Research Article

Received: 17 July 2008, Revised: 17 November 2008, Accepted: 8 January 2009 Published online 16 March 2009 in Wiley Interscience

(www.interscience.wiley.com) DOI 10.1002/pca.1119

Characterisation of Tannin-Containing Herbal

John Wiley & Sons, Ltd.

Drugs by HPLC
Charlotte Møller, Steen Honoré Hansen* and Claus Cornett
Characterisation of tannin-containing herbal drugs by HPLC

ABSTRACT:
Introduction – Herbal drugs containing tannins are characterised in the European Pharmacopoeia by their tannin content
analysed by the old, nonspecific, colorimetric Folin–Ciocalteu method. The result of the analysis is a single figure relating to
the content of tannins, but this does not provide much information on the identity or status of the herbal drug.
Objective – In the present paper methods for obtaining more detailed information of the constituents in these herbal drugs
are described.
Methodology – The methods developed are based on a reversed-phase gradient HPLC system coupled to DAD, fluorescence,
electrochemical and MS detectors.
Results – The HPLC system developed provides characteristic fingerprints of the herbal drugs when using UV detection at
250 nm. The fingerprints may be used for identification of tannin-containing herbal drugs. Methanolysis of the herbal drug
generated methyl gallate and ellagic acid, which were analysed in the HPLC system. The molar ratio between methyl gallate
and ellagic acid may also be used for the identification of herbal drugs.
Conclusions – An HPLC system equipped with selective detectors was shown to be valuable in the identification of herbal tan-
nin. Most promising was fingerprinting using UV detection, but methanolysis followed by HPLC also proved useful. Copyright
© 2009 John Wiley & Sons, Ltd.

Keywords: HPLC; electrochemical detection; tannins; proanthocyanidins; methanolysis; gallotannins; ellagitannins

Introduction
The term ‘tannins’ is often used as a general term for a large
group of plant polyphenols. Tannins are polyphenols containing
a number of phenolic and sometimes also carboxylic acid groups
that facilitate strong complexation to proteins and other macro-
molecules as well as to minor cations. Tannins are widely distributed
in plants, and their presence is considered a defence mechanism
against pathogens and herbivores.
Several definitions of tannins exist. In 1962, Bate-Smith and
Swain defined tannins as, ‘Water soluble phenolic compounds
having molecular weight between 500 and 3000 and, besides
giving the usual phenolic reactions, they have special properties
such as the ability to precipitate alkaloids, gelatine and other
protein’ (Bate-Smith and Swain, 1962). This definition is still
generally accepted. Later on, tannins were divided into groups
according to their structures (Haslam, 1989), and this was further
elaborated by Khanbabaee and Van Ree, who subdivided the
tannins into gallotannins, ellagitannins, proanthocyanidins and
more complex tannins (Khanbabaee and Van Ree, 2001). Examples
of some of the types of tannins are given in Fig. 1 together with
the structure of ellagic acid.
Gallotannins consist of gallic acid esterified to a polyol. Ella-
gitannins are formed from oxidative coupling of two galloyl
groups in a gallotannin to form hexahydroxydiphenic acid (HHDP).
Hydrolysis will transform gallotannins and ellagitannins into
gallic acid and ellagic acid, respectively. Proanthocyanidins are
typically dimers, trimers or polymers of catechin and epicat-
echin. Complex tannins are built from gallotannins or ellagitannins
together with catechin or other minor polyphenols.
The pharmacological aspects of polyphenols and tannins are
the major threats to human health (Anonymous, WHO, 2004). Poly-
phenols are assumed to decrease the mortality rate related to
these diseases (De Curtis et al., 2005; Yamada and Watanabe,
2007). The mechanism is considered to be related to the anti-
oxidant properties of the polyphenols, regulating the amount of
free radicals and singlet oxygen in the body (De Bruyne et al.,
1999). However, polyphenols may also have a negative effect on
health because they form complexes with iron and therefore
reduce the bioavailability of this essential element (Santos-Buelga
and Scalbert, 2000).
A number of herbal drugs included in The European Pharma-
copoeia (Ph. Eur.) are quantified only by their tannin content
(Anonymous, 2007). The content is given as a minimum amount
(as a percentage) determined using the relative nonspecific
Folin-Ciocalteu method—an old colorimetric method involving
redox chemistry (Folin and Ciocalteu, 1927). The content is
calculated relative to a pyrogallol standard and expressed as
percent of pyrogallol. However, tannins are a large and diverse
group of compounds and the tannin assay gives no information
about the actual content of the different tannins in herbal drugs.
The purpose of the present study is to obtain more detailed
knowledge about the content of tannins and phenolic compounds
* Correspondence to: S. H. Hansen, Department of Pharmaceutics and Analyt-
ical Chemistry, Faculty of Pharmaceutical Sciences, University of Copenha-
gen, Universitetsparken 2, DK-2100 Copenhagen, Denmark. E-mail:
shh@farma.ku.dk
Department of Pharmaceutics and Analytical Chemistry, Faculty of Pharma-
ceutical Sciences, University of Copenhagen, Universitetsparken 2, DK-2100
231

often discussed in relation to cardiac diseases and cancer, two of Copenhagen, Denmark

Phytochem. Anal. 2009; 20: 231–239 Copyright © 2009 John Wiley & Sons, Ltd.

Figure 1. The structure of ellagic acid together with examples of gallotannins, ellagitannins and
proanthocyanidins.
in herbal drugs using HPLC with various detection techniques.
The results obtained from the HPLC analysis are compared
with the results obtained with the method from the Ph. Eur. The
chromatograms of the herbal drugs are compared in order
to explore their use as fingerprints for the identification of
tannin-containing herbal drugs.
Experimental
Plant Material. The herbal drugs used for the experiments are
listed in Table 1. They were either purchased in local Danish
drugstores or donated by NaturDrogeriet A/S (Hørning, Denmark)
or Finzelberg GmbH and Co (Andernach, Germany).
Reagents. Pyrogallol, gallic acid, methyl gallate, chlorogenic acid,
catechin hydrate, ellagic acid, 4-aminobenzoic acid, salicylamide
and acetyl chloride were from Sigma/Aldrich (Buchs, Switzer-
land). Hide powder was from Research Institute for Leather and
Plastic Sheeting (Freiburg, Germany), Folin–Ciocalteu Phenol
Reagent and formic acid were from Merck (Darmstadt, Germany)
232
www.interscience.wiley.com/journal/pca Copyright © 2009 John Wiley & Sons, Ltd.
C. Møller et al.
and methanol was from VWR (Leicester, UK). All chemicals were
of the highest analytical grade available. Anhydrous methanolic
hydrochloric acid was prepared by adding 1 vol. of acetyl
chloride to 4 vols of methanol cooled on an ice bath.
HPLC. An Agilent (Walbronn, Germany) 1100 HPLC system,
equipped with a G1379A on-line degasser, a G1313A autosam-
pler, a G1316A column oven and a G1376A binary pump, was
connected to three detectors in series, namely, a G1365B multi-
wavelength detector operated at 250, 280 and 320 nm, a G1321A
fluorescence detector operated at the excitation wavelength
280 nm and emission wavelength 318 nm, and an electrochemi-
cal amperometric detector II from Antec Leyden (Zoeterwoude,
Holland) operated at 0.6 V. All data were collected using Chem-
station software. A Phenomenex (Torrance, CA, USA) Luna C18(2)
column (250 × 4.6 mm i.d.; 5 μm) was eluted in gradient mode
with mobile phase A consisting of methanol, water and formic
acid (5:95:0.1 v/v/v) and mobile phase B consisting of methanol
and formic acid (100:0.1 v/v). Linear gradients between the time
points were as follows: 0 min, 5% B; 7 min, 18% B; 11 min, 18% B;
Phytochem. Anal. 2009; 20: 231–239
Characterisation of tannin-containing herbal drugs by HPLC

Table 1. Tannin content of the dried herbal drug determined according to Ph. Eur. 2.8.14. and calculated as percentage pyro-
gallol. The assay was performed in duplicate and all results are given

Herbal drug Batch 1 Batch 2 Batch 3 Batch 4 Ph. Eur.

Requirement
English Ph. Eur. name/Latin plant name
Agrimony
Agrimonia eupatoria L. 4.1; 3.9 3.9; 3.7 4.3; 3.8 Min. 2.0
Alchemilla
Alchemilla vulgaris L. 7.1; 7.3 9.4; 10.0 7.6; 7.5 7.8; 7.7 Min. 6.0
Bilberry fruit, dried
Vaccinium myrtillus L. 1.1; 1.0 1.1; 1.6 1.0; 1.0 1.4; 0.8 Min 1.0
Oak bark
Quercus robur, Q. petraea and Q. pubescens 3.1; 3.3 3.9; 4.2 3.4; 3.0 Min. 3.0
Hamamelis leaf
Hamamelis virginiana L. 6.2; 7.2 4.0; 3.9 3.2; 3.8 Min 3.0
Loosestrife
Lythrum salicaria L. 5.2; 5.1 Min 5.0
Rhatany Root
Krameria triandra Ruiz et Pavon 7.0; 6.6 6.2; 6.1 Min 5.0
Tormentil
Potentilla erecta Raeusch 11.3; 11.4 10.3; 10.8 11.9; 12.0 11.1; 10.8 Min. 7.0

15 min, 25% B; 20 min, 50% B; 25 min, 75% B; 26 min, 100% B; acid was, however, dissolved in DMSO. From the stock solution,
29 min, 100% B; and 30 min, 5% B. The flow rate was 1.0 mL/min five calibration standards were prepared by dilution with 50%
and the column temperature was maintained at 40ºC. methanol–water.

HPLC-MS. An Agilent G2446A 1100 LC/MSD Trap equipped with
a G1322A on-line degasser, a G1313A autosampler, a G1316A
column oven, a G1312A binary pump and a G1315B photodiode
array detector was used. The separation method was as described
under HPLC. The ion-trap was operated with electrospray ionisa-
tion in the negative mode using the following parameters: HV
capillary, 4000 V; dry gas (nitrogen), 12 L/min at 350°C; nebuliser,
70 psig; and trap drive, 70.
Spectrophotometry. The colorimetric assay of tannins was
performed according to Ph. Eur. 2.8.14. ‘Determination of
tannins in herbal drugs’, but before the aqueous extraction,
4-aminobenzoic acid (as HPLC internal standard) was added
to the mixture to give a final concentration of 0.01 mg/mL. The
extraction was scaled down by a factor of 10 and centrifugation
of the extract was used instead of filtration. Measurement of
UV absorption in the Ph. Eur. assay was performed using a Shimadzu
UV 1700 instrument (Columbia, MD, USA) with a detection wave-
length of 760 nm.
Loss on Drying. All results given in this article are calculated
based on the amount of dry herbal drug. Loss on drying was
performed in accordance with assay 2.2.32 ‘Loss on drying’,
method (d), in Ph. Eur. (drying in an oven at 105°C). The drying
time for the herbal drugs is stated in the respective monograph.
Sample Preparation for HPLC. The aqueous herbal drug extract
prepared for the Ph. Eur. assay (i.e. extraction with water on a water
bath (100°C) for 30 min) was used without further sample prep-
aration.
Calibration Standards. Stock solutions with a concentration of
0.5 mg/mL were prepared in 95% methanol and 5% water. Ellagic
Methanolysis. A 100 mg sample of the herbal drug was weighed
into a Teflon container (CEM Corporation, Matthews, NC, USA),
and 10 mL of anhydrous methanolic hydrochloric acid was
added together with 500 μL of 2 mg/mL salicylamide in methanol
used as an internal standard. The capped container was heated
for 2 h on a steam bath. After cooling to room temperature, the
contents were transferred quantitatively to a 50 mL measure-
ment glass with 50% methanol–water. The contents were centri-
fuged before HPLC analysis. The calibration standards for methyl
gallate and ellagic acid were prepared as described under Cali-
bration standards.
Results and Discussion
The Ph. Eur. Assay for Tannins
The spectrophotometric assay of tannins using Folin–Ciocalteu
reagent is based on the reduction of a phospho–molybdate–
wolframate complex to give a blue colour as the phenolic com-
pounds are oxidised. However, the assay does not discriminate
between tannins and other small oxidisable compounds. There-
fore, the assay is designed as a differential method, subtracting
the value corresponding to the presence of small molecular
phenols from the total value. This is performed by using hide
powder, a collagen product obtained from bovine hide. Having
obtained a total value for the content of tannins and oxidisable
compounds, the major tannins are removed by adsorption to
the hide powder. The small phenolic compounds will remain in
solution, and the amount of small phenolic compounds is
measured and the value obtained is subtracted from the total
value.
The amounts of tannins found in a number of herbal drugs
are given in Table 1, together with the requirements stated in
233

Phytochem. Anal. 2009; 20: 231–239 Copyright © 2009 John Wiley & Sons, Ltd. www.interscience.wiley.com/journal/pca
 C. Møller et al.

Ph. Eur. A number of the herbal drugs are represented by differ-
ent batches. All of the batches analysed fulfilled the require-
ments of Ph. Eur.
The repeatability of the method, expressed as the relative
standard deviation (RSD) of the pyrogallol calibration standard,
was found to be 4.9% (n = 9) (data not shown). The variation
may be larger when analysing the herbal drugs due to the lack
of homogeneity of the plant materials.
HPLC Detection
An HPLC method was developed in order to obtain more detailed
information on the tannins present in the herbal drugs. The
HPLC method was optimised in order to separate and quantify
the following phenolic compounds: ellagic acid, gallic acid,
chlorogenic acid, catechin and methyl gallate. 4-aminobenzoic
acid was chosen as the internal standard because it does not
interfere with any of the analytes, is water soluble, does not give
a response in the Ph. Eur. assay for tannins, and is detectable
using UV spectrophotometry as well as fluorescence. In Fig. 2,
the chromatograms of the phenolic reference standards, obtained
using different detection methods, are shown.
Catechin and epi-catechin may be detected selectively using
fluorescence, but the most convenient overall detection for all
standards was obtained using UV absorption at 250 nm. Repre-
sentative chromatograms of eight herbal drugs obtained with
UV detection at 250 nm are shown in Fig. 3.
Quantification
The quantitative content of a number of compounds has been
determined as given in Table 2. All samples were analysed in
duplicate, and all data are given in order to give an impression
of the repeatability of the method. The relative standard devia-
tion (n = 6) was calculated for one batch of Tormentil, and was
found to be between 2.6 and 6%, depending on the quantified
compound in question (data not shown). The observed variance
may be ascribed mainly to the lack of homogeneity of the samples.
Unfortunately, no reference standards for all of the proantho-
cyanidins were available. The proanthocyanidins were identified
by mass spectrometry, and the quantitative content of these
substances was calculated using fluorescense detection and
catechin as a standard. The proanthocyanidins were identified
using single-ion extraction of the following m/z: 289 (monomer),
577 (dimer), 865 (trimer), 1153 (tetramer) and 720 (pentamer,
double charged). The mass spectra were compared with those
published by Wollgast (2004). The total peak area for all of the
peaks in the chromatogram monitored by electrochemical
detection at 0.6 V was expressed as ‰ pyrogallol. The results
indicate the amount of oxidisable groups in the herbal drug,
and may also give an indication of the antioxidative effect of the
herbal drug.
The results given in Table 2 show major differences in the
tannins and phenolic compounds present in the different herbal
drugs. Minor differences were found between the concentrations
of each compound determined within different batches of the
herbal drug, whereas there were marked differences between
different herbal drugs. The differences between the herbal drugs
are also deducible from the chromatograms depicted in Fig. 3,
and therefore the chromatograms may be used as fingerprints
for the characterisation and identification of the herbal drugs.
The results were obtained with a simple aqueous extraction and
thus only show the content of easily extractable compounds
that are not bound tightly to the plant material.
Hide Powder
Hide powder is used the in the Ph. Eur. assay to remove tannins
in order to be able to measure the contribution from small phe-
nolic compounds to the total value. The ability of hide powder
to remove tannins from the extract was evaluated using HPLC
by analysing the extracts before and after adding hide powder.

Figure 2. Chromatograms of a standard mixture containing five phenolic compounds. (1) Gallic acid; (2)
catechin; (3) methyl gallate; (4) chlorogenic acid; (5) ellagic acid. The chromatograms were obtained using
gradient elution as described under HPLC. Detection with (A) UV 250 nm; (B) UV 280 nm; (C) electrochemi-
234

cal detection, 0.6 V; (D) fluorescence 280/318 nm; (E) UV 320 nm.

www.interscience.wiley.com/journal/pca Copyright © 2009 John Wiley & Sons, Ltd. Phytochem. Anal. 2009; 20: 231–239
Characterisation of tannin-containing herbal drugs by HPLC

Table 2. Quantitative data expressed as weight (‰) of the dried herbal drug obtained using HPLC-UV-Flu-ED

Herbal drug Batch ED Pyrogallola Gallic Ellagic Chlorogenic Catechin Epicatechinb Proanthocyanidinsb
(‰) acid (‰) acid (‰) acid (‰) (‰) (‰ catechin) (‰ catechin)
Agrimony A 1.32 n.d. 2.09 n.d.c 5.68 0.47 5.44
Agrimony A 1.32 n.d. 1.88 n.d. 6.18 0.51 6.53
Agrimony B 1.39 n.d. 1.46 0.25 0.72 1.75 8.28
Agrimony B 1.33 n.d. 1.19 0.23 0.64 1.61 8.09
Agrimony C 1.42 0.33 3.53 n.d. 2.68 1.74 5.23
Agrimony C 1.25 0.30 3.14 n.d. 2.44 1.49 4.78
Agrimony C 1.19 0.28 2.96 n.d. 2.58 1.64 5.49
Alchemilla A 4.83 1.19 15.34 n.d. 0.36 n.d. 0.46
Alchemilla A 5.44 1.49 17.80 n.d. 0.37 n.d. 0.49
Alchemilla B 4.99 n.d. 17.79 n.d. n.d. n.d. 1.30
Alchemilla B 5.77 n.d. 19.77 n.d. n.d. n.d. 1.21
Alchemilla C 4.65 0.71 24.74 1.16 0.19 n.d. 0.60
Alchemilla C 4.58 0.81 24.55 1.03 0.24 n.d. 0.67
Alchemilla D 4.22 0.71 25.69 0.49 0.15 n.d. 0.85
Alchemilla D 3.94 0.66 23.43 0.49 0.14 n.d. 0.81
Bilberry fruit, dried A 1.24 0.21 n.d. 8.87 0.06 0.12 0.64
Bilberry fruit, dried A 1.14 0.20 n.d. 8.24 0.05 0.09 0.50
Bilberry fruit, dried B 1.50 0.26 n.d. 5.23 n.d. 0.11 0.90
Bilberry fruit, dried B 1.38 0.18 n.d. 4.67 n.d. 0.11 0.66
Bilberry fruit, dried C 1.38 0.26 n.d. 4.00 n.d. 0.27 0.46
Bilberry fruit, dried C 1.06 0.18 n.d. 2.81 n.d. 0.19 0.72
Bilberry fruit, dried D 1.11 0.51 n.d. 1.36 n.d. 0.12 0.77
Bilberry fruit, dried D 1.05 0.48 n.d. 1.26 n.d. 0.12 0.72
Oak bark A 1.99 0.64 0.54 n.d. 2.19 n.d. 3.15
Oak bark A 1.79 0.81 0.44 n.d. 2.56 n.d. 3.91
Oak bark B 1.33 0.55 0.99 n.d. 3.68 0.34 1.32
Oak bark B 1.37 0.60 0.79 n.d. 3.88 0.46 1.35
Oak bark C 1.89 1.48 n.d. n.d. 2.27 0.12 3.46
Oak bark C 1.92 1.40 n.d. n.d. 2.02 0.17 3.67
Oak bark D 2.91 0.81 0.37 n.d. 2.32 n.d. 6.05
Oak bark D 2.69 0.93 0.38 n.d. 2.34 n.d. 5.56
Hamamelis leaf A 7.08 5.40 1.47 n.d. 0.40 n.d. 0.29
Hamamelis leaf A 6.09 5.13 1.57 n.d. 0.39 n.d. 0.41
Hamamelis leaf B 8.60 7.70 0.87 1.40 1.34 n.d. 0.81
Hamamelis leaf B 9.20 6.83 0.94 1.48 1.33 n.d. 0.58
Hamamelis leaf C 5.76 5.52 1.33 0.65 0.29 n.d. 0.23
Hamamelis leaf C 5.27 7.13 1.48 0.74 0.35 n.d. 0.27
Loosestrife A 1.87 0.49 7.39 n.d. n.d. n.d. 0.43
Loosestrife A 1.11 0.57 2.77 n.d. n.d. n.d. 0.61
Rhatany Root A 1.14 n.d. n.d. 3.84 0.46 1.27 1.35
Rhatany Root A 0.98 n.d. n.d. 3.74 0.50 1.36 1.39
Rhatany Root B 0.96 n.d. 0.23 2.18 0.66 1.37 1.89
Rhatany Root B 0.94 n.d. 0.19 2.05 0.61 1.29 1.62
Tormentil A 3.94 n.d. 9.95 n.d. 18.07 1.01 21.40
Tormentil A 3.58 n.d. 9.12 n.d. 17.85 1.23 23.07
Tormentil B 2.87 n.d. 5.21 n.d. 17.86 0.89 16.08
Tormentil B 2.78 n.d. 3.77 n.d. 16.00 0.71 15.80
Tormentil C 4.98 n.d. 7.33 n.d. 13.69 0.95 28.51
Tormentil C 4.43 n.d. 5.89 n.d. 13.52 0.51 30.54
Tormentil D 4.49 n.d. 8.68 n.d. 15.54 0.96 24.13
Tormentil D 4.08 n.d. 6.60 n.d. 14.79 0.62 26.08
a
The total area under the curve from the electrochemical detector, expressed as weight (‰) pyrogallol. b Calculation based on the
calibration curve for catechin. c n.d., not detected.
235

Phytochem. Anal. 2009; 20: 231–239 Copyright © 2009 John Wiley & Sons, Ltd. www.interscience.wiley.com/journal/pca
 C. Møller et al.

Figure 3. Chromatograms of the eight different herbals drugs using UV detection at 250 nm. As internal
standard (IS), 4-aminobenzoic acid was used.

Figure 4 shows the chromatograms of Alchemilla and Tormentil
before and after adding hide powder. The chromatograms
clearly demonstrate that hide powder removes a large quantity
of ellagic acid and some catechin. Only 8% ellagic acid (UV
detection at 250 nm) and around 75% of catechin (fluorescence
detection) were left in solution after the addition of hide powder.
HPLC-MS was used to evaluate the removal of tannins by extract-
236
783 (corresponding to a single charged as well as a double
charged ellagitannin) and at 934 (probably double charged
agrimoniin). A nearly complete removal of gallotannins and
ellagitannins was obtained using hide powder (data not shown).
The dimer proanthocyanidin shown in Fig. 4 was, however,
only partly removed using hide powder. Binding of ellagic acid,
catechin and other small phenolic compounds to hide powder

ing the mass to charge ratios corresponding to the tannins at leads to an overestimation of the tannin contents in the herbal

www.interscience.wiley.com/journal/pca Copyright © 2009 John Wiley & Sons, Ltd. Phytochem. Anal. 2009; 20: 231–239
Characterisation of tannin-containing herbal drugs by HPLC
Figure 4. Chromatograms of (A) Tormentil before addition of hide powder; (B) Tormentil after treatment with hide powder; (C) Alchemilla before
addition of hide powder; (D) Achemilla after treatment with hide powder. Chromatograms A and B obtained using fluorescence detection; chromato-
grams C and D obtained using UV detection at 250 nm.
drugs. The result of the assay will remain consistent as the same
amount of small phenolic compounds is removed every time.
The results cannot be used for comparing the tannin contents
among different herbal drugs as the tannin content is affected
by the presence of other phenolic compounds. It was further
shown that centrifugation can be employed instead of filtration
without affecting the results (data not shown).
Methanolysis
Data concerning the total amount of gallotannins and ellagitan-
nins present in herbal drugs may be obtained using methanolysis.
The treatment of the herbal drug with anhydrous methanolic hydro-
chloric acid transformed gallotannins and ellagitannins into methyl
gallate and ellagic acid, and these two compounds were quantified
using HPLC. The difference between the contents of gallic acid
(methyl gallate) and ellagic acid before and after methanolysis
originates from the decomposition of gallotannins and ellagi-
tannins. Thus, the data in Table 3 are corrected for the amounts
of free gallic acid and ellagic acid obtained by direct extraction.
Release of tightly bound gallic acid or ellagic acid during
the methanolysis or decomposition of ellagic glycosides may
also contribute to the amount of methyl gallate and ellagic acid
found after methanolysis. Furthermore, the solubility of ellagic acid
is much higher in methanol compared with water, which may
also result in a higher concentration of ellagic acid determined
after methanolysis. Methanolysis is, however, the preferable method
for determining whether gallotannins or ellagitannins are the
dominating type of tannins in the herbal drug.
Phytochem. Anal. 2009; 20: 231–239 Copyright © 2009 John Wiley & Sons, Ltd.
A single herbal drug may also be characterised by the molar
ratio between ellagic acid and methyl gallate along with the
total content of these two compounds. The ratio and the total
amount of ellagic acid and methyl gallate seem to be character-
istic for the herbal drug. Only minor variations are seen between
different batches of a herbal drug.
Correlation between Ph. Eur. and HPLC Results with
Electrochemical Detection
The results from the HPLC analysis after extraction with water
and using electrochemical detection at 0.6 V were assumed to
show some correlation with the results obtained from the Ph.
Eur. assay as both methods rely on redox processes. The final results
from the Ph. Eur. is, however, a measure of the tannin content.
Therefore the correlation is based on the results obtained from
the measurement of the total content of tannins and other oxi-
disible compounds before adding hide powder. No correlation
between the methods was found, which is surprising since Blasco
et al. (2005) revealed a correlation between the results from the
Folin–Ciocalteu method and the FIA-ED in a study based on 11
different fruits, juice, wine and beans.
Conclusions
The characterisation of tannin-containing herbal drugs was con-
ducted using HPLC with UV, fluorescence and electrochemical
detection, and the results were compared with data obtained
237
using the tannin assay performed according to Ph. Eur. The HPLC
www.interscience.wiley.com/journal/pca
 C. Møller et al.

Table 3. The content of methyl gallate and ellagic acid after methanolysis expressed as μmol/mg dried herbal drug. Data are cor-
rected for the amounts of free gallic acid and ellagic acid obtained by direct extraction

Herbal drug Batch Methyl gallate Ellagic acid Molar Sum

(μmol/mg) (μmol/mg) ratioa (μmol/mg)b
Agrimony A 1.6 20 12.8 22
Agrimony A 1.7 22 12.7 24
Agrimony B 2.1 13 5.9 15
Agrimony B 3.6 19 5.2 23
Agrimony C 3.1 26 8.4 29
Agrimony C 2.3 23 9.7 25
Agrimony C 2.2 23 10.4 25
Alchemilla A 18 240 12.9 258
Alchemilla A 19 290 14.5 309
Alchemilla B 22 280 12.8 302
Alchemilla B 17 260 14.8 277
Alchemilla C 21 180 8.6 201
Alchemilla C 18 190 10.6 208
Alchemilla D 19 240 12.6 259
Alchemilla D 21 270 13.3 294
Bilberry fruit, dried A n.d. n.d. n.d.c n.d.
Bilberry fruit, dried A n.d. n.d. n.d. n.d.
Bilberry fruit, dried B 3.7 n.d. n.d. 3.7
Bilberry fruit, dried B 4.2 n.d. n.d. 4.2
Bilberry fruit, dried C 4.2 n.d. n.d. 4.2
Bilberry fruit, dried C 2.5 n.d. n.d. 2.5
Bilberry fruit, dried D n.d. n.d. n.d. n.d.
Bilberry fruit, dried D n.d. n.d. n.d. n.d.
Oak bark A 79 25 0.3 104
Oak bark A 80 27 0.3 107
Oak bark B 29 32 1.1 61
Oak bark B 29 32 1.1 61
Oak bark C 46 20 0.4 66
Oak bark C 45 21 0.5 66
Oak bark D 84 32 0.4 116
Oak bark D 85 33 0.4 118
Hamamelis leaf A 370 3.5 0.009 374
Hamamelis leaf A 380 2.8 0.007 383
Hamamelis leaf B 360 1.3 0.004 361
Hamamelis leaf B 350 0.89 0.003 361
Hamamelis leaf C 350 3.4 0.010 354
Hamamelis leaf C 360 8.7 0.024 369
Loosestrife A 5.1 55 10.9 60
Loosestrife A 5.5 67 12.3 73
Rhatany Root A n.d. 8.8 n.d. 8.8
Rhatany Root A n.d. 9.8 n.d. 9.8
Rhatany Root B n.d. 13 n.d. 13
Rhatany Root B n.d. 13 n.d. 13
Tormentil A 4.4 100 23.3 105
Tormentil A 3.9 110 28.8 114
Tormentil B 5.9 99 16.9 105
Tormentil B 5.3 110 19.8 115
Tormentil C 5.2 160 30.3 165
Tormentil C 3.6 150 40.2 154
Tormentil D 5.5 150 27.6 156
Tormentil D 5.6 150 25.9 156
a
Molar ratio = molellagic acid/molmethyl gallate; b sum of ellagic acid and methyl gallate; c n.d., not detected.
238

www.interscience.wiley.com/journal/pca Copyright © 2009 John Wiley & Sons, Ltd. Phytochem. Anal. 2009; 20: 231–239
Characterisation of tannin-containing herbal drugs by HPLC
data showed that there were major differences in the content
of polyphenols in the herbal drugs. HPLC with electrochemical
detection as well as the Ph. Eur. assay for the determination of
tannins were based on redox processes, but no correlation between
the two sets of data was found.
The HPLC fingerprint obtained of aqueous extracts of the herbal
drugs using UV detection at 250 nm showed distinct differences
between each herbal drug and also a high repeatability of the
fingerprints between different batches of the same herbal drug.
As the number of individuals familiar with the pharmacognostic
method of plant identification using microscopy is decreasing,
another possible method for identifying herbals drug is to use
chromatography to obtain a fingerprint of the herbal drug. TLC
may be used, but an even more reliable technique involves
the use HPLC with UV, fluorescence and/or electrochemical
detection. However, further experiments are required to estab-
lish whether HPLC can be used as an identification method,
especially regarding closely related herbal drugs.
The quantification of proanthocyanidins using catechin as the
standard is, of course, subject to some uncertainty as the proan-
thocyanidins constitute a major group of compounds probably
with major difference in fluorescence response. However, as
no reference standards were available catechin was used for the
calculations.
Methanolysis of the herbal drugs provides data concerning
the total content of gallotannins and ellagitannins determined
as methyl gallate and ellagic acid. The molar ratio between the
two analytes together with the total content may be used for
the characterisation/identification of herbals drugs. A further
investigation of the possibility of using methanolysis as a way of
identification should also be conducted.
Phytochem. Anal. 2009; 20: 231–239 Copyright © 2009 John Wiley & Sons, Ltd.
Acknowledgements
NaturDrogeriet A/S (Hørning, Denmark) and Finzelberg GmbH &
Co (Andernach, Germany) are acknowledged for donating some
of the herbal drugs used in these experiments.
References
Anonymous. 2004. WHO World Health Report. WHO Media Centre.
Anonymous. 2007. European Pharmacopoeia, 6th edn. European Directorate
of the Quality of Medicine.
Bate-Smith EC and Swain T. 1962. Flavonoid compounds. In Comparative
Biochemistry Academic Press: New York.
Blasco AJ, Rogerio MC, González MC and Escarpa A. 2005.
‘Electrochemical index’ as a screening method to determine ‘total
polyphenolics’ in foods: a proposal. Anal Chim Acta 539: 237–244.
De Bruyne T, Pieters L, Deelstra H and Vlietinck A. 1999. Condensed
vegetable tannins: biodiversity in structure and biological activities.
Biochem Syst Ecol 27: 445–459.
De Curtis A, Murzilli S, Di Castelnuovo A, Rotilio D, Donati MB and De
Gaetano G, et al. 2005. Alcohol-free red wine prevents arterial
thrombosis in dietary-induced hypercholesterolemic rats: Experimental
support for the ‘French paradox’. J Thromb Haemost 3: 346–350.
Folin O and Ciocalteu V. 1927. On tyrosine and tryptophane
determinations in protein. J Biol Chem 73: 627–650.
Haslam E. 1989. Plant Polyphenols—Vegetable Tannins Revisited—
Chemistry and Pharmacology of Natural Products. Cambridge University
Press: Cambridge.
Khanbabaee K and Van Ree T. 2001. Tannins: Classification and
definition. Nat Prod Rep 18: 641–649.
Santos-Buelga C and Scalbert A. 2000. Proanthocyanidins and tannin-
like compounds—nature, occurrence, dietary intake and effects on
nutrition and health. J Sci Food Agric 80: 1094–117.
Wollgast J. 2004. The contents and effects of polyphenols in chocolate,
Thesis, Justus Liebig University Giessen, Germany.
Yamada H and Watanabe H. 2007. Tea polyphenols in preventing
cardiovascular diseases. Cardiovasc Res 73: 439–440.
239
www.interscience.wiley.com/journal/pca