Horse Chestnut Seed Extract For Chronic Venous Insufficiency Cochrane Pittler2012

Horse chestnut seed extract for chronic venous insufficiency
(Review)
Pittler MH, Ernst E
This is a reprint of a Cochrane review, prepared and maintained by The Cochrane Collaboration and published in The Cochrane Library
2012, Issue 12
http://www.thecochranelibrary.com
Horse chestnut seed extract for chronic venous insufficiency (Review)

Copyright © 2012 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
TABLE OF CONTENTS
HEADER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
PLAIN LANGUAGE SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
BACKGROUND . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
OBJECTIVES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Figure 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
AUTHORS’ CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
CHARACTERISTICS OF STUDIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
DATA AND ANALYSES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
WHAT’S NEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
HISTORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
CONTRIBUTIONS OF AUTHORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
DECLARATIONS OF INTEREST . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
SOURCES OF SUPPORT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
INDEX TERMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Horse chestnut seed extract for chronic venous insufficiency (Review) i

[Intervention Review]
Horse chestnut seed extract for chronic venous insufficiency
Max H Pittler1 , Edzard Ernst2
1
German Cochrane Centre, Institute of Medical Biometry and Medical Informatics University Medical Center Freiburg and Comple-
mentary Medicine, Peninsula Medical School, University of Exeter, UK, Freiburg, Germany. 2 Complementary Medicine, Peninsula
Medical School, Exeter, UK
Contact address: Max H Pittler, German Cochrane Centre, Institute of Medical Biometry and Medical Informatics University Medical
Center Freiburg and Complementary Medicine, Peninsula Medical School, University of Exeter, UK, Berliner Allee 29, Freiburg,
79110, Germany. [email protected].
Editorial group: Cochrane Peripheral Vascular Diseases Group.

Publication status and date: Stable (no update expected for reasons given in ’What’s new’), published in Issue 12, 2012.
Review content assessed as up-to-date: 19 June 2012.
Citation: Pittler MH, Ernst E. Horse chestnut seed extract for chronic venous insufficiency. Cochrane Database of Systematic Reviews
2012, Issue 11. Art. No.: CD003230. DOI: 10.1002/14651858.CD003230.pub4.
ABSTRACT
Background
Conservative therapy of chronic venous insufficiency (CVI) consists largely of compression treatment. However, this often causes
discomfort and has been associated with poor compliance. Therefore, oral drug treatment is an attractive option. This is an update of
a Cochrane review first published in 2002 and updated in 2004, 2006, 2008 and 2010.
Objectives
To review the efficacy and safety of oral horse chestnut seed extract (HCSE) versus placebo, or reference therapy, for the treatment of
CVI.
Search methods
For this update the Cochrane Peripheral Vascular Diseases Review Group searched their Specialised Register (last searched June 2012)
and CENTRAL (Issue 5, 2012). For the previous versions of the review the authors searched AMED (inception to July 2005) and
Phytobase (inception to January 2001) for randomised controlled trials (RCTs) of HCSE for CVI. Manufacturers of HCSE preparations
and experts on the subject were contacted for published and unpublished material. There were no restrictions on language.
Selection criteria
RCTs comparing oral HCSE mono-preparations with placebo, or reference therapy, in people with CVI. Trials assessing HCSE as one
of several active components in a combination preparation, or as a part of a combination treatment, were excluded.
Data collection and analysis
Both authors independently selected the studies and, using a standard scoring system, assessed methodological quality and extracted
data. Disagreements concerning evaluation of individual trials were resolved through discussion.
Horse chestnut seed extract for chronic venous insufficiency (Review) 1
Main results
Overall, there appeared to be an improvement in CVI related signs and symptoms with HCSE compared with placebo. Leg pain was
assessed in seven placebo-controlled trials. Six reported a significant reduction of leg pain in the HCSE groups compared with the
placebo groups, while another reported a statistically significant improvement compared with baseline. One trial suggested a weighted
mean difference (WMD) of 42.4 mm (95% confidence interval (CI) 34.9 to 49.9) measured on a 100 mm visual analogue scale. Leg
volume was assessed in seven placebo-controlled trials. Six trials (n = 502) suggested a WMD of 32.1ml (95% CI 13.49 to 50.72) in
favour of HCSE compared with placebo. One trial indicated that HCSE may be as effective as treatment with compression stockings.
Adverse events were usually mild and infrequent.
Authors’ conclusions
The evidence presented suggests that HCSE is an efficacious and safe short-term treatment for CVI. However, several caveats exist and
larger, definitive RCTs are required to confirm the efficacy of this treatment option.
PLAIN LANGUAGE SUMMARY
Horse chestnut seed extract for long-term or chronic venous insufficiency
Poor blood flow in the veins of the legs, known as chronic venous insufficiency, is a common health problem, particularly with ageing. It
can cause leg pain, swelling (oedema), itchiness (pruritus) and tenseness as well as hardening of the skin (dermatosclerosis) and fatigue.
Wearing compression stockings or socks helps but people may find them uncomfortable and do not always wear them. A seed extract
of horse chestnut (Aesculus hippocastanum L.) is a herbal remedy used for venous insufficiency. Seventeen randomised controlled trials
were included in the review. In all trials the extract was standardised to escin, which is the main active constituent of horse chestnut
seed extract.
Overall, the trials suggested an improvement in the symptoms of leg pain, oedema and pruritus with horse chestnut seed extract when
taken as capsules over two to 16 weeks. Six placebo-controlled studies (543 participants) reported a clear reduction of leg pain when
the herbal extract was compared with placebo. Similar results were reported for oedema, leg volume, leg circumference and pruritis.
The other studies which compared the extract with rutosides (four trials), pycnogenol (one trial) or compression stockings (two trials)
reported no significant differences between the therapies for leg pain or a symptom score that included leg pain. The herbal extract was
equivalent to rutosides, pycnogenol and compression on the other symptoms with the exception that it was inferior to pycnogenol on
oedema.
The adverse events reported (14 trials) were mild and infrequent. They included gastrointestinal complaints, dizziness, nausea, headache
and pruritus, from six studies.
BACKGROUND
results in leg oedema (swelling), dermatosclerosis (hardening of
the skin) and feelings of pain, fatigue and tenseness in the lower
extremities (Spraycar 1995). Patients often require hospitalisation
Description of the condition and surgery, for instance, for symptomatic varicose veins (London
Chronic venous insufficiency (CVI) is one of the commonest con- 2000; Rigby 2002).
ditions afflicting humans. About 10-15% of men and 20-25% of
women present signs and symptoms consistent with the diagno-
sis of CVI, indicating that being female is an important risk fac-
tor, as well as age, geographical location and race (Callam 1992;
Description of the intervention
Callam 1994). This condition is characterised by chronic inade- Mechanical compression is the treatment of choice for this condi-
quate drainage of venous blood and venous hypertension, which tion (Partsch 1991). However, compression therapy, for example,
using compression stockings often causes discomfort and has been Types of interventions
associated with poor compliance. Oral drug treatment is therefore Trials were included if they compared oral preparations contain-
an attractive option. ing HCSE as the only active component (mono-preparation) with
placebo or reference therapy. Trials assessing HCSE as one of sev-
eral active components in a combination preparation or as a part
How the intervention might work of a combination treatment were excluded.
Horse chestnut (Aesculus hippocastanum L.) has traditionally been
used as a herbal remedy for treating CVI (Bombardelli 1996). Types of outcome measures
The seed extract of Aesculus hippocastanum L. (HCSE) contains
Trials using clinical outcome measures were included. Studies fo-
escin, a triterpenic saponin, as its active component (Guillaume
cusing exclusively on physiological parameters were excluded.
1994; Lorenz 1960; Schrader 1995). Escin has been shown to in-
hibit the activity of hyaluronidase, an enzyme involved in pro-
teoglycan degradation (Facino 1995). The accumulation of leu- Primary outcomes
cocytes (white blood cells) in CVI-affected limbs (Moyses 1987; The primary outcome measures were CVI-related symptoms (e.g.
Thomas 1988) and subsequent activation and release of such en-
leg pain, pruritus (itching), oedema (swelling)).
zymes (Sarin 1993) is considered to be an important pathophysi-
ological mechanism of CVI.
Secondary outcomes
Secondary outcomes were, leg volume and leg circumference at
Why it is important to do this review ankle and calf. Adverse events were assessed as reported in the
included trials.
Regardless of the postulated mechanism of action, the most im-
portant clinical questions are whether HCSE is safe and efficacious
for treating patients with CVI.
Search methods for identification of studies
Electronic searches
OBJECTIVES
For this update the Cochrane Peripheral Vascular Diseases Group
To review the evidence from rigorous clinical trials assessing the Trials Search Co-ordinator (TSC) searched the Specialised Regis-
efficacy and safety of HCSE versus placebo, or reference therapy, ter (last searched June 2012) and the Cochrane Central Register of
for the symptomatic treatment of CVI. Controlled Trials (CENTRAL) 2012, Issue 5, part of The Cochrane
Library, www.thecochranelibrary.com. See Appendix 1 for de-
tails of the search strategy used to search CENTRAL. The Spe-
METHODS cialised 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
Criteria for considering studies for this review
have been searched, as well as the search strategies used are de-
scribed in the Specialised Register section of the Cochrane Periph-
eral Vascular Diseases Group module in The Cochrane Library (
Types of studies www.thecochranelibrary.com).
Randomised, controlled trials (RCTs), i.e. trials with a randomised For the original review Phytobase inception to January 2001 and
generation of allocation sequences. Studies assessing acute effects AMED were searched using the terms listed in Appendix 2 .
only were excluded. No restrictions regarding the language of pub-
lication were imposed (Egger 1997).
Searching other resources
Manufacturers of HCSE preparations and experts on the subject
Types of participants were contacted and asked to contribute published and unpub-
Studies were included if participants were patients with CVI. Stud- lished material. Furthermore, our own files were scanned. The
ies that did not use adequate diagnostic criteria (e.g. Widmer 1978) bibliographies of the studies retrieved were searched for further
were excluded. trials.

Data collection and analysis information was insufficient. The Cochrane Collaboration sug-
gests imputing the variance of the change by assuming a correla-
Max Pittler and Edzard Ernst independently screened and selected
tion factor between pre-intervention and post-intervention values.
trials for inclusion, assessed their methodological quality and ex-
The variance of the change was imputed using a correlation factor
tracted data. Disagreements at any of these stages were resolved by
of 0.4, which was then used to assign a weight to the mean of the
discussion.
within-study treatment effect.
Selection of studies
Assessment of heterogeneity
Trials were selected according to the criteria outlined above under
The chi-square test for heterogeneity tested whether the distribu-
Criteria for considering studies for this review.
tion of the results was compatible with the assumption that inter-
trial differences were attributable to chance variation alone.
Data extraction and management
Data were extracted independently by both authors using a data Data synthesis
extraction form. Disagreements were resolved by discussion. The
Data-pooling of continuous data was performed using the
following data were extracted:
weighted mean difference; for dichotomous data the odds ratio
1. Participant characteristics: age, gender.
was used. Summary estimates of the treatment effect were calcu-
2. Methods used: randomisation, double-blinding,
lated using a random effects model.
concealment of treatment allocation, description of drop outs.
3. Interventions: oral preparations containing HCSE as the
only active component (mono-preparation), compared with Subgroup analysis and investigation of heterogeneity
placebo or comparator medication(s). There are no planned subgroup analyses as yet. Subgroup analyses
4. Outcome measures: CVI-related symptoms (e.g. leg pain, to investigate possible heterogeneity will be performed in future if
pruritus, oedema), leg volume, circumference at ankle and calf, more data become available.
and adverse events.
Sensitivity analysis
Assessment of risk of bias in included studies
There are no planned sensitivity analyses as yet. Sensitivity analyses
Methodological quality was assessed using the Jadad score (Jadad to test the robustness of the main analysis will be performed in
1996) and the Cochrane risk of bias tool. The Jadad score was future if more data become available.
applied independently by both authors and disagreements were
resolved by discussion. The Cochrane risk of bias tool was applied
by the first author only.
RESULTS
Measures of treatment effect
The effect measures of choice in case of dichotomous data were
odds ratio (improvement of leg pain, improvement of oedema, Description of studies
improvement of pruritus), in case of continuous data the mean See: Characteristics of included studies; Characteristics of excluded
difference (reduction of leg pain, reduction of oedema, reduction studies.
of lower leg volume, reduction of circumference at ankle, reduc-
tion of circumference at calf, improvement of symptom score, leg
volume). Included studies
Seventeen trials met the above mentioned inclusion criteria
(Cloarec 1992; Diehm 1992; Diehm 1996a; Diehm 2000; Erdlen
Unit of analysis issues 1989; Erler 1991; Friederich 1978; Kalbfleisch 1989; Koch 2002;
There were no special issues such as carry-over effects or period Lohr 1986; Morales 1993; Neiss 1976; Pilz 1990; Rehn 1996;
effects with the analysis of the three cross over trials included in Rudofsky 1986; Steiner 1986; Steiner 1990a). Of these, ten were
this review. placebo-controlled; two compared HCSE against reference treat-
Statistical analysis was performed using RevMan Analyses 1.0.4. ment with compression stockings and placebo (Diehm 1996a;
It uses the inverse of the variance to assign a weight to the mean of Diehm 2000); four were controlled against reference medication
the within-study treatment effect. For most studies, however, the with O-ß-hydroxyethyl rutosides (HR) (Erdlen 1989; Erler 1991;

Kalbfleisch 1989; Rehn 1996) and one was controlled against combination treatments (Boehm 1989; Coninx 1974; Dols 1987;
medication with pycnogenol (Koch 2002). In all trials the extract Dustmann 1984; Hirsch 1982; Neumann-Mangoldt; Zuccarelli
was standardised to escin which is the main active constituent of 1986); and two focused exclusively on physiological parameters
HCSE. (Bisler 1986; Lochs 1974). The four additional trials identified
through update searches were excluded because they used topical
treatment (Marhic 1986; Paciaroni 1982), did not asses clinical
Excluded studies outcomes (Nill 1970) or tested a combination preparation (Krc¡lek
1973).
Fourteen trials were excluded (Bisler 1986; Boehm 1989; Coninx
1974; Dols 1987; Dustmann 1984; Hirsch 1982; Krc¡lek 1973;
Lochs 1974; Marhic 1986; Nill 1970; Neumann-Mangoldt;
Paciaroni 1982; Pauschinger 1987; Zuccarelli 1986). The trial by
Risk of bias in included studies
(Pauschinger 1987;) used non-clinical outcome measures; seven Figure 1 displays the risk of bias associated with the included
tested HCSE as a component in combination preparations or studies.
Figure 1. Methodological quality graph: review authors’ judgements about each methodological quality
item presented as percentages across all included studies.
Key data from the included trials, including scores for quality
and allocation concealment, are presented in the Characteristics minimal.
of included studies table.
Allocation
Most trials did not report on allocation concealment and it is Incomplete outcome data
therefore unclear to what extent bias was introduced. Only three In all but three studies (Friederich 1978; Steiner 1990a; Rehn
trials (Erdlen 1989; Pilz 1990; Steiner 1986) reported adequate 1996) it is unclear whether incomplete outcome data were ad-
allocation concealment. dressed. It is therefore unclear to what extent bias was introduced.
Blinding
Selective reporting
Only one of the included RCTs was not double-blinded (Koch
2002). All of the included studies administered HCSE in capsules, None of the included trials showed evidence of selective reporting
permitting the preparation of adequate placebos. The likelihood and therefore it is unlikely that bias was introduced here. In the
that bias was introduced through blinding or the lack thereof is included trials, all of the pre-stated outcomes were reported.
Effects of interventions trials (Steiner 1986; Steiner 1990a) suggested a statistically signif-
icant difference in favour of HCSE compared with baseline (P <
The majority of the included studies diagnosed the patients ac-
0.05). Another trial (Kalbfleisch 1989), which compared HCSE
cording to the classification by Widmer (Widmer 1978). Fourteen
with HR, but failed to include a placebo group, seemed to cor-
trials reported inclusion criteria for CVI patients relating to this
roborate these findings. A further trial (Diehm 2000) reported no
classification. Eighty-two percent of the participants in these trials
significant differences for a score including the symptom pruritus
were categorised into CVI stages I, II or I-II. Three trials, com-
compared with compression.
prising 22% of the total number of participants did not refer to
this classification. Overall, the included placebo-controlled trials
suggested an improvement in the CVI related symptoms of leg
pain, oedema and pruritus. Leg volume
Leg volume was assessed in seven placebo-controlled trials (Diehm
1992; Diehm 1996a; Diehm 2000; Lohr 1986; Rudofsky 1986;
Leg pain
Steiner 1986; Steiner 1990a). All of these studies used water
Leg pain was assessed in seven placebo-controlled trials (Cloarec displacement plethysmometry to measure this outcome. Meta-
1992; Friederich 1978; Lohr 1986; Morales 1993; Neiss 1976; analysis of six trials (Diehm 1992; Diehm 1996a; Diehm 2000;
Rudofsky 1986; Steiner 1990a). Six studies (n = 543) reported a Rudofsky 1986; Steiner 1986; Steiner 1990a; n = 502) suggested
statistically significant reduction (P < 0.05) of leg pain on various a WMD of 32.1ml (95% CI 13.49 to 50.72) in favour of HCSE
measurement scales in participants treated with HCSE compared compared with placebo (Analysis 1.6) (pooled standardised mean
with placebo, while another reported an improvement compared difference 0.34; 95% CI 0.15 to 0.52). One trial (Rehn 1996)
with baseline (Steiner 1990a). One study (Cloarec 1992), reported reported findings suggesting that HCSE was equivalent to HR,
adequate data which could be included within RevMan Analyses and another (Diehm 1996a, n = 194) suggested that it may be as
(Analysis 1.2), assessed on a 100 mm VAS, suggesting a weighted efficacious as treatment with compression stockings (WMD -2.90
mean difference (WMD) of 42.40 mm (95% confidence interval ml; 95% CI -30.42 to 24.62).
(CI) 34.90 to 49.90). Other studies which compared HCSE with
HR (Kalbfleisch 1989), pycnogenol (Koch 2002) or compression Significant beneficial effects for CVI patients were reported in tri-
(Diehm 2000) reported no significant inter group differences for als which administered HCSE standardised to 100-150 mg escin
leg pain or a symptom score including leg pain. daily. Three studies, using 100 mg escin daily, reported a statis-
tically significant reduction of mean leg volume after two weeks
of treatment compared with placebo (P < 0.01) (Rudofsky 1986;
Oedema
Steiner 1986; Steiner 1990a). Persistence of treatment effects was
Oedema was assessed in six placebo-controlled trials (Cloarec suggested by one study (Rehn 1996). At the end of a six-week
1992; Friederich 1978; Lohr 1986; Morales 1993; Neiss 1976; follow-up period mean leg volume was similar to post-treatment
Steiner 1990a). Four trials (n = 461) reported a statistically sig- values.
nificant reduction of oedema in participants treated with HCSE
compared with placebo, whilst one (Steiner 1990a) reported an
improvement compared with baseline. One study (Cloarec 1992)
reported adequate data suggesting a WMD of 40.10 mm (95% CI Circumference
31.60 to 48.60) in favour of HCSE assessed on a 100 mm VAS. Circumference at calf and ankle was assessed in seven placebo-
Another study (Koch 2002) reported that HCSE was inferior to controlled trials (Cloarec 1992; Diehm 1992; Lohr 1986; Pilz
pycnogenol, whereas a further trial (Diehm 2000) reported no 1990; Rudofsky 1986; Steiner 1986; Steiner 1990a). Five studies
significant differences for a score including the symptom oedema (n = 172) suggested a statistically significant reduction at the ankle,
compared with compression. Oedema provocation before and af- and three (n = 112) at the calf in favour of HCSE compared
ter treatment with HCSE revealed oedema protective effects (Erler with placebo. At the ankle, meta-analysis of three trials (Cloarec
1991). 1992; Pilz 1990; Steiner 1986), which reported adequate data
suggested a statistically significant reduction in favour of HCSE
compared with placebo (WMD 4.71 mm; 95% CI 1.13 to 8.28;
Pruritus pooled standardised mean difference 0.60; 95% CI 0.15 to 1.05)
Pruritus was assessed in eight placebo-controlled trials (Diehm (Analysis 1.7). At the calf, the pooled analysis of three trials (
1992; Friederich 1978; Lohr 1986; Morales 1993; Neiss 1976; Cloarec 1992; Pilz 1990; Steiner 1986), suggested a statistically
Rudofsky 1986; Steiner 1986; Steiner 1990a). Four trials (n = 407) significant reduction in favour of HCSE compared with placebo
suggested a statistically significant reduction of pruritus in partici- (WMD 3.51 mm; 95% CI 0.58 to 6.45; pooled standardised mean
pants treated with HCSE compared with placebo (P < 0.05). Two difference 0.42; 95% CI -0.04 to 0.88).

Adverse events The mechanism of action involved in the observed effects when
Fourteen studies reported on adverse events. Four studies (Cloarec HCSE is administered may be of interest. The active component
1992; Diehm 1996a; Pilz 1990; Rudofsky 1986) reported that of HCSE is the saponin escin (Lorenz 1960). This has been shown
there were no treatment-related adverse events in the HCSE group. to inhibit the activity of elastase and hyaluronidase in vitro. Both
Gastrointestinal complaints, dizziness, nausea, headache and pru- these enzymes are involved in proteoglycan degradation (proteo-
ritus were reported as adverse events in six studies (Diehm 2000; glycan constitutes part of the capillary endothelium and is the
Friederich 1978; Morales 1993; Neiss 1976; Rehn 1996; Steiner main component of the extravascular matrix) (Facino 1995). The
1990a). The frequency ranged from 1 to 36% of treated patients. accumulation of leucocytes (white blood cells) in CVI-affected
Four other studies (Diehm 1992; Koch 2002; Lohr 1986; Steiner limbs (Moyses 1987; Thomas 1988), and subsequent activation
1986) reported good tolerability with HCSE. and release of such enzymes (Sarin 1993), is considered to be an
important pathophysiological mechanism of CVI. An earlier study
found increased serum activity of proteoglycan hydrolases in pa-
tients with CVI that were reduced with HCSE (Kreysel 1983).
HCSE treatment may shift the equilibrium between degradation
DISCUSSION and synthesis of proteoglycans towards a net synthesis, thus pre-
venting vascular leakage. This hypothesis has been supported by
animal experiments (Enghofer 1984). Using electron microscopy,
Summary of main results the author demonstrated a marked reduction in vascular leakage
after treatment with HCSE. Uncertainty exists regarding the ef-
The results of our systematic review suggest, overall, that com- fects of HCSE on venous tone. In vitro, HCSE increases venous
pared with placebo and reference treatment, HCSE is an effec- pressure of normal and pathologically altered veins. This is corrob-
tive treatment option for CVI. The adverse events reported in the orated by studies in laboratory animals, which demonstrate an in-
reviewed trials were mild and infrequent. Thus, according to the crease in venous pressure and venous flow after HCSE administra-
available data the risk/benefit ratio of HCSE for the short term tion (Guillaume 1994). However, studies on humans have failed to
treatment of CVI is positive. replicate effects on venous capacity (Bisler 1986; Rudofsky 1986).
Overall completeness and applicability of Quality of the evidence

evidence
All randomised double-blind trials included in this review scored
In an attempt to locate all randomised trials of oral preparations at least one out of five points for methodological quality. Nonethe-
containing HCSE, 29 trials were identified of which 17 could be less, the extent of methodological rigour varied between studies.
included. It is noteworthy that one unpublished trial was supplied Only three trials (Diehm 1996a; Rehn 1996; Rudofsky 1986) re-
by a manufacturer of HCSE-containing preparations, while a sec- ported data indicating that compliance was monitored. The ma-
ond unpublished trial was identified in a report by another author jority of studies suffered from a small sample size with drop-out
(Diehm 2000). The search strategy for this review involved sev- rates ranging from zero to 19.5%.
eral databases including those with a focus on the European and
American literature, as well as manual searching and contact with
experts and manufacturers. Moreover, searching was not restricted
in terms of publication language.
Potential biases in the review process
The conservative treatment of CVI comprises a number of other Despite systematic efforts to find all studies on the subject, it
therapeutic modalities. Compression therapy improves venous re- is conceivable that some were not uncovered. Several forms of
turn and is widely accepted as the treatment of choice (Tooke publication and location bias exist (Egger 1998) including the
1996). In combination with heparin, it prevents venous stasis and tendency for negative trials to remain unpublished (Easterbrook
reduces the risk of deep vein thrombosis. O-ß-hydroxyethyl ruto- 1991), for positive findings to be published in English language
sides are reported to have beneficial short-term effects by reducing journals (Egger 1997) and for some European journals to not be
oedema and relieving symptoms of CVI. However, their efficacy indexed in major medical databases (Nieminen 1999). There is
during long-term use has yet to be established (Wadworth 1992). also evidence that positive findings may be over-represented in
Ruscus extract decreases capillary filtration rate in healthy volun- complementary medicine journals (Ernst 1997; Schmidt 2001)
teers and people with CVI (Rudofsky 1991). A review has con- and that these journals favour positive conclusions at the expense
cluded that combined treatment using oedema protective agents of methodological quality (Pittler 2000). Therefore, there is a pos-
and compression therapy improves CVI to a greater extend than sibility that treatment effects are exaggerated. Overall, we are con-
either treatment alone (Diehm 1996b). fident that the search strategy that we used and the therefore the

completeness of the evidence minimised bias. However, more tri- Implications for practice
als assessing the efficacy of HCSE in larger patient samples using
The evidence presented suggests that HCSE is an efficacious and
adequate outcome measures and systematic investigations of its
safe short-term treatment for CVI. However, caveats exist and
safety are still required (Ernst 2001).
more rigorous, large RCTs are required to assess the efficacy of this
treatment option.
Implications for research

Agreements and disagreements with other Future studies should be rigorously executed and reported in a uni-
studies or reviews form manner following the CONSORT statement (Moher 2001).
These findings update and extend the findings of previous system- Detailed description of randomisation and double-blinding pro-
atic reviews (Pittler 1998; Pittler 2004; Siebert 2002). In the re- cedures should be included in the report. More controlled clinical
viewed trials adverse events were mild and infrequent, which sup- trials are needed, which should include larger numbers of partic-
ports the findings of post-marketing surveillance studies (Greeske ipants and assess HCSE particularly for long-term use and as an
1996; Leskow 1996) reporting pruritus, nausea, gastrointestinal adjunct to compression treatment.
complaints, headache and dizziness in 43 of 6183 patients (0.7%)
treated with HCSE.
ACKNOWLEDGEMENTS
The Cochrane Peripheral Vascular Diseases Group ran the searches
of CENTRAL and the Specialised Register. The Plain Language
AUTHORS’ CONCLUSIONS Summary was provided by the Cochrane Consumer Network.
REFERENCES
References to studies included in this review Erler 1991 {published data only}
Erler M. Horse chestnut seed extract in the therapy of
Cloarec 1992 {published and unpublished data} peripheral venous edema - clinical therapies in comparison
Cloarec, M. Study on the effect of a new vasoprotective [Roßkastaniensamenextrakt bei der Therapie peripherer
Venostasin administered over a period of 2 months in venöser Ödeme – ein klinischer Therapievergleich].
chronic venous insufficiency of the lower limb (data from Medizinische Welt 1991;42(7):593–6.
1992). Data on file.
Friederich 1978 {published data only}
Diehm 1992 {published data only} Friederich HC, Vogelsberg H, Neiss A. Evaluation of
Diehm C, Vollbrecht D, Amendt K, Comberg HU. Medical internally effective venous drugs [Ein Beitrag zur Bewertung
edema protection - Clinical benefit in patients with chronic von intern wirksamen Venenpharmaka]. Zeitschrift fur
deep vein incompetence. VASA 1992;21(2):188–92. Hautkrankheiten 1978;53(11):369–74.
Diehm 1996a {published data only} Kalbfleisch 1989 {published data only}
Diehm C, Trampisch HJ, Lange S, Schmidt C. Comparison Kalbfleisch W, Pfalzgraf H. Ödemprotektiva. Äquipotente
of leg compression stocking and oral horse-chestnut seed Dosierung - Roßkastaniensamenextrakt und O-ß-
extract therapy in patients with chronic venous insufficiency. Hydroxyethylrutoside im Vergleich. Therapiewoche 1989;
Lancet 1996;347(8997):292–4. 39:3703–7.
Diehm 2000 {unpublished data only} Koch 2002 {published data only}
Diehm C, Schmidt C. Venostasin retard gegen Plazebo Koch R. Comparative study of venostasin and pycnogenol
und Kompression bei Patienten mit CVI II/IIIA. Final in chronic venous insufficiency. Phytotherapy Research 2002;
Study Report. Klinge Pharma GmbH Munich, Germany. 16(Suppl 1):S1–S5.
Reported in: Ottillinger B et al. BMC Cardiovascular Lohr 1986 {published data only}
Disorders 2001;1:5. Lohr E, Garanin G, Jesau P, Fischer H. Anti-oedemic
Erdlen 1989 {published data only} treatment in chronic venous insufficiency with tendency
Erdlen F. Clinical efficacy of Venostasin: A double blind to formation of oedema [Ödempräventive Therapie
trial [Klinische Wirksamkeit von Venostasin retard im bei chronischer Veneninsuffizienz mit Ödemneigung].
Doppelblindversuch]. Medizinische Welt 1989;40(36): Münchener Medizinische Wochenschrift 1986;128(34):
994–6. 579–81.
Morales 1993 {published data only} Roßkastaniensamenextrakt auf die transkapilläre Filtration
Morales Paris CA, Barros Soares RM. Efficacy and safety bei chronisch venöser Insuffizienz]. Deutschs Medizinische
on use of dried horse chestnut extract in the treatment of Wochenschrift 1986;111(35):1321–9.
chronic venous insufficiency of the limbs. Revista Brasileira Boehm 1989 {published data only}
de Medicina 1993;50(11):1563–5. Bohm, C. Venodiuretics: a new combination and oedema
Neiss 1976 {published data only} protective drug [Venodiuretikum – Kombination eines
Neiss A, Böhm C. Demonstration of the effectiveness Diuretikums mit einem Oedemprotektivum]. Medizinische
of horse chestnut seed extract in the varicose Welt 1989;40(30-31):887–8.
syndrome complex [Zum Wirksamkeitsnachweis
Coninx 1974 {published data only}
von Roßkastaniensamenextrakt beim varikösen Coninx S. Supplementary drug therapy in the treatment
Symptomenkomplex]. Münchener Medizinische
of venous insufficiency. Results of a double blind study
Wochenschrift 1976;118(7):213–6. [Medikamentoese Zusatztherapie bei der Behandlung der
Pilz 1990 {published data only} venoesen Insuffizienz]. Fortschritte der Medizin 1974;92
Pilz E. Oedemas in venous disease [Ödeme bei (18):792–4.
Venenerkrankungen]. Medizinische Welt 1990;41(12): Dols 1987 {published data only}
1143–4. Dols W, Fiala G. Treatment of varicosis in patients with
Rehn 1996 {published data only} chronic venous insufficiency and peripheral oedema
Rehn D, Unkauf M, Klein P, Jost V, Lücker PW. [Therapie der Varikosis bei chronischer Varikosis und
Comparative clinical efficacy and tolerability of oxerutins peripheren Oedemen]. Therapiewoche 1987;37(38):
and horse chestnut extract in patients with chronic venous 3601–4.
insufficiency [Vergleich der klinischen wirksamkeit und Dustmann 1984 {published data only}
vertraglichkeit von oxerutin und Rosskastanien–extrakt Dustmann HO, Godolias G, Seibel K. Foot volume with
bei patienten mit chronischer venoser Insuffizienz]. chronic venous insufficiency while standing; effect of a
Arzneimittel-Forschung 1996;46(5):483–7. [MEDLINE: new treatment [Verminderung des Fußvolumens bei der
1996168792] chronischen venösen Insuffizienz im Stehversuch durch eine
Rudofsky 1986 {published data only} neue Wirkstoffkombination]. Therapiewoche 1984;34(36):
Rudofsky G, Neiss A, Otto K, Seibel K. Oedema-protective 5077–86.
effect and clinical efficacy of horse chestnut seed extract Hirsch 1982 {published data only}
in a double blind study [Ödemprotektive Wirkung und Hirsch J. The effect of Essaven ultra in chronic venous
klinische Wirksamkeit von Roßkastaniensamenextrakt im insufficiency. Fortschritte der Medizin 1982;100(10):436–8.
Doppeltblindversuch]. Phlebologie und Proktologie 1986;15
(2):47–54. Krc¡lek 1973 {published data only}
Krc¡lek A, Smejkal V. [Therapeutic effects of venotonics in
Steiner 1986 {published data only}
clinical pharmacotherapeutical evaluations by double-blind
Steiner M. Evaluation of the oedema protective effect of
tests]. Casopis Lekaru Ceskych 1973;112:930–3.
horse chestnut seed extract [Ausmaß der ödemprotektiven
Wirkung von Roßkastaniensamenextrakt]. Vasa 1991;20 Lochs 1974 {published data only}
(Supplement 33):S217. Lochs H, Baumgartner H, Konzett H. Effect of horse
Steiner M. Investigation into the oedema reducing chestnut seed extract on venous tone [Zur Beeinflussung
and oedema protective effects of horse chestnut seed des Venetonus durch Rosskastanienextrakte]. Arzneimittel-
extract [Untersuchungen zur ödemvermindernden und Forschung 1974;24(9):1347–50.
ödemprotektiven Wirkung von Roßkastaniensamenextrakt]. Marhic 1986 {published data only}
Phlebologie und Proktologie 1990;19(5):239–42. Marhic C, Anglade JP. [Analgesic effect of rap cream on
Steiner M, Hillemanns HG. Tests for anti-oedema action venous insufficiency pain syndrome of the lower limbs].
of a venous therapy [Untersuchung zur oedemprotektiven Phlebologie 1986;39(4):1011–5.
Wirkung eines Venentherapeutikums]. Munchener
Neumann-Mangoldt {published data only}
Medizinische Wochenschrift 1986;128(31):551–2.
Neumann-Mangoldt P. Experiences in the use of Essaven
Steiner 1990a {published data only} capsules in the treatment of venous leg diseases. Results
Steiner M, Hillemanns HG. Venostasin retard in the of a double blind study [Erfahrungen in der Behandlung
management of venous problems during pregnancy. venoeser Beinleiden mit Essaven–Kapseln]. Fortschritte der
Phlebology 1990;5(1):41–4. Medizin 1979;97(45):2117–20.
References to studies excluded from this review Nill 1970 {published data only}
Nill HJ, Fischer H, Nill HJ, Fischer H. [Comparative
Bisler 1986 {published data only} investigations concerning the effect of extract of horse
Bisler H, Pfeifer R, Klüken N, Pauschinger P. Effect chestnut upon the pressure-volume-diagramm of patients
of horse chestnut seed extract on transcapillary with venous disorders]. [German]. Arztliche Forschung
filtration in chronic venous insufficiency [Wirkung von 1970;24(5):141–3.
Paciaroni 1982 {published data only} Ernst 1997
Paciaroni E, Marini M. Topical therapy for phlebopathies. Ernst E, Pittler MH. Alternative therapy bias. Nature 1997;
Results of a controlled clinical study. Policlinico - Sezione 385(6616):480.
Medica 1982;89(3):255–64. Ernst 2001
Pauschinger 1987 {published data only} Ernst E, Pittler MH. Assessment of therapeutic safety in
Pauschinger P. Clinical investigation into the effects of systematic reviews: Literature review. BMJ 2001;323
horse chestnut seed extract on transcapillary filtration (7312):546.
and intravenous volume in patients with chronic venous Facino 1995
insufficiency [Klinisch experimentelle Untersuchungen Facino RM, Carini M, Stefani R, Aldini G, Saibene L. Anti-
zur Wirkung von Roßkastaniensamenextrakt auf die elastase and anti-hyaluronidase activities of saponins and
transkapilläre Filtration und das intravasale Volumen an sapogenins from Hedera helix, Aesculus hippocastanum,
Patienten mit chronisch venöser Insuffizienz]. Phlebologie and Ruscus aculeatus: factors contributing to their efficacy
und Proktologie 1987;16:57–61. in the treatment of venous insufficiency. Archiv der
Pharmazie 1995;328(10):720–4.
Zuccarelli 1986 {published data only}
Zuccarelli F. Study of the clinical efficacy of escin plus Greeske 1996
metescufylline in painful manifestations of chronic venous Greeske K, Pohlmann B-K. Horse chestnut seed extract -
insufficiency [Etude de l’efficacite clinique du Veinotonyl an effective therapy principle in practice. Drug therapy of
sur les manifestations douloureuses de l’insuffisance veineuse chronic venous insufficiency [Rosskastaniensamenextrakt –
chronique]. Gazette Medicale 1986;93(42):67–70. ein wirksames Therapieprinzip in der Praxis]. Fortschritte
der Medizin 1996;114(15):42–6.
Additional references Guillaume 1994
Guillaume M, Padioleau F. Veinotonic effect, vascular
Bombardelli 1996 protection, antiinflammatory and free radical scavenging
Bombardelli E, Morazzoni P, Griffini A. Aesculus properties of horse chestnut extract. Arzneimittel-Forschung
hippocastanum L. Fitoterapia 1996;67(6):483–511. 1994;44(1):25–35.
Callam 1992 Jadad 1996
Callam M. Prevalence of chronic leg ulceration and severe Jadad AR, Moore A, Carroll D, Jenkinson C, Reynolds
chronic venous disease in western countries. Phlebology DJM, Gavaghan DJ, et al.Assessing the quality of reports of
1992;7(Suppl 1):6–12. randomized clinical trials: Is blinding necessary?. Controlled
Clinical Trials 1996;17:1–12.
Callam 1994
Callam MJ. Epidemiology of varicose veins. British Journal Kreysel 1983
of Surgery 1994;81(2):167–73. Kreysel HW, Nissen HP, Enghofer E. Raised lysosomal
enzyme activities in serum of patients with varicosity
Diehm 1996b [Erhöhte Serumaktivität lysomaler Enzyme bei Varikosis].
Diehm C. The role of oedema protective drugs in the Therapiewoche 1983;33(9):1098–104.
treatment of chronic venous insufficiency: a review of
Leskow 1996
evidence based on placebo-controlled trials with regard to
Leskow P. Effective treatment with horse chestnut seed
efficacy and tolerance. Phlebology 1996;11(1):23–9.
extract in chronic venous insufficiency [Theraoie von
Easterbrook 1991 Venenleiden. Rosskastanienpraeparat gut wirksam].
Easterbrook PJ, Berlin JA, Gopalan R, Matthews DR. Therapiewoche 1996;16(16):874–7.
Publication bias in clinical research. Lancet 1991;337 London 2000
(8746):867–76. London NJM, Nash R. ABC of arterial and venous disease.
Egger 1997 Varicose veins. BMJ 2000;320(7246):1391–4.
Egger M, Zellweger-Zahner T, Schneider M, Junker Lorenz 1960
C, Lengeler C, Antes G. Language bias in randomised Lorenz D, Marek ML. Das therapeutische wirksame Prinzip
controlled trials published in English and German. Lancet der Rosskastanie (Aesculus hippocastanum). Arzneimittel-
1997;350(9074):326–9. Forschung 1960;10:263–72.
Egger 1998 Moher 2001
Egger M, Davey Smith G. Meta-analysis: Bias in location Moher D, Schulz KF, Altman DG, Lepage L. The
and selection of studies. BMJ 1998;316(7124):61–6. CONSORT statement: Revised recommendations
Enghofer 1984 for improving the quality of reports of parallel-group
Enghofer E, Seibel K, Hammersen F. Antiexudative effect randomised trials. Lancet 2001;357(9263):1191–4.
of Aesculus hippocastanum L. (horse chestnut) extract [Die Moyses 1987
antiexsudative Wirkung von Rosskastaniensamenextrakt]. Moyses C, Cederholm-Williams SA, Michel CC.
Therapiewoche 1984;34(27):4130–44. Haemoconcentration and accumulation of white cells

in the feet during venous stasis. International Journal Siebert 2002
of Microcirculation: Clinical & Experimental 1987;5(4): Siebert U, Brach M, Sroczynski G, Ueberla K. Efficacy,
311–20. routine effectiveness and safety of horsechestnut seed extract
Nieminen 1999 in the treatment of chronic venous insufficiency. Angiology
Nieminen P, Isohanni M. Bias against European journals in 2002;21:305–15.
medical publication databases. Lancet 1999;353(9164): Spraycar 1995
1592. Spraycar M, editor. Stedman’s Medical dictionary. 26th
Partsch 1991 Edition. Baltimore: Williams and Wilkins, 1995.
Partsch H. Compression therapy of the legs. A review. Thomas 1988
Journal of Dermatology, Surgery & Oncology 1991;17(10): Thomas PR, Nash GB, Dormandy JA. White cell
799–805. accumulation in dependent legs of patients with venous
Pittler 2000 ulceration: a possible mechanism for trophic changes in the
Pittler MH, Abbot NC, Harkness EF, Ernst E. Location bias skin. British Medical Journal Clinical Research Edition 1988;
in controlled clinical trials of complementary/alternative 296(6638):1693–5.
therapies. Journal of Clinical Epidemiology 2000;53(5): Tooke 1996
485–9. Tooke JE, Lowe GDO. A textbook of vascular medicine.
Rigby 2002 London: Arnold, 1996.
KA Rigby, SJ Palfreyman, C Beverley, JA Michaels. Surgery
Wadworth 1992
for varicose veins: use of tourniquet. Cochrane Database
Wadworth AN, Faulds D. Hydroxyethylrutosides: A review
of Systematic Reviews 2002, Issue 2. [DOI: 10.1002/
of its pharmacology and therapeutic efficacy in venous
14651858.CD001486]
insufficiency and related disorders. Drugs 1992;44(6):
Rudofsky 1991 1013–32.
Rudofsky G. Influence of Phlebodril on the capillary
Widmer 1978
filtration rate [Beeinflussung der kapillären Filtrationsrate
Widmer LK, Stähelin HB. Peripheral venous disorders Basel
durch Phlebodril]. Phlebologie 1991;20(1):14–6.
III. Bern: Huber, 1978.
Sarin 1993
Sarin S, Andaz A, Shields DA, Scurr JH, Coleridge Smith References to other published versions of this review
PD. Neutrophil activation in venous disease. Journal of
Vascular Surgery 1993;17:444. Pittler 1998
Pittler MH, Ernst E. Horse chestnut seed extract for chronic
Schmidt 2001
venous insufficiency. A criteria-based systematic review.
Schmidt K, Pittler MH, Ernst E. Bias in alternative
Archives of Dermatology 1998;134(11):1356–60.
medicine is still rife but is diminishing. BMJ 2001;323
(7320):1071. Pittler 2004
Schrader 1995 Pittler MH, Ernst E. Horse chestnut seed extract for
Schrader E, Schwankl W, Sieder CH, Christoffel V. chronic venous insufficiency. Cochrane Database of
Comparative study of the bioavailability of beta-escin Systematic Reviews 2004, Issue 2. [DOI: 10.1002/
after single oral administration of two different drug 14651858.CD003230.pub2]
formulations containing an extract of horse chestnut Pittler 2006
seeds [Vergleichende Untersuchung zur Bioverfügbarkeit Pittler MH, Ernst E. Horse chestnut seed extract for
von ß–Aescin nach oraler Einmalverabreichung zweier chronic venous insufficiency. Cochrane Database of
Rosskastaniensamenextrakt enthaltender, galenisch Systematic Reviews 2006, Issue 1. [DOI: 10.1002/
unterschiedlicher Darreichungsformen]. Pharmazie 1995; 14651858.CD003230.pub3]
50(9):623–7. ∗
Indicates the major publication for the study

CHARACTERISTICS OF STUDIES
Characteristics of included studies [ordered by study ID]
Cloarec 1992
Methods Study design: 2 parallel arms, randomised, double-blind.

Method of randomisation: not reported.
Exclusion post randomisation: none.
Losses to follow up: none.
Quality score = 3.
Participants Country: France.

Setting: hospital.
No: 30 entered, 0 drop outs.
Age: (mean) 45.5 and 47.7 years in HCSE and placebo group, respectively
Sex: males 15; females 15.
Inclusion criteria: patients with functional symptoms due to CVI at least in one leg;
patients with impression oedema at least in one leg
Exclusion criteria: systolic blood pressure ankle/arm > 0.9; acute or precedent (< 1 month)
thrombophlebitis; leg ulcer of venous origin; cardiac, renal or orthopaedic oedema
Interventions Treatment: 1 capsule HCSE (standardised to 50 mg escin) twice daily

Control: placebo.
Duration: 4 weeks.
Outcomes Primary: (not explicitly stated).

Secondary: (not explicitly stated).
1) circumference (mm)
2) leg pain (mm)
3) oedema (mm)
Notes Standardised mean difference (95% CI):

1) circumference
a) ankle 0.57 (-0.16 to 1.30)
b) calf 0.26 (-0.46 to 0.98)
2) 3.93 (2.65 to 5.22)
3) 3.28 (2.14 to 4.43)
Risk of bias
Bias Authors’ judgement Support for judgement
Allocation concealment (selection bias) Unclear risk Allocation concealment not reported
Blinding (performance bias and detection Low risk “In a double blind study Venostasin versus
bias) placebo was studied in 30 cases ...”
All outcomes

Cloarec 1992 (Continued)
Incomplete outcome data (attrition bias) Unclear risk It is unclear how incomplete outcome data
All outcomes were addressed; intention to treat analysis
Selective reporting (reporting bias) Low risk No evidence of selective reporting
Other bias Unclear risk No evidence of other biases
Diehm 1992

Method of randomisation: block randomisation.
Exclusion post randomisation:
one patient.
Quality score = 4.
Participants Country: Germany.

Setting: hospital.
No: 40 entered, 1 excluded post randomisation.
Age: (mean) 53 and 48 years in treatment and control groups, respectively
Sex: reported males 9; females 29.
Inclusion criteria:
CVI stage 2 according to Hach, venous flow impairment, oedema, possible trophic skin
changes, venous capacity and / or venous return outside normal limits
Exclusion criteria: CVI liable to venous compression, acute venous inflammation, acute
thrombosis, venous ulceration, oedema due to other conditions than CVI

Control: placebo.
Duration: 6 weeks.
Outcomes Primary: leg volume (ml).

Secondary:
1) circumference
2) pruritus

Primary: 0.30 (-0.33 to 0.94)
Secondary: 1), 2) Not enough data provided for effect size calculation
Risk of bias

Diehm 1992 (Continued)
Blinding (performance bias and detection Low risk “The double blind nature of the trial as as-
bias) sured by using placebos which in terms of
All outcomes outer appearance and taste were identical
to verum”
All outcomes were addressed
Other bias Unclear risk No evidence of further biases
Diehm 1996a
Methods Study design: 3 parallel arms, randomised, double-blinded (for placebo and HCSE only)
Losses to follow up: not reported.
Quality score = 2.

Setting: hospital.
No: 240 entered, drop outs not reported.
Age: (mean) 52 years.
Sex: not reported.
Inclusion criteria:
oedema due to CVI (confirmed by medical history, clinical findings, venous Doppler
and duplex sonography
Exclusion criteria: venotherapeutic drugs within the last 6 weeks before run-in

Control: placebo or compression stockings.
Duration: 12 weeks.
Outcomes Primary: (not explicitly stated) leg volume (ml).

Secondary: (not explicitly stated) circumference, symptoms .

Primary:
HCSE versus placebo
0.49 (0.14 to 0.85)
HCSE versus compression
-0.03 (-0.31 to 0.25)
Risk of bias

Diehm 1996a (Continued)
Blinding (performance bias and detection Low risk “Patients were treated over a period of 12
bias) weeks in a randomised partially blinded
All outcomes placebo controlled paralles study”
Incomplete outcome data (attrition bias) Unclear risk It is unclear how incomplete data were ad-
All outcomes dressed
Diehm 2000

Exclusion post randomisation: not reported.
Losses to follow up: 69.
Quality score = 2.

Setting: unclear.
No: 355 entered, drop outs 69.
Age: not reported.
Sex: not reported.
Inclusion criteria:
CVI stage II and IIIA.
Exclusion criteria: venotherapeutic drugs within the last 6 weeks, patients with oedema
of non-venous origin

Control: placebo or compression stockings.
Duration: 16 weeks.

Secondary: symptom score computed from:
1) feeling of swelling
2) tiredness in the leg
3) itching
4) leg cramps
5) paraesthesia
6) plantar burning
7) unspecific complaints

Primary:
HCSE versus placebo 0.26 (-0.03 to 0.54)

Diehm 2000 (Continued)
HCSE versus compression 0.70 (-0.94 to 0.46)

Secondary:
HCSE versus compression 0.06 (-0.17 to 0.29)
Risk of bias
Blinding (performance bias and detection Low risk “The study was double-blind regarding al-
bias) location to HCSE or placebo and open
All outcomes regarding allocation to the compression
group”
Other bias Unclear risk Reported within a review article only
Erdlen 1989
Methods Study design: 2 parallel arms, randomised double-blind.

Method of randomisation: Central randomisation by company.
Quality score = 4.

Setting: GP setting.
Age: (mean) 55 years in treatment group; no data for control
Inclusion criteria: varicosis due to CVI, peripheral venous oedema
Exclusion criteria: oedema due to other conditions than CVI, vasoactive medication,
compression treatment, venous ulcers

Control: rutoside.
Duration: 4 weeks.
Outcomes Primary: circumference (mm).

Secondary: not reported.

Erdlen 1989 (Continued)

Primary:
ankle 0.0 (-0.72 to 0.72)
Risk of bias
Allocation concealment (selection bias) Low risk Block randomisation was done. Random
codes were kept in sealed envelopes
Blinding (performance bias and detection Low risk Treatment and placebo capsules were indis-
bias) tinguishable in terms of outer appearance
All outcomes
Incomplete outcome data (attrition bias) Unclear risk It is unclear how incomplete data were ad-
All outcomes dressed
Erler 1991

Quality score = 3.

Setting: hospital.
Age: (mean) 55.5 and 53.9 years in treatment and control group, respectively
Inclusion criteria: oedema due to CVI.
Exclusion criteria: oedema due to other conditions than CVI, vasoactive medication,
compression treatment, venous ulcers

Control: O-beta-hydroxyethyl rutosides (2 g daily).
Duration: 8 weeks.
Outcomes Primary: circumference before and after oedema provocation.

Secondary: (not explicitly defined) symptoms (leg pain, oedema, pruritus, fatigue)

Erler 1991 (Continued)

Primary: Not enough data provided for effect size calculation
Secondary: Not enough data provided for effect size calculation
Risk of bias
Blinding (performance bias and detection Low risk Neutrally coated capsules which were indis-
bias) tinguishable in terms of outer appearance
All outcomes
Friederich 1978
Methods Study design: crossover, randomised,

double-blind.
Quality score = 4.

Setting: hospital.
Age: (mean) 48 and 47 years in men and women, respectively.
Inclusion criteria: oedema, leg pain, pruritus, feeling of tenseness and fatigue
Exclusion criteria: not reported.

Control: placebo.
Duration: 20 days.
Outcomes Primary: (not explicitly defined) symptoms

1) leg pain
2) oedema
3) pruritus
Secondary: (not explicitly defined) patients’ impression of effectiveness

Friederich 1978 (Continued)

Primary:
1), 2), 3) Not enough data provided for effect size calculation
Secondary:
Not enough data provided for effect size calculation.
Risk of bias
Blinding (performance bias and detection Low risk Treatment and placebo capsules were iden-
bias) tical in terms of outer appearance
All outcomes
Kalbfleisch 1989
Methods Study design: 2 parallel arms, randomised,

double-blind.
Losses to follow up: three.
Quality score = 4.

Setting: GP practice.
Age: “18 years and over”.
Sex: male and female (numbers not reported).
Inclusion criteria:
CVI and oedema.
Exclusion criteria: cardiac and hepatic oedema, patients with kidney and liver dysfunc-
tions, venous ulcers, vasoactive medication, NSAIDs, glucosides
Interventions Treatment: 1 capsule HCSE (standardised to 50 mg escin) once daily

Control: O-beta-hydroxyethyl rutosides (50 mg daily).
Duration: 8 weeks.

Kalbfleisch 1989 (Continued)
Outcomes Primary: (not explicitly stated)

circumference (mm).
Secondary: (not explicitly stated)
1) leg pain (mm)
2) oedema (mm)
3) pruritus

Primary:
a) ankle 2.13 (1.20 to 3.06)
b) calf 1.83 (0.95 to 2.21)
Secondary:
1) 0.19 (-0.54 to 0.91)
2) -0.25 (-0.97 to 0.48)
3) Not enough data provided for effect size calculation.
Risk of bias
Blinding (performance bias and detection Low risk Treatment and placebo capsules were neu-
bias) trally coated
All outcomes
Koch 2002
Methods Study design: 2 parallel arms, randomised, open.

Exclusion post randomisation: one.
Quality score = 1.

Age: (mean) 56 and 59 years in HCSE and pycnogenol group respectively
Inclusion criteria: CVI.

Koch 2002 (Continued)
Interventions Treatment: 1 capsule HCSE (standardised to 50 mg escin twice daily)

Control: pycnogenol (360 mg daily).
Duration: 4 weeks.
Outcomes Primary: (not explicitly defined)

1) Symptoms
a) leg pain
b) oedema
c) cramps
d) feeling of heaviness
e) leg reddening
2) Circumference (mm).
Secondary: (not explicitly defined)
Serum cholesterol.
Notes 1), 2) Not enough data provided for effect size calculation.
Risk of bias
Blinding (performance bias and detection High risk “In an open controlled comparative study
bias) 40 patients with diagnosed CVI were
All outcomes treated.”
Incomplete outcome data (attrition bias) Unclear risk It is not clear how incomplete outcome data
Lohr 1986

Quality score = 3.

Age: (mean) 54 years in total patient sample.
Inclusion criteria: CVI.

Lohr 1986 (Continued)

Control: placebo.
Duration: 8 weeks.
Outcomes Primary:
leg volume
Secondary:
1) circumference
2) leg pain
3) oedema
4) pruritus
Notes Primary and secondary outcomes: Not enough data provided for effect size calculation
Risk of bias
Blinding (performance bias and detection Low risk The study was conducted randomised and
bias) double blind
All outcomes
Incomplete outcome data (attrition bias) Unclear risk It is not clear how incomplete outcome data
Morales 1993

double-blind.
Losses to follow up: three.
Quality score = 3.
Participants Country: Brazil.

Setting: hospital.
Age: (mean) 40 years in total patient sample.
Inclusion criteria: oedema, varicosis, venous ulcers.

Morales 1993 (Continued)
Exclusion criteria: diabetes mellitus, oedema of other origin, peripheral arterial disease,
diuretic medication

Control: placebo.
Duration: 20 days.
Outcomes Primary: (not explicitly defined)

oedema
Secondary: (not explicitly defined)
1) leg pain
2) pruritus
Notes Primary and secondary outcomes: Not enough data provided for effect size calculation
Risk of bias
Blinding (performance bias and detection Low risk “This is a double blind randomised placebo
bias) controlled parallel study of the ise of dried
All outcomes horse chestnut extract (Venostasin retard)
in chronic venous insufficiency of the
limbs:”
Selective reporting (reporting bias) High risk No evidence for selective reporting
Neiss 1976

double-blind.
Method of randomisation:
not reported.
Losses to follow up: seven.
Quality score = 3.

Age: (mean) 56 and 55 in women and men, respectively.

Neiss 1976 (Continued)

Inclusion criteria: CVI with symptoms including oedema, leg pain, pruritus fatigue and
tenseness, calf cramps
Exclusion criteria: concomitant medication or physical treatments

Control: placebo.
Duration: 20 days.
Outcomes Primary: (not explicitly defined) symptoms

1) leg pain
2) oedema
3) pruritus
4) feeling of fatigue and tenseness
5) calf cramps
Secondary: not described.
Notes 1), 2), 3), 4), 5) Not enough data provided for effect size calculation
Risk of bias
Blinding (performance bias and detection Low risk “...the study was carried out in a double-
bias) blind design.”
All outcomes
Pilz 1990

double-blind.
Exclusion post randomisation: two.
Quality score = 4.


Pilz 1990 (Continued)
Age: (mean) 46 in total patient sample.

Inclusion criteria: Symptoms of CVI with peripheral leg oedema
Exclusion criteria: patients under 20 and over 70 years of age, less than 2 symptoms
of CVI, leg ulcers, oedema or leg pain of other origin than CVI, rheumatic diseases,
concomitant medication, compression treatment

Control: placebo.
Duration: 20 days.
Outcomes Primary: (not explicitly stated).

Secondary: (not explicitly stated).

Primary: circumference (mm).
a) ankle 0.70 (-0.04 to 1.45)
b) calf 0.86 (0.11 to 1.61)
Secondary: (not explicitly stated) adverse events.
none.
Risk of bias
Allocation concealment (selection bias) Low risk Block randomisation and allocation of pa-
tients to treatment and control groups was
performed centrally by Klinge Pharma.
The random code was stored in sealed en-
velopes
Blinding (performance bias and detection Low risk Verum and placebo were indistinguishable
bias) in terms of outer appearance and taste
All outcomes

Rehn 1996

double-blind.
Quality score = 4.

Setting: not reported.
Age: mean 58.4 years ß-hydroxyethyl-rutosides (1 g daily) group; 62.8 years ß-hydrox-
yethyl-rutosides (1 to 0.5 g daily) group; 59.0 years HCSE group
Sex: all females.
Inclusion criteria: uni- or bilateral CVI stage II, doppler sonographic assessment within
the past 6 months
Exclusion criteria: oedema due to other conditions than CVI, over 70 years of age, current
acute phlebitis or thrombosis, concomitant medication, compression treatment

Control: ß-hydroxyethyl-rutosides (1 g daily) or ß-hydroxyethyl-rutosides (1 to 0.5 g
daily)
Duration: 12 weeks.

Secondary:
Symptoms: tired, heavy legs (VAS (mm)).

Primary:
HR (1g): -0.17 (-0.56 to 0.22)
HR (1 to 0.5g): 0.05 (-0.38 to 0.48)
Secondary:
(mean, SD) 4.1, 2.9; 3.8, 2.6; 3.0, 2.2 at baseline for beta-HR 1 g, beta-HR 1 to 0.5
g and HCSE respectively. -1.5, 3.0; -1.0, 3.3; -0.2, 2.5 are the respective changes from
baseline (no formal statistical analysis)
Risk of bias
Blinding (performance bias and detection Low risk “According to the double dummy proce-
bias) dure both for oxerutin film tablets and
All outcomes horse chestnut extract capsules identically
appearing placebo tablets or capsules were
used.”

Rehn 1996 (Continued)
Incomplete outcome data (attrition bias) Low risk “Missing data were interpolated if possible
All outcomes or the method of last value carry forward
was used.”
Rudofsky 1986

double-blind.
Method of randomisation: random number generator.
Quality score = 5.

Setting: hospital.
No: 40 entered, 1 drop out.
Age: (mean) 41 and 38 years in treatment and placebo groups, respectively
Sex: males 14, females 25.
Inclusion criteria: clinical signs of CVI (e.g. varicosis, hyperpigmentation), symptoms
(e.g. leg pain, pruritus), venous capacity of over 6 ml per 100 ml tissue, venous pressure
(dorsum pedis) of at least 60 mmHg
Exclusion criteria: CVI stage III, acute phlebitis, oedema of other origin than CVI,
concomitant medication (e.g. diuretics, vasoactive drugs)

Control: placebo.
Duration: 4 weeks.
Outcomes Primary: (not explicitly stated)

leg volume (ml)
Secondary: (not explicitly stated)
1) circumference
2) leg pain
3) pruritus

Primary:
0.46 (-0.18 to 1.10)
Secondary:
Not enough data provided for effect size calculation.
Risk of bias

Rudofsky 1986 (Continued)
Blinding (performance bias and detection Low risk Verum and placebo capsules were indistin-
bias) guishable .... Thus it was impossible for
All outcomes physician and patient to determine whether
they received the true or placebo medica-
tion
Other bias Unclear risk There is no evidence of other biases
Steiner 1986

double-blind.
Quality score = 4.

Setting: hospital.
No: 20 entered, drop outs none.
Age: (range) 20 to 40 years in total patient sample.
Sex: all females.
Inclusion criteria: CVI stage I, peripheral venous oedema.
Exclusion criteria: Patients in third trimenon, CVI stages II and III, diuretics, vasoactive
medication

Control: placebo.
Duration: 2 weeks.
Outcomes Primary:
leg volume (ml)
Secondary:
2) symptoms (e.g. pruritus).

Primary:
0.41 (-0.48 to 1.30)
Secondary:
1)

Steiner 1986 (Continued)
a) ankle 0.48 (-0.41 to 1.38)

b) calf 0.07 (-0.81 to 0.95)
Risk of bias
Allocation concealment (selection bias) Low risk The random code was kept in sealed en-
velopes (information from duplicate pub-
lication Steiner 1990b)
Blinding (performance bias and detection Low risk Verum and placebo capsules were indis-
bias) tinguishable in terms of colour and taste
All outcomes (information from duplicate publication
Steiner 1990b)
Other bias Unclear risk No evidence of other bias
Steiner 1990a

double-blind.
Losses to follow up: two.
Quality score = 4.

Setting: hospital.
Age: (mean) not reported.
Sex: all females.
Inclusion criteria: over 18 years of age, varicose veins and clinically detectable oedema,
CVI had to be confirmed by at least two of either Doppler sonography, plethysmography
venous pressure measurements or light reflection rheography
Exclusion criteria: not described.

Control: placebo.
Duration: 2 weeks.

Steiner 1990a (Continued)
Outcomes Primary:
1) leg volume (ml)
Secondary: symptoms: leg pain, pruritus, oedema, fatigue.

Primary:
1) 0.15 (-0.40 to 0.71)
Secondary: Not enough data provided for effect size calculation
Risk of bias
Blinding (performance bias and detection Low risk “Patients were given either one capsule of
bias) Venostasin retard twice daily or an identical
All outcomes placebo”
All outcomes were addressed. Drop outs are described:
“Of the 52 patients who were entered two
patients discontinued the study; one had
to undergo an operation and the other was
lost to follow-up”
Other bias Unclear risk There is no evidence of other biases
Characteristics of excluded studies [ordered by study ID]
Study Reason for exclusion
Bisler 1986 Used non-clinical outcome measures.
Boehm 1989 HCSE applied as part of a combination preparation.
Coninx 1974 HCSE applied as part of a combination preparation.
Dols 1987 HCSE applied as part of a combination preparation.
Dustmann 1984 HCSE applied as part of a combination preparation.

(Continued)
Hirsch 1982 HCSE applied as part of a combination preparation.
Krc¡lek 1973 HCSE applied as part of a combination preparation.
Lochs 1974 Trial performed on healthy volunteers, not people with CVI.
Marhic 1986 Used cream, not oral preparation.
Neumann-Mangoldt HCSE applied as part of a combination preparation.
Nill 1970 Used non-clinical outcome measures.
Paciaroni 1982 Used cream, not oral preparation.
Pauschinger 1987 Used non-clinical outcome measures.
Zuccarelli 1986 HCSE applied as part of a combination preparation.

DATA AND ANALYSES
Comparison 1. HCSE versus placebo
No. of No. of
Outcome or subgroup title studies participants Statistical method Effect size
1 Improvement of leg pain 1 Odds Ratio (M-H, Random, 95% CI) Totals not selected
(responder ratio)
2 Reduction of leg pain (100 mm 1 Mean Difference (IV, Random, 95% CI) Totals not selected
VAS)
3 Reduction of oedema (100 mm 1 Mean Difference (IV, Random, 95% CI) Totals not selected
VAS)
4 Improvement of oedema 1 Odds Ratio (M-H, Random, 95% CI) Totals not selected
(responder ratio)
5 improvement of pruritus 1 Odds Ratio (M-H, Random, 95% CI) Totals not selected
(responder ratio)
6 Reduction of lower leg volume 6 502 Mean Difference (IV, Random, 95% CI) 32.10 [13.49, 50.72]
(ml)
7 Reduction of circumference at 3 80 Mean Difference (IV, Random, 95% CI) 4.71 [1.13, 8.28]
ankle (mm)
8 Reduction of circumference at 3 80 Mean Difference (IV, Random, 95% CI) 3.51 [0.58, 6.45]
calf (mm)
Comparison 2. HCSE versus compression
No. of No. of
1 Reduction of lower leg volume 2 479 Mean Difference (IV, Random, 95% CI) -37.34 [-104.07, 29.
(ml) 39]
2 Improvement of symptom score 1 Mean Difference (IV, Random, 95% CI) Totals not selected
(40 point scale)
Comparison 3. HCSE versus ß-hydroxyethyl-rutosides
No. of No. of
1 Reduction of circumference at 2 60 Mean Difference (IV, Random, 95% CI) 2.38 [-1.47, 6.23]
ankle (mm)
2 Reduction of circumference at 1 Mean Difference (IV, Random, 95% CI) Totals not selected
calf (mm)
3 Reduction of leg pain (VAS) 1 Mean Difference (IV, Random, 95% CI) Totals not selected
4 Leg volume (ml) 1 Mean Difference (IV, Random, 95% CI) Totals not selected
5 Reduction of oedema (VAS) 1 Mean Difference (IV, Random, 95% CI) Totals not selected
WHAT’S NEW
Last assessed as up-to-date: 19 June 2012.
Date Event Description
23 October 2012 Review declared as stable No new included studies have been identified since 2005. This Cochrane review has
been marked stable and will only be updated when new studies are identified
HISTORY
Protocol first published: Issue 3, 2001
Review first published: Issue 1, 2002
Date Event Description
19 June 2012 New search has been performed Searches re-run, no new trials found. The review was
assessed as up to date
19 June 2012 New citation required but conclusions have not Searches re-run, no new trials found. Minor copy edits
changed made, conclusions not changed
27 July 2010 New search has been performed Searches re-run and four additional studies were ex-
cluded from the review. Risk of bias tables added to
the Included studies and minor changes made to the
text of the review
22 September 2008 New search has been performed Searches re-run, no new trials found. Minor changes
to the text of the review
22 September 2008 Amended Converted to new review format.
15 February 2007 Amended Search dates changed, no new trials found. Plain lan-
guage summary added and minor copy edits
15 November 2005 New citation required but conclusions have not Substantive amendment. One additional trial included
changed but no change to conclusions
25 February 2004 New citation required but conclusions have not Substantive update. One additional trial included but
changed no change to conclusions

CONTRIBUTIONS OF AUTHORS
Conception and design: MH Pittler, E Ernst
Analysis and interpretation of the data: MH Pittler, E Ernst
Drafting of the article: MH Pittler, E Ernst
Critical revision of the article for important intellectual content: MH Pittler, E Ernst
Final approval of the article: MH Pittler, E Ernst
DECLARATIONS OF INTEREST
None known
SOURCES OF SUPPORT
Internal sources
• No sources of support supplied
External sources
• Chief Scientist Office, Scottish Government Health Directorates, The Scottish Government, UK.
The PVD Group editorial base is supported by the Chief Scientist Office.
INDEX TERMS
Medical Subject Headings (MeSH)

∗
Aesculus [adverse effects]; ∗ Seeds; Administration, Oral; Chronic Disease; Leg [∗ blood supply]; Pain [drug therapy]; Phytotherapy
[adverse effects; ∗ methods]; Plant Extracts [adverse effects; ∗ therapeutic use]; Randomized Controlled Trials as Topic; Treatment
Outcome; Venous Insufficiency [∗ drug therapy]
MeSH check words

Humans


Horse Chestnut Seed Extract For Chronic Venous Insufficiency Cochrane Pittler2012

Uploaded by

Copyright:

Available Formats

Horse Chestnut Seed Extract For Chronic Venous Insufficiency Cochrane Pittler2012

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Horse Chestnut Seed Extract For Chronic Venous Insufficiency Cochrane Pittler2012

Uploaded by

Copyright:

Available Formats

Horse chestnut seed extract for chronic venous insufficiency

Pittler MH, Ernst E

Horse chestnut seed extract for chronic venous insufficiency (Review)

Horse chestnut seed extract for chronic venous insufficiency (Review) i

Horse chestnut seed extract for chronic venous insufficiency

Max H Pittler1 , Edzard Ernst2

Editorial group: Cochrane Peripheral Vascular Diseases Group.

Data collection and analysis

PLAIN LANGUAGE SUMMARY

Horse chestnut seed extract for long-term or chronic venous insufficiency

Horse chestnut seed extract for chronic venous insufficiency (Review) 3

Horse chestnut seed extract for chronic venous insufficiency (Review) 4

Horse chestnut seed extract for chronic venous insufficiency (Review) 6

Overall completeness and applicability of Quality of the evidence

Horse chestnut seed extract for chronic venous insufficiency (Review) 7

Implications for research

Horse chestnut seed extract for chronic venous insufficiency (Review) 10

Horse chestnut seed extract for chronic venous insufficiency (Review) 11

Characteristics of included studies [ordered by study ID]

Methods Study design: 2 parallel arms, randomised, double-blind.

Participants Country: France.

Interventions Treatment: 1 capsule HCSE (standardised to 50 mg escin) twice daily

Outcomes Primary: (not explicitly stated).

Notes Standardised mean difference (95% CI):

Bias Authors’ judgement Support for judgement

Horse chestnut seed extract for chronic venous insufficiency (Review) 12

Selective reporting (reporting bias) Low risk No evidence of selective reporting

Other bias Unclear risk No evidence of other biases

Methods Study design: 2 parallel arms, randomised, double-blind.

Participants Country: Germany.

Interventions Treatment: 1 capsule HCSE (standardised to 75 mg escin) twice daily

Outcomes Primary: leg volume (ml).

Notes Standardised mean difference (95% CI):

Bias Authors’ judgement Support for judgement

Horse chestnut seed extract for chronic venous insufficiency (Review) 13

Selective reporting (reporting bias) Low risk No evidence of selective reporting

Other bias Unclear risk No evidence of further biases

Participants Country: Germany.

Interventions Treatment: 1 capsule HCSE (standardised to 50 mg escin) twice daily

Outcomes Primary: (not explicitly stated) leg volume (ml).

Notes Standardised mean difference (95% CI):

Bias Authors’ judgement Support for judgement

Horse chestnut seed extract for chronic venous insufficiency (Review) 14

Selective reporting (reporting bias) Low risk No evidence of selective reporting

Other bias Unclear risk No evidence of other biases

Methods Study design: 3 parallel arms, randomised, double-blind.

Participants Country: Germany.

Interventions Treatment: 1 capsule HCSE (standardised to 50 mg escin) twice daily

Outcomes Primary: leg volume (ml).

Notes Standardised mean difference (95% CI):

Horse chestnut seed extract for chronic venous insufficiency (Review) 15

HCSE versus compression 0.70 (-0.94 to 0.46)

Bias Authors’ judgement Support for judgement

Selective reporting (reporting bias) Low risk No evidence of selective reporting

Other bias Unclear risk Reported within a review article only

Methods Study design: 2 parallel arms, randomised double-blind.

Participants Country: Germany.

Interventions Treatment: 1 capsule HCSE (standardised to 50 mg escin) twice daily

Outcomes Primary: circumference (mm).