Flavonoids for treating venous leg ulcers

  • Review
  • Intervention

Authors


Abstract

Background

Venous leg ulcers are a major health burden: annually, in the UK alone, they contribute an estimated cost to the NHS of GBP 400 million. Flavonoids are a diverse group of naturally-occurring venotonic compounds that address certain microcirculatory parameters involved in venous leg ulcer pathophysiology.

Objectives

To evaluate the clinical effects of flavonoids on the healing of venous leg ulcers.

Search methods

In February 2013 we searched the Cochrane Wounds Group Specialised Register; The Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library 2013, Issue 1); Ovid MEDLINE; Ovid MEDLINE (In-Process & Other Non-Indexed Citations); Ovid EMBASE; and EBSCO CINAHL. No date or language restrictions were applied. We checked reference lists of included trials, and contacted pharmaceutical companies.

Selection criteria

Randomised controlled trials (RCTs) that investigated the efficacy of any flavonoid-containing compound on venous leg ulcer healing in adults.

Data collection and analysis

Two review authors independently assessed trials for the review and disagreements were referred to a third author. All rejected articles were double-checked by a third author. Assessment of risk of bias and data extraction were performed independently by two authors, discrepancies were resolved by referring to the third author.

Main results

Of the nine studies (1075 participants): five investigated Micronised Purified Flavonoid Fraction (MPFF), and four investigated hydroxyethylrutosides (HR).

Meta-analysis involving 723 participants from five trials - four of which were characterised by poor reporting - showed more venous leg ulcers were healed in the MPFF groups than in the control groups (RR 1.36; 95% CI 1.07 to 1.74). However, the most rigorously conducted trial, which was at low risk of bias, did not show any additional benefit of MPFF (RR 0.94; 95% CI 0.73 to 1.22). Since this trial was unpublished, the possibility of publication bias in trials involving flavonoids must be acknowledged. Overall, the quality of reporting of trials on HR was also poor. Pooling three trials, all at unclear risk of bias, involving 279 participants showed a statistically significant effect in favour of HR with respect to number of ulcers healed (RR 1.70; 95% CI 1.24 to 2.34).

Authors' conclusions

Although the overall estimate of the number of healed ulcers appeared to show a significant effect in favour of flavonoids (both MPFF and HR), this result needs to be interpreted cautiously, as most of these trials were poorly reported, and so had an unclear risk of bias for randomisation, allocation concealment, blinding and methods for addressing incomplete outcome data. There was also a possibility of publication bias.

Résumé scientifique

Les flavonoïdes pour le traitement des ulcères de jambe veineux

Contexte

Les ulcères de jambe veineux sont un problème majeur de santé : chaque année, dans le seul Royaume-Uni, ils coûtent environ 400 millions de GBP au NHS. Les flavonoïdes sont un groupe diversifié de composés veinotoniques naturels qui influencent certains paramètres de la microcirculation veineuse impliqués dans la physiopathologie des ulcères de jambe.

Objectifs

Évaluer les effets cliniques des flavonoïdes sur la cicatrisation des ulcères de jambe veineux.

Stratégie de recherche documentaire

En février 2013 nous avons effectué des recherches dans le registre spécialisé du groupe Cochrane sur les plaies et contusions, le registre Cochrane des essais contrôlés (CENTRAL) (The Cochrane Library 2013, numéro 1), Ovid MEDLINE, Ovid MEDLINE (In-Process & Other Non-Indexed Citations), Ovid EMBASE et EBSCO CINAHL. Aucune restriction n'a été appliquée sur la date ou la langue. Nous avons effectué des recherches dans les bibliographies des essais inclus et contacté des sociétés pharmaceutiques.

Critères de sélection

Des essais contrôlés randomisés (ECR) ayant examiné l'efficacité d'un composé contenant des flavonoïdes sur la cicatrisation des ulcères de jambe veineux chez l'adulte.

Recueil et analyse des données

Deux auteurs de la revue ont indépendamment évalué les essais pour la revue et les désaccords ont été renvoyés à un troisième auteur. Tous les articles rejetés ont été revérifiés par un troisième auteur. L'évaluation du risque de biais et l'extraction des données ont été réalisées indépendamment par deux auteurs, les divergences ont été résolues par renvoi au troisième auteur.

Résultats principaux

Sur les neuf études (soit 1075 participants) : cinq avaient examiné la Fraction Flavonoïde Purifiée Micronisée (FFPM) et quatre les hydroxyéthylrutosides (HR).

La méta-analyse portant sur 723 participants à cinq essais - dont quatre étaient caractérisés par un mauvais compte-rendu - a montré que plus d'ulcères de jambe veineux avaient cicatrisé dans les groupes à FFPM que dans les groupes témoins (RR 1,36 ; IC 95% 1,07 à 1,74). Toutefois, l'essai le plus rigoureusement mené, qui présentait un faible risque de biais, n'avait pas mis en évidence de bénéfice supplémentaire de la FFPM (RR 0,94 ; IC 95% 0,73 à 1,22). Comme cet essai n'avait pas été publié, il faut admettre la possibilité d'un biais de publication dans les essais impliquant des flavonoïdes. Dans l'ensemble, la qualité de compte-rendu des essais sur les HR était aussi médiocre. Le regroupement de trois essais, tous à risque incertain de biais, impliquant 279 participants a montré un effet statistiquement significatif en faveur des HR quant au nombre d'ulcères cicatrisés (RR 1,70 ; IC 95% 1,24 à 2,34).

Conclusions des auteurs

Bien que l'estimation globale du nombre d'ulcères cicatrisés semblait montrer un effet significatif en faveur des flavonoïdes (tant FFPM que HR), ce résultat doit être interprété avec prudence car ces essais avaient pour la plupart des comptes-rendus médiocres et présentaient donc un risque incertain de biais quant à la randomisation, l'assignation secrète, la mise en aveugle et les méthodes pour gérer les données de résultat incomplètes. Il y avait aussi une possibilité de biais de publication.

アブストラクト

静脈性下腿潰瘍を治療するためのフラボノイド

背景

静脈性下腿潰瘍は、大きな健康上の負担となっている。英国だけでも、英国の国営医療制度(NHS)の推計400,000,000英国ポンド(GBP)が使われている。フラボノイドは、静脈性下腿潰瘍の病態生理に関わる特定の微小循環パラメーターを扱う多様なな天然の静脈強壮化合物のグループである。

目的

静脈性下腿潰瘍の治癒に対するフラボノイドの臨床効果を評価すること。

検索戦略

2013年2月、Cochrane Wounds Group Specialised Register、The Cochrane Central Register of Controlled Trials (CENTRAL) (「コクラン・ライブラリ」2013年第1号)、Ovid MEDLINE、Ovid MEDLINE (In-Process & Other Non-Indexed Citations)、Ovid EMBASEおよびEBSCO CINAHLを検索した。 日付や言語の制限は設けなかった。選択した試験の参考文献一覧を確認し、製薬会社に連絡を取った。

選択基準

成人の静脈性下腿潰瘍治癒に対するフラボノイド含有化合物の有効性を調べるランダム化比較試験(RCT)。

データ収集と分析

2名のレビュー著者が独立してレビュー用試験を評価し、意見の相違点は3人目の著者に問い合わせた。レビュー対象から外した論文は、すべて3人目の著者によって二重チェックされた。バイアスのリスクの評価およびデータ抽出は2名の著者が独立して実施したが、相違点は3人目の著者に問い合わせて解決した。

主な結果

9件の研究(1075例の参加者)のうち、5件の研究ではマイクロピュア・フラボノイド[Micronised Purified Flavonoid Fraction (MPFF)]、4件の研究ではヒドロキシエチルルトシド(HR)が調べられていた。

723例の参加者を対象とした5件の試験からメタアナリシスを実施したが、5件のうち4件の試験は質が悪かった。このメタアナリシスの結果、コントロール群と比較してMPFF群の方が静脈性下腿潰瘍の治癒率が高いことが示された(RR 1.36、95%CI 1.07~1.74)。しかし、バイアスのリスクが低く最も厳密に実施された試験では、MPFF群にさらなる利益は認められなかった(RR 0.94、95%CI 0.73~1.22)。この試験については未発表なので、フラボノイドに関する出版バイアスの可能性を認識しておかなければならない。HRに関する試験の報告の質も、総体的に低かった。バイアスのリスクがすべて不明な3件の試験を集積すると(279例の参加者を対象)、治癒した潰瘍の数についてHRを支持する統計学的に有意な効果が認められた(RR 1.70、95%CI 1.24~2.34)。

著者の結論

治癒した潰瘍の数の総体的推定は、フラボノイド(MPFFとHRの両者)を支持する有意な効果を示しているように思える。しかし、この結果は慎重に解釈する必要がある。これらの試験のほとんどの報告は質が悪く、ランダム化、割り付けの隠蔵化、盲検化および不完全なアウトカム・データの処理方法のバイアスのリスクが不明であった。出版バイアスの可能性もあった。

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/)[2016.1.1]
《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

Flavonoids for treating venous leg ulcers

Venous leg ulcers (which appear as open sores) can be caused by a blockage or breakdown in the veins of the legs. Compression of the leg, using bandages or hosiery (stockings), is known to help heal venous ulcers. Flavonoids, which are commonly used as food supplements, occur in a variety of plant-based foods and beverages, including cocoas, chocolates, teas and red wines. They are sometimes used to try to promote ulcer healing in the leg. This review of clinical trials concluded that there is some evidence to show that flavonoids can help heal venous leg ulcers, however, many trials were not reported well, and we could not know for certain whether the apparently beneficial effects were real or not. This meant that we could not draw firm conclusions, or recommend routine use of flavonoids for people with leg ulcers. Larger and better conducted trials are needed to assess the true clinical effect of flavonoids.

Résumé simplifié

Les flavonoïdes pour le traitement des ulcères de jambe veineux

Les ulcères de jambe veineux (qui ont l'apparence de plaies ouvertes) peuvent être provoqués par une obstruction ou une perforation des veines des jambes. La compression de la jambe, au moyen de bandages ou de compressions élastiques (bas), est réputée pour aider à la cicatrisation des ulcères. Les flavonoïdes, qui sont couramment utilisés comme compléments alimentaires, se rencontrent dans toute une variété d'aliments et de boissons à base de plantes, comme le cacao, le chocolat, le thé et le vin rouge. Ils sont parfois utilisés pour tenter de favoriser la cicatrisation des ulcères de jambe. Cette revue d'essais cliniques a conclu qu'il existe des preuves que les flavonoïdes peuvent aider à guérir les ulcères de jambe veineux ; toutefois, le compte-rendu de nombreux essais n'avait pas été bien fait, et nous ne pouvions pas savoir avec certitude si les effets bénéfiques apparents étaient réels ou non. Cela signifiait que nous n'avons pas pu tirer de conclusions définitives, ni recommander une utilisation systématique des flavonoïdes pour les personnes souffrant d'ulcères de jambe. Des essais de plus grande taille et mieux menés seront nécessaires pour évaluer le véritable effet clinique des flavonoïdes.

Notes de traduction

Traduit par: French Cochrane Centre 3rd June, 2013
Traduction financée par: Pour la France : Minist�re de la Sant�. Pour le Canada : Instituts de recherche en sant� du Canada, minist�re de la Sant� du Qu�bec, Fonds de recherche de Qu�bec-Sant� et Institut national d'excellence en sant� et en services sociaux.

平易な要約

静脈性下腿潰瘍を治療するためのフラボノイド

静脈性下腿潰瘍(開放創として出現)は、下肢の静脈の閉塞または機能不全が原因となる。包帯や靴下(ストッキング)を用いた下肢の圧迫は、静脈性潰瘍の治癒を促進することで知られている。フラボノイドは、一般に補助食品として使用される。ココア、チョコレート、お茶、赤ワインなどの植物をベースにして作られた食品や飲料に含まれる。潰瘍治癒の促進目的で使用されることもある。臨床試験の本レビューでは、フラボノイドが静脈性下腿潰瘍の治癒を促進できることを示すエビデンスが一部あると結論づけた。しかし、多くの試験の報告の質は低く、有益と思われる効果が本当に存在するか否かは確信できなかった。つまり、確固たる結論を出すことはできず、下腿潰瘍のある人にフラボノイドをルーチンで使用することを推奨できない。フラボノイドの真の臨床効果を評価するには、より大規模で質の良い試験を実施する必要がある。

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/)[2016.1.1]
《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

Venous leg ulcers - prevalence and burden

Leg ulceration is defined as the loss of skin below the knee on the leg or foot which takes more than six weeks to heal (Dale 1995; NHS CRD 1997). Many factors can be associated with leg ulcer development. The four most common causes of leg ulceration are venous insufficiency (venous disease), arterial insufficiency, nerve damage (often due to diabetes mellitus), and prolonged pressure and ischaemia (low oxygen levels) (Araujo 2003). Several studies found that about 60 to 80% of chronic leg ulcers were venous in origin (Baker 1991; Baker 1992; Bello 1998). In the UK, data from prevalence surveys suggest a much broader range of uncertainty around the prevalence of uncomplicated venous leg ulceration, i.e. between 51% to 85% of all leg ulcers (Cornwall 1986 Callam 1987; Salaman 1995; Graham 2003), while recent pragmatic audits suggest that the prevalence of  uncomplicated venous leg ulcers is much lower at between 38% to 40% (Srinivasaiah 2007; Vowden 2009). 

An estimated 5% to 8% of the world's population is afflicted with venous disease (Trent 2005). While venous disease affects a large proportion of the population, however, only a small percentage proceed to the most severe form of venous disease, which is ulceration. Several population-based prevalence studies reported prevalence rates of open leg ulcers that ranged from 0.12% to 1.1% of the population (Baker 1994; Nelzen 1996; Walker 2002). Within the UK population, the estimated prevalence of leg ulcers is 1.5 to 3 per 1000 (NHS CRD 1997). Prevalence, however, increases with age, reaching 20 per 1000 in people over 80 years old (Margolis 2002).

Venous leg ulcers are widely considered to be a major health problem, and social and economic burden (Simka 2003). The estimated cost of venous ulceration to the UK's National Health Service (NHS) is GBP 400 million a year (Simon 2004; Konstantinos 2005), and in the UK, France and Germany, the cost of venous ulcers is estimated to be 1.5% to 2.0% of the total annual healthcare budget (Trent 2005). As the proportion of elderly citizens in western European countries continues to rise, these figures can be expected to increase.

Venous leg ulcers - pathological mechanism of development

The precise pathological mechanism of development of venous ulcers is contentious (Valencia 2001), although in most cases the root cause can be traced back to the inefficient activity of the calf muscle 'pump' (Orsted 2001), and malfunctioning valves in the veins of the lower leg (Dale 1995). These lead to venous hypertension and, subsequently, chronic venous insufficiency (Dale 1995), the most severe stage of which is marked by the venous leg ulcer (Glinski 1999). Indeed, venous ulceration has been allocated to the most advanced categories (C5 and C6) of the CEAP (Clinical-Etiology-Anatomy-Pathophysiology) classification system (Lyseng 2003).

Venous disease is characterised by ambulatory venous hypertension (abnormally sustained elevation of the venous pressure upon walking) (Araujo 2003; Simon 2004). Normally, blood in the leg veins is propelled to the heart by the activity of the calf muscle pump. The pump is activated by walking and, when operating efficiently, the venous pressure should drop substantially upon ambulation (Dale 1995).

Ambulatory venous hypertension occurs when a vital component of the calf muscle pump - the valves in the veins of the lower leg - malfunction; this is otherwise known as venous valvular incompetence (Howlader 2004). Back flow (reflux) of blood from the deep veins in the leg occurs, resulting in increased hydrostatic pressure in the superficial veins, which is then transmitted to the capillaries. The increase in intra-capillary pressure is known as chronic venous insufficiency (CVI) (Dale 1995). In the US, approximately one million of the seven million people with CVI develop venous leg ulcers (Araujo 2003).

Venous hypertension causes the single-celled capillary walls to become stretched, resulting in an increase in the permeability of the venous wall. This in turn provokes the leakage of fluid, containing plasma protein deposits and red blood cells, into the tissues and skin. This induces pericapillary oedema, overloading of the lymphatic network, and inhibits perfusion of oxygen and nutrients to skin cells. Such 'microcirculatory dysfunction' leads to capillary damage, tissue destruction and ultimately, venous ulceration (Dale 1995; Glinski 1999; Konstantinos 2005).

Struckmann 1999 has also suggested that capillary blocking by trapped white blood cells, which initiates an inflammatory response, is a prominent factor in ulcer development. Indeed, a recent review suggested that the trapping and activation of white blood cells (leukocytes) was the major factor associated with the inflammatory reaction that may ultimately lead to ulceration (Pascarella 2005). This hypothesis has become increasingly popular since the discovery that shortage of oxygen is not the primary cause of venous ulceration (Pascarella 2005).

Treating venous leg ulcers

Several authors have summarised the evidence for existing therapies that have been proven to be, or are likely to be, therapeutically beneficial in the treatment of venous ulceration. These interventions include multilayer elastic compression bandaging, limb elevation, and surgical therapy (Margolis 1994; Araujo 2003; Nelson 2006; O'Meara 2009). There has also been considerable interest in the role of pharmacological agents to address the microcirculatory dysfunction that gives rise to venous ulceration (Coleridge-Smith 2003; Howlader 2004; Pascarella 2005).

A number of drugs have been implicated as adjunctive therapies in leg ulcer treatment; including aspirin, sulodexide, flavonoids, granulocyte-macrophage colony-stimulating factor, and systemic mesoglycan (Nelson 2006). Additionally, a Cochrane Review found pentoxifylline to be an effective adjunct in venous leg ulcer healing, both in the presence and absence of compression therapy (Jull 2009). Araujo 2003, however, stated that most of these drugs are efficacious only when used in conjunction with multilayer compression bandaging. Despite ongoing efforts to develop effective pharmacological agents, the nucleus of venous ulcer treatment continues to be physical in nature.

Description of the intervention

Flavonoids, first isolated in the 1930s, are a large and diverse family of naturally-occurring compounds. Flavonoid compounds occur in a variety of plant-based foods and beverages, including cocoas, chocolates, teas and red wines (Hammerstone 2000). They are categorised into six subclasses according to their chemical structure (flavonols, flavones, isoflavones, flavonones, catechins and anthocyanidins)(Manach 2004).

Most flavonoids are administered orally, however, they can also be administered in the form of a cream (Tajana 1988; Hasanoglu 2001), or intravenously (Howlader 2004). Micronised Purified Flavonoid Fraction (MPFF) is the most extensively studied flavonoid formulation, and has been reported to demonstrate efficacy in venous ulcer healing (Guilhou 1997; Glinski 1999; Shephard 2002). Another two flavonoid preparations which have been associated with the treatment of venous ulceration are hydroxyethylrutosides (HR), also known as troxerutin, and Pycnogenol® (Stegmann 1986; Zuccarelli 1996; Belcaro 2005). Several studies have demonstrated the efficacy of Pycnogenol® in reducing the symptoms of chronic venous insufficiency (Arcangeli 2000; Petrassi 2000; Koch 2002; Riccioni 2004). The other flavonoid compounds which have been investigated for their anti-ulcer activities are catechins and epicatechins which have been tested in mice (Sannomiya 2005).

How the intervention might work

Flavonoids have been widely used in the management of the symptoms of venous disease for many years (Coleridge-Smith 2003). They can address certain microcirculatory deficiencies involved in ulcer pathophysiology. These include decreasing leukocyte (white blood cell) adhesion and free radical formation; decreasing the permeability and fragility of the vein valves and venous wall (and therefore decreasing abnormal leakage in the lower limbs); and increasing venous flow (Pascarella 2005). They cannot, however, attack the underlying cause of venous hypertension (Howlader 2004). MPFF, marketed as Daflon® 500 mg, has been strongly implicated in the treatment of venous ulceration (Struckmann 1999; Lyseng 2003; Bergan 2005; Katsenis 2005), although its mode of efficacy is incompletely understood (Howlader 2004).

Why it is important to do this review

An industry-sponsored meta-analysis of randomised controlled trials (RCTs) was previously undertaken by Coleridge-Smith 2005 to assess the effects of MPFF (Daflon® 500mg) in the treatment of venous leg ulcers. Unlike that meta-analysis, which included only trials of MPFF (Daflon® 500 mg), our review aims to assess the effects of all possible flavonoid formulations for leg ulcer healing.

Another systematic review that was published on the Cochrane Library assessed the effectiveness of phlebotonics for venous insufficiency (including flavonoids and venous leg ulcers) (Martinez-Zapata 2005); this review included two trials of phlebotonics for venous ulceration (Schultz-Ehrenburg 1993a; Guilhou 1997). As both these reviews were undertaken some years ago, a current review of the available evidence is due.

Objectives

To evaluate the clinical effects of flavonoids on the healing of venous leg ulcers.

Methods

Criteria for considering studies for this review

Types of studies

All RCTs investigating the effects of all flavonoids, or flavonoid-containing compounds, on venous leg ulcer healing.

Types of participants

All adults (aged over 18 years) with clinically diagnosed venous leg ulceration, in any care setting. Studies should specify that venous ulcers were diagnosed to differentiate them from ulceration of alternative aetiology (causes). Methods of diagnosis may include the ankle-brachial index or colour duplex ultrasonography (Araujo 2003).

Types of interventions

Any flavonoid (natural, semi-synthetic or synthetic) or flavonoid formulation, as well as any pharmacological compound that contains an active flavonoid ingredient, whether administered orally, intravenously or topically.

Comparisons include:

  1. Flavonoids (with or without compression bandaging) compared with placebo (with or without compression bandaging).

  2. Flavonoids compared with compression bandaging alone.

  3. Comparison of different dosages of the same flavonoid or comparisons of different types of flavonoid.

Types of outcome measures

Primary outcomes
  • Total number of completely healed ulcers within the trial period.

  • Change in ulcer surface area, however reported.

  • Time for ulcers to achieve complete healing.

Secondary outcomes
  • Recurrence (rates, new ulcer episodes, number of ulcer-free weeks or months).

  • Quality of Life.

  • Cost-effectiveness of the intervention.

  • Episodes of hospitalisation.

  • Pain, mobility, functional status.

  • Adverse effects or events.

Search methods for identification of studies

Electronic searches

In February 2013 we searched the following electronic databases to find reports of relevant RCTs:

  • The Cochrane Wounds Group Specialised Register (searched 11 February 2013);

  • The Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library 2013, Issue 1);

  • Ovid MEDLINE (1950 to January Week 5 2013);

  • Ovid MEDLINE (In-Process & Other Non-Indexed Citations, February 8, 2013);

  • Ovid EMBASE (1980 to 2013 Week 06);

  • EBSCO CINAHL (1982 to 8 February 2013).

The following search strategy was used in The Cochrane Central Register of Controlled Trials (CENTRAL):

#1 MeSH descriptor: [Leg Ulcer] explode all trees 1052
#2 (varicose next ulcer*) or (venous next ulcer*) or (leg next ulcer*) or (stasis next ulcer*) or (crural next ulcer*) or "ulcus cruris":ti,ab,kw 1495
#3 #1 or #2 1905
#4 MeSH descriptor: [Flavonoids] explode all trees 1547
#5 (flavonoid or bioflavonoid* or flav*none* or flavone* or flav*nol*) 401
#6 hesper*din* or hesper*tin 100
#7 diosmin* 103
#8 barosmin or "buchu resin" or venosmine 2
#9 querc*tin* or rutin* or oxerutin* 308
#10 hydroxyethylrutoside* or O-s-hydroxyethyl-rutoside* or B-hydroxyethyl next rutoside* 113
#11 daflon 77
#12 etralex or capiven or "arvenum 500" or ardium or alvenor 3
#13 variton or venitol or flebotropin or elatec 4
#14 MPFF 19
#15 troxerutin* 80
#16 trihydroxyethylrutoside* 3
#17 paroven 19
#18 epicatechin 56
#19 chalcone* or proanthocyanidin* or anthocyanin* or anthocyanidin* 162
#20 catechin* or procyanidolic oligomers or flavon-3-ol* or benzoflavone* 231
#21 flavanolignan* or flav*n* or isoflavone* 1732
#22 #4 or #5 or #6 or #7 or #8 or #9 or #10 or #11 or #12 or #13 or #14 or #15 or #16 or #17 or #18 or #19 or #20 or #21 2462
#23 #3 and #22 47

The search strategies for Ovid MEDLINE, Ovid EMBASE and EBSCO CINAHL can be found in Appendix 1, Appendix 2 and Appendix 3 respectively. The Ovid MEDLINE search was combined with the Cochrane Highly Sensitive Search Strategy for identifying randomised trials in MEDLINE: sensitivity- and precision-maximizing version (2008 revision) (Lefebvre 2011). The Ovid EMBASE and EBSCO CINAHL searches were combined with the trial filters developed by the Scottish Intercollegiate Guidelines Network (SIGN) (SIGN 2011). There were no restrictions on the basis of date, or language, of publication.

Searching other resources

We scrutinised the reference lists of all relevant review articles and included studies found through the database searches, for further studies. To minimise publication bias, efforts were made to identify unpublished material by contacting manufacturers of flavonoid drugs (e.g. Servier Internationale), as well as communicating with experts in the field and authors who have published on this subject.

Data collection and analysis

Selection of studies

Two review authors independently applied a priori inclusion criteria to all titles and abstracts (where available) from the electronically-generated citation lists. Irrelevant citations were identified and discarded during two phases of citation-screening. Disagreements were referred to a third independent review author. The third author double-checked all excluded citations. Full texts of potentially relevant studies were retrieved, regardless of language, and assessed using a study screening sheet. The Department of Language and Linguistic Science at the University of York was approached to translate studies in German or French. Full texts that met all the inclusion criteria were included in the review. Excluded studies were documented, with the reasons for their exclusion. Disagreements about inclusion were referred to the third author.

Data extraction and management

Two review authors independently extracted data (including risk of bias items) from the included studies, using a pre-defined paper data extraction sheet. Any discrepancies were resolved by referring to the third review author.

Data extracted were categorised into:

  • Verification of study eligibility.

  • Trial details (e.g. authors, journal and date of publication, study type, number of trial arms, treatment setting, geographical location).

  • Participants (e.g. numbers in each group, exclusion and inclusion criteria, age and gender).

  • Interventions (e.g. type of flavonoid, composition, dosage, duration of interventions, type of control intervention).

  • Primary and secondary outcomes.

  • Analyses conducted.

  • Results.

  • Risk of bias.

Where data were missing, attempts were made to contact the study authors. In the case of duplicate publications, data were extracted from the most complete article and counted only once, using the largest, most complete report with the longest follow-up (Khan 2003).

Assessment of risk of bias in included studies

For this review two review authors independently assessed the risk of bias of the included studies, as recommended by the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). This tool addresses six specific domains, namely sequence generation, allocation concealment, blinding, incomplete outcome data, selective outcome reporting and other issues (e.g. extreme baseline imbalance). Blinding and completeness of outcome data were assessed for each outcome separately. We completed a risk of bias table for each eligible study. We discussed any disagreement amongst all review authors to achieve a consensus.

We presented an assessment of risk of bias using a 'risk of bias summary figure' (Figure 1), which presents all of the judgements in a cross-tabulation of study by entry. This display of internal validity indicates the weight the reader may give the results of each study.

Figure 1.

Methodological quality summary: review authors' judgements about each methodological quality item for each included study.

Measures of treatment effect

The studies included in the review were combined by narrative overview with a quantitative summary of the results of similar trials, where appropriate. For each trial with dichotomous outcomes, e.g. were ulcers healed? (yes or no), a risk ratio (RR) of healing with 95% CIs was calculated. Statistical heterogeneity was assessed using the Chi² and I² tests (Higgins 2003). The I² indicates the percentage of between-study variation explained by true heterogeneity rather than chance effect. In the absence of significant heterogeneity (i.e. where I² was less than 25%), we used a fixed-effect model when combining studies. Where heterogeneity was moderate (i.e. I² was between 25% and 75%), we used a random-effects model. If statistical heterogeneity was considerable (i.e. I² was greater than 75%), and there was substantial clinical and methodological heterogeneity - or where there was substantial clinical and methodological heterogeneity alone - then pooling was not attempted, and a narrative overview was presented instead.

Mean differences and 95% CIs were calculated for continuous data (e.g. change in ulcer surface area). Where outcomes for continuous variables were presented without confidence intervals, standard deviations, or some measure of the precision of the result, the data were entered into the analysis table and not used in data pooling. We assessed the possibility of pooling studies, for which some measure of the precision of the result was available.

We intended to investigate publication bias by using a funnel plot, however, this was not possible because an insufficient number of trials was included in the review. Funnel plots have been shown to be unhelpful in reviews that have only a small number of included studies (Tang 2000). Similarly, sensitivity analyses were not possible, due to the small number of included studies.

Results

Description of studies

Six studies were excluded. One trial did not report the condition of interest (Gilly 1994), another trial examined the prevention of ulcer recurrence in patients with recently healed ulcers (Wright 1991), one trial considered mixed wounds and did not report the data on venous leg ulcers separately (Shrivastava 2011), one trial did not include patients with venous leg ulcers (Balmer 1980), and two trials were non-randomised clinical trials (Mann 1981; Belcaro 2005) (see table Description of studies).

Nine trials met the inclusion criteria for the review (Stegmann 1986; Schultz-Ehrenburg 1993a; Schultz-Ehrenburg 1993b; Zuccarelli 1996; Guilhou 1997; Glinski 1999; Roztocil 2002; Saveljev 2002; Zuccarelli 2004).

When the French pharmaceutical company Servier Internationale (manufacturers of MPFF preparation Daflon® 500 mg) was contacted it provided detailed unpublished data from two trials of MPFF (Saveljev 2002; Zuccarelli 2004). No other information was forthcoming from any other manufacturers. We contacted the lead author of two of the included studies (Schultz-Ehrenburg 1993a; Schultz-Ehrenburg 1993b), as he had reported both trials in a single publication that was only an extended abstract and contained limited information. The author replied that the abstract had not been published as a full paper, and he had no additional data available. Finally, the results of another included trial were only available as a conference abstract (Zuccarelli 1996). No contact details could be found for any of its authors, so no additional data were identified.

The trials were published between 1986 and 2004. Five trials investigated MPFF, four trials investigated hydroxyethylrutosides (HR), and no other trials involving other flavonoid preparations that met the inclusion criteria were identified. All flavonoid interventions were delivered orally. All but one of the MPFF trials had an intervention duration of six months: the intervention duration of the trial by Guilhou 1997 was only two months, because physicians wished to have access to skin grafting for those patients who still had unhealed ulcers after this time. The trials investigating the efficacy of HR ranged from six weeks to three months.

In total, 552 patients received a flavonoid agent in conjunction with standard therapy (compression bandaging and local therapy), and 523 patients received standard therapy alone, or standard therapy with placebo. Most studies used similar criteria for diagnosis of venous leg ulceration except for Saveljev 2002, which was a translated paper and classified the ulcers as venous trophic ulcers.

Risk of bias in included studies

Incomplete reporting was a concern in the assessment of risk of bias in most of the studies.Three studies were not available as full publications (Schultz-Ehrenburg 1993a; Schultz-Ehrenburg 1993b; Zuccarelli 1996). Details are presented in separate risk of bias tables for each of the nine included studies, and there is also a summary table (Figure 1).

Allocation

Only two trials reported adequate randomisation: participants were randomly distributed to an intervention or control group according to either a centrally-generated randomisation list (Zuccarelli 2004), or a list prepared by a biostatistician (Glinski 1999). All other trials had unclear randomisation methods.

Only one trial had adequate allocation concealment (Zuccarelli 2004); one trial was judged to be at high risk of bias (Glinski 1999), while the rest had unclear allocation concealment.

Blinding

One trial had adequate blinding on all three primary outcomes (Zuccarelli 2004), while Guilhou 1997 was adequately blinded for only one outcome (surface area). Four studies did not assess one primary outcome (time to complete healing), and blinding was unclear for the other two primary outcomes (Stegmann 1986; Schultz-Ehrenburg 1993a; Schultz-Ehrenburg 1993b; Glinski 1999). Blinding in the remaining studies was unclear for all the primary outcomes.

Incomplete outcome data

In judging the risk of bias for this item, we considered whether missing data were imputed appropriately and whether an intention-to-treat analysis was reported for the primary outcomes.

Outcomes: numbers of ulcer healed

Three trials were considered to have performed an intention-to-treat analysis: Glinski 1999 reported no loss to follow-up; Zuccarelli 2004 imputed missing data using the method of last observation carried forward; while Stegmann 1986 reported one loss to follow-up from each group, with the missing data being balanced across the two groups.

In Saveljev 2002, the loss to follow-up was high for the control group at 21% (13/62), compared with 8% (5/62) in the flavonoid group. Since missing data in the study were considered and analysed as "ulcer not healed," this gave a greater effect size estimate in favour of the flavonoid group. This trial was judged to be at high risk of bias. It was unclear whether bias due to incomplete outcome data was present in the other trials.

Outcomes: changes in surface area

Only two trials were judged not to be at risk of bias due to incomplete outcome data; Glinski 1999 had no loss to follow-up, while Zuccarelli 2004 imputed the missing data appropriately using the method of last observation carried forward. It was unclear whether bias due to incomplete outcome data was present in the other trials.

Outcomes: time to achieve healing

Four trials did not assess time to achieve healing (Stegmann 1986; Schultz-Ehrenburg 1993a; Schultz-Ehrenburg 1993b; Glinski 1999). One trial imputed the missing data using the method of last observation carried forward (Zuccarelli 2004), and this trial was judged to have a low risk of bias. It was unclear whether bias due to incomplete outcome data was present in the other trials.

Selective reporting

In judging the risk of bias for selective reporting, we were unable to assess the trial protocols, and, therefore, assessed the studies on the basis of the pre-specified outcome measures reported in the methods section of the trial reports. Four trials reported pre-specified outcome measures as well as other secondary outcomes and were, therefore, judged to be at a low risk of bias (Glinski 1999; Roztocil 2002; Saveljev 2002; Zuccarelli 2004). Stegmann 1986, however, omitted pre-specified outcome measures, and was thus considered to be at a high risk of bias. The remaining trials did not provide sufficient information to permit judgements to be made about bias due to selective reporting.

Other potential sources of bias

Most of the included studies did not have sufficient information to assess for other potential biases.There were significant baseline difference between the two groups in Saveljev 2002. Even though the authors reported that adjustment was made to take account of the differences, it was not clear how the calculation was done. Calculation of percentage reduction without proper adjustment in wound area from baseline would have given favourable result to the flavonoid group as it had a smaller baseline ulcer area.

Effects of interventions

1. Micronised purified flavonoid fraction (MPFF) compared with control

1.1 Number of ulcers completely healed

All five studies provided data for this outcome (Guilhou 1997; Glinski 1999; Roztocil 2002; Saveljev 2002; Zuccarelli 2004), and the data were sufficiently homogenous clinically to be used in a quantitative synthesis. All were parallel group RCTs investigating ulcer healing in adults receiving two 500 mg tablets of MPFF a day in addition to standard treatment. All authors reported using similar patient inclusion criteria and standardised local therapy.

Pooling of the five trials using a random-effects model (I² = 60%) showed a statistically significant effect in favour of MPFF (RR 1.36; 95% CI 1.07 to 1.74) (Analysis 1.1). The evidence of heterogeneity (I² = 60%) across studies was probably due to one trial (Zuccarelli 2004), in which the effect was in the opposite direction to that in the other four studies. It should be noted that Zuccarelli 2004 was at low risk of bias and individually did not demonstrate any benefit in favour of MPFF; the other four trials were all at unclear risk of bias overall.

1.2 Change in ulcer surface area (cm²)

There was substantial variability in the reporting of this outcome across the five studies that investigated MPFF. Two studies did not provide any measure of the precision of the estimates (Guilhou 1997; Glinski 1999).

It was judged to be inappropriate to pool two trials, Roztocil 2002 and Saveljev 2002. In Saveljev 2002 the mean ulcer area at baseline differed significantly between the MPFF group (4.76 cm²) and the control group (8.01 cm²). This has important consequences for the validity of the trial's results. A comparison of end scores would favour the MPFF group (0.85 cm²) over the control group (2.53 cm²), however, comparison of change in ulcer area would favour the control group (5.48 cm² versus 3.91 cm²). A comparison of percentage reduction from baseline would again change the results, with the MPFF group showing a greater reduction of 82.1% compared to 68.4% in the control group. In a trial with poor comparability of wound area at baseline, and when the outcome is based on change in wound area, the result can only be considered valid where percentage area change and absolute area change are in the same direction (Bradley 1999). Therefore, we reported the mean change in ulcer surface area separately for the two trials. Roztocil 2002 reported that the mean difference (MD) between the two groups was statistically significant in favour of MPFF (MD 0.93; 95% CI 0.21 to 1.65) (Analysis 1.2), whereas in Saveljev 2002 the difference was not significant (MD -1.57; 95% CI -3.42 to 0.28) (Analysis 1.3).

We could not present the comparisons in the study by Zuccarelli 2004 graphically as area change from baseline because the standard deviation for changes from baseline values was not available. Since ulcer surface area was comparable between treatment groups at baseline, and because this was the most rigorous trial, we felt it was appropriate to illustrate graphically the difference in ulcer surface area between treatment groups at trial end (after six months of treatment) (Analysis 1.4). At trial end, the mean difference in ulcer surface area between the two groups was not significant (MD 0.22; 95% CI -1.84 to 2.28).

Glinski 1999 reported a greater reduction in ulcer area in the MPFF group compared to the control group (80% and 65% respectively). While Guilhou 1997 presented mean baseline ulcer surface areas for both groups, no data at endpoint were presented in the report. The change in ulcer surface area over time was reported to be similar in the two groups.

1.3 Time to achieve complete healing

Glinski 1999 did not report this outcome. All the other trials except for Zuccarelli 2004 reported significantly faster healing time with MPFF. In Guilhou 1997, the time to achieve healing was reported to be shorter in the MPFF group than in the placebo group, but no other data were provided. In Roztocil 2002, the mean healing time was 137 days for the MPFF group compared to 166 days for the control group, while in Saveljev 2002, the median healing time was 90 days in the MPFF group versus 119 days in the control group. Based on the reported data, it was decided it was not appropriate to pool studies.

1.4 Recurrence rates

None of the trials reported this outcome.

1.5 Quality of life

None of the trials reported this outcome.

1.6 Cost-effectiveness

Only two of the five included studies on MPFF addressed cost (Glinski 1999; Saveljev 2002). Glinski 1999 reported that MPFF, in addition to standard therapy, was more effective and cheaper than standard therapy alone. The cost effectiveness ratio (cost per healed ulcer) in the MPFF group was EUR1026.2 compared with EUR 1871.8 in the control group. In the analysis, a social point of view was chosen and costs and consequences of the therapy in the two groups were taken into account. Costs covered direct medical costs, such as drugs, dressings, laboratory tests, hospitalisation caused by the deterioration of ulceration and adverse events, while effectiveness was measured by the number of patients healed. In contrast, Saveljev 2002 concluded that it was more expensive to treat with MPFF and standard therapy than standard therapy alone. The incremental cost-effectiveness ratio for MPFF group for the additional healing of another ulcer was RUB 10906.2, in comparison with standard therapy alone.

1.7 Hospitalisation

Four of the five trials reported on hospitalisation. Guilhou 1997 reported that there was no significant difference in duration of time in hospital between the MPFF group (7.5 days) and the control group (14.5 days). Three trials provided data on the number of patients hospitalised, which we pooled (Glinski 1999; Saveljev 2002; Zuccarelli 2004) (Analysis 1.5). The pooled analysis shows no statistically significant difference in number of hospitalisations in the two groups (RR 0.76; 95% CI 0.47 to 1.23).

1.8 Pain and mobility

Two trials reported no significant difference in severity and reduction of pain between the two groups (Glinski 1999; Saveljev 2002). Two other trials did not assess pain specifically (Guilhou 1997; Zuccarelli 2004), while Roztocil 2002 reported that there was no significant difference between groups for the presence of night cramps and painful sensations.

1.9 Adverse events

Roztocil 2002 reported that no adverse events were observed in the trial, while Glinski 1999 did not report on this outcome. Significantly, the pooled analysis of three trials showed that 52% more participants in the MPFF group experienced adverse events compared to participants in the control group (RR 1.52; 95% CI 1.01 to 2.30) (Analysis 1.6). Adverse events were reported to be mild to moderate across studies. The most common events were skin changes (including eczema), gastrointestinal disturbances (including diarrhoea) and hypertension.

2 Hydoxyethylrutosides (HR) compared with control

2.1 Number of ulcers completely healed

One trial did not provide data on the proportion of ulcers healed (Schultz-Ehrenburg 1993b), but reported that there was no significant difference in the number of ulcers healed between the two groups.Three other trials were sufficiently homogenous clinically and methodologically for their estimated effects to be pooled (Stegmann 1986; Schultz-Ehrenburg 1993a; Zuccarelli 1996). Stegmann 1986 reported the proportion of ulcers healed for patients with existing suprafacial chronic venous insufficiency and those with post-thrombotic syndrome separately; we combined the data from the two subgroups of patients to give the proportion of total ulcers healed. Pooling the three trials using a fixed-effect model (I² = 0%) showed a statistically significant effect in favour of HR (RR 1.70; 95% CI 1.24 to 2.34) (Analysis 2.1). The three trials, however, had a limited number of patients and an unclear risk of bias regarding method of randomisation, allocation concealment, blinding and handling of missing data.

2.2 Change in ulcer surface area (cm²)

All four trials assessed change in ulcer surface area, although none of the publications provided sufficient data for pooling. Three studies reported significantly greater reduction in ulcer size in groups treated with HR compared with control groups (Stegmann 1986; Schultz-Ehrenburg 1993a; Zuccarelli 1996), but did not provide sufficient data to allow this review to reproduce these analyses. Schultz-Ehrenburg 1993b reported that there was no difference in the reduction of ulcer surface area in the two groups. No other data were provided.

2.3 Time to achieve complete healing

Zuccarelli 1996reported a significantly faster healing time in patients in the treatment group compared with those in the control group (log rank test: P value < 0.010). None of the other four trials reported time to complete healing.

2.4 Recurrence rates

None of the trials reported this outcome.

2.5 Quality of life

None of the trials reported this outcome.

2.6 Cost-effectiveness

None of the trials reported this outcome.

2.7 Hospitalisation

None of the trials reported this outcome.

2.8 Pain and mobility

Ulcer pain was not measured in Zuccarelli 1996, while Stegmann 1986 reported no significant differences between groups for venous symptoms (including pain). Pain was mentioned in trials by Schultz-Ehrenburg 1993a and Schultz-Ehrenburg 1993b; however, the results of pain were not reported in the publication (extended abstract).

2.9 Adverse events

Adverse events were generally reported poorly. Two trials (Stegmann 1986; Zuccarelli 1996) did not report on adverse events. Schultz-Ehrenburg 1993a and Schultz-Ehrenburg 1993b merely reported that adverse events were evenly distributed between the groups and none was serious.

Discussion

Nine RCTs involving 1075 patients were included in this review. Unlike the previous systematic review (Coleridge-Smith 2005), which included only trials on micronised purified flavonoid fraction (MPFF), this review considered all flavonoid preparations that have been used for venous leg ulcers. Beside MPFF, two other flavonoid preparations, namely hydroxyethylrutosides (HR) and Pycogenol® have been assessed for their effects on leg ulcers, however, one trial on Pycogenol® was excluded because it was not an RCT(Belcaro 2005).

Although the overall estimate on the number of healed ulcers showed significant effects of flavonoids - both MPFF and HR - on ulcer leg healing, this result needs to be interpreted with caution. The meta-analysis result was not based on data from all trials. Several trials either did not report on the outcome of interest, or did not provide sufficient data to allow pooling of estimates. Additionally, pooling of data was from trials with a limited number of patients that were either at high risk of bias or had an unclear risk of bias regarding method of randomisation, allocation concealment, blinding and handling of missing data.

Trials which showed significant effects in favour of the MPFF group were judged either to be at high risk of bias in at least one of the bias domains of the Cochrane Risk of Bias Assessment Tool, or did not provide sufficient information to allow us to assess the risk of bias. In contrast, the trial which was considered to be at low risk of bias showed no significant difference in effects between MPFF and control groups. There is evidence to show that studies that are inadequately concealed and not blinded can lead to overestimation of treatment effects by up to 41% and 17% respectively (Schulz 1995).

The most rigorously conducted trial with the largest sample size was unpublished (Zuccarelli 2004). This raises the possibility of publication bias, as its result was not favourable for flavonoids. There is substantial evidence that studies with significant or favourable results are more likely to be published than those with non-significant or unfavourable results (Easterbrook 1991; Dickersin 1993; Song 2000), and, therefore, the possibility of publication bias in this review cannot be dismissed. Even though we have undertaken extensive searches for published material, we still could not exclude the possibility that studies with negative findings remain unpublished. We did not generate a funnel plot to assess the degree of publication bias in this systematic review, because of the small number of included studies. Publication bias should be taken into consideration when interpreting the present findings. The small number of trials identified, and the generally low methodological quality of these trials, also prevented us from performing a sensitivity analysis to check the robustness of the results of the review by excluding trials with a high risk of bias from the analysis.

Three out of four trials investigating HR reported on the number of completely healed ulcers, with one study demonstrating a significant effect in favour of HR (Zuccarelli 1996). The pooled estimate of these three trials showed that more ulcers healed in the HR group than the control group. This finding should also be interpreted with caution, as all these trials had unclear risk of bias for randomisation, allocation concealment and blinding. A previous review of oedema protective drugs (including HR) concluded that such agents "are probably beneficial for patients with acute venous ulcers in addition to compression therapy" (Diehm 1996), however, more research is needed to confirm the efficacy of HR.

The quality of reporting of secondary outcomes in included studies was varied; secondary outcomes of interest were not reported in all trials. Overall, painful sensations were less frequently experienced among groups receiving flavonoids. In contrast, adverse events were reported more often in those treated with flavonoids, although very few severe events were observed that could be related to treatment. More research is needed to establish the cost-effectiveness of flavonoids; two of the trials conducted economic evaluations, but their findings on cost were contradictory and hence, inconclusive.

Quality of the evidence

The shortcomings of the nine trials are outlined in the risk of bias section of the results. Overall, the quality of reporting was poor. Many trials omitted important methodological information, such as method of randomisation, allocation concealment, blinding and completeness of outcomes, which are essential for assessing risk of bias.

Schulz 1995 showed that studies which are not randomised and not blinded can lead to overestimation of treatment effects. Only two studies included in the review were judged to have adequate randomisation and to be sufficiently blinded (Glinski 1999; Zuccarelli 2004), and only one trial was considered to have adequate allocation concealment (Zuccarelli 2004). Thus, it is likely that bias due to lack of randomisation, allocation concealment and blinding have exaggerated the treatment effect obtained in this review. Additionally, in one of the studies (Saveljev 2002), the loss to follow-up was high for the control group 21% (13/62) compared with 8% (5/62) in the MPFF group. Since missing data in the study were considered and analysed as "ulcer not healed," this has overestimated the treatment effect in favour of the flavonoids. One unpublished study included in the review which was clearly the most rigorously conducted and thus has high internal validity compared to all other studies did not show any benefit in favour of the flavonoids (Zuccarelli 2004).

Authors' conclusions

Implications for practice

At present, there is some evidence of the beneficial effects of flavonoids (both micronised purified flavonoid fraction (MPFF) and hydroxyethylrutosides (HR)) for venous leg ulcers from trials that we judged to be suffering from some degree of bias. Since there is at least one rigorously-conducted unpublished trial that did not show any evidence of beneficial effects with MPFF, we concluded that the overall evidence is not adequate to support the use of flavonoids in practice. There is some evidence that MPFF caused significantly more adverse events than control treatments, however, these adverse events were reported to be mild to moderate, including events such as skin changes (including eczema) and gastrointestinal disturbances (including diarrhoea).

Implications for research

Based on the assessment of the included trials, which were small and had methodological problems, we recommend that future trials will require:

1.  participant numbers that ensure sufficient statistical power to detect true treatment effects;
2.  use of true randomisation with allocation concealment (e.g. computer-generated codes);
3.  measures to help ensure comparability of treatments at baseline (e.g. stratification for ulcer size);
4.  blinded outcome assessment;
5.  use of objective outcome measurement (e.g. ulcer area, number of wounds completely healed in a pre-specified time period); and
6.  use of intention-to-treat analysis.

Randomised controlled trials should be adequately reported. The CONSORT 2010 statement (Moher 2010) lists 25 items that need to be reported to show readers whether or not a trial is likely to produce valid and reliable results. Further research into the relationship of flavonoids and venous ulcer healing needs to be reported in accordance with this guideline.

Acknowledgements

Thanks to Dr Philip Coleridge-Smith, Dr Catherine Lok, and Mr Francoise Pitsch for their assistance in making available unpublished material from the French pharmaceutical company Servier International.

Appreciation also extends to the following authors for replying to e-mail queries and sending full-text copies of published papers: Gianno Belcaro; Andrea Nelson; Professor Michael Junger; Robert S Kirsner; Dr AA Ramelet; Denis Clement; Professor Michel R Boisseau; Marian Simka; John J Bergan; Professor Ulrich Schultz-Ehrenburg; and Philip Coleridge-Smith.

The authors would also like to acknowledge the contribution of the referees: Wounds Group Editors (Andrew Jull, David Margolis, Gill Worthy) and peer referees (Una Adderley, Jane Burch, Nikki Stubbs) and to Elizabeth Royle for copy editing the review.

Data and analyses

Download statistical data

Comparison 1. MPFF versus control
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Number of ulcers healed5723Risk Ratio (M-H, Random, 95% CI)1.36 [1.07, 1.74]
2 Mean change in ulcer surface area from baseline (cm2)1150Mean Difference (IV, Fixed, 95% CI)0.93 [0.21, 1.65]
3 Mean change in ulcer surface area from baseline (cm2)1124Mean Difference (IV, Fixed, 95% CI)-1.57 [-3.42, 0.28]
4 Ulcer surface area (cm2) at trial-end1199Mean Difference (IV, Fixed, 95% CI)0.22 [-1.84, 2.28]
5 Hospitalisation3466Risk Ratio (M-H, Fixed, 95% CI)0.76 [0.47, 1.23]
6 Adverse events3431Risk Ratio (M-H, Fixed, 95% CI)1.52 [1.01, 2.30]
Analysis 1.1.

Comparison 1 MPFF versus control, Outcome 1 Number of ulcers healed.

Analysis 1.2.

Comparison 1 MPFF versus control, Outcome 2 Mean change in ulcer surface area from baseline (cm2).

Analysis 1.3.

Comparison 1 MPFF versus control, Outcome 3 Mean change in ulcer surface area from baseline (cm2).

Analysis 1.4.

Comparison 1 MPFF versus control, Outcome 4 Ulcer surface area (cm2) at trial-end.

Analysis 1.5.

Comparison 1 MPFF versus control, Outcome 5 Hospitalisation.

Analysis 1.6.

Comparison 1 MPFF versus control, Outcome 6 Adverse events.

Comparison 2. Hydroxyethylrutosides (HR) versus control
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Number of ulcers healed3279Risk Ratio (M-H, Fixed, 95% CI)1.70 [1.24, 2.34]
Analysis 2.1.

Comparison 2 Hydroxyethylrutosides (HR) versus control, Outcome 1 Number of ulcers healed.

Appendices

Appendix 1. Ovid MEDLINE search strategy

1 exp Leg Ulcer/
2 (varicose ulcer* or venous ulcer* or leg ulcer* or stasis ulcer* or crural ulcer* or ulcus cruris).tw.
3 or/1-2
4 exp Flavonoids/
5 (flavonoid* or bioflavonoid* or flav?none* or flavone* or flav?nol*).tw.
6 (hesper?din* or hesper?tin).tw.
7 diosmin*.tw.
8 (barosmin or buchu resin or venosmine).tw.
9 (querc?tin* or rutin* or oxerutin*).tw.
10 (hydroxyethylrutoside* or O-s-hydroxyethyl-rutoside* or B-hydroxyethyl rutoside*).tw.
11 daflon.tw.
12 (detralex or capiven or arvenum 500 or ardium or alvenor).tw.
13 (variton or venitol or flebotropin or elatec).tw.
14 MPFF.tw.
15 troxerutin*.tw.
16 trihydroxyethylrutoside*.tw.
17 paroven.tw.
18 epicatechin$.tw.
19 (chalcone* or proanthocyanidin* or anthocyanin* or anthocyanidin*).tw.
20 (catechin* or procyanidolic oligomers or flavon-3-ol* or benzoflavone*).tw.
21 (flavanolignan* or flav?n* or isoflavone*).tw.
22 or/4-21
23 3 and 22

Appendix 2. Ovid EMBASE search strategy

1 exp Leg Ulcer/
2 (varicose ulcer* or venous ulcer* or leg ulcer* or stasis ulcer* or crural ulcer* or ulcus cruris).tw.
3 or/1-2
4 exp Flavonoid/
5 exp Flavanone/
6 (flavonoid* or bioflavonoid* or flav?none* or flavone* or flav?nol*).tw.
7 (hesper?din* or hesper?tin).tw.
8 diosmin*.tw.
9 (barosmin or buchu resin or venosmine).tw.
10 (querc?tin* or rutin* or oxerutin*).tw.
11 (hydroxyethylrutoside* or O-s-hydroxyethyl-rutoside* or B-hydroxyethyl rutoside*).tw.
12 daflon.tw.
13 (detralex or capiven or arvenum 500 or ardium or alvenor).tw.
14 (variton or venitol or flebotropin or elatec).tw.
15 MPFF.tw.
16 troxerutin*.tw.
17 trihydroxyethylrutoside*.tw.
18 paroven.tw.
19 epicatechin*.tw.
20 (chalcone* or proanthocyanidin* or anthocyanin* or anthocyanidin*).tw.
21 (catechin* or procyanidolic oligomers or flavon-3-ol* or benzoflavone*).tw.
22 (flavanolignan* or flav?n* or isoflavone*).tw.
23 or/4-22
24 3 and 23

Appendix 3. EBSCO CINAHL search strategy

S23 S4 and S22
S22 S5 or S6 or S7 or S8 or S9 or S10 or S11 or S12 or S13 or S14 or S15 or S16 or S17 or S18 or S19 or S20 or S21
S21 TI ( flavanolignan* or flavon* or flavan* or isoflavone* ) or AB ( flavanolignan* or flavon* or flavan* or isoflavone* )
S20 TI ( catechin* or procyanidolic oligomers or flavon-3-ol* or benzoflavone* ) or AB (catechin* or procyanidolic oligomers or flavon-3-ol* or benzoflavone*)
S19 TI ( chalcone* or proanthocyanidin* or anthocyanin* or anthocyanidin* ) or AB ( chalcone* or proanthocyanidin* or anthocyanin* or anthocyanidin* )
S18 TI epicatechin* or AB epicatechin*
S17 TI paroven or AB paroven
S16 TI trihydroxyethylrutoside* or AB trihydroxyethylrutoside*
S15 TI troxerutin* or AB troxerutin*
S14 TI MPFF or AB MPFF
S13 TI ( variton or venitol or flebotropin or elatec ) or AB ( variton or venitol or flebotropin or elatec )
S12 TI ( detralex or capiven or arvenum 500 or ardium or alvenor ) or AB ( detralex or capiven or arvenum 500 or ardium or alvenor )
S11 TI daflon or AB daflon
S10 TI ( hydroxyethylrutoside* or O-s-hydroxyethyl-rutoside* or B-hydroxyethyl rutoside* ) or AB ( hydroxyethylrutoside* or O-s-hydroxyethyl-rutoside* or B-hydroxyethyl rutoside* )
S9 TI ( rutin* or oxerutin* or quercetin ) or AB ( rutin* or oxerutin* or quercetin )
S8 TI ( diosmin* or barosmin or buchu resin or venosmine ) or AB ( diosmin* or barosmin or buchu resin or venosmine )
S7 TI hesper* or AB hesper*
S6 TI ( flavonoid* or bioflavonoid* or flavanone* or flavone* or flavanol* or flavonone* or flavone* or flavonol* ) or AB ( flavonoid* or bioflavonoid* or flavanone* or flavone* or flavanol* or flavonone* or flavone* or flavonol* )
S5 (MH "Flavonoids+")
S4 S1 or S2 or S3
S3 lower extremity N3 ulcer* or AB lower extremity N3 ulcer*
S2 TI (varicose ulcer* or venous ulcer* or leg ulcer* or foot ulcer* or (feet N1 ulcer*) or stasis ulcer* or crural ulcer*) or AB (varicose ulcer* or venous ulcer* or leg ulcer* or foot ulcer* or (feet N1 ulcer*) or stasis ulcer* or crural ulcer*)
S1 (MH "Leg Ulcer+")

History

Protocol first published: Issue 2, 2007
Review first published: Issue 5, 2013

DateEventDescription
11 February 2010AmendedConverted to new review format.
1 February 2006New citation required and conclusions have changedSubstantive amendment

Contributions of authors

Chris Scallon wrote the first draft of the review, screened searches for the review, undertook data extraction and an initial quality assessment.
Sally Bell-Syer screened searches for the review, advised on methodology, checked the data extraction, reviewed the text and analyses and supervised the risk of bias assessment.
Zoriah Aziz screened searches for the review, completed the risk of bias assessment, dealt with the peer referee comments and input data into the review.

Contributions of editorial base:

Nicky Cullum: edited the review, advised on methodology, interpretation and review content.
Andrea Nelson Editor: approved the final review prior to submission.
Sally Bell-Syer: coordinated the editorial process. Advised on methodology, interpretation and content.
Ruth Foxlee: designed the search strategy, ran the searches and edited the search methods section.
Rachel Richardson: edited the review, checked the content.

Declarations of interest

NONE

Sources of support

Internal sources

  • Department of Health Sciences, University of York, York, UK.

External sources

  • NIHR/Department of Health (England), (Cochrane Wounds Group), UK.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Glinski 1999

MethodsRCT: multi-centre trial involving 10 centres in Poland. Central randomisation list.
Participants140 (36 men and 104 women) patients aged over 18 years; 1 or more venous ulcers on lower extremities; ulcers 2-10 cm in diameter; ulceration present for at least 3 months; ankle-arm index (Doppler) > 0.9.
InterventionsGroup 1: Detralex® 500 mg containing diosmin 450 mg and hesperidin 50 mg (2 tablets/day) + local treatment + compression bandaging (n = 71).
Group 2: standard compression therapy and local treatment alone (n = 69).
Outcomes

Pre-specified outcomes
Proportion of fully healed ulcers, variation of ulcer surface area, cost-effectiveness, discomfort, pain, heavy legs, night cramps.

Primary outcomes
A. Proportion of healed ulcers:
Group 1: 33/71
Group 2: 19/69

B. Ulcer surface area (% reduction):
Group 1: 80%
Group 2: 65%

C. Time for ulcers to achieve complete healing: this outcome was not examined in the trial.

Secondary outcomes
D. Cost-effectiveness:
The cost effectiveness ratio (cost per healed ulcer) of MPFF group was EUR 1026.2 compared with EUR 1871.8 in control group. Treatment with MPFF and standard therapy was more effective and cheaper than standard therapy alone.

E. Hospitalisation (%), (mean number of days hospitalised):
Group 1: 14 (45)
Group 2: 19 (34)

F. Pain (% reduction):
Group 1: 57%
Group 2: 43% (difference not significant)

G. Adverse effects:
Slightly more adverse effects were reported in the control group than the MPFF group. The difference was not statistically significant. No other data were provided.

All other secondary outcomes were not reported.

NotesTreatment duration = 6 months.
Loss to follow-up = None.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskQuote: "randomisation list prepared by the Biostatistics ....."
Allocation concealment (selection bias)High riskQuote: "....the allocation of each patient was known to the investigator from the randomisation list"
Blinding (performance bias and detection bias)
number of ulcers healed
Unclear riskQuote:" the patients and doctors were aware of the treatment group to which the patient was allocated."
Comment: There are insufficient data to judge whether the outcome assessor was blinded in this study, however, the non-blinding of participants and personnel was unlikely to introduce bias for this outcome.
Blinding (performance bias and detection bias)
wound surface area
Unclear riskQuote: "the patients and doctors were aware of the treatment group to which the patient was allocated."
Comment: There are insufficient data to judge whether the outcome assessor was blinded in this study.
Blinding (performance bias and detection bias)
time to achieve complete healing
Unclear riskOutcome not assessed
Incomplete outcome data (attrition bias)
number of ulcers healed
Low riskComment: no loss to follow-up. Data were analysed according to ITT.
Incomplete outcome data (attrition bias)
wound surface area
Low riskComment: no loss to follow-up. Data were analysed according to ITT.
Incomplete outcome data (attrition bias)
time to achieve complete healing
Unclear riskOutcome not assessed
Selective reporting (reporting bias)Low riskAll pre-specified outcomes were reported.
Other biasUnclear riskNo other bias detected

Guilhou 1997

MethodsRCT: multi-centre trial involving 9 centres in France. Randomisation of treatment stratified according to size of ulcers (<10 cm or >10 cm).
Participants107 (41 men and 66 women) aged 18-85 years, presenting with leg ulcers of any size for 3 months. Ulcers of venous origin; open for at least 3 months; systolic pressure index (ankle-arm) > 0.8; patients willing to have elastic compression therapy.
InterventionsGroup 1: Daflon® 500 mg containing diosmin 450 mg, hesperidin 50 mg (2 tablets/day) + standardised local care + compression therapy (n = 55).
Group 2: placebo (2 tablets/day) + standardised local care + compression therapy (n = 52).
Outcomes

Pre-specified outcomes
Proportion of fully healed ulcers, time to complete healing, variation in ulcer surface area, appearance of ulcer and peri-ulcerous skin, evolution of symptoms, hospitalisation duration, adverse events.

Primary outcomes (The paper states that an ITT analysis was carried out on 105 participants, but it is not clear how missing data imputed)
A. Proportion of healed ulcers:
Group 1: 14/53
Group 2: 6/52

B. Ulcer surface area:
Did not present any data, however, the paper reported that the change in surface area over time was similar in the 2 groups.

C. Time for ulcers to achieve complete healing:
Time to healing was significantly shorter in the MFF group than in the placebo group (P value = 0.037).

Secondary outcomes
D. Hospitalisation:
Group 1: 7.5 days
Group 2: 14.5 days (P value = 0.113)

E. Adverse events:
Group 1; eczema (n = 2); urticaria (n = 1); pruritus of the scalp (n = 1); local pain (n = 1)
Group 2: skin changes around ulcer (n = 1); asthenia (n = 1); headaches (n = 1); exacerbation of chronic colopathy (n = 1).
All other secondary outcomes were not reported.

NotesTreatment duration = 2 months.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNo description of the method used to generate allocation sequence.
Allocation concealment (selection bias)Unclear riskNo description of the method used to conceal the allocation sequence.
Blinding (performance bias and detection bias)
number of ulcers healed
Unclear riskComment: the study was described as a double-blind, randomised controlled trial, however, there was no description of how the participants, personnel and outcome assessor were blinded.
Blinding (performance bias and detection bias)
wound surface area
Low riskQuote: "Surface area was assessed by a computerised calculation blindly performed in one centre."
Comment: the study was described as a double-blind, randomised controlled trial, however, there was no description of how the participants and personnel were blinded. Nevertheless, the outcome assessor was blinded, and the non-blinding of others was unlikely to introduce bias for this outcome.
Blinding (performance bias and detection bias)
time to achieve complete healing
Unclear riskComment: there was no description of how the participants, personnel and outcome assessor were blinded.
Incomplete outcome data (attrition bias)
number of ulcers healed
Unclear riskQuote: "107 patients were enrolled; two had no data available while under treatment. An ITT was, therefore, performed for 105 patients...."
Comment: it was not clear whether 2 patients were lost before, or after, randomisation.
Incomplete outcome data (attrition bias)
wound surface area
Unclear riskQuote: "107 patients were enrolled; two had no data available while under treatment. An ITT was, therefore, performed for 105 patients...."
Comment: it was not clear whether 2 patients were lost before, or after, randomisation.
Incomplete outcome data (attrition bias)
time to achieve complete healing
Unclear riskComment: it was not clear whether 2 patients were lost before, or after, randomisation.
Selective reporting (reporting bias)Unclear riskComment: insufficient information for the ulcer surface area outcome. All other pre-specified outcomes were reported.
Other biasUnclear riskNo other bias detected

Roztocil 2002

MethodsRCT: multi-centre trial involving 17 centres in the Czech and Slovak Republics.
Participants150 (48 men and 102 women) aged over 18 years; venous ulcers only (2-10 cm diameter); ulcers present for at least 3 months; ankle-arm pressure index > 0.9; no history of skin grafting within 6 months prior to study.
InterventionsGroup 1: Daflon® containing diosmin 450 mg, hesperidin 50 mg (2 tablets/day) + standard local therapy + compression bandaging (n = 82).
Group 2: compression bandaging + standard local therapy alone (n = 68).
Outcomes

Pre-specified outcomes
Proportion of fully healed ulcers, time to complete healing, variation of the ulcer surface area, appearance of the skin in the ulcer area, clinical assessment of CVI evolution (leg heaviness, night cramps and pain).

Primary outcomes (The paper states that an ITT analysis was carried out, but it is not clear how missing data imputed)
A. Proportion of ulcers healed:
Group 1: 53/82
Group 2: 28/68

B. Ulcer surface area (% reduction from baseline):
Group 1: 77%
Group 2: 69%

C. Time for ulcers to achieve complete healing (mean):
Group 1:137 days
Group 2:166 days (P value 0.022)

Secondary outcomes
D. Pain:
Quote: "there was no significant difference between groups on the presence of night cramps and painful sensations." Other details not reported.

E. Adverse events:
Quote "No side effects in relation with treatment were observed".

All other secondary outcomes were not reported.

NotesTreatment duration = 6 months.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNo description of the method used to generate allocation sequence.
Allocation concealment (selection bias)Unclear riskNo description of the method used to conceal the allocation sequence.
Blinding (performance bias and detection bias)
number of ulcers healed
Unclear riskDescribed as a blinded study, but no information on blinding.
Blinding (performance bias and detection bias)
wound surface area
Unclear riskDescribed as a blinded study, but no information on blinding.
Blinding (performance bias and detection bias)
time to achieve complete healing
Unclear riskDescribed as a blinded study, but no information on blinding.
Incomplete outcome data (attrition bias)
number of ulcers healed
Unclear riskComments: imbalance in number of dropouts across the group: Treatment group 2/82; Control group 11/68. No description of how the trialists dealt with missing data.
Incomplete outcome data (attrition bias)
wound surface area
Unclear riskVariation in wound surface area was pre-specified as secondary outcome. No descriptions of how the trialists dealt with missing data.
Incomplete outcome data (attrition bias)
time to achieve complete healing
Unclear riskNo descriptions of how the trialists dealt with missing data.
Selective reporting (reporting bias)Low riskAll pre-specified outcomes were reported.
Other biasUnclear riskNo other bias detected

Saveljev 2002

MethodsOpen RCT: multi-centre trial involving 3 centres in Moscow, Russia. Central randomisation list.
Participants124 (53 men and 71 women) aged 20-74 years, with venous leg ulcers only (2-10 cm diameter); ulcer present for at least 3 months.
InterventionsGroup 1: Detralex® 500 mg containing diosmin 450 mg and hesperidin 50 mg (2 tablets/day) + standard local therapy + compression bandaging (n = 62).
Group 2: standard therapy (elastic compression + local treatment) alone (n = 62).
Outcomes

Pre-specified outcomes
Proportion of fully healed ulcers, time to complete healing, variation of ulcer surface area, appearance of ulcers, symptoms and signs of vein diseases (heavy feeling in legs, pain, night cramps etc.), cost-effectiveness, adverse events.

Primary outcomes
A. Proportion of ulcers healed (paper states that missing data treated as if ulcer not healed)
Group 1:54/62
Group 2: 40/62

B. Ulcer surface area (% reduction from baseline):
Group 1: 82.1%
Group 2: 68.4%

C. Time to achieve complete healing (median):
Group 1: 90 days
Group 2: 119 days (P value 0.001)

Secondary outcomes
D. Cost-effectiveness:
The incremental cost-effectiveness ratio for MPFF group for the additional healing of another ulcer was RUB 10906.2, in comparison with standard therapy alone.

E. Hospitalisation:
Group 1: 1/62
Group 2: 3/62

F. Pain:
Reported no significant difference in severity or intensity in pain and frequency of night cramps between the two groups.

G. Adverse events:
Group 1: Arterial hypertension (4.8%); reduction in body mass (3.2%) - Total unwanted effects 21.0% (13/62)
Group 2: Reduction in body mass (1.6%) -Total unwanted effects 11.3% (7/62)

All other secondary outcomes were not reported.

NotesTreatment duration = 6 months.
Loss to follow up:
Group 1: 5/62
Group 2: 13/62
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskQuote: "In accordance with the codes of the randomised list, all patients were randomly divided...."
Comment: insufficient information on how the randomised list was generated.
Allocation concealment (selection bias)Unclear riskThe method of concealment was not described.
Blinding (performance bias and detection bias)
number of ulcers healed
Unclear riskThis was an open randomised trial.
Comment: there are insufficient data to judge whether the outcome assessor was blinded in this study, however, the non-blinding of participants and personnel was unlikely to introduce bias for this outcome.
Blinding (performance bias and detection bias)
wound surface area
Unclear riskThis was an open randomised trial.
Comment: there were insufficient data to judge whether the outcome assessor was blinded in this study.
Blinding (performance bias and detection bias)
time to achieve complete healing
Unclear riskComment: there were insufficient data to judge whether the outcome assessor was blinded in this study, however, the non-blinding of participants and personnel was unlikely to introduce bias for this outcome.
Incomplete outcome data (attrition bias)
number of ulcers healed
High risk

Quote: "5 patients in the main group and 13 patients in the control group left the trial early. Where this occurred, the patients were included in the category "has not healed"."

Comment: the missing data were greater in the control group (13/62) than the treatment group (5/62), and, since missing data were considered as "ulcer not healed", this would have given a greater treatment effect in favour of the treatment group.

Incomplete outcome data (attrition bias)
wound surface area
Unclear riskInsufficient reporting on how the trialists dealt with missing data.
Incomplete outcome data (attrition bias)
time to achieve complete healing
Unclear riskInsufficient reporting of how the trialists dealt with missing data.
Selective reporting (reporting bias)Low riskAll pre-specified outcomes were reported.
Other biasUnclear riskComment: there was significant baseline difference between the 2 groups. Even though the trialists reported that adjustment was made to take account of the differences, it was not clear how the calculation was done. Calculation of percentage reduction without proper adjustment in wound area from baseline would have given a favourable result to the flavonoid group which had a smaller baseline ulcer area.

Schultz-Ehrenburg 1993a

MethodsRCT: multi-centre trial in 12 centres in Germany and Belgium.
Participants55 patients with CVI; minimum ulcer size 1.5 cm in 1 axis; ulcer persisted for > 1 year; minimum ankle-arm index 0.8.
InterventionsGroup 1: hydroxyethylrutosides (1 g/day) + compression bandaging + local therapy (n = 28).
Group 2: placebo tablets + compression bandaging + local therapy (n = 27).
Outcomes

Pre-specified outcomes
Proportion of fully healed ulcers, variation of ulcer surface area, maximal calf and minimal ankle circumferences, ulcer pain, orthostatic complaints, adverse events.

Primary outcomes
A. Proportion of wound healed:
Group 1: 12/23
Group 2: 7/25

B. Ulcer surface area (mean % reduction from baseline (SD)):
Group 1: 81 (29)
Group 2: 52 (54)

C. Time for ulcers to achieve complete healing:
This outcome was not examined.

Secondary outcomes
D. Pain:
No reference to pain was made in the publication, even though the authors reported that this outcome would be reported elsewhere.

E. Adverse events:
Adverse effects were reported to be evenly distributed between the groups and none was serious.

All other secondary outcomes were not reported.

NotesTreatment duration = 3 months.
Loss to follow up:
Group 1: 5/28
Group 2: 2/27
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: the study was described as randomised, but information about the sequence generation process was not provided.
Allocation concealment (selection bias)Unclear riskThe method of concealment was not described.
Blinding (performance bias and detection bias)
number of ulcers healed
Unclear riskComment: the study was described as double-blind but information about blinding was not provided.
Blinding (performance bias and detection bias)
wound surface area
Unclear riskComment: the study was described as double-blind but information about blinding was not provided.
Blinding (performance bias and detection bias)
time to achieve complete healing
Unclear riskOutcome not assessed
Incomplete outcome data (attrition bias)
number of ulcers healed
Unclear riskComment: unbalanced withdrawals from the two groups.
Incomplete outcome data (attrition bias)
wound surface area
Unclear riskComment: unbalanced withdrawals from the two groups.
Incomplete outcome data (attrition bias)
time to achieve complete healing
Unclear riskOutcome not assessed
Selective reporting (reporting bias)Unclear riskQuote: "Results of the secondary end-points will be reported in a full publication elsewhere."
Comment: the authors did not report secondary outcomes, such as pain, probably because the paper is an extended abstract.

Schultz-Ehrenburg 1993b

MethodsRCT: multi-centre trial in 12 centres in Germany and Belgium.
Participants64 patients with CVI; minimum ulcer size 1.5 cm in 1 axis; ulcer persisted for > 1 year; minimum ankle-arm index 0.8.
InterventionsGroup 1: hydroxyethylrutosides (2 g/day) + compression bandaging + local therapy (n = 33).
Group 2: placebo tablets + compression bandaging + local therapy (n = 31).
Outcomes

Pre-specified outcomes
Proportion of fully healed ulcers, variation of ulcer surface area, maximal calf and minimal ankle circumferences, ulcer pain, orthostatic complaints, adverse events.

Primary outcomes
A. Proportion of wound healed:
No data provided. Reported as no difference between groups.

B. Ulcer surface area:
No data provided. Reported as no difference between groups.

C. Time for ulcers to achieve complete healing:
This outcome was not reported.

Secondary outcomes
D. Pain:
No reference to pain was made in the publication, even though the authors reported that this outcome would be reported elsewhere.

E. Adverse events:
Adverse events were reported to be evenly distributed between the groups, and none was serious.

All other secondary outcomes were not reported

NotesTreatment duration = 3 months.
Loss to follow up:
Group 1: 5/33
Group 2: 2/31
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: the study was described as "randomised" but information about the sequence generation process was not provided.
Allocation concealment (selection bias)Unclear riskThe method of concealment was not described.
Blinding (performance bias and detection bias)
number of ulcers healed
Unclear riskComment: the study was described as double-blind but information about blinding was not provided.
Blinding (performance bias and detection bias)
wound surface area
Unclear riskComment: the study was described as double-blind but information about blinding was not provided.
Blinding (performance bias and detection bias)
time to achieve complete healing
Unclear riskOutcome not assessed
Incomplete outcome data (attrition bias)
number of ulcers healed
Unclear riskComment: unbalanced withdrawals from the two groups.
Incomplete outcome data (attrition bias)
wound surface area
Unclear riskComment: unbalanced withdrawals from the two groups.
Incomplete outcome data (attrition bias)
time to achieve complete healing
Unclear riskOutcome not assessed
Selective reporting (reporting bias)Unclear riskQuote: "Results of the secondary end-points will be reported in a full publication elsewhere."
Comment: the authors did not provide data for primary outcomes and secondary outcomes.
Other biasUnclear riskNo other bias detected

Stegmann 1986

MethodsRCT: multi-centre trial in 3 centres in Germany.
Participants107 men and women aged approximately 58 years.Venous ulcers; ambulatory venous pressure > 55 mm Hg; Refilling time < 8 seconds; lesion present.
InterventionsGroup 1: Paroven® containing hydroxyethylrutosides (2 × 500 mg tablets/day) + compression bandaging + local therapy (n = 55).
Group 2: compression bandaging and local topical therapy alone (n = 52).
Outcomes

Pre-specified outcomes
Proportion of fully healed ulcers, variation of ulcer surface area, circumference measurements of ankle, phlebitis induration, eczema, purpura, pains, night cramps, heavy and swollen legs, tingling sensations, paraesthesia.

Primary outcomes
A. Proportion of ulcers healed:
Group 1: 25/54
Group 2: 17/51

B. Ulcer surface area:
No data provided.

C. Time for ulcers to achieve complete healing:
This outcome was not examined.

Secondary outcomes:
D. Pain:
Pain was not specifically assessed, however, the study reported no significant differences between groups for venous symptoms (including pain).

All other secondary outcomes were not reported.

NotesTreatment duration = 6 weeks.
Loss to follow up:
Group 1:1/55
Group 2: 1/52
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: the study reported that patients received random allocation, but information about the sequence generation process was not provided.
Allocation concealment (selection bias)Unclear riskThe method of concealment was not described.
Blinding (performance bias and detection bias)
number of ulcers healed
Unclear riskThe method of blinding was not described.
Blinding (performance bias and detection bias)
wound surface area
Unclear riskThe method of blinding was not described.
Blinding (performance bias and detection bias)
time to achieve complete healing
Unclear riskOutcome not assessed
Incomplete outcome data (attrition bias)
number of ulcers healed
Low riskComment: missing outcome data balanced across the two groups (one missing from each group).
Incomplete outcome data (attrition bias)
wound surface area
Unclear riskThis outcome was not reported.
Incomplete outcome data (attrition bias)
time to achieve complete healing
Unclear riskOutcome not assessed
Selective reporting (reporting bias)High riskData for variation in ulcer surface area and several pre-specified secondary outcomes were omitted.
Other biasUnclear riskNo other bias detected

Zuccarelli 1996

MethodsRCT: study conducted in France.
Participants126 patients with venous ulcers; ankle-arm index > 0.9; no other disease impairing wound healing.
InterventionsGroup 1: high dosage of troxerutin (4,000 mg/day) + standard compression + local therapy (n = 63).
Group 2: placebo + standard compression + local therapy (n = 63).
Outcomes

Pre-specified outcomes
Proportion of fully healed ulcers, time to complete healing, variation of ulcer surface area.

Primary outcomes
A. Proportion of ulcers healed:
Group 1: 30/63
Group 2: 15/63

B. Ulcer surface area cm2 (mean change from baseline (SD)):
Group 1: 6.2 (9.4)
Group 2: 1.6 (3.1)

C. Time for ulcers to achieve complete healing:
Faster time to complete healing observed in treatment group (logrank test:P value < 0.01).

Secondary outcomes
No secondary outcomes were reported.

NotesTreatment duration = 3 months.
Loss to follow up: no reference.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: the study was described as "randomised", but information about the sequence generation process was not provided.
Allocation concealment (selection bias)Unclear riskThe method of concealment was not described.
Blinding (performance bias and detection bias)
number of ulcers healed
Unclear riskComment: the study was described as double-blind, but information about blinding was not provided.
Blinding (performance bias and detection bias)
wound surface area
Unclear riskComment: the study was described as double-blind, but information about blinding was not provided.
Blinding (performance bias and detection bias)
time to achieve complete healing
Unclear riskComment: the study was described as double-blind, but information about blinding was not provided.
Incomplete outcome data (attrition bias)
number of ulcers healed
Unclear riskComment: no information was given about the number lost to follow-up, even though the data were reported to be analysed according to ITT (it is not clear what assumptions were made for this ITT analysis).
Incomplete outcome data (attrition bias)
wound surface area
Unclear riskComment: no information on the number lost to follow-up was given, even though the data were reported to be analysed according to ITT (it is not clear what assumptions were made for this ITT analysis).
Incomplete outcome data (attrition bias)
time to achieve complete healing
Unclear riskComment: no information on the number lost to follow-up was given, even though the data were reported to be analysed according to ITT (it is not clear what assumptions were made for this ITT analysis).
Selective reporting (reporting bias)Unclear riskOutcomes of interest (number with complete healing and time to healing) were incompletely reported.
Other biasUnclear riskNo other bias detected

Zuccarelli 2004

  1. a

    Abbreviations

    CVI = chronic venous insufficiency
    EUR = Euros
    ITT = intention-to-treat (analysis)
    MPFF = Micronised Purified Flavonoid Fraction
    RUB = Roubles

MethodsRCT: multi-centre trial involving 19 centres in France (12) and Germany (7).
Block randomisation from a central randomisation list, with stratification by country.
Participants202 (73 men and 129 women) patients aged 35-86 years; with venous ulcers 2-10 cm diameter in 1 or both legs; ulceration present for at least 3 months; irrespective of the stage of debridement.
InterventionsGroup 1: Daflon® 500 mg containing diosmin 450 mg, hesperidin 50 mg (2 tablets/day) + standardised local care + compression therapy (n = 103).
Group 2: placebo (2 tablets/day) + standardised local care + compression therapy (n = 99).
Outcomes

Pre-specified outcomes
Proportion of fully healed ulcers, time to complete healing, variation of the ulcer surface area, aspect of peri-ulcerous skin, number and duration of hospitalisations, adverse events.

Primary outcomes (ITT analysis)
A. Proportions of healed ulcers:
Group 1: 53/103
Group 2: 54/99

B. Ulcer surface area (mean (SD) mm2):
Group 1: changed from 7.04 (6.83) to 2.97 (7.09) at endpoint.
Group 2: changed from 7.43 (9.40) to 2.75 (7.71) at endpoint.

C. Time to complete healing:
No statistical difference was observed between the two groups.

Secondary outcomes
D. Hospitalisation:
Group 1: 14/103
Group 2: 16/99 (P value not reported).

E. Adverse events:
Group 1: 29/103
Group 2: 19/99

All other secondary outcomes were not reported.

Notes

Treatment duration = 6 months.

Loss to follow up:
Group 1: 17/103
Group 2: 13/99

Pains: not specifically assessed.

Cost-effectiveness: not specifically assessed.

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskCentrally-generated randomisation list.
Allocation concealment (selection bias)Low riskSequentially-coded, sealed envelopes used to conceal allocation.
Blinding (performance bias and detection bias)
number of ulcers healed
Low riskQuote: "Daflon 500 mg and placebo were supplied as identical tablets in appearance and taste: salmon pink film-coated tablets."
Comment: participants and study personnel (also outcome assessor) were blinded.
Blinding (performance bias and detection bias)
wound surface area
Low riskQuote: "Daflon 500 mg and placebo were supplied as identical tablets in appearance and taste: salmon pink film-coated tablets."
Quote: "The surface of the ulcer was assessed blindly by a computerised calculation."
Comment: participants, study personnel and outcome assessor were blinded.
Blinding (performance bias and detection bias)
time to achieve complete healing
Low riskComment: participants and study personnel (also outcome assessor) were blinded.
Incomplete outcome data (attrition bias)
number of ulcers healed
Low riskMissing data were imputed using method of last observation carried forward.
Incomplete outcome data (attrition bias)
wound surface area
Low riskMissing data were imputed using method of last observation carried forward.
Incomplete outcome data (attrition bias)
time to achieve complete healing
Low riskMissing data were imputed using method of last observation carried forward.
Selective reporting (reporting bias)Low riskAll pre-specified outcomes were reported.
Other biasUnclear riskNo other bias detected

Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion
Balmer 1980Paper in German. English abstract. Participants did not have venous leg ulcers.
Belcaro 2005Not a randomised trial.
Gilly 1994Participants did not have venous leg ulcers.
Mann 1981Not a randomised trial.
Neumann 1996After trying several sources we were unable to obtain the full text of this paper for appraisal.
Shrivastava 2011Trial examined mixed wounds and did not present data on venous leg ulcers separately.
Wright 1991This trial examined the prevention of ulcer recurrence in patients with recently healed ulcers.

Ancillary