See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/47814245

Botanical Extracts as Anti-Aging Preparations for the Skin

Article  in  Drugs & Aging · December 2010

DOI: 10.2165/11584420-000000000-00000 · Source: PubMed

CITATIONS READS

17 4,173

3 authors, including:

Katherine Hunt Edzard Ernst

University of Southampton University of Exeter
33 PUBLICATIONS   718 CITATIONS    2,346 PUBLICATIONS   52,574 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Palliative Care conversations in COPD View project

All content following this page was uploaded by Katherine Hunt on 03 June 2014.

The user has requested enhancement of the downloaded file.

 Drugs Aging 2010; 27 (12): 973-985
REVIEW ARTICLE 1170-229X/10/0012-0973/$49.95/0

ª 2010 Adis Data Information BV. All rights reserved.

Botanical Extracts as Anti-Aging

Preparations for the Skin
A Systematic Review
Katherine J. Hunt,1,2 Shao Kang Hung1 and Edzard Ernst1
1 Complementary Medicine, Peninsula College of Medicine and Dentistry, University of Exeter, Devon, UK
2 Faculty of Health Sciences, University of Southampton, Southampton, UK

Contents
Abstract. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 973
1. Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 974
2. Methodological Considerations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 974
2.1 Literature Search Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 974
2.2 Inclusion/Exclusion Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 974
2.3 Extraction and Evaluation of Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 975
3. Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 976
3.1 Trials of Oral Treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 976
3.2 Trials of Topical Treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 976
3.3 Methodological Quality of Included Trials. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 980
4. Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 983
5. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 984

Abstract Although topical creams and other anti-aging products purport to reduce
the appearance of aging and skin wrinkling, there has been no critical analysis
in the scientific literature of their effectiveness.
This systematic review critically evaluates the evidence for the effectiveness
or efficacy of botanical treatments in reducing skin aging and wrinkling.
MEDLINE, Embase, CINAHL, CENTRAL and AMED databases were
searched from their inception until October 2009. Reference lists of retrieved
articles were hand-searched. Manufacturers and professional associations
were contacted in order to identify further non-published studies. No lan-
guage restrictions were applied. Only randomized clinical trials or controlled
clinical trials assessing the effectiveness of botanical extracts in reducing
wrinkling and aging of the skin were included. Data were extracted by two
independent reviewers and methodological quality was assessed using the
Jadad score and key aspects of the Cochrane risk of bias tool.
Of 36 potentially relevant studies, 11 trials of botanical extracts for reduc-
ing skin wrinkling and the appearance of aging met all the inclusion criteria.
No trials were identified following contact with anti-aging and cosmetic or-
ganizations, companies and professional bodies. A significant reduction in
skin wrinkling was noted for date kernel extract, cork extract, soy extract,
974 Hunt et al.

Rosaceae and peony extract. No significant reduction was noted for green
tea, Vitaphenol (a combination of green and white teas, mangosteen and
pomegranate extract) or maca root. All trials were of poor methodological
quality. Adverse effects were frequently not reported.
In summary, there is some weak evidence to suggest that several botanical
extracts may be effective in reducing the appearance of skin aging but no
evidence that this effect is enduring. Independent replications with larger,
more diverse samples, longer treatment durations and more rigorous study
designs are required to validate these preliminary findings.

1. Background 2.1 Literature Search Methodology

Over the last decade there has been an increase
in scientific interest in reducing the appearance of
aging. Although it has been argued that there is
no evidence to suggest that aging can be halted or
reversed,[1,2] there is a plethora of products on the
market that claim to be able to reduce both the
physical and aesthetic effects of aging.[3]
It has been estimated that 50% of the US public
are aware of anti-aging therapies,[4] and in 2000,
sales of topical anti-aging preparations exceeded
over $US2 billion in the US,[5] a figure that seems
to be rising.[6] Botanical products represent a large
proportion of this market.[6-11]
Many of the companies that sell these products
make grossly exaggerated claims about their effec-
tiveness that are emotionally charged and frequently
not founded on clinical evidence.[3,9,12] The safety
of these products is often unknown; contaminants
have been found in several anti-aging treatments,
which might put the public at risk of both physical
and economic harm.[1] Despite this, there has, to our
knowledge, been no systematic review of the litera-
ture on the effectiveness of botanical preparations
as anti-aging treatments. We therefore aimed to sys-
tematically review all randomized controlled trials
(RCTs) and controlled clinical trials (CCTs) testing
the efficacy and effectiveness of botanical extracts
compared with placebo in reducing the appearance
of dermal aging in humans.
2. Methodological Considerations
All methods used to conduct this review were
based on a predefined protocol which is available
from the authors.
A search of the following databases (from
their inception to October 2009) was conducted:
AMED, CENTRAL (Cochrane Central Register
of Controlled Trials), Embase, MEDLINE (using
the Ovid interface) and CINAHL (via the EBSCO
interface) using a combination of MeSH and key-
word terms. Previous systematic reviews of herbal
medicine or anti-aging were obtained and their
search strategies studied. The search terms were
constructed to cover herbal/botanical terms, anti-
aging terms and terms describing clinical trial
design (see table S1, Supplemental Digital Con-
tent 1, which details the MEDLINE search strat-
egy, http://links.adisonline.com/DAZ/A7) and were
adapted to run in the other databases. No language
restrictions were applied. Additional trials were
sought by searching the reference lists of retrieved
articles and by contacting anti-aging and cosmetics
organizations, companies and professional bodies
(figure 1).
2.2 Inclusion/Exclusion Criteria
Only RCTs or CCTs that tested the efficacy or
effectiveness of botanical extracts and prepara-
tions in reducing the appearance of aging were
included in this review. Thus, we excluded any
studies that, based on the title or abstract, clearly
did not meet the following inclusion criteria:
 RCTs or CCTs of topical or oral treatments
that claimed to reduce/ameliorate the appear-
ance of dermal aging in humans (‘appearance
of dermal aging’ defined as wrinkling, the pres-
ence of age spots, and loss of skin elasticity).
 ‘Quasi’-RCTs that utilized a ‘split-face’ design
whereby experimental cream was applied to

ª 2010 Adis Data Information BV. All rights reserved. Drugs Aging 2010; 27 (12)
Botanical Extracts as Anti-Aging Skin Preparations
2243 original articles located
after duplicates removed
36 articles retrieved in full
11 RCTs included
Fig. 1. Flow diagram of study selection. RCT = randomized controlled trial.
one side of the face and placebo cream to the
other. Such trials are commonly used in the
cosmeceutical industry and were included be-
cause of the nature of the research in question.
 Trials in which the interventions were described
as herbal/botanical extracts/preparations and
were compared with placebo treatments (topical
creams or oral medications) or no treatment.
 Trials that measured changes in skin para-
meters from baseline.
 Trials that assessed the impact of interventions
attempting to reduce the appearance of dermal
aging.
Trials were also excluded if they:
 Assessed preparations designed to prevent the
appearance of dermal aging.
 Included control interventions that contained
botanical/herbal preparations or other active
ingredients that were not also present in the ex-
perimental treatment (e.g. vitamin E, retinol).
 Included non-botanical/herbal ingredients in
the experimental preparation that were not
also present in the control intervention.
The two latter exclusion criteria reflected the
fact that it would not have been possible in such
circumstances to isolate the effects of the bota-
nical ingredients.
All studies that met these criteria were re-
trieved in full and considered for inclusion by two
ª 2010 Adis Data Information BV. All rights reserved.
975
No articles identified through liaison
with anti-aging and cosmetics organizations,
companies and professional bodies
2207 articles excluded on basis of abstract
(not RCTs of botanicals for anti-aging)
25 articles excluded:
• 3 not for anti-aging
• 4 not botanical ingredients
• 2 not for the skin
• 9 with other, non-botanical active
ingredients in the experimental
preparation but not in the
placebo preparation
• 7 not RCTs
of the authors (KJH and SKH). As noted in the
previous section, the reference lists of these arti-
cles were searched for additional studies. Any
disagreements between the authors regarding
the inclusion of studies were resolved through
discussion.
2.3 Extraction and Evaluation of Data
Two authors (KJH and SKH) independently
extracted data on PICO (patient/problem, inter-
vention, comparison and outcome) questions,
adverse effects, appropriateness of statistical
methods, and funding sources into pre-piloted
data extraction forms and rated the methodo-
logical quality and risk of bias of the trials (at
study level) using the Jadad score.[13] The Jadad
score has a maximum of 5. Trials are allocated
one point for each of the following: (i) the study is
described as randomized; (ii) the randomization
procedure is described as appropriate, with one
point being deducted if the described procedure is
inappropriate; (iii) the study is described as double-
blind (both participant and treatment provider
blinded to group assignment); (iv) the double-
blinding procedure is described and appropriate,
with one point being deducted if it is not appro-
priate; and (v) the number of and reasons for with-
drawals and drop-outs are provided. In addition,
Drugs Aging 2010; 27 (12)
976
key elements of the Cochrane risk of bias tool
were used to further assess the methodological
quality of the trials.[14] Studies were rated on the
heterogeneity of the groups at baseline, the ap-
propriateness of allocation concealment, whether
a power calculation and intention-to-treat anal-
ysis were conducted, and the appropriateness of
practitioner training. Disagreements between the
two reviewers were resolved through discussion.
If trials included between-group comparisons, we
present results according to the original report.
Where available, data were summarized as differ-
ence in means/proportion improved (table I).
3. Findings
After the removal of duplicates, our electronic
searches identified 2243 articles. Eleven trials as-
sessing the effectiveness of oral and topical treat-
ments as anti-aging skin treatments were included
in this review (figure 1). The trials were not suffi-
ciently homogeneous to allow meta-analysis, and
for all trials there were insufficient data for us to
perform any further secondary analyses. No trials
were identified following contact with the anti-
aging and cosmetic organizations, companies and
professional bodies. The methodological character-
istics and main results of the included trials are
detailed in table I.
3.1 Trials of Oral Treatments
Oral supplementation of fermented soy extract
(isoflavone aglycone) for facial dermal aging re-
sulted in a significant decrease in fine wrinkles in the
treatment group compared with the control group
at 12 weeks.[20] At 8 weeks, there was a significant
improvement in skin microrelief but not at any
other timepoint or at the end of treatment, which
may suggest that this was a chance finding. No sig-
nificant differences in linear wrinkles were reported.
However, this trial was methodologically weak.
For instance, even though the trial was described
as randomized, participants’ group allocation was
based on extent of linear wrinkles, thus using quota
sampling and possibly introducing selection bias.
The effectiveness of oral supplementation with
green tea extract to reduce skin aging was as-
ª 2010 Adis Data Information BV. All rights reserved.
Hunt et al.
sessed in an RCT of 18 volunteers who received
both green tea oral supplements and green tea
topical cream for 8 weeks.[16] Although there was
a significant improvement in elastin content of
skin biopsies in the treatment group compared
with the control group, no improvement in clinical
parameters was visible. In addition, self-report
data showed a significant increase in skin dryness
in the treatment group compared with the control
group: three participants discontinued the cream
because of irritation, perhaps suggesting that the
concentration of the active ingredient was too
high or that it was not well tolerated, particularly
given that these effects were not noted with oral
supplementation.
A similar study of oral green tea supplementa-
tion (no topical treatment), but with an extended
study duration (2 years), also met our inclusion
criteria.[21] The study was of moderate methodo-
logical quality. No significant differences were
noted between groups on any parameter after
2 years of treatment or at any other timepoint.
3.2 Trials of Topical Treatments
The effectiveness of a topical gel-based green
tea extract (2% epigallocatachin gallate) compared
with placebo gel was assessed in 60 women.[25]
Improvement in wrinkling was evident in 83.3%
of the treatment group compared with 10.0% in
the control group but no statistics were provided
to summarize this difference, and it is not clear
on which outcome measure this is based or who
conducted the clinical assessments. Because this
trial was available only as a conference report,
very little information on the conduct and design
of the study or data was presented, which pre-
cluded any meaningful interpretation.
The hemi-face/split-face design was employed
in five of the included trials. In this design, treat-
ment was randomized to one-half of the face and
control cream to the other. The first of these trials
tested a cream containing date kernel.[15] Analysis
of silicon replicas of the eye region demonstrated
a significant reduction in total surface and depth
of wrinkles at the crow’s foot area, and 50% of
volunteers self-reported improvement on the treat-
ment side. However, the proportion who reported
Drugs Aging 2010; 27 (12)
ª 2010 Adis Data Information BV. All rights reserved.

Botanical Extracts as Anti-Aging Skin Preparations

Table I. Methodological characteristics and main results of included trials
Study; study design Participants Intervention group Control group Assessments and outcomes Main results
Bauza et al.[15] 10 healthy female volunteers Test cream containing 5% date Randomized to Clinical examination under Data very poorly presented.
(2002); RCT left- aged 46–58 y (no mean) kernel extract as active apply ‘placebo magnifier of face and crow’s foot Clinical examination showed
right hemi-face ingredient. cream’ to left or area: skin relief, no. and depth of improvements in 60% of parti-
study (1 group), Participants randomized to right side of face. wrinkles, skin hydration and cipants and no change on placebo
10-wk duration apply to ‘crow’s foot area’ right or Applied bid for 5 wk firmness on 1–5 scale (5 = good side (no statistics to support).
left side of face bid for 5 wk improvement). After 5 wk of treatment, total
Questionnaire of self-reported surface wrinkles (p < 0.05) and
change in wrinkling. wrinkle depth (p < 0.05) evident on
Skin microrelief examination using silicon replicas significantly
silicon replicas of crow’s feet taken reduced in experimental compared
at baseline and study end: no., with control group.
complete surface, depth and 50% reported improvement with
length of wrinkles verum, although proportion who
noted improvement on placebo
side unknown
Chiu et al.[16] 40 women (no age N = 18 N = 22 Dermatologist rated before and In test group, three discontinued
(2005); RCT 2 information) with moderate Test cream containing 10% Placebo cream bid after scores on 0–5 scale cream but continued taking oral
groups, 8-wk photoaging graded 2–3 on green tea extract bid to face and to face and arms; (increasing severity) for fine and supplements and one applied
duration Glogau scale and Fitzpatrick arms; 300 mg green tea oral oral tablet of similar coarse wrinkling, roughness, cream to wrong site but continued
skin phototype I–III. supplement bid. weight and feel bid. dryness, skin tone, keratoses. taking oral supplements.
Women taking Told to use a facial cleanser and Told to use a facial 4 mm skin biopsies at baseline and No significant differences were
corticosteroids/retinoids/HRT sunscreen (no specification). cleanser and 8 wk, graded by blinded noted between groups on any of
excluded Required to refrain from new sunscreen (no dermatologist on -3 to +3 scale the dermatologist-rated
medications, other topical specification). (0 = no change) for epidermal parameters.
treatments and new nutritional Required to refrain thickness, perivascular Significantly greater improvement
supplements from new inflammation, stratum corneum in elastin content of skin biopsies in
medications, other structure. treatment group compared with
topical treatments Self-report scale (0–5, increasing control (p < 0.05). No significant
and new nutritional severity) for roughness, dryness differences in any other
supplements and overall appearance histological parameter.
From self-report data, significantly
greater increase in dryness noted
in treatment group compared with
control (p < 0.01) and improvement
in overall appearance (p < 0.05)
Coquet et al.[17] 2 men and 13 women aged 1 application of gel formula 1 application of Average and maximum skin rough- Significant reduction in roughness
(2005); RCT left- 22–52 y (mean 32.4 y) containing 3% cork extract to placebo gel ness measured from photographs of and smoothing effect at 1 h in
Drugs Aging 2010; 27 (12)

right arm study, delimited area on forearm. formula silicon replicas of treated areas at 73.3% of women (p < 0.01).
2-h duration Assigned to receive baseline and 1 and 2 h after Significant reduction in surface
treatment/placebo to right/left treatment; report does not mention peaks at 2 h in 53.3% of women
arm at random who performed these readings. (p = 0.05)
Self-reported improvement

Continued next page

977
 Table I. Contd
ª 2010 Adis Data Information BV. All rights reserved.

978
Study; study design Participants Intervention group Control group Assessments and outcomes Main results
Garnier et al.[18] 19 women (mean age 60 y). 2% maca root (5% dry weight) in Vehicle only to Cutaneous relief measured by Significant decrease in no.,
(2005); CCT left- No information on inclusion vehicle to blinded half of face bid other half of face image analysis of silicon replica – surface, length and depth of
right hemi-face criteria or skin state at for 8 wk analysed by microscopy. wrinkles at 4 and 8 wk whilst
study, 8-wk baseline Clinical evaluation. vehicle had no effect – unclear
duration Self-evaluation. although probably intra-group
All measured at 4 and 8 wk analyses
Hsu et al.[19] 20 women aged 35–65 y with Vitaphenol skin cream Placebo skin Fine lines and wrinkles around No significant differences between
(2007); RCT left- demonstrable facial wrinkling (containing green and white cream to other half eyes and mouth evaluated by a groups for skin topography
right hemi-face (Rao Goldan score ‡2). teas, mangosteen and of face bid for 60 d blinded dermatologist using Rao compared with baseline.
study, 74-d active Not permitted to use facial pomegranate extract) to Goldman score (0–5). No significant differences in Rao
treatment, 14-d cleaners, a-hydroxy acids, randomized half of face bid for Skin surface topography Goldman score for eye or mouth
follow-up retinol, retinoic acids, salicylic 60 d measured by PRIMOS optical 3D area between groups.
acids, vitamins C/D for 7 d skin device for digital photographs. Twice as many in treatment group
prior to study Self-report on improvement rated enhancement of skin texture
with use of verum compared with
placebo – no inferential statistics
provided
Izumi et al.[20] 26 women aged 30–40 y with N = 13 (mean – SD age N = 13 (mean – SD Wrinkle area and area ratio of eye Significant decrease in fine
(2007); RCT 2 fine wrinkles around the eyes 40.1 – 1.0 y). age 40.5 – 0.95 y). wrinkles based on Kligman wrinkles at 12 wk (but not at any
groups, 12-wk 4 · 25 mg tablets of fermented 4 tablets of colour- classification of linear or fine other timepoint) in treatment group
duration soy extract (containing 10 mg matched placebo wrinkles assessed at baseline, 4, compared with control group
isoflavone aglycone) daily for for 12 wk 8, 12 wk. (p < 0.05).
12 wk Skin microrelief area ratio and skin No between-group differences for
elasticity linear wrinkles.
Significant improvement in skin
microrelief at 8 wk in treatment
group compared with control group
but not at any other timepoint
Janjua et al.[21] 56 female volunteers aged N = 29 (mean – SD age N = 27 (mean – SD Visia Complexion analysis No significant difference between
(2009); RCT 2 25–75 y with Facial Glogau 50.9 – 10.2 y). age 48.3 – 12.7 y). system (by blinded dermatologist groups at any timepoint for any
groups, 2-y Photoaging Scale II 1 capsule containing 250 mg oral 1 placebo capsule at 0, 6, 12, 24 mo using digital left- parameter.
duration (moderate) or III (advanced) green tea polyphenols (70% bid for 2 y sided photos): for fine and coarse Within-group analyses presented
and Fitzpatrick skin types I–III catechins, 99.5% caffeine free) wrinkles, hyperpigmentation, pore
bid for 2 y size, roughness, erythema and
overall damage.
Histological assessment of 4 mm
punch biopsies by blinded
Drugs Aging 2010; 27 (12)

histopathologist.
Global response to treatment
assessed by blinded dermatologist.
Self-reported change on 100-point

Hunt et al.
scale

Continued next page

 Table I. Contd
ª 2010 Adis Data Information BV. All rights reserved.

Botanical Extracts as Anti-Aging Skin Preparations

Study; study design Participants Intervention group Control group Assessments and outcomes Main results
Kim et al.[22] 20 female volunteers aged Test cream containing 0.03% Placebo cream not Following measurements were Significantly greater improvement
(2008); RCT left- 35–53 y (no mean), with dry to ziyuglycoside I (from containing active taken at 0, 4, 8, 12 wk by 2 blinded in photodamage score at the
right hemi-face very dry skin Sanguisorba officinalis) applied ingredient dermatologists: photodamage crow’s foot area in the treatment
study (1 group) to randomized side of face (did (did not state score 0–7 (0 = none); wrinkle group compared with the control
12-wk duration not state frequency of frequency of peaks and depths measured on 3D group at 12 wk (p < 0.05) but not at
applications) applications) skin system programme; skin any other timepoint.
Visiometer SV 600 for skin Significantly reduced roughness
roughness of silicon replica of crow’s foot
area in treatment group compared
with control (p < 0.05) as per
skin Visiometer SV 600.
No such differences at other
timepoints

Lee et al.[23] (2006); 20 women aged 30–56 y 0.5% paeoniflorin (white peony ‘Control cream’ Skin replicas of eye region Significantly reduced wrinkles
RCT left-right (mean age 42.5 y) with extract) face cream applied to applied to other constructed at baseline, 4 wk and (surface roughness) in treatment
hemi-face study moderate to fine wrinkling randomized side of face bid for half of face bid for 8 wk – scanned for 3D profiles to group compared with control group
(1 group), around the eyes 8 wk 8 wk measure surface roughness at 8 wk (p < 0.05) but not at 4 wk.
8-wk duration Significantly reduced wrinkles
(as evaluated by dermatologist)
in treatment group compared
with control group (p < 0.05)
although neither the scale used
to measure wrinkles nor the
data obtained were documented
in the study

Martelli et al.[24] 20 women aged 20–25 y with Cream containing sylibin (milk
(2000); RCT left- healthy skin thistle) and Centella asiatica
right arm study (gotu kola) extracts and soya
(1 group), 1-mo bean non-saponifiable lipids
duration applied to randomized left/right
arm bid for 1 mo.
No other skincare products
permitted during study period.
Required to wash with synthetic
detergent (provided)
‘Placebo emulsion’ Clinical examination: dryness, No between-group analyses
applied to arm bid irritation, desquamation. provided and insufficient data for
for 1 mo. Biophysical non-invasive re-analysis.
Same skincare measurements. Significant increase in elasticity
regimen as for Skin hydration with corneometer and firmness in treatment group
treatment group on three areas of investigated site. (p < 0.02) – intra-group
Elasticity and firmness with suction
device

Syed et al.[25] 60 female volunteers aged N = 30 N = 30 ‘Dermatological evaluations’: Marked improvement in skin
Drugs Aging 2010; 27 (12)

(2002); RCT 2 25–60 y (no mean), with
groups, visible signs of damaged
4-wk duration facial skin observed by
uneven texture and large
pores
2% epigallocatechin gallate
(found in green tea) in a
hydrophilic gel.
Women given 50 g tube
and self-applied gel bid
for 4 wk
Placebo gel in photography; skin topography texture noted in 46.7% of subjects:
identical 50 g tube analysis 83.3% of experimental group
improved; 10.0% of control group
improved; no statistics provided

979
bid = twice daily; CCT = controlled clinical trial; HRT = hormone replacement therapy; RCT = randomized controlled trial; SD = standard deviation.
980
improvement after the placebo cream was not
described and the study was poorly reported with
no information on the qualifications of the asses-
sors, and few data presented. Furthermore, it is
unclear whether all patients completed the study.
Ziyuglycoside I (0.03%) extracted from San-
guisorba officinalis (Rosaceae) was tested on 20
females using the same hemi-face design.[22] Treat-
ment continued for 12 weeks and silicone skin
replicas were used to determine changes in wrin-
kle peaks and depths as well as in skin roughness
from baseline. However, there was no information
provided on the dose regimen, the vehicle, or the
content of the placebo cream. p-Values (without
effect sizes and lacking information on which
statistical test was used) suggested a significant
improvement in photodamage score (p < 0.05) and
average roughness (p < 0.05) in the treatment
group compared with the control group.
A further randomized hemi-face study assessed
the effectiveness of 0.5% paeoniflorin (extracted
from the root of Paeonia albiflora Pall ) face cream
applied twice daily for 8 weeks.[23] Paeoniflorin
induced a significant reduction in wrinkles (from
skin replicas) at 8 weeks in the treatment group
compared with the control group (p < 0.05). A sig-
nificant reduction in wrinkles in the treatment group
compared with the control group was also noted
by a blinded dermatologist (p < 0.05), although the
data were not provided and the scale used to mea-
sure wrinkles was not described or validated.
Using the same design, the effectiveness of
maca root (obtained from maca flour) for facial
wrinkles was investigated.[18] The most notable
methodological problem in this trial was the
absence of randomization. However, for topical
skin treatments that use the hemi-face design,
randomization plays little role. Normally, ran-
domization to a group helps to eliminate bias, yet
if the participant receives both active and control
treatment concomitantly (i.e. not in a crossover
design) and both the participant and outcome
assessor are blinded to allocation, randomization
to either the left or the right has little merit.
However, no between-group analyses were per-
formed by the authors and there were insufficient
data provided for us to perform these analyses, so
the effectiveness of maca root is still unknown.
ª 2010 Adis Data Information BV. All rights reserved.
Hunt et al.
A hemi-face trial with a longer treatment dura-
tion (74 days) tested Vitaphenol (cream con-
taining green and white teas, mangosteen and
pomegranate extract) in 20 women but found no
significant differences between groups in improve-
ment of wrinkles around the eyes or mouth.[19]
However, there were inconsistencies in some of
the other reported results. For instance, the re-
sults state that 41% preferred the verum whilst
only 0.06% preferred the placebo cream, which is
implausible in a sample of 20.
The left-right design was used in two further
studies in which creams were applied to random-
ized sections of the arms rather than the face.[17,24]
In the first of these, young, healthy women were
randomized to apply milk thistle, gotu kola ex-
tracts and soya bean non-saponifiable lipids or
‘vehicle’ (placebo) to a part of the left and right
arms.[24] However, because only within-group an-
alyses were provided and there were insufficient
data provided in the report for us to conduct re-
analyses, it is not possible to draw conclusions on
the effectiveness of these extracts. In the second
study,[17] males and females applied one applica-
tion of a cream containing either 3% cork extract
or placebo to a delimited area of the right or left
arm. Analysis of silicon replicas revealed a sig-
nificant improvement in skin surface roughness
and surface peaks.
3.3 Methodological Quality of Included Trials
Overall, the 11 included trials were of poor
methodological quality (table II). Jadad scores
ranged from 0 to 3 and reports were frequently
lacking essential information, for instance, that
outlined in the Consolidated Standards of Re-
porting Trials (CONSORT) statement.[27] Allo-
cation to group was not concealed in any study
and the randomization procedure was described
in only three studies.[15,17,21] Blinding was also an
issue. Participants were blinded to group assign-
ment in all included trials but blinding of the
administering researcher was achieved in none.
In three studies, the reports stated that the codes
for placebo and verum allocation were disclosed
at the end of the study but there was no descrip-
tion of blinding procedures.[15,24,25] However,
Drugs Aging 2010; 27 (12)
ª 2010 Adis Data Information BV. All rights reserved.

Botanical Extracts as Anti-Aging Skin Preparations

Table II. Methodological quality and risk of bias in included trials
Study (year) JS Sequence generation Allocation Who Homogeneity Training of Power ITT analysis CONSORT AE reporting
approp.? adequately was between groups assessor calc. conducted? guidelines for
concealed? blinded? at baseline? approp.? described? herbal products[26]
Bauza et al.[15] 2 Yesa NM Unclear NM (although NM No No Fully met: 4E NM
(2002) split-face Partially met: 4C1,
design) 4C2/3

Chiu et al.[16] 1 NM NM P+OCA NM Yes No No Fully met: 4B1, Three discontinued

(2005) 4B2, 4E, 4F cream due to irritation,
Partially met: 4A1, no other AEs discussed
4C1, 4C3

Coquet et al.[17] 2 Yesa NM P+OCA NM (although NM No No Fully met: 4B2, 4B3 NM

(2005) left-right arm Partially met: 4A1,
design) 4C1, 4C2, 4D1

Garnier et al.[18] 0 NM NM P+OCA NM (although NM NM NM, unclear if all Fully met: 4B2 NM
(2005) split-face completed Partially met: 4A1,
design) 4C1, 4E

Hsu et al.[19] 2 NM NM P+OCA NM (although Yes No No (and does not Fully met: 4A2, 4B1 NM
(2007) split-face mention how Partially met: 4A1,
design) many completed) 4C1

Izumi et al.[20] 1 Quota sample (but None P+OCA Yes (for age and NM No N/A (all Fully met: 4A2, No AEs in either group
(2007) described as number and participants 4B1, 4B2, 4E
randomized) area of wrinkles) completed study) Partially met: 4C1
[inapprop.]

Janjua et al.[21] 3 Computer-generated NM P+OCA No, but adjusted Yes NM No Fully met: 4B1, No significant
(2009) randomization with ANCOVA 4B2, 4C1, 4C2, differences between
[approp.] 4C3, 4E, 4F1 groups in number or
Partially met: 4A1, severity of AEs
4B3, 4D1 (including loose stool,
chest infections)
Drugs Aging 2010; 27 (12)

Kim et al.[22] 1 NM NM P+OCA NM (although Yes No No Fully met: 4B1, 4E NM

(2008) split-face Partially met: 4A1,
design) 4C2, 4C3

Continued next page

981
982 Hunt et al.

the majority of the trials did blind the outcome as-

AE = adverse effect; ANCOVA = analysis of covariance; approp. = appropriate; calc. = calculation; CONSORT = Consolidated Standards of Reporting Trials; inapprop. = inappropriate;
NM, although authors
concluded treatment
sessor,[16-24] thus minimizing bias to some extent.
Although some of the trials provided demogra-
AE reporting

phic information for the sample, this information

was ‘safe’
was rarely provided by group (see table I) and no
None

trials provided information on co-morbidities.

NM

There was, however, a concerted effort by two

of the trials to exclude participants on the basis
Partially met: 4C1,
herbal products[26]

Partially met: 4A1,

Partially met: 4A1,

Fully met: 4B1,

of co-morbidity or previous cosmetic surgery/

Fully met: 4B2
4B2, 4C1, 4E
guidelines for

Fully met: nil

treatments of the face.[19,20] In addition, two RCTs

CONSORT

4C1, 4C2
4B2, 4B3

required all participants to refrain from changing

their current skincare routine, use of supplements
4E

or medication,[16,21] and one provided a mild

cleanser for all participants.[24] However, no stud-
ITT analysis

ies provided information on the use of other

conducted?

ITT = intent-to-treat; JS = Jadad score; N/A = not applicable; NM = not mentioned; P+OCA = participant and outcome assessor.

skincare products that may have had an impact

on skin texture and relief such as sunscreens,
NM
No

No

moisturizers and cleansers. Whilst one RCT of

described?

soy did account for oral intake,[20] this was not

Training of Power

accounted for in the other two RCTs of green

calc.

NM
No

No

tea.[16,21]
Description of drop-outs was cursory. Nine
assessor
approp.?

trials simply did not mention drop-outs[15,17-20,22-25]

Yesb

even though, in some cases, tables in the report

NM

NM

indicated that not all volunteers completed the

left-right design)
between groups

trial.[17] None of the trials conducted intent-to-

NM (although

NM (although
Homogeneity

blinded? at baseline?

treat analyses and there was an absence of power

split-face
design)

calculations in all included trials, so the role of

NM

chance was not quantified, thus reducing the re-

Could not find any information on method ‘Hazard number generator’.

liability of the results. All trials failed to define

P+OCA

P+OCA

the primary outcome, and it was often unclear

Who
was

NM

which results came from which outcomes.

Moreover, data were often poorly described.
concealed?
adequately
JS Sequence generation Allocation

This manifested as an absence of descriptions of

Unclear

statistical analyses in three trials,[18,19,22] no statis-

NM

NM

tical tests conducted in one trial,[25] no standard

deviations for the means in nine trials,[15,17-20,22-25]
and a dearth of data provided in the reports of
five trials.[17,19,20,22,23] This meant that we were
unable to conduct secondary analyses, something
approp.?

that would have been most useful, particularly

NM

NM

NM

given that suboptimal statistical tests were con-

ducted in some RCTs.[16,17,20] For instance, in
‘Dermatologist’.
1

Martelli et al.[24] 1

one trial,[16] a matched pairs test was performed,

Table II. Contd

Syed et al.[25]

yet there was no information provided on the

Study (year)

Lee et al.[23]

factors used to match pairs. In addition, partici-

(2006)

(2000)

(2002)

pants were excluded from the control group be-

cause of drop-outs from the treatment group,
a
b

ª 2010 Adis Data Information BV. All rights reserved. Drugs Aging 2010; 27 (12)
Botanical Extracts as Anti-Aging Skin Preparations
which meant that for one analysis only four pa-
tients were included. Another trial used a paired
t-test for between-group differences but did not
measure change from baseline between groups.[20]
Questionable use of statistics was also observed
in another trial,[17] in which the McNemar matched
pairs test was used. Although this test is per-
formed on dichotomous data, the outcome mea-
sures produced continuous data and there was no
explanation of the cut-offs used to transform the
variables. A paired two-sample t-test would have
been a much more appropriate test for both of
these trials.
Another important omission in the trial reports
was a thorough description of the botanicals tested.
The CONSORT guidelines for the reporting of
herbal products[26] recommend provision of detailed
information on the extraction methods, place or
origin of products. However, our analyses of the
included trials showed that none satisfactorily
complied with the CONSORT guidelines.[26] Some
of the trials fared better than others[16,20,21,23] but
still failed to provide sufficient information for
the trials to be accurately replicated and to pro-
vide standardization of ingredients. In addition
to poor description of the botanical ingredients,
there was inadequate description of the placebo
creams and vehicles used to suspend the botanical
ingredients.
Lastly, there was little attention paid to the
reporting of adverse effects (AEs). Of the 11 in-
cluded RCTs, only five mentioned AEs at
all.[16,20,21,24,25] For one of these trials, the authors
did not report on the prevalence of AEs despite
concluding that the treatment was ‘safe’.[25]
4. Discussion
Eleven trials met the inclusion criteria for this
review. All were relatively recent publications.
Although there were three trials of green tea, all
other nine trials tested different botanical agents,
which means independent replications are lack-
ing. For green tea, two studies reported no sig-
nificant differences between groups, and the third
did not conduct inter-group analyses. However,
date kernel extract, cork extract, soy extract,
Rosaceae and peony extract seemed to signif-
ª 2010 Adis Data Information BV. All rights reserved.
983
icantly reduce skin wrinkling compared with
placebo. For these extracts, a greater improve-
ment was noted after a longer treatment dura-
tion, suggesting that time is required for the
agents to take full effect. This was not the case for
green tea, however: no significant differences be-
tween groups were noted after 2 years of treat-
ment. No significant reduction in wrinkling was
noted after treatment with Vitaphenol or maca
root and we are unable to comment on the ef-
fectiveness of the combination treatment con-
taining sylibin, gotu kola and soy because it was
not statistically compared with placebo and in-
sufficient data were provided for us to perform
those analyses. The methodological quality and
reporting of the trials was moderate to poor,
which reduces confidence in any conclusions about
the effectiveness of the botanicals tested.
There was an overwhelmingly negative response
from the cosmetics companies who were con-
tacted for relevant data on their products. None
of the companies approached were prepared to
provide any information for this review. How-
ever, some of the studies included in this review
were conducted by cosmetics companies, poten-
tially leading to conflicts of interest and reducing
the reliability of the results, particularly given the
lack of transparency in the reports. In addition,
the absence of complete data and use of statistical
tests in many of the trials may suggest that some
selective reporting might have taken place.
Seven of the included trials employed the split-
face or split-arm design. This entailed partici-
pants being randomized to receive treatment to
one side of the face/arm and placebo to the other
side of the face/other arm. Although this has sev-
eral methodological advantages, including being
able to account for skin type and administration
technique (how the participants apply the cream),
this study design has a major flaw. Participants
are likely to use the same hand for both the treat-
ment and placebo creams (no reports documented
that participants were asked to wear gloves or wash
hands between cream changes) which means that
there would be contamination of products, and
treatment and placebo would not be accurately
or consistently demarcated. To some extent, this
latter issue was accounted for by focusing outcome
Drugs Aging 2010; 27 (12)
984
measurements on the distal part of the eyes (crow’s
feet), and not the facial mid-line, but contamination
was still not accounted for in these studies.
In addition to methodological problems, sev-
eral other factors bring into question the external
validity of the findings. The trials usually re-
cruited all-female samples and it is unknown how
the products would fare on other skin types such
as those from ethnic minority groups. Samples
were small, not racially diverse and were under-
powered, so even if samples were diverse, sub-
analyses could not be performed. Moreover,
most of the trials reported on different botanical
products, with the exception of soya and green
tea, which means external replication of the stud-
ies would need to be performed before conclusions
could be accurately drawn on the effectiveness of
individual botanicals.
This systematic review has several limitations.
Although the search strategy was thorough, some
clinical trials may not have been identified. How-
ever, our systematic and detailed search strategy
should have assisted in identifying all trials and in
reducing bias. It has been suggested that publi-
cation bias exists in complementary medicine,[28]
and our communications with relevant experts
did not identify any further trials, which may
suggest an effect of publication bias, particularly
given that cosmetics companies refused to present
their data. Another limitation is that our stan-
dardized quality assessment and data analyses
were based on the information and data provided
in the report and not the study itself. In the case
of studies where only a conference report was
available, trials may have been unfairly judged.
Furthermore, we included one CCT, and because
it has been shown that non-randomized studies
are more likely to produce greater effect estimates
than randomized trials,[29] this is a further cause
of bias. However, because the trial did not pre-
sent inter-group analyses we were unable to draw
conclusions on the effectiveness of its active in-
gredient. Nonetheless, we acknowledge the lim-
itations of the trial and suggest that the left-right
forearm design somewhat reduced the impor-
tance of randomization given that treatment and
placebo are taken contemporaneously and parti-
cipants were blinded to treatment allocation.
ª 2010 Adis Data Information BV. All rights reserved.
Hunt et al.
Most of the included trials eschewed safety
issues by not mentioning AEs and not doc-
umenting reasons that subjects dropped out of
the study. This is a frequent phenomenon in
complementary and alternative medicine re-
search because there is a common misconception
that ‘natural’ means ‘safe’.[30] Reporting of AEs
in trials of herbs and botanicals is scarce[31] and,
as we have shown, trials do not include the
necessary information on the extraction and origin
of the botanical products that would be required
to suitably compare trials in terms of effective-
ness. Researchers investigating botanical products
should comply with the CONSORT guidelines
for the reporting of herbal products[26] and greater
emphasis needs to be placed on the importance of
AE documentation in complementary and alter-
native medicine.
5. Conclusions
Our analysis showed that while there is some
preliminary evidence that certain botanical ex-
tracts may reduce the appearance of skin aging,
there is no evidence to suggest that this effect is
enduring. Furthermore, the low methodological
quality of the included trials limited the capacity
to draw conclusions on the effectiveness of the
botanicals tested. Lastly, very few preparations
have been shown to penetrate sufficiently deep into
the dermis to permanently reduce skin wrinkling,[7]
which may indicate that any improvement in
wrinkling elicited by these botanicals is superficial.
Acknowledgements
The contribution of S.K. Hung to this review was funded
by Schwabe Pharmaceuticals, which researches and develops
plant-based medicines and healthcare products but is not a
manufacturer of skincare or anti-aging products. No other
external sources of funding were used to assist in the conduct
or writing of this review. The authors have no conflicts of
interest that are directly relevant to the content of this review.
The authors thank Leala Watson for advice and help in de-
signing the literature search strategy.
References
1. GAO. Health products for seniors: ‘anti-aging’ products
pose potential for physical and economic harm. Report
Drugs Aging 2010; 27 (12)
Botanical Extracts as Anti-Aging Skin Preparations
to Chairman, Special Committee on Aging, US Senate.
Washington, DC: Government Accountability Office,
2001 Sep
2. Olshansky SJ, Hayflick L, Carnes BA. Position statement on
human ageing. J Gerontol A Biol Sci Med Sci 2002; 57A:
B292-7
3. Perls TT. Hope drives antiaging hype. Cleveland Clin J Med
2006; 73: 1039-44
4. Harris Interactive Inc. Anti-aging medicine, vitamins, minerals
and food supplements: a public opinion survey conducted
for the International Longevity Center. J Anti Aging Med
2003; 6: 83-90
5. Gorman C. Face-lift in a jar? Time 2000; 14 Aug: 48-52
6. Thornefeldt C. Cosmeceuticals containing herbs: fact, fiction
and future. Dermatol Surg 2005; 37: 873-80
7. Baumann LS. Less-known cosmeceuticals. Dermatol Ther
2007; 20: 330-42
8. Bogdan Allemann I, Baumann L. Antioxidants used in
skincare formulations. Skin Ther Lett 2008; 13: 1-12
9. Chiu A, Kimball AB. Topical vitamins, minerals and botanical
ingredients as modulators of environmental and chrono-
logical skin damage. Br J Dermatol 2003; 149: 681-91
10. Kamel NS, Gammack J, Cepeda O, et al. Antioxidants and
hormones as antiaging therapies: high hopes, disappoint-
ing results. Cleveland Clin J Med 2006; 73: 1049-56
11. Gao XH, Zhang L, Wei H, et al. Efficacy and safety of in-
novative cosmeceuticals. Clin Dermatol 2008; 26: 367-74
12. Calasanti T, King N. ‘Beware of the estrogen assault’: ideals
of old manhood in anti-aging advertisements. J Aging
Studies 2007; 21: 357-68
13. Jadad AR, Moore RA, Carroll D, et al. Assessing the quality
of reports of randomized clinical trials: is blinding necess-
ary? Control Clin Trials 1996; 17: 1-12
14. Cochrane Collaboration. Risk of bias tool [online]. Avail-
able from URL: http://www.ohg.cochrane.org/forms/Risk
%20of%20bias%20assessment%20tool.pdf [Accessed 2010
Aug 12]
15. Bauza E, Dal Farra C, Berghi A, et al. Date palm kernel
extract exhibits antiaging properties and significantly re-
duces skin wrinkles. Int J Tissue React 2002; 24: 131-6
16. Chiu AE, Chan JL, Kern DG, et al. Double-blinded, placebo-
controlled trial of green tea extracts in the clinical and
histological appearance of photoaging skin. Dermatol Surg
2005; 31: 855-9
17. Coquet C, Bauza E, Oberto G, et al. Quercus suber cork
extract displays a tensor and smoothing effect on human
skin: an in vivo study. Drugs Exp Clin Res 2005; 31: 89-99
18. Garnier S, Bredif S, Legrand JL, et al. Oligosaccharidic and
peptidic extract from maca root: a new cell energizer with
clinical anti-aging properties [abstract no. P02.18]. In: 14th
Congress of the European Academy of Dermatology and
ª 2010 Adis Data Information BV. All rights reserved.
View publication stats
985
Venereology; 2005 Oct 12-15; London. J Eur Acad Der-
matol Venereol 2005; 19 Suppl. 2: 47
19. Hsu J, Skover G, Goldman MP. Evaluating the efficacy in
improving facial damage with a mixture of topical anti-
oxidant. J Drugs Dermatol 2007; 6: 1141-8
20. Izumi T, Saito M, Obata A, et al. Oral intake of soy iso-
flavone aglycone improves the aged skin of adult women.
J Nutr Sci Vitaminol 2007; 53: 57-62
21. Janjua R, Munoz C, Gorell E, et al. A two-year, double-
blind, placebo-controlled trial of oral green tea polyphenols
on the long-term clinical and histological appearance of
photoaging skin. Dermatol Surg 2009; 35: 1-9
22. Kim YH, Chung CB, Kim JG, et al. Anti-wrinkle activity of
Ziuglycoside I isolated from Sanguisorba officianalis root
extract and its application as a cosmeceutical ingredient.
Biosci Biotechnol Biochem 2008; 72: 303-11
23. Lee S, Lim JM, Jin MH, et al. Partially purified paeoniflorin
exerts protective effects on UV-induced DNA damage and
reduces facial wrinkles in human skin. J Cosmet Sci 2006;
57: 57-64
24. Martelli L, Beradesca E, Martelli M. Topical formulation of
a new plant extract complex with refirming properties:
clinical and non-invasive evaluation in a double-blind trial.
Int J Cosmet Sci 2000 Jun; 22 (3): 201-6
25. Syed TA, Ahmad SA, Wong W, et al. Clinical evaluation
of 2 polyphenone (green tea extract) in a hydrophilic gel
to assess improvement in aged and damaged facial skin:
a placebo-controlled, double-blind study [abstract]. 20th
World Congress of Dermatology; 2002 Jul 1-5; Paris
26. Gagnier JJ, Boon H, Rochon P, et al. Reporting randomized-
controlled trials of herbal interventions: an elaborated
CONSORT statement. Ann Intern Med 2006; 144: 364-7
27. Moher D, Schultz KF, Altman DG. The CONSORT state-
ment: revised recommendations for improving the quality
of reports of parallel-group randomised trials. Lancet 2001;
357: 1191-4
28. Ernst E, Pittler M. Alternative therapy bias [letter]. Nature
1997; 385: 480
29. Cochrane Collaboration. Cochrane handbook: defining types
of study [online]. Available from URL: http://www.
cochrane-handbook.org/ [Accessed 2009 Feb 27]
30. Ernst E. ‘First do no harm’ with complementary and alter-
native medicine. Trends Pharmacol Sci 2007; 28: 48-50
31. Chan TYK. Monitoring the safety of herbal medicines.
Drug Saf 1997; 17: 209-15
Correspondence: Katherine J Hunt, Complementary Medicine,
Peninsula College of Medicine and Dentistry, University of
Exeter, 25 Victoria Park Road, Exeter, EX2 4NT, UK.
E-mail: katherine.hunt@pms.ac.uk
Drugs Aging 2010; 27 (12)