Compression stockings for the initial treatment of varicose veins in patients without venous ulceration(review)

Transcription

1 Cochrane Database of Systematic Reviews Compression stockings for the initial treatment of varicose veins in patients without venous ulceration(review) Shingler S, Robertson L, Boghossian S, Stewart M Shingler S, Robertson L, Boghossian S, Stewart M. Compression stockings for the initial treatment of varicose veins in patients without venous ulceration. Cochrane Database of Systematic Reviews 2013, Issue 12. Art. No.: CD DOI: / CD pub3. Compression stockings for the initial treatment of varicose veins in patients without venous ulceration(review) Copyright 2013 The Cochrane Collaboration. Published by John Wiley& Sons, Ltd.

2 T A B L E O F C O N T E N T S HEADER ABSTRACT PLAIN LANGUAGE SUMMARY BACKGROUND OBJECTIVES METHODS RESULTS Figure Figure DISCUSSION AUTHORS CONCLUSIONS ACKNOWLEDGEMENTS REFERENCES CHARACTERISTICS OF STUDIES DATA AND ANALYSES APPENDICES WHAT S NEW CONTRIBUTIONS OF AUTHORS DECLARATIONS OF INTEREST SOURCES OF SUPPORT DIFFERENCES BETWEEN PROTOCOL AND REVIEW INDEX TERMS i

3 [Intervention Review] Compression stockings for the initial treatment of varicose veins in patients without venous ulceration Sarah Shingler 1, Lindsay Robertson 2, Sheila Boghossian 3, Marlene Stewart 4 1 Oxford Outcomes, Oxford, UK. 2 Department of Vascular Surgery, The Freeman Hospital, Newcastle upon Tyne, UK. 3 Public Health Sciences, The Medical School, The University of Edinburgh, Edinburgh, UK. 4 Centre for Population Health Sciences, University of Edinburgh, Edinburgh, UK Contact address: Sarah Shingler, Oxford Outcomes, Seacourt Tower, West Way, Oxford, OX2 0JJ, UK. shinglersl@hotmail.com. Editorial group: Cochrane Vascular Group. Publication status and date: New search for studies and content updated (no change to conclusions), published in Issue 12, Review content assessed as up-to-date: 20 August Citation: Shingler S, Robertson L, Boghossian S, Stewart M. Compression stockings for the initial treatment of varicose veins in patients without venous ulceration. Cochrane Database of Systematic Reviews 2013, Issue 12. Art. No.: CD DOI: / CD pub3. Background A B S T R A C T Compression hosiery or stockings are often the first line of treatment for varicose veins in people without either healed or active venous ulceration. Evidence is required to determine whether the use of compression stockings can effectively manage and treat varicose veins in the early stages. This is an update of a review first published in Objectives To assess the effectiveness of compression stockings for the only and initial treatment of varicose veins in patients without healed or active venous ulceration. Search methods For this update the Cochrane Peripheral Vascular Diseases Group Trials Search Co-ordinator searched the Specialised Register (last searched August 2013) and CENTRAL (2013, Issue 5). Selection criteria Randomised controlled trials (RCTs) were included if they involved participants diagnosed with primary trunk varicose veins without healed or active venous ulceration (Clinical, Etiology, Anatomy, Pathophysiology (CEAP) classification C2 to C4). Included trials assessed compression stockings versus no treatment, compression versus placebo stockings, or compression stockings plus drug intervention versus drug intervention alone. Trials comparing different lengths and pressures of stockings were also included. Trials involving other types of treatment for varicose veins (either as a comparator to stockings or as an initial non-randomised treatment), including sclerotherapy and surgery, were excluded. Data collection and analysis Two authors assessed the trials for inclusion and quality (SS and LR). SS extracted the data, which were checked by LR. Attempts were made to contact trial authors where missing or unclear data were present. 1

4 Main results Seven studies involving 356 participants with varicose veins without healed or active venous ulceration were included. Different levels of pressure were exerted by the stockings in the studies, ranging from 10 to 50 mmhg. One study assessed compression hosiery versus no compression hosiery. The other six compared different types or pressures of stockings. The methodological quality of all included trials was unclear, mainly because of inadequate reporting. The symptoms subjectively improved with the wearing of stockings across trials that assessed this outcome, but these assessments were not made by comparing one randomised arm of a trial with a control arm and are therefore subject to bias. Meta-analyses were not undertaken due to inadequate reporting and actual or suspected high levels of heterogeneity. Authors conclusions There is insufficient, high quality evidence to determine whether or not compression stockings are effective as the sole and initial treatment of varicose veins in people without healed or active venous ulceration, or whether any type of stocking is superior to any other type. Future research should consist of a large RCT of participants with trunk varices either wearing or not wearing compression stockings to assess the efficacy of this intervention. If compression stockings are found to be beneficial, further studies assessing which length and pressure is the most efficacious could then take place. P L A I N L A N G U A G E S U M M A R Y Wearing stockings to provide compression for the treatment of varicose veins Evidence from randomised controlled trials is not sufficient to determine if compression stockings as the only and initial treatment are effective in managing and treating varicose veins in the early stages. Varicose veins are widened veins that twist and turn and are visible under the skin of the leg. They generally do not cause medical problems although many sufferers seek medical advice. Symptoms that may occur include pain, ankle swelling, tired legs, restless legs, night cramps, heaviness, itching and distress from their cosmetic appearance. Complications such as oedema, pigmentation, inflammation and ulceration can also develop. Compression stockings are often the first line of treatment and come in a variety of lengths, knee length to full tights, and apply different pressures to support the flow of blood in the veins. Seven studies involving 356 participants with varicose veins and who had not experienced venous ulceration were included in this review. One study assessed compression hosiery versus no compression hosiery. The other six compared different types or pressures of stockings, ranging from 10 to 50 mmhg. The methodological quality of the included trials was unclear and not all studies assessed the same outcomes. One study included only pregnant women whilst other studies included participants who were on surgical waiting lists, that is, people who had sought medical intervention for their varicose veins. The participants subjective symptoms, and foot swelling and blood flow (physiological measures) improved in all of the studies that assessed these outcomes when stockings were worn, but these assessments were not made by comparing one randomised arm of the trial with a control arm in the same study. Conclusions from the individual studies regarding the optimum pressure provided by stockings were conflicting, although the results of one study suggested that lower pressured stockings (20 mmhg) may be as effective as higher pressured stockings (30 to 40 mmhg) for relieving symptoms. Conclusions regarding the optimum length of the stockings were inconclusive. No severe or long lasting side effects were noted. B A C K G R O U N D Description of the condition Varicose veins are tortuous, widened veins in the subcutaneous tissue of the lower limb (Campbell 2006). Varicose veins that 2

5 have arisen as a result of a known cause, such as a previous deep vein thrombosis, are known as secondary varicose veins (Golledge 2003). Theories on the cause of primary varicose veins range from initial structural weakness within the vein wall, which leads to vein dilation, to valve incompetence leading to pooling of the blood and vein dilation (Golledge 2003; London 2000). The Clinical, Etiology, Anatomy, Pathophysiology (CEAP) classification for chronic venous disorders was developed in 1994 by the American Venous Forum and was revised and disseminated in 2004 (Eklöf 2004). It serves as a systematic guide to the diagnosis and classification of chronic venous disorders, which was previously lacking. Today, most published papers on chronic venous disorders will use all or some of the CEAP classification (Eklöf 2004). There are six categories, ranging from small superficial thread like veins (C1) to active venous ulcers (C6). A classification of C0 is given where no clinical findings of venous disease are found. The CEAP classification is therefore used to classify and diagnose varicose veins, which if uncomplicated are classed as grade C2. A classification of C3 indicates varicose veins with oedema and a classification of C4 indicates varicose veins with skin changes due to venous disorders (Padberg 2005). In 2004, it was also decided to define a varicose vein as greater than 3 mm in diameter, to differentiate from reticular veins (Eklöf 2004). Although varicose veins are associated with low morbidity and mortality, many sufferers seek medical advice and there is some evidence that quality of life can be improved following treatment. Approximately 2% of NHS resources in the UK are spent on managing venous disease, with an estimated 20 to 25 million (excluding non-hospital costs) spent on operations for varicose veins in the year 2001 (Beale 2005). Many patients with varicose veins are asymptomatic, but some experience aching, dullness in the lower limbs, itching, throbbing, distress about the cosmetic appearance and swelling (Metcalfe 2008). Around 5% will develop complications including haemorrhage, thrombophlebitis, oedema, skin pigmentation, atrophie blanche, varicose eczema, lipodermatosclerosis and ulceration (Beale 2005). Prevalence of the condition varies widely between studies from different countries (up to 56% in men and 60% in women) and few studies have measured the incidence within the general population (Robertson 2008). Currently there is geographical variation within the UK as to whether varicose veins (CEAP classification C2 to C4) are treated on the NHS, which is often dependent on financial restrictions (Edwards 2009). Description of the intervention Lower limb compression has been used to treat varicose veins since biblical times (Pierson 1983). Nowadays, compression hosiery or stockings (knee and thigh length and full tights) are often first line treatment for varicose veins, especially in primary care, even though evidence supporting their use is limited (Ramelet 2002; Tisi 2007). Although compression may be achieved by different modalities, this review is restricted to compression stockings. There are currently five classifications of compression stockings based on the different manufacturers, who apply different thresholds of pressure as measured at the ankle for different classes: the British Standard, German Standard, French Standard, draft European standard and USA Standard. The draft European standard was developed to try and ensure consistency within the European Union but, due to lack of consensus, this has not occurred (Rajendran 2007). Each classification has between three and four classes according to differing levels of pressure as set by the manufacturer (Palfreyman 2009). This sub-bandage pressure range varies greatly between different countries, leading to variation in the different norms for stockings internationally. For example, for class one alone, the British Standard bandage pressure can be between 14 and 17 mmhg whereas the German Standard can be between 18 and 21 mmhg. Stockings come in a variety of lengths, knee length, thigh length and full tights. The optimum length and pressure are not known, but studies suggest that a lower compression pressure (20 to 30 mmhg) is tolerated better than are higher levels (Beale 2005). How the intervention might work Optimum management of varicose veins requires accurate diagnosis and identification of the source of venous incompetence, and treatment should aim to abolish venous reflux and relieve symptoms (Beale 2005). Graded compression works by applying a controlled pressure to the skin. This supports the superficial venous system by exerting an external pressure that is greatest at the ankle (minimum of 14 mmhg) and decreases up the leg (Johnson 2002). By reducing venous capacity and increasing venous velocity in the deeper veins, venous stasis and reflux are reduced helping to reduce the severity of varicose veins and the associated symptoms (Walker 2007). It is, however, important to be aware that the effect of compression stockings is influenced by many factors. For example, the material (both compression class and elasticity) of the stocking, the size and shape of the leg, and the activity of the wearer can all influence the compression applied (Rabe 2008). Why it is important to do this review Although compression therapy is widely used in the management of varicose veins, there is still conflicting, poor quality evidence on its effectiveness. There is evidence that patients with varicose veins without healed or active venous ulceration can go on to develop severe complications and consume a large amount of NHS resources. If treatments such as compression stockings can be used to treat varicose veins in the early stages, morbidity and resource use may ultimately be reduced. This review is important to assess the efficacy of compression stockings in the treatment of varicose veins in people without healed or active venous ulceration, including 3

6 the optimum length and pressure of stockings and whether the use of stockings has any effect on the symptoms related to varicose veins. O B J E C T I V E S To assess the effectiveness of compression stockings for the only and initial treatment of varicose veins in patients without healed or active venous ulceration. M E T H O D S Criteria for considering studies for this review Types of participants We included adults (aged 18 years and above) of both sexes who suffered from varicose veins without having healed or active venous ulceration in the lower limb(s). The classification of varicose veins was performed according to CEAP. If methods such as duplex scanning were used to assist in the diagnosis, a clinical diagnosis of trunk or stem varicose veins (CEAP C2) was also required, with or without ankle oedema or minor skin changes (CEAP C3 & C4). Where venous leg ulcers (CEAP C5 and C6) or other severe complications were present, the participants were not included. In studies where the CEAP classification was not used, in early studies for example, a diagnosis by a clinician of trunk varices or varicose veins was sufficient. Co-morbidity or pregnancy did not influence the decision to include or exclude a trial. Trials including participants who had bilateral varicose veins but where compression was only applied unilaterally were included. Types of studies We considered all randomised controlled trials (RCTs) in people with varicose veins and without healed or active venous ulceration (CEAP classification C2 to C4) that assessed: compression stockings versus no treatment, compression stockings versus placebo compression stocking (Tubigrip, for example, which does not provide graduated pressure), or compression plus a drug intervention versus the drug intervention alone. We also included studies that compared different lengths and pressures of compression stockings. We included parallel group and cross-over trials. There were no language restrictions. Trials in which compression stockings were compared with interventions other than no treatment, placebo compression stockings or stockings of another length or pressure (for example sclerotherapy, surgery, laser therapy, exercise, balneotherapy or hydrotherapy) were excluded. Existing Cochrane reviews have assessed the treatment of varicose veins using surgical techniques including sclerotherapy (Rigby 2004; Tisi 2006) as well as assessing the best form of prevention and management of venous disease in certain subsets of the population such as pregnant women (Bamigboye 2007) and those travelling by air (Clarke 2006). Trials in which compression stockings were assessed against no treatment or a placebo stocking as an adjuvant to primary (initial) treatment with surgery, sclerotherapy or laser therapy were also excluded, as were those assessing other methods of compression such as pneumatic compression techniques. Studies of participants with a CEAP classification of C5 or C6, indicating venous ulcers, were also excluded due to overlap with an existing Cochrane review (Kolbach 2003). Types of interventions Interventions included compression from any type of hosiery that exerted a graduated pressure on the lower limb(s). We included any length (knee, thigh, full tights) and grade of pressure. Compression had to be from a graduated pressure stocking and not from a more general support bandage such as Tubigrip. We compared the intervention with no intervention at all, a placebo stocking such as Tubigrip, or both. We also included trials that involved participants taking a drug intervention when the drug alone was compared against drug plus compression stocking, allowing the effect of the compression stocking to be assessed. We assessed comparisons between different lengths or different grades of pressure, or both, where trials compared stockings in two arms of the same trial. Types of outcome measures Primary outcomes Change in symptoms such as aching and itching, reported by participants Physiological measures used to monitor the impact of the intervention on varicose vein status, such as venous pressure at the ankle Secondary outcomes Complications: these included direct complications or side effects from wearing the stockings, e.g. itching, increase in temperature 4

7 Compliance: this included ensuring the patient was wearing the stockings and that they were being worn correctly, measured subjectively e.g. a report from the patient that they had stopped wearing the stockings Quality of life: where this was measured using a standard questionnaire One review author (SS) attempted to contact three trial authors or co-authors for additional information in order to allow a decision regarding inclusion or exclusion of their studies in this review. This included not being able to tell if all participants had varicose veins or whether subgroup analysis had been performed. Only one study author provided additional information to permit a judgment to be made (Benigni 2003). Search methods for identification of studies Electronic searches For this update the Cochrane Peripheral Vascular Diseases Group Trials Search Co-ordinator (TSC) searched the Specialised Register (last searched August 2013) and the Cochrane Central Register of Controlled Trials (CENTRAL) 2013, Issue 5, part of The Cochrane Library, ( See Appendix 1 for details of the search strategy used to search CENTRAL. The Specialised Register is maintained by the TSC and is constructed from weekly electronic searches of MEDLINE, EMBASE, CINAHL, AMED, and through handsearching relevant journals. The full list of the databases, journals and conference proceedings which have been searched, as well as the search strategies used are described in the Specialised Register section of the Cochrane Peripheral Vascular Diseases Group module in The Cochrane Library ( Searching other resources For the original version of the review (Shingler 2011) we contacted companies, specialists in the field and journal article authors in order to obtain information regarding unpublished data, or to clarify information where needed. We checked reference lists of appropriate cited studies and reviewed conference proceedings and abstracts from relevant organisations. Data collection and analysis Selection of studies We used the search strategy described to obtain titles and abstracts of studies that were potentially relevant to this review. Two review authors (SS and LR) independently screened the titles and abstracts to identify studies and select trials for possible inclusion in the review. Full text articles were obtained where the above inclusion criteria were met, where further clarification was required, or if translation was necessary. Any disagreements were resolved first by discussion between review authors and, where required, through consultation with the third author (SB). Data extraction and management Two review authors (SS and LR) independently reviewed, extracted and summarised information from the studies using standardised data extraction forms specific to the Cochrane Peripheral Vascular Disease Group. The studies were checked for accuracy by a third review author (SB) where required. The following information was collected: 1) methods (study design, method of randomisation, concealment of allocation, blinding, power calculations, and source of funding); 2) participants (number, age, sex, inclusion and exclusion criteria); 3) interventions (treatment, control or placebo, duration); 4) outcomes (primary and secondary). Assessment of risk of bias in included studies Risk of bias tables were completed for each included study along with narrative in the text to assess the quality of the included studies. Tables were completed independently by two review authors (SS and LR) with a third (SB) available to resolve disagreements, in accordance with the Cochrane Handbook for Systematic Reviews of Interventions version 5.0.2, Chapter 8 (Higgins 2009). The minimum level of study quality included in this review was a RCT. Where this level of quality was not met, the trials were excluded. If enough high quality studies were found, sensitivity analysis was planned. The Risk of bias tables included six domains: adequate sequence generation, allocation concealment, blinding, incomplete outcome data addressed, free of selective reporting and free of other bias. Each study was independently reviewed by SS and LR and given a judgement of low risk of bias, high risk of bias, or unclear for each domain, along with a description as to why the judgement was made. Measures of treatment effect In line with the Cochrane Handbook for Systematic Reviews of Interventions, for dichotomous outcomes, results were expressed as risk ratio (RR) with 95% confidence interval (CI). For continuous scales of measurement, the mean difference (MD) was used, or the standardised mean difference (SMD) if different scales were used (Higgins 2009). 5

8 Unit of analysis issues All types of randomised controlled trials were included. In the case of a cross-over trial, data from all periods were included. Dealing with missing data Where data were missing or unclear, attempts were made to contact the author of the study to obtain the information. Where we were unable to obtain missing data, assumptions such as assuming normal distribution to use the median value as a mean value were used. Any assumptions made were described with the relevant analysis. Assessment of heterogeneity Where applicable, meta-analyses were considered for studies that showed sufficient homogeneity in terms of participants, interventions and outcomes. Where appropriate, assessment of heterogeneity between trials was assessed in the meta-analysis by the Chi² test and the I² statistic. Results for each study were expressed as MD with 95% CI and combined for meta-analysis with RevMan software (version ) (RevMan 2008). Data synthesis Meta-analyses were planned where appropriate data were available. Ultimately, no studies were entered into meta-analyses due to the lack of compatible data from the different studies. If meta-analyses had been completed, many assumptions would have had to be made due to the poor levels of reporting. Subgroup analysis and investigation of heterogeneity No subgroup analyses were pre-planned. Sensitivity analysis If sufficient high quality trials were available, sensitivity analysis was planned to determine the robustness of the review findings. R E S U L T S Assessment of reporting biases Where applicable, it was planned to assess publication bias via a funnel plot of outcomes. These can show publication bias by an asymmetrical output (Higgins 2009). Ultimately, a funnel plot was not carried out in this review due to the very small number of studies included in the statistical analysis. Description of studies Results of the search See Figure 1. 6

9 Figure 1. Study flow diagram. 7

10 For this update there was one additional study identified from the Specialised Register and 194 records from CENTRAL. Following screening of titles one additional study was identified for possible inclusion (Mosti 2011). Included studies (See also the Characteristics of included studies tables) A total of nine reports of seven studies were included. The seven included studies were published between 1980 and 2001 and in total involved 356 participants with varicose veins without healed or active venous ulceration. All the studies were set in Europe: five of the seven studies were in the United Kingdom (Anderson 1990; Chant 1985; Chant 1989; Coughlin 2001; Jones 1980), the study by (Jungbeck 1997) was set in Sweden, and the study by (Chauveau 2000) was set in France. All included studies were randomised controlled trials but three had a cross-over design (Anderson 1990; Chauveau 2000; Jones 1980). Types of participants All participants were diagnosed with primary varicose veins without healed or active venous ulceration as specified in the inclusion criteria. None of the studies used the CEAP classification; mostly this was because they were conducted prior to the development of the classification. The studies varied in the amount of detail provided to describe their exclusion and inclusion criteria, from providing no detail at all to being very specific. In six of the studies all participants had varicose veins and were randomised into different groups either to compare types of stocking or different pressures, or to compare wearing stockings versus not wearing stockings. Only one trial (Jones 1980) included other groups of participants who were not diagnosed with varicose veins. This trial was included because the groups were clearly distinguished at entry and in the results, so those participants with varicose veins could be identified for inclusion in our analyses. Four of the studies recruited patients from surgical waiting lists (Anderson 1990; Chant 1985; Chant 1989; Jones 1980), two from outpatient departments (Chauveau 2000; Jungbeck 1997) and one from a maternity setting as they were pregnant (Coughlin 2001). All included participants were outpatients. Five of the studies included both men and women; the other two included only female participants (Chauveau 2000; Coughlin 2001), one because the study was on pregnant women. Only two of the studies stated the numbers of male and female participants (Jones 1980; Jungbeck 1997), with the others stating that randomised groups were evenly matched for sex. In the two studies where the participants sex was stated, there were more women (87) than men (10), possibly reflecting a greater prevalence of symptomatic varicose veins in women or that women seek medical help for varicose veins more readily than men. The age range also varied between studies. Some studies did not specify age (Chant 1985; Chant 1989; Coughlin 2001; Jones 1980) and in the others the age ranged from 20 to 82 years, with mean ages of 40 years (Anderson 1990) and 52 years (Chauveau 2000) in two of the studies. Interventions Five of the studies used knee length graduated compression stockings (Chant 1985; Chant 1989; Chauveau 2000; Jones 1980; Jungbeck 1997). Coughlin 2001 used compression tights and Anderson 1990 used full length stockings. Anderson 1990 also used other treatments in the trial, including Paroven and a placebo tablet. This trial assessed the effect of compression stockings and Paroven, alone and in combination, and the groups were identifiable for analyses. A majority of the other studies assessed one type or pressure of stocking against another (Chant 1985; Chant 1989; Chauveau 2000; Jones 1980; Jungbeck 1997), or both. Only one study (Coughlin 2001) randomised participants to wear one type of compression tights or not and then followed up participants to assess compliance. The following compression stockings or tights were used in the studies: class 1 compression tights (Coughlin 2001), full length hosiery giving a pressure of 30 to 40 mm Hg at the ankle (Anderson 1990), Sigvaris stockings giving a pressure of 30 to 40 mmhg at the ankle or 40 to 50 mmhg at the ankle (Chant 1985), an Eesiness NHS two-way stocking (20 mmhg at the ankle) and Sigvaris medium stocking (30 to 40 mmhg at the ankle) (Jones 1980), French class 1 (10 to 15 mmhg) and French class 2 (15 to 20 mmhg) below-knee stockings (Chauveau 2000), Jobst Medical Leg wear knee-hi class 1 (20 mmhg at the ankle) and Jobst Medical Leg wear knee-hi class 2 (30 mmhg at the ankle) (Jungbeck 1997), and Sigvaris and Medi Plus support hose (Chant 1989). Some of the studies did not provide the amount of pressure provided by the hosiery or the class of stocking. It was therefore difficult to make comparisons between the types of stocking. Outcomes Some studies assessed more than one outcome. Four of the included studies used the change in symptoms as an outcome measure (Anderson 1990; Chant 1985; Jones 1980; Jungbeck 1997). Two of these used a visual analogue scale (VAS) to determine change in symptoms before and after wearing stockings (Anderson 1990; Jungbeck 1997). Only Anderson 1990 reported the results of the VAS at the level of individual symptoms. Jungbeck 1997 grouped all symptoms together and reported the results before and after intervention for the two different classes of stocking used in the trial. The other two studies used self-reporting of symptoms as their outcome measure (Chant 1985; Jones 8

11 1980). Although Jones 1980 reported on symptoms, it was not the study s primary outcome and no data were reported, just a line stating patients with varicose veins were relieved of aching symptoms by both sorts of stocking. Chant 1985 stated the number of participants in each stocking group who reported an improvement in symptoms, as well as the number of participants that were removed from the surgical waiting list. Three studies reported the change in physiological measures (Chauveau 2000; Jones 1980; Jungbeck 1997). Jungbeck 1997 and Jones 1980 used foot volumetry before and after wearing stockings. Chauveau 2000 used air plethysmography to look for a tourniquet effect of stockings. The only study to assess complications of wearing compression stockings as their main outcome was Chauveau This study determined whether below-knee stockings (class 1 (10 to 15 mmhg) and class 2 (15 to 20 mmhg)) impeded venous return by a tourniquet effect using air plethysmography with venous occlusion. Anderson 1990 described side effects of headache, abdominal pain and nausea, which occurred throughout the trial, but these could have been caused by the other interventions that were given, which included Paroven and a placebo tablet. Chant 1989 assessed compliance with compression stockings and also recorded the reasons why participants did not wear them. These included caused irritation, which affected four of the 40 participants. No trials revealed severe side effects. Two studies assessed the outcome measure of compliance (Chant 1989; Coughlin 2001). Coughlin 2001 investigated the acceptability and usage of compression tights in pregnant women at various time points up to six weeks post partum via self-reporting by the participants. Chant 1989 assessed the compliance of two different types of stocking (Sigvaris and Medi Plus) by prescribing participants the stockings and then visiting them at home on a random basis, on average six weeks later, to determine whether or not they were wearing the stockings. No study assessed the outcome measure of quality of life. Length of studies Studies varied in length from one day (wearing each of the two types of stocking for 15 minutes) (Chauveau 2000) to throughout pregnancy and up to six weeks post partum (Coughlin 2001). For a majority of the studies, it was difficult to tell for how long participants actually wore the stockings, and in his review we used the time period taken at when the assessors reviewed participants. However, this relied on the participants actually wearing the stockings between assessments. Excluded studies (See also the Characteristics of excluded studies tables) For this update one additional study (Mosti 2011) was identified for possible inclusion and subsequently excluded. In total forty-three articles of 40 studies were identified as excluded, some for more than one reason. Seventeen of the articles included participants who had a surgical intervention or sclerotherapy, either as a randomised alternative to stockings or as an initial treatment prior to adjuvant therapy with compression (Abramowitz 1973; Biswas 2007; Bond 1997; Bond 1999; Hamel-Desnos 2008; Hamel-Desnos 2010; Houtermans-Auckel 2009; Isiklar 2003; Kline 1972; Makin 1982; Mariani 2011; Melrose 1979; Mosti 2009; O Hare 2010; Perhoniemi 1983; Raraty 1999; Weiss 1999). Of the 40 excluded studies, 15 were not randomised controlled trials (RCTs) (Acsady 1996; Hirai 2002; Horvath 1983; Ibegbuna 1997; Isiklar 2003; Kakkos 2001; Kline 1972; Lascasas 2009; Leon 1993; Mauss 1969; Norgren 1988; Pierson 1983; Raju 2007; Szendro 1992; Zhang 2004). Six of the articles (Austrell 1995; Derman 1989; Guest 2003; Rabe 2010; Schul 2009; Thaler 2001) did not include participants with diagnosed varicose veins at recruitment; generally, the participants in these articles were either healthy or were diagnosed with other forms of venous disease, such as ulceration. Some of the excluded articles did not use hosiery as their method of compression, Griffin 2007 and Kakkos 2001 used pneumatic compression as their intervention, and Acsady 1996 did not use compression at all. The study by Benigni 2003 would have met the inclusion criteria of this review but the trial included participants with CEAP classification C1 to 3. The author was contacted but reported no subgroup analysis of participants with a CEAP classification of C2 or above. Therefore, this study had to be excluded. This was the same for the study by Gandhi 1984 in which participants included people with varicose veins and clinical gravitational disease. Those with varicose veins could not be separated in the results, thereby precluding inclusion of the study in this review. The final study which could potentially have been included was by Mosti However, this study included participants with CEAP classification C2 to C5 and no subgroup analysis of participants with a CEAP classification of C2 to C4 was reported. The study author was contacted to see if this analysis had been undertaken, but no reply was received. Risk of bias in included studies See also Risk of bias tables and Figure 2 9

12 Figure 2. Methodological quality summary: review authors judgements about each methodological quality item for each included study. 10

13 Allocation None of the included studies described the allocation sequence or the concealment of allocation sufficiently to make a clear judgement and were therefore marked as unclear. Only the study by Chant 1985 described how allocation was decided: Patients were selected sequentially from a surgical waiting list, starting with those added most recently. However, no other information on the type of sequence or how the sequence was produced was provided and, therefore, it was unclear as to whether it was adequate or not. Blinding It is difficult to blind participants when using compression stockings as an intervention. It is obvious to the participant whether they are wearing a compression stocking or not. It is likely that it was for this reason that blinding of the participants was rarely stated in the studies. If a study compared different types of stocking, blinding of the participants was not assumed unless stated. If a study used only one type and pressure of stocking, and participants had been randomised to wear or not wear stockings, it was assumed that the participants were not blinded. Only one study stated that the participants were assessed blind (Chant 1985). However, the study did not state whether participants were also blinded, and as participants were randomised to two different pressures of stocking it cannot be assumed that they were or were not blinded. In all other included studies there was no discussion around blinding and, therefore, all the studies were marked as unclear. Incomplete outcome data For nearly all of the included studies there were insufficient data available to make a low or high risk of bias judgement on the completeness of outcome data. Therefore, six out of the seven included studies were marked as unclear. In some of the studies, although reasons for drop-outs or withdrawals were given, it was not possible to tell from the results whether all participants were included or had completed the course of treatment. For example, in the studies by Chauveau 2000 and Anderson 1990 the results tables provided mean measures, not permitting the reader to tell if all participants results were included. The only study that allowed the reader to determine that all outcome data were accounted for was Chant This study provided results for the outcome compliance by providing the actual number of participants who were wearing stockings after six weeks. Data for this outcome were complete. Selective reporting None of the included studies provided sufficient information to permit a low or high risk of bias judgement on selective reporting and were therefore classed as unclear. None of the study protocols were available and text reporting pre-specified outcomes was not clear. Other potential sources of bias All studies were marked as unclear. Although none appeared to have major risks of bias, due to insufficient reporting there was not enough information to permit a judgement of low risk of bias. Nearly all the included studies gave acknowledgement to companies providing stockings: Chant 1985 thanked Sigvaris Company; Jones 1980 thanked Ganzoni and Cie AG, Switzerland; Anderson 1990 thanked Zyma (UK) and Ganzoni (Switzerland) for providing Sigvaris stockings, Paroven and placebo capsules; and Jungbeck 1997 thanked Beiersdorf for supplying stockings and their contribution to the study. The participants in Chant 1989 were prescribed stockings tailored to their degree of venous insufficiency, which could mean people were given different types and pressures of stocking. This may have had an impact on their compliance. In studies by Chant 1985 and Coughlin 2001, there were high drop-out rates following the initial prescription of stockings (33% and 32% respectively), which may have affected the results. All the included studies were relatively small. The largest study (Chant 1985) included only 104 participants. Two studies ( Chauveau 2000; Jones 1980) had very small numbers of participants with varicose veins (both n = 10). Effects of interventions The main objective of this review was to assess the effectiveness of compression stockings in the initial treatment of varicose veins in individuals without venous ulceration. Seven studies assessed effectiveness using one or more of the pre-specified outcomes for this review. Primary outcomes Change in symptoms Four studies assessed this outcome (Anderson 1990; Chant 1985; Jones 1980; Jungbeck 1997). Two used a visual analogue scale (VAS) to report and evaluate symptoms including pain, ankle swelling, tired legs, restless legs, night cramps, heaviness, itching and distress from the cosmetic appearance. One study combined 11

14 their results, providing only a median percentage change in all symptoms before and after wearing stockings (Jungbeck 1997), while the other study provided VAS results for change in each of the individual symptoms using means and standard errors (Anderson 1990). The other two studies (Chant 1985; Jones 1980) used selfreporting of symptoms and gave a more general result as to whether symptoms had improved or not at different time points in each study. No objective measure was used. Therefore, the results of the studies could not be combined. Individually, all studies reported a subjective improvement in symptoms by the end of the trial, but these were not analysed comparing the two randomised arms of the trials and were therefore subject to bias. Jungbeck 1997 provided limited reporting of symptom change (all subjective symptoms) combined in participants wearing class 1 (20 mmhg) versus class 2 (30 mmhg) stockings. Results were given as patients assessments of subjective symptoms expressed as median of VAS (in per cent). For class 1 stockings, there was a reduction of 28.2%, and for class 2 stockings there was a reduction of 31.3%. There was no significant difference between the two groups. Anderson 1990 found no statistically significant differences in VAS scores after any of the four treatments. However, the use of Paroven and compression hosiery on their own appeared to reduce symptoms more than the placebo, as seen in the VAS scores. For example, for the symptom of swelling the mean VAS score for the placebo was 35.3; the mean score for hosiery plus placebo was 28.2; the mean score for Paroven was 31.5; and the mean score for hosiery plus Paroven was These observed improvements were true for all symptoms with the exception of distress from cosmetic appearance. Chant 1985 provided limited information and data, reporting only the number of participants who experienced improved symptoms with wearing 30 to 40 mmhg stockings (42 out of 53 participants) and 40 to 50 mmhg stockings (40 out of 51 participants). Jones 1980 provided no data for changes in symptoms comparing the different intervention groups of the trial. One study was a cross-over trial (Jones 1980) where participants wore each type of stocking for three weeks, with a week of not wearing stockings in between. The other trial (Jungbeck 1997) was a parallel group study, where participants were randomised to one of two different types of stocking for eight weeks. Because there were only two studies that assessed these measures, and they were of different methodology, it was decided not to combine them in a comparison table or meta-analysis. Both studies also had poor levels of reporting: Jungbeck 1997 used only medians (and 95% confidence intervals (CI)) to report initial values and values at eight weeks for both class 1 (20 mmhg) and class 2 (30 mmhg) stockings. The results in Jones 1980 required the reader to read values off a graph for each physiological measure. These results would have to be used very cautiously if a meta-analysis had been produced. Jones 1980 reported in their varicose vein group (n = 10) an improved performance for EV (20% (20 mmhg) and 8% (30 to 40 mmhg)), EVR (30% and 15%), Q (50% and 20%) and Q/EVR (20% and 35%) compared with the baseline values. Jungbeck 1997 reported significant improvement in EV and EVrel after wearing class 1 (20 mmhg) stockings for eight weeks compared with the baseline values (EV: baseline median 12.4, 95% CI 10.5 to 12.7 versus median 14.7, 95% CI 12.8 to 16.4, P < 0.05; EVrel: median 1.09, 95% CI 0.87 to 1.23 versus median 1.38, 95% CI 1.10 to 1.56, P < 0.01). Jungbeck 1997 reported no significant differences for Q and Q/EVrel or for EV, EVrel, Q and Q/EVrel in class 2 (30 mmhg) stockings. One study used air plethysmography to measure arterial inflow (AI), venous volume (V50) and maximum venous outflow (MVO) as part of the assessment for a tourniquet effect whilst wearing stockings (Chauveau 2000). No significant changes were found in any of these measures whilst wearing class 1 (10 to 15 mmhg) stockings. Significant reductions were found for arterial inflow (mean value of 1.22 ml/s to 0.93 ml/s) and venous volume (98 ml to 80 ml) whilst wearing class 2 (15 to 20 mmhg) stockings. However, this study was not designed to compare the effectiveness of the stocking. Physiological measures Three studies assessed this outcome (Chauveau 2000; Jones 1980; Jungbeck 1997). Two studies used foot volumetry to measure expelled volume (EV), the relative expelled volume related to 100 ml of foot volume (EVR or EVrel), rate of refilling of the foot after exercise (Q), and the quotient combining both aspects of functional assessment (Q/ EVR or Q/EVrel). This quotient is a measure of the blood expelled by the musculovenous pump in the leg and the prevention of reflux by venous valves, both of which are important for the prevention of varicose veins. Other measures were taken but only these four were consistent between the two studies (Jones 1980; Jungbeck 1997). Both studies investigated differences between stockings providing 20 mmhg pressure and those providing 30 to 40 mmhg pressure. Secondary outcomes Complications Only one small study (n = 10) assessed this as a main outcome (Chauveau 2000). Two different pressures of stocking (class 1 (10 to 15 mmhg) and class 2 (15 to 20 mmhg)) were tested to assess if there was any difference in producing a tourniquet effect whilst they were being worn. A tourniquet effect is a possible side effect of wearing compression stockings, which could impede venous return (Chauveau 2000). For people who already suffer from arteriopathy such a side effect could potentially lead to ischaemia of the surrounding tissues. No tourniquet effect of class 1 or 2 belowknee stockings was found. 12

15 Two other studies reported on side effects. Anderson 1990 reported the side effects of abdominal pain, headache and nausea; however this was a four-arm cross-over trial that also included taking Paroven and a placebo tablet, which may have caused the reported side effects. The study reported that the side effects were similar in all four arms and none were severe or long lasting, but no results were presented. In Coughlin 2001, in which compliance was the main outcome, irritation was given as a reason why some participants stopped wearing stockings but no information regarding the severity was given. The authors stated that abdomen size was likely to be the cause of the discomfort because the participants were pregnant women and compression tights were provided as the intervention. Compliance Two studies reported on compliance (Chant 1989; Coughlin 2001). Both studies had a high initial drop-out rate. In Coughlin 2001, 33% of participants (pregnant women with varicose veins) did not want to take part in the study once randomised to wearing compression tights. In Chant 1989, 30% of randomised participants (22 out of 66 patients wanting symptomatic relief of varicose veins prior to elective surgery for the condition) did not collect the compression stockings they were randomised to. Compliance was assessed differently in the two trials, with Coughlin 2001 relying on self-reports from the women as to when they stopped wearing the tights. As their pregnancies progressed, increasing numbers of women rejected the compression tights: 7 women wore them up to 20 weeks, 4 to 28 weeks, 14 to 34 weeks, 2 to term and 4 to 6 weeks post partum. No statistical calculations were performed. Chant 1989 completed random visits to participants houses approximately six weeks after the stockings had been prescribed to assess whether participants were wearing the stockings. Overall compliance with treatment for the 66 patients entered into the trial was 32%. Both studies found a general low compliance, but in the study by Chant 1989 neither type of stocking studied produced significantly more or less compliance than the other (Chi 2 = 1.70). Generally, non-compliance appeared to be greatest when the hosiery was initially prescribed, and reasons given by participants for not wearing them included discomfort, application, appearance, not effective and that they caused irritation. Coughlin 2001 concluded that tights were unacceptable in pregnant women due to their abdominal size. Meta-analysis was not carried out because of the described differences in compliance assessments between the studies. Quality of life No studies assessed or reported on this outcome directly. All the included studies had an insufficient level of reporting and relatively small sample sizes. Therefore this review is unable to make comment with regards to the effect of compression stockings for varicose veins in people without healed or active venous ulceration for the outcomes assessed: change in symptoms, change in physiological measures, complications, compliance and quality of life. D I S C U S S I O N Summary of main results This review has summarised the evidence for the use of graduated compression hosiery or stockings for the treatment of primary varicose veins in patients without venous ulceration. Of the 79 articles identified, only seven studies were eventually included, with a total of 356 participants with varicose veins without healed or active venous ulceration. All but one of the outcomes of the review (quality of life) were assessed in one or more of the included studies. No severe or long lasting side effects were noted. Subjectively, participants symptoms and physiological measures improved in all of the studies that assessed these outcomes when stockings were worn, but these assessments were not made by comparing one randomised arm of the trial with a control arm, and so are subject to bias. No conclusions regarding the optimum length of compression stocking can be made as there were no conclusive results from the included studies. Conclusions from individual studies regarding the optimum pressure provided by stockings are conflicting, although the results of one study (Jungbeck 1997) suggest that lower pressured stockings (20 mmhg) may be as effective for relieving symptoms as higher pressured stockings (30 to 40 mmhg). Overall completeness and applicability of evidence No studies assessed symptom change when wearing stockings compared to not wearing stockings. All the studies that assessed the outcome of symptom change included participants in whom two different types of stocking were compared. There is insufficient, high quality evidence to determine whether or not compression stockings are effective in the sole and initial treatment of varicose veins in people without healed or active venous ulceration or whether any type of stocking is superior to any other type. To answer the objective of this review, adequately powered trials that compare wearing stockings versus not wearing stockings would be required. Only one included study randomised participants to wearing hosiery or not wearing hosiery (Coughlin 2001). However, there were no data for the group that did not wear hosiery and the outcome of this study was compliance in pregnant women, not improvement of symptoms or physiological measures. 13