants were not blinded but outcomes unlikely to be affected by this
Blinding of outcome assessment (detection bias)
All outcomes
Low riskPersonnel assessing outcome stated to be "blinded"
Incomplete outcome data (attrition bias)
All outcomes
Low risk32 consecutive participants referred to a vascular laboratory were randomised. There were no exclusions post randomisation or withdrawals during follow-up
Selective reporting (reporting bias)Low riskAll prespecified outcomes are reported in the prespecified manner
Other biasLow riskNo other obvious sources of bias

Stricker 2008

Methods

Study design: randomised controlled trial

Method of randomisation: randomised in blocks of 4 (using www.randomisation.com)

Concealment of allocation: unclear

Blinded: double-blind

Power calculation: no

Number of participants randomised: 40

Number of participants analysed: 40

Number of exclusions post randomisation: not stated (implied 0)

Number of withdrawals and reasons: not stated (implied 0)

Intention-to-treat analysis: yes

Source of funding: Fondo Balli, Locarno, Switzerland

Participants

Country: Switzerland

Setting/location: hospital

Number of centres: 1

Number of participants: 40 (20 in each group)

Age: 50+ years (mean 66.8 intervention; 63.7 control)

Sex: 27 men, 13 women (canola oil group: 14 men, 6 women) (sunflower oil group: 13 men, 7 women)

Inclusion criteria: chronic PAD defined as ABI < 0.9 plus > 50% stenosis or occlusion in a leg artery (on duplex or angiogram)

Exclusion criteria: acute, intermittent illness; thromboangiitis obliterans; renal insufficiency (creatinine > 130 µmol/L); acute stroke or MI in previous 2 months; current oral anticoagulation medicine; liver cirrhosis; presence of a malignant tumour

Interventions

Treatment: 2 tablespoons (35 mL) a day of canola oil (= 2.24 g of alpha-linolenic acid, C18 3n-3 omega-3 fatty PUFA)

Control/placebo: 2 tablespoons (35 mL) a day of sunflower oil (= 16.24 g of linolenic acid, C18 2n-6 omega-6 PUFA)

Duration: 8 weeks

Outcomes

1. Cholesterol (total, HDL, LDL)

2. Triglycerides

Notes-
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskRandomised in blocks of 4 using a randomisation website (www.randomization.com)
Allocation concealment (selection bias)Unclear riskNo description of allocation concealment
Blinding of participants and personnel (performance bias)
All outcomes
Low risk"Patients and doctors blinded". The bottles containing the treatment and placebo were packaged similarly and rendered visually indistinguishable
Blinding of outcome assessment (detection bias)
All outcomes
Low riskAny lack of blinding of assessors is unlikely to have affected the outcomes
Incomplete outcome data (attrition bias)
All outcomes
Unclear risk40 participants were randomised. 40 were analysed so it is implied that there were no exclusions or withdrawals; however, there is no explicit statement to confirm this
Selective reporting (reporting bias)Low riskAll prespecified outcomes are reported in the prespecified way
Other biasLow riskNo other obvious source of bias

Woodcock 1984

  1. a

    ABI: ankle brachial pressure index;
    bd: twice daily;
    BMI: body mass index;
    BP: blood pressure;
    CV: cardiovascular;
    DBP: diastolic blood pressure;
    DHA: docosahexaenoic acid;
    DM: diabetes mellitus;
    EPA: eicosapentaenoic acid;
    GLA: gamma-linolenic acid;
    HbA1c: haemoglobin A1c;
    HDL: high-density lipoprotein;
    IDDM: insulin-dependent diabetes mellitus;
    LDL: low-density lipoprotein;
    m: metres;
    MI: myocardial infarction;
    MWD: maximal walking distance;
    NIDDM: non-insulin-dependent diabetes mellitus;
    PAD: peripheral arterial disease;
    PFWD: pain-free walking distance;
    PUFA: polyunsaturated fatty acid;
    QoL: quality of life;
    SD: standard deviation;
    SE: standard error;
    USS: ultrasound scan.

Methods

Study design: randomised controlled trial

Method of randomisation: random table of numbers

Concealment of allocation: not stated

Blinded: double-blind

Power calculation: no

Number of participants randomised: 19

Number of participants analysed: 19

Number of exclusions post randomisation: none stated

Number of withdrawals and reasons: none stated

Intention-to-treat analysis: no

Source of funding: supported by research grants from the British Heart Foundation and the special trustees for the Former United Sheffield Hospitals

Participants

Country: UK

Number of participants: 19 (10 in intervention; 9 in control)

Age: 56-75 years

Sex: men 15; women 4. Fish oil group: men 7; women 3; corn/olive oil group: men 8; women 1

Inclusion criteria: intermittent claudication. Presence and site of disease confirmed by Doppler USS. Claudication distance 14-1600 m, ABI 0.34-0.78. All participants attended vascular clinic and were ineligible for surgery

Exclusion criteria: none described

Interventions

Treatment: 5 capsules of fish oil bd (= 1.8 g EPA per day)

Control: placebo 5 (identical) capsules of corn oil (linoleic acid) and olive oil (oleic acid) bd

Duration: 7 weeks

Outcomes1. Walking distance
2. Doppler USS
3. ABI
4. Platelets
5. Blood viscosity
6. Cholesterol
7. Triglycerides
NotesNo participants with DM
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskTable of random numbers
Allocation concealment (selection bias)Unclear riskMethod of allocation concealment not stated
Blinding of participants and personnel (performance bias)
All outcomes
Low riskTreatment and control both provided in identical capsules
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskUnclear whether the outcome assessors were blinded or not. This is unlikely to have affected the majority of outcomes, which were subjective but may have had a potential effect on the more subjective outcomes such as PFWD
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNumber of participants randomised not stated. No description of withdrawals/losses to follow-up
Selective reporting (reporting bias)High riskAll prespecified outcomes are mentioned in the results section; however, for the majority of outcomes no data were provided to back up statements made in the text
Other biasLow riskNo other obvious sources of bias

Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion
Berrettini 1996Only 5 people in intervention group and 2 in control group with peripheral arterial disease. Data requested but not provided for this subset. Therefore, no means of identifying which participants had peripheral arterial disease
Ishikawa 2010Unclear whether subjects had intermittent claudication
MacKay 2012No relevant outcomes
Moller 1998Unclear whether subjects were randomly assigned to groups
Olsson 1984Deemed after translation to be a controlled clinical trial and not a randomised controlled trial
Ramirez-Tortosa 1999Non-randomised, non-blinded cross-over study

Characteristics of ongoing studies [ordered by study ID]

Leyva 2011

Trial name or titleFLAX-PAD Randomized Controlled Trial: The Effect of Dietary Flaxseed on Improving Symptoms of Cardiovascular Disease in Patients with Peripheral Artery Disease
MethodsRandomised prospective, double-blind controlled trial
Participants110 people over 40 years old with peripheral arterial disease
Interventions30 g of milled flaxseed (or placebo) per day
OutcomesPrimary endpoints are incidence of myocardial infarction and stroke
Secondary measures include: requirement for surgical interventions, exercise and cardiopulmonary performance, cardiac arrhythmias, serum lipid profile, arterial sufficiency, blood pressure, inflammatory profile, platelet function, changes in drug dosage levels, and nutrigenomic and biomarker profiles in the blood
Starting dateOctober 2008
Contact informationDelfin Rodriguez Levya, Cardiovascular Research Division, V.I. Lenin University Hospital, Holguin, 80100, Cuba
Notes-

Ancillary