Institute of Cancer Research Repository
Please direct all emails to:
This is an author produced version of an article that appears in:
The internet address for this paper is:
RADIOTHERAPY AND ONCOLOGY
S Brooker, S Martin, A Pearson, D Bagchi, J Earl, L Gothard, E
Hall, L Porter, J Yarnold (2006) Double-blind, placebo-
controlled, randomised phase II trial of IH636 grape seed
proanthocyanidin extract (GSPE) in patients with radiation-
induced breast induration, Radiotherapy and Oncology, Vol.
Double-blind, placebo-controlled, randomised phase II trial of IH636 grape seed
proanthocyanidin extract (GSPE) in patients with radiation-induced breast induration
aSonja Brooker, aSusan Martin, aAnn Pearson, bDebasis Bagchi, cJudith Earl,
aLone Gothard, dEmma Hall, dLucy Porter, aJohn Yarnold
aDepartment of Radiotherapy, Royal Marsden Hospital, Sutton, Surrey, UK
bDepartment of Pharmacy Sciences, Creighton University Medical Center, Omaha, Nebraska,
cPharmacy Department, Royal Marsden Hospital, Sutton, Surrey, UK
dClinical Trials & Statistics Unit (ICR-CTSU), Institute of Cancer Research, Sutton, Surrey, UK
John Yarnold, Professor of Clinical Oncology
Academic Radiotherapy Department, Royal Marsden Hospital
Downs Road, Sutton, Surrey SM2 5PT, UK
Tel: +44 (0)20 8661 3388 / Fax: +44 (0)20 8661 3107 / Email: email@example.com
Total number of pages: 20
Running head: Grape seed extract in patients with radiation-induced fibrosis
Key words: Oligomeric proanthocyanidins, randomised clinical trial,
radiotherapy complications, fibrosis
Background and purpose: Tissue hardness (induration), pain and tenderness are common late
adverse effects of curative radiotherapy for early breast cancer. The purpose of this study was
to test the efficacy of IH636 grape seed proanthocyanidin extract (GSPE) in patients with
tissue induration after high-dose radiotherapy for early breast cancer in a double-blind
placebo-controlled randomised phase II trial.
Patients and methods: Sixty-six eligible research volunteers with moderate or marked breast
induration at a mean 10.8 years since radiotherapy for early breast cancer were randomised to
active drug (n = 44) or placebo (n = 22). All patients were given grape seed proanthocyanidin
extract (GSPE) 100 mg three times a day orally, or corresponding placebo capsules, for 6
months. The primary endpoint was percentage change in surface area (cm2) of palpable breast
induration measured at the skin surface 12 months after randomisation. Secondary endpoints
included change in photographic breast appearance and patient self-assessment of breast
hardness, pain and tenderness.
Results: At 12 months post-randomisation, ≥50% reduction in surface area (cm2) of breast
induration was recorded in13/44 (29.5%) GSPE and 6/22 (27%) placebo group patients (NS).
At 12 months post-randomisation, there was no statistically significant difference between
treatment and control groups in terms of external assessments of tissue hardness, breast
appearance or patient self-assessments of breast hardness, pain or tenderness.
Conclusions: The study failed to show efficacy of orally-administered GSPE in patients with
breast induration following radiotherapy for breast cancer.
The possibility of inducing a durable healing response in some radiation-damaged tissues is
supported by trials of hyperbaric oxygen (HBO) prior to salvage facial surgery for
radiotherapy complications. These reported highly significant reductions in the incidence of
wound dehiscence, infection and delayed healing [19; 20]. In addition, high rates of
improvement in long-standing haemorrhagic cystitis have been reported in non-randomised
studies of patients with late complications of pelvic radiotherapy . Two recent studies offer
preliminary evidence of a volume response to HBO in chronic radiation-induced arm
lymphoedema [16; 24]. The reversibility of radiation fibrosis is supported in humans by
regression of subcutaneous induration in response to bovine liposomal Cu/Zn superoxide
dismutase . However, experience with high dose alpha-tocopherol and pentoxifylline, both
having antioxidant properties, also suggests therapeutic effects. A case report by Gottlober et
al described marked regression of fibrosis 17 years after post-mastectomy radiotherapy in a
woman treated with alpha-tocopherol 400 mg once daily and pentoxifylline 400 mg three
times daily for several months . In a phase II trial testing alpha tocopherol 500 IU twice
daily and pentoxifylline 400 mg twice daily for 6 months in 43 patients with marked
subcutaneous fibrosis after radiotherapy, regression and functional improvement at 6 months
were reported in all assessable sites of induration . The same group reported significant
mean radiation-induced surface regression at 6 months with combined pentoxifylline/vitamin
E versus double placebo in a randomised study of 24 patients . A single case report of a
woman treated with pentoxifylline 400 mg three times daily and vitamin E 200 mg twice
daily for 18 months reported almost complete healing of an ulcerated radiation-induced
fibrosis after therapy for breast cancer . Further evidence of effect came from a pilot
study testing a combination of pentoxifylline 800 mg once daily, tocopherol (vitamin E) 1000
IU daily and clodronate reported clinical benefit with more than 50% regression of
progressive osteoradionecrosis observed at 6 months in 12 patients .
Free radicals and their metabolites released by ionisation of water mediate the classical
cytotoxic effects of ionising radiation via DNA strand breakage. The postulated anti-fibrotic
effects of antioxidants prescribed years after exposure cannot be related to this phase of free
radical attack. Other spheres of free radical effects are legion, in lipid membranes and
aqueous environments, inside cells and outside. Reactive oxygen species interact randomly
with molecular targets, but the biological effects are not merely stochastic, as in the case of
cell-signalling mediated by redox reactions . Processes modulated by antioxidants are
unlikely to be specific to radiotherapy, since the fibrotic response to ionising radiation has
much in common with responses to a range of chronic physical, chemical or biological
damaging agents. Reactive oxygen species may drive obliterative endarteritis, as suggested
by elevated levels of oxidised tissue methionine and VEGF in bronchial lavage fluid in
patients given combined chemo-radiotherapy for non-small cell lung cancer . Downstream
in the fibrogenic pathway, oxidant-induced hydrolysis of matrix-bound TGF-ß1 latent
complex with release of active TGF-ß1 may contribute to upregulated collagen synthesis .
Grape seed proanthocyanidin extract (GSPE, commercially available ActiVin from
InterHealth Nutraceuticals Incorporated, Benicia, CA) is a rich source of polyphenolic
antioxidants, a naturally occurring family of oligomeric proanthocyanidins found in a wide
range of fruit and vegetables [1; 22; 23]. A number of studies have demonstrated superior free
radical scavenging ability of GSPE as compared to vitamins C, E and β-carotene [1; 2]. For
example, superoxide anion levels measured by chemoluminescence response and cytochrome
c reduction were scavenged by GSPE (200 mg/l) in vitro with a maximum inhibition of
approximately 90% compared to vitamin C (19% at 100mg/l, 500 μM) and vitamin E
succinate (44% at 75 mg/l, 141.3 μM) . Similar inhibitions were observed in case of
hydroxyl radicals . Potential problems affecting in vivo testing of naturally occurring
extracts include uncertainties about whether i) relevant ingredients vary in concentration in
different batches of extract, ii) relevant ingredients are absorbed and/or modified after oral
ingestion, iii) bioactive molecules reach target tissues in sufficient molar concentrations to
exert an effect and iv) observed clinical responses are related to the postulated ingredients (in
this case proanthocyanidins).
Absorption and distribution of GSPE have been demonstrated in animal models. Oral extract
of GSPE fed to adult male ICR(CD-1) mice at a dose of 100 mg/kg for 7-10 days offers a
high degree of protection against acetaminophen-induced hepatic toxicity, amiodarone-
induced lung toxicity and doxorubicin-induced cardiac toxicity . For example, 4
experimental groups of mice were fed saline alone, GSPE alone (100 mg/Kg/day po for 9
days prior to sacrifice), doxorubicin alone (20 mg/Kg ip 48 hours prior to sacrifice) and
GSPE plus doxorubicin (the doxorubicin administered on day 7 of GSPE medication, 48
hours prior to sacrifice on day 9). Doxorubicin alone caused a six-fold rise in serum creatine
kinase, gross internucleosomal DNA fragmentation and the appearance of apoptotic nuclei on
tissue sections, changes not recorded in the animals fed GSPE plus doxorubicin . From the
in vivo effects noted in multiple organs, there seems little doubt that active ingredients are
absorbed and distributed in a wide range of mouse tissues.
In humans, 40 research subjects with hypercholesterolaemia (total cholesterol 210-300
mg/dL) were randomised in a double-blind, placebo-controlled trial to placebo 100 mg BD,
GSPE 100 mg BD, chromium polynicotinate 200 μg BD, or a combination of both . One
patient in the combination group experienced sinusitis, otitis media and recurrent migraines
following the ingestion of migraine-triggering foods, but no toxicity (predominantly
gastrointestinal disturbance) were reported in the GSPE alone group. With regard to efficacy,
there was a significant reduction in total and LDL cholesterol in the combined group
compared to placebo. The only other report of GSPE use in humans relates to 3 patients with
advanced chronic pancreatitis, in whom GSPE 100 mg po BD improved abdominal pain
without any recorded toxicity . In conclusion, the rodent data offer reliable data relating to
the efficacy of oral GSPE in preventing oxidative damage in a wide range of tissues. Limited
data raise no concerns about the safety of administering GSPE to humans.
Against this background, a double-blind, placebo-controlled randomised phase II trial was
conducted to test the efficacy of oral IH636 grape seed proanthocyanidin extract (GSPE) in
patients with tissue induration after high-dose radiotherapy for early breast cancer.
Patients and methods
Eligibility and pretreatment assessment
Eligibility criteria included palpable breast induration (tissue hardness) due to radiotherapy
for early breast cancer delivered at least 24 months prior to trial entry, freedom from cancer
recurrence, availability for follow-up and written informed consent. Pre-treatment baseline
assessments included surface measurement and clinical grading of palpable breast induration,
clinical photographs and a patient self-assessment questionnaire.
Clinical assessments of tissue induration and photographic breast appearance
Clinical examination was undertaken by an experienced research nurse (SB) to confirm trial
eligibility. SB identified an area of palpable tissue hardness (induration) in the treated breast
and outlined this in ink. The outline was traced onto transparent film, and the image was
scanned onto a computer and the area measured electronically by planimetry. SB did not have
access to baseline measurements of induration at the time of assessments 6 and 12 months
post-randomisation. Response was defined as a reduction in surface area of palpable
induration ≥50% at 12 months (e.g. a change corresponding to a plaque of induration
reducing from 6 cm to 4 cm in diameter), since this was thought to represent a clinically
worthwhile improvement that could not easily be attributed to measurement error.
The degree of breast hardness was also scored clinically by palpation using a graded scale 0 -
3 (0 = none, 1 = mild, 2 = moderated, 3 = marked). Partial response was defined as a
minimum improvement of one level of effect at 12 months and complete response as grade 0
induration at 12 months post randomization. SB did not have access to the earlier gradings of
induration at follow-up assessments. Clinical photographs were taken of the upper body
(hands on hips and above head) at baseline and 12 months post-randomisation. Follow-up
photographs were compared to baseline and scored by three independent observers blind to
treatment allocation using a graded scale (‘better’, ‘no change’ or ‘worse’ compared to
baseline photographs). Response was defined as an improvement in breast appearance at 12
Patients were asked to complete a self-assessment questionnaire evaluating a maximum of 3
sites of breast hardness, pain and tenderness before they started therapy. All three parameters
were identified and scored separately for each site on a scale 0-3 (none = 0, mild = 1,
moderate = 2, marked = 3). From the areas scored in each questionnaire, an average was
calculated, and a partial response was defined as a minimum improvement of one level of
effect at 12 months, complete response as grade 0 (no hardness/pain/tenderness) at 12
Urine and plasma analyses
Urine and blood samples were collected from all volunteers at baseline and 1 - 2 weeks
before the end of 6 months therapy, and stored in liquid nitrogen. Urinary 8-OH-
deoxyguanosine , urinary creatinine, plasma lipid peroxides  and plasma cholesterol
levels were determined as crude measures of antioxidant activity.
Trial Entry and Follow-up
Patients were randomised by a member of the Pharmacy Department at the Royal Marsden
Hospital via a telephone call to the Clinical Trials and Statistics Unit at the Institute of
Cancer Research. Patients were randomised in a 2:1 ratio (active drug:placebo) to 6 months
of IH636 grape seed proanthocyanidin extract (GSPE) 100 mg three times a day orally versus
placebo starting from day of randomisation. Neither the patients nor the investigators were
informed of the treatment allocation. Clinical assessment of breast induration was repeated at
6 and 12 months after randomisation; clinical photographs were repeated at 12 months only,
whilst patient self-assessment questionnaires were completed at 3, 6, 9 and 12 months after
start of therapy.
The primary endpoint was defined as change in area (cm2) of skin markings outlining
induration 12 months after randomisation measured by planimetry coded as a binary response
variable with response defined as a reduction in area of at least 50%. Secondary endpoints
included i) external grading of hardness, ii) photographic breast appearance, iii) patient self-
assessment of breast hardness, pain and tenderness.
Sample size and analysis
2:1 randomisation was used in order to optimise information gained on the experimental arm.
It was considered that randomisation of 72 patients (48 GSPE; 24 placebo) would allow a
30% response rate to be detected compared to zero response in the placebo group with 90%
power at α = 0.05; 2-sided. This would allow the response rate in the GSPE group to be
estimated with a standard error of 6.6%. The shelf-life of the trial medication expired before
target accrual was achieved. We recruited 66 eligible volunteers, 44 to the GSPE group and
22 to the placebo group. The recruitment of 66 volunteers reduced the power of the study
from 90% to 87%.
The primary endpoint was analysed using Fisher’s exact test. Analyses of all endpoints were
stratified by years since radiotherapy (≤10 years since radiotherapy at trial entry vs. >10
years). Analyses were also stratified by age group (≤65 years of age at trial entry vs. >65
years), since difference in age between groups was shown to be statistically significant.
Mantel-Haenszel tests for heterogeneity were used to determine if information across
stratification levels could be combined. These tests were statistically non-significant. There
was no formal statistical analysis of the randomised comparison of secondary endpoints.
There was a significant difference between treatment groups in the mean age of the
volunteers at the start of therapy, with a mean age of 62.4 years (SD = 9.2) in the treatment
group and 67.8 years (SD = 9.6) in the placebo group (p = 0.03). All volunteers had
undergone lumpectomy and radiotherapy for breast cancer, and the mean time from
radiotherapy to trial entry was 10.1 years (SD = 5.4) in the treatment group, 12.2 years (SD =
7.3) in the placebo group (p = 0.19).
61/66 (92.4%) volunteers completed their trial medication, attended 6 and 12 months
assessments and completed self-assessment questionnaires at baseline, 3, 6, 9 and 12 months
post randomisation. 5/66 (7.6%) volunteers were not evaluable at 12 months: 4/44 (9.1%)
patients in the treatment arm, 1/22 (4.5%) patients in the placebo arm. One volunteer in the
placebo group developed bone metastases two months after randomisation and was
withdrawn from the study. Three volunteers in the GSPE group withdrew from the study
before their 6-month assessment due to personal reasons unrelated to the study as far as we
can tell, and one was lost to follow-up after her 6-month assessment.
Clinical assessments of tissue induration and photographic breast appearance
At 12 months post-randomisation, 13/44 (29.5%) volunteers in the GSPE group were found
to have a reduction in induration area of ≥50%, compared to 6/22 (27.3%) in the placebo
group (Table 1). This was not statistically significant (p = 1.00). The mean reduction in
surface area of induration between the placebo group and GSPE group was –4.50 (-17.40,
8.39; p = 0.487) at 12 months post randomisation and –1.30 (-13.00, 10.40; p = 0.825) at the
end of treatment, 6 months post randomisation. In a planned subgroup analysis, 8/25 (32.0%)
volunteers in the treatment group who had radiotherapy ≤10 years before trial entry were
found to have a reduction of ≥50% in surface of palpable induration area at 12 months,
compared to 2/11 (18.2%) in the placebo group (p = 0.69). In the subgroup having
radiotherapy >10 years prior to trial entry, 5/19 (26.3%) in the treatment group were found to
have a reduction of ≥50% in induration area at 12 months, compared to 4/11 (36.4%) in the
placebo group (p = 0.69).
1/44 (2.3%) volunteers in the GSPE group was found to have an improvement in palpable
induration of 2 grades 12 months after randomisation, compared to none in the placebo group
(Table 2). 13/44 (29.5%) patients in the GSPE group had an improvement of 1 grade,
compared to 6/22 (27.3%) in the placebo group. 1/66 (1.5%) volunteer with a clinical grade
of induration = 1 at baseline was found to have a complete response at 12 months, i.e. a
clinical grade of induration = 0. This patient was in the placebo group.
At 12 months only 1/66 (1.5%) patients scored an improvement in breast appearance, and this
person was in the placebo group. 37/44 (84.1%) volunteers in the GSPE group scored no
change, compared to 18/22 (81.8%) in the placebo group, while the remaining 2/44 (4.5%)
patients in the treatment group scored a worse cosmetic outcome at 12 months than at
Patient self-assessments of tissue hardness, pain and tenderness
The difference in the baseline and 12 months averages was used as this secondary endpoint.
At 12 months post randomisation, 1/44 (2.3%) evaluable patients in the GSPE group and 1/22
(4.5%) in the placebo group scored an improvement of more than 2 grades (Table 3). 22/44
(50%) in the treatment group scored an improvement of less than 1 grade, compared to 10/22
(45.5%) in the placebo group, whilst 11/44 (25.0%) volunteers in the GSPE group and 9/22
(40.9%) in the placebo group scored no change or a deterioration of their tissue hardness.
Analyses of the change in maximum areas of tissue hardness from baseline to 12 months did
not show any evidence for an improvement (data not shown).
Table 4 shows results of patients’ self-assessment of breast pain and tenderness. 5/44 (11.4%)
volunteers in the GSPE group and 2/22 (9.1%) in the placebo group recorded a change of
between 1 and 2 grades. 12/44 (27.3%) in the treatment arm scored between 0 and 1 grade
change compared to 7/22 (31.8%) in the placebo arm, whilst the remaining 22/44 (50.0%) of
the patients in the treatment arm and 12/22 (54.5%) in the placebo arm scored no change or
an increase in pain and tenderness.
Serum & urinary analyses
Tests for antioxidant effect in serum and urine were performed on all women allocated study
drug in whom samples were collected and in 10 controls, 39 volunteers in all (Table 5).
The study failed to demonstrate efficacy of IH636 grape seed proanthocyanidin extract
(GSPE) in patients with radiation-induced breast induration, although a planned subgroup
analysis of the primary endpoint by time raises the possibility of an effect in women treated
within 10 years of radiotherapy. Analysis of the secondary endpoints viz, external
assessments of breast hardness and patient self-assessments of breast hardness, pain and
tenderness raise no concerns that a therapeutic effect has been missed. According to informal
patients’ verbal comments recorded at the time of follow up assessments, 11/44 (25.0%)
patients in the GSPE group and 4/22 (18.1%) in the placebo group reported tissue softening,
whilst 6/44 (13.6%) women in the treatment group commented on a reduction in pain and
tenderness compared to 2/22 (9.1%) in the placebo group.
A limitation of this and other studies testing clinical modulation of late normal tissue
responses in the breast after radiotherapy is the scarcity of reliable endpoints. It could be
argued that since volunteers were unable to notice and record any changes in breast
symptoms, it is unlikely that worthwhile treatment effects have been missed. It is striking that
the primary endpoint (clinical assessment of palpable induration) was associated with a
placebo response of 27%. This reflects gross inconsistencies in marking out the area of
palpable radiation-induced breast induration on the skin surface, and is consistent with
reports from a previous trial . No suggestion of an effect was seen at 6 months post-
randomisation, so it is unlikely that a temporary therapeutic response has been missed (Tables
The serum and urine assays were introduced to gain some estimate of compliance and bio-
availability, although it is recognized that inter-assay and inter-individual variation are wide.
Although no significant differences between treatment arms were seen, this does not exclude
a high level of compliance and the absorption of biologically active metabolites. Where the
latter is concerned, there is evidence in humans that these involve metabolism by intestinal
microflora of high molecular weight proanthocyanidins to phenolic acids representing the
biologically active components .
In conclusion, the study fails to show efficacy of GSPE in patients with breast induration
following radiotherapy for breast cancer. Although the numbers are small and the effect is not
statistically significant, there is a possibility that GSPE may benefit patients if administrated
within a few years of radiotherapy. This suggestion of effect would require independent
confirmation before being accepted as real.
We are grateful for the commitment shown by the volunteers participating in this study, and
we acknowledge Cancer Research UK for funding the work. Particular thanks go to Prof
Frank Kelly and Ms Chrissi Dunster of Kings College, London who performed the analyses
of urinary 8-OH-deoxyguanosine, urinary creatinine, plasma lipid proxides and plasma
1. Bagchi D, Bagchi M, Stohs SJ, et al. Free radicals and grape seed proanthocyanidin
extract: importance in human health and disease prevention. Toxicology 2000; 148(2-
Bagchi D, Garg A, Krohn RL, et al. Oxygen free radical scavenging abilities of
vitamins C and E, and a grape seed proanthocyanidin extract in vitro. Res Commun
Mol Pathol Pharmacol 1997; 95(2): 179-189.
Bagchi D, Ray SD, Patel D, Bagchi M. Protection against drug- and chemical-
induced multiorgan toxicity by a novel IH636 grape seed proanthocyanidin extract.
Drugs Exp Clin Res 2001; 27(1): 3-15.
Banerjee B, Bagchi D. Beneficial effects of a novel IH636 grape seed
proanthocyanidin extract in the treatment of chronic pancreatitis. Digestion 2001;
Barcellos-Hoff MH, Dix TA. Redox-mediated activation of latent transforming
growth factor-beta 1. Mol Endocrinol 1996; 10(9): 1077-1083.
Beinert T, Binder D, Oehm C, et al. Further evidence for oxidant-induced vascular
endothelial growth factor up-regulation in the bronchoalveolar lavage fluid of lung
cancer patients undergoing radio-chemotherapy. J Cancer Res Clin Oncol 2000;
Bevers RF, Bakker DJ, Kurth KH. Hyperbaric oxygen treatment for haemorrhagic
radiation cystitis. Lancet 1995; 346(8978): 803-805.
Cordis GA, Maulik G, Bagchi D, et al. Detection of oxidative DNA damage to
ischemic reperfused rat hearts by 8-hydroxydeoxyguanosine formation. J Mol Cell
Cardiol 1998; 30(10): 1939-1944.
Delanian S, Baillet F, Huart J, et al. Successful treatment of radiation-induced fibrosis
using liposomal Cu/Zn superoxide dismutase: clinical trial. Radiother Oncol 1994;
Delanian S, Depondt J, Lefaix JL. Major healing of refractory mandible
osteoradionecrosis after treatment combining pentoxifylline and tocopherol: a phase II
trial. Head Neck 2005; 27(2): 114-123.