Food Additives & Contaminants: Part A

ISSN: 1944-0049 (Print) 1944-0057 (Online) Journal homepage: http://www.tandfonline.com/loi/tfac20

Pyrrolizidine and tropane alkaloids in teas and the

herbal teas peppermint, rooibos and chamomile in
the Israeli market

Jakob Avi Shimshoni, Arne Duebecke, Patrick P.J. Mulder, Olga Cuneah &
Shimon Barel

To cite this article: Jakob Avi Shimshoni, Arne Duebecke, Patrick P.J. Mulder, Olga Cuneah
& Shimon Barel (2015): Pyrrolizidine and tropane alkaloids in teas and the herbal teas
peppermint, rooibos and chamomile in the Israeli market, Food Additives & Contaminants:
Part A, DOI: 10.1080/19440049.2015.1087651

To link to this article: http://dx.doi.org/10.1080/19440049.2015.1087651

Accepted author version posted online: 12

Sep 2015.
Published online: 15 Oct 2015.

Submit your article to this journal

Article views: 36

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=tfac20

Download by: [Arne Dübecke] Date: 20 October 2015, At: 05:37

 Food Additives & Contaminants: Part A, 2015
http://dx.doi.org/10.1080/19440049.2015.1087651

Pyrrolizidine and tropane alkaloids in teas and the herbal teas peppermint, rooibos
and chamomile in the Israeli market
Jakob Avi Shimshonia*, Arne Duebeckeb, Patrick P.J. Mulderc, Olga Cuneaha and Shimon Barela
a
Kimron Veterinary Institute, Department of Toxicology, Bet Dagan, Israel; bQuality Services International, Bremen, Germany;
c
RIKILT-Wageningen UR, Wageningen, The Netherlands
(Received 13 July 2015; accepted 23 August 2015)

Dehydro pyrrolizidine alkaloids (dehydro PAs) are carcinogenic phytotoxins prevalent in the Boraginaceae, Asteraceae and
Fabaceae families. Dehydro PAs enter the food and feed chain by co-harvesting of crops intended for human and animal
consumption as well as by carry-over into animal-based products such as milk, eggs and honey. Recently the occurrence of
dehydro PAs in teas and herbal teas has gained increasing attention from the EU, due to the high levels of dehydro PAs
found in commercially available teas and herbal teas in Germany and Switzerland. Furthermore, several tropane alkaloids
(TAs, e.g. scopolamine and hyoscyamine) intoxications due to the consumption of contaminated herbal teas were reported
Downloaded by [Arne Dübecke] at 05:37 20 October 2015

in the literature. The aim of the present study was to determine the dehydro PAs and TAs levels in 70 pre-packed teabags of
herbal and non-herbal tea types sold in supermarkets in Israel. Chamomile, peppermint and rooibos teas contained high
dehydro PAs levels in almost all samples analysed. Lower amounts were detected in black and green teas, while no dehydro
PAs were found in fennel and melissa herbal teas. Total dehydro PAs concentrations in chamomile, peppermint and rooibos
teas ranged from 20 to 1729 μg/kg. Except for black tea containing only mono-ester retrorsine-type dehydro PAs, all other
teas and herbal teas showed mixed patterns of dehydro PA ester types, indicating a contamination by various weed species
during harvesting and/or production. The TA levels per teabag were below the recommended acute reference dose;
however, the positive findings of TAs in all peppermint tea samples warrant a more extensive survey. The partially high
levels of dehydro PAs found in teas and herbal teas present an urgent warning letter to the regulatory authorities to perform
routine quality control analysis and implement maximum residual levels for dehydro PAs.
Keywords: pyrrolizidine alkaloids; tropane alkaloids; tea; herbal tea; LC-MS/MS

Introduction by cytochrome P450 (Schmeller et al. 1997; Fu et al.

Pyrrolizidine alkaloids (PAs) are a group of naturally
occurring phytotoxins exclusively biosynthesised by
angiosperms and used by the plants as a defence mechan-
ism against various herbivores (Bull et al. 1968; Mattocks
1986; Rizk 1991; El-Shazly & Wink 2014). PAs are found
primarily in the plant families of Boraginaceae (all tribes),
Compositae (Eupatorieae and Senecioneae-tribe) and
Leguminosae (Crotolarieae-tribe) (Bull et al. 1968; Rizk
1991; El-Shazly & Wink 2014). PAs are composed of a
core necine base condensed with one or two necic acids
(Figure 1; Bull et al. 1968; Rizk 1991; El-Shazly & Wink
2014). Only PAs composed of C1–C2 unsaturated necine
bases (dehydro PAs) such as retronecine-, heliotridine-,
otonecine- and supinidine-types were shown to exert
hepatotoxic, carcinogenic, genotoxic and teratogenic
effects (Figure 1; EFSA 2011; Edgar et al. 2014;
El-Shazly & Wink 2014). Moreover, PAs frequently
co-occur in two forms, their N-oxides (PA-N-oxide) and
as their tertiary base, the former being the most prevalent
form in plants (Figure 1). Dehydro PAs are pro-toxins,
generating the toxic intermediate dihydropyrrolizine
(DHP) esters by hepatic metabolic activation mediated
2004; Ruan et al. 2014). Upon formation, the reactive
DHP ester reacts primarily at C-7 and/or C-9 with tissue
nucleophiles such as proteins and nucleic acids, resulting
in cell damage or death (Schmeller et al. 1997; Fu et al.
2004; Ruan et al. 2014). The primary target organ in
mammals is the liver; however, pulmonary toxicity was
also observed mainly in laboratory animals following the
ingestion of specific alkaloids such as monocrotaline
(Mattocks 1986; EFSA 2011; Edgar et al. 2014).
Observed carcinogenesis in laboratory animals following
exposure to dehydro PAs was demonstrated in numerous
studies as being directly associated with DHP-DNA
adducts (Fu et al. 2004; Xia et al. 2013; Ruan et al.
2014). Notwithstanding, carcinogenesis could not be
explicitly established in field cases of dehydro PAs poi-
soning in livestock or in humans acutely or chronically
exposed to dehydro PAs in food and beverages (EFSA
2011). In humans, acute intoxications with dehydro PAs
are characterised by veno-occlusive changes, haemorrha-
gic liver necrosis, hepatomegaly and ascites, with a high
mortality rate occurring anywhere between several days to
over 2 years following exposure (Ghanem & Hershko

*Corresponding author. Email: jakobshimshoni@gmail.com

© 2015 Kimron Veterinary Institute, Israel. Published by Taylor & Francis.

 2 J.A. Shimshoni et al.
Downloaded by [Arne Dübecke] at 05:37 20 October 2015
Figure 1. Structural features of some dehydro pyrrolizidine alkaloids and tropane alkaloids.
1981; Kumana et al. 1985; Chen & Huo 2010; Edgar et al.
2014). Survivors of an acute dehydro PAs intoxication as
well as humans and livestock chronically exposed to sub-
toxic dehydro PAs levels develop veno-occlusive disease
characterised by biliary endothelial proliferation, hepatic
vein occlusions and extensive fibrosis eventually leading
to liver cirrhosis (Ghanem & Hershko 1981; Chen & Huo
2010; Edgar et al. 2014). Moreover, some dehydro PAs
may induce pulmonary arterial hypertension due to pul-
monary endothelial proliferation, which may result in
acute right-sided heart failure (Ghanem & Hershko 1981;
Chen & Huo 2010; Edgar et al. 2014).
Dehydro PAs usually enter food and feeds chain by
co-harvesting of crops intended for human and animal
consumption with various grains, salad greens and teas
containing noxious weeds or weed seeds producing dehy-
dro PAs, as well as by carry-over of dehydro PAs into
animal-based products such as milk, eggs and honey
(EFSA 2011; Molyneux et al. 2011; Bodi et al. 2014).
Numerous case reports have been published in the last
decade of fatal dehydro PAs intoxications in humans and
farm animals due to the ingestion of food, medicinal
herbs or tea-based infusions, and feed contaminated
with dehydro PAs (Harper et al. 1985; Roulet et al.
1988; Sperl et al. 1995; Hill et al. 1997; Kakar et al.
2010; EFSA 2011; Shimshoni et al. 2015). Although the
health hazards associated with acute and chronic expo-
sure to dehydro PAs are well known, no regulations have
been established in the European Union, the USA or in
Israel imposing limits for dehydro PAs levels in food and
beverages for human consumption. Notwithstanding, sev-
eral European risk assessment agencies have established
various recommendations for limiting the total daily
dehydro PAs uptake (EFSA 2011; Bodi et al. 2014).
The UKs Committee on Toxicity of Chemicals in Food
Consumer Products and Environment and the German
Federal Institute for Risk Assessment recommended a
maximum tolerable daily intake of 0.007 µg of dehydro
PAs/kg body weight (bw), whereas the corresponding
Austrian Agency for Health and Food Safety did not
authorise any PA residues in the final food products
(EFSA 2011).

Recently, the occurrence of dehydro PAs in teas and
medicinal herbs has gained increasing attention from
several national authorities in Europe due to the very
high dehydro PAs levels found in commercially avail-
able popular teas and medicinal herbs in Germany and
Switzerland, with levels of dehydro PAs up to 5.6 and
0.47 mg/kg, respectively (Bodi et al. 2014; Griffin et al.
2014; Mathon et al. 2014; Mädge et al. 2015). All tea
products are manufactured from the leaves of the ever-
green shrub Cammellia sinensis native to East, South
and South East Asia and cultivated mainly in tropical
and subtropical climate (Hicks 2009; Hayat et al. 2015).
The tea leaves are usually handpicked every 1–2 weeks.
Most tea consumed in Israel as well as in Europe and
the USA is produced on large plantations in Asia and
Africa, and is destined to be sold to large international
businesses (Hicks 2009). The Israeli tea market is domi-
nated by two international global brands, namely
Downloaded by [Arne Dübecke] at 05:37 20 October 2015
Wissotzky (~80%) and Lipton (~12%) (Rivlin 2005).
According to the Food and Agricultural Organization
of the United Nations report on world black tea market,
the Israeli market consumes about 1800 tonnes annually
which is 0.05% from the global consumption
(Committee on Commodity Problems 2012). With the
rise of health and wellness trends in Israel, tea and
herbal teas such as peppermint, chamomile and rooibos
are increasingly being promoted as a healthier alterna-
tive to coffee and other caffeinated beverages
(Euromonitor 2015).
An additional hazardous group of alkaloids found to
contaminate food and feed commodities are the tropane
alkaloids (TAs), naturally occurring secondary metabolites
of several plant families including Brassicaceae, Solanaceae
and Erythroxylaceae (EFSA 2013). Although more than 200
different TAs have been identified, data on their occurrence
in food and feed are limited (EFSA 2013). The most fre-
quently occurring TAs in plant-derived food and feed pro-
ducts are (−)-hyoscyamine and (−)-scopolamine, which in
contrast to their corresponding (+)-enantiomers are formed
naturally (Figure 1; Adamse et al. 2014; EFSA 2013). The
racemic mixture of (−)-hyoscyamine and (+)-hyoscyamine is
called atropine, which is only formed during the extraction
and isolation procedure by racemisation of (−)-hyoscyamine
(Adamse et al. 2014). TAs such as (−)-hyoscyamine and
(−)-scopolamine are also known to co-occur with their cor-
responding N-oxides in plants (Mroczek et al. 2006;
Phillipson & Handa 1975, 1976). The wide distribution of
TA-producing weeds in temperate and tropical regions
worldwide results in occasional contamination of agricul-
tural crops such as linseed, soybean, millet, sunflower and
buckwheat (EFSA 2013). The toxicological effects of
atropine and scopolamine are mediated by excessive
anti-muscarinic activity in the central and autonomic
nervous system. Recently, the EFSA Panel established
a group acute reference dose (a dose which upon
Food Additives & Contaminants: Part A 3
ingestion over a short period of time does not pose
appreciable health risks to humans) of 0.016 μg/kg
bw expressed as the sum of (−)-hyoscyamine and
(−)-scopolamine, assuming equivalent potency (EFSA
2013).
Several TA intoxications due to consumption of con-
taminated herbal teas were reported in the literature, i.e.
burdock root tea (Bryson 1978), nettle tea (Hks &
Zingerle 1980), comfrey tea (Galizia 1983; Routledge &
Spriggs 1989), marshmallow tea and Paraguay tea (EFSA
2013).
The objective of the present study was to determine
the levels of dehydro PAs and TAs (in the form of total
atropine and (–)-scopolamine and their corresponding
N-oxides) in pre-packed teabags of popular herbal and
non-herbal tea types sold in supermarkets in Israel and to
compare the obtained levels of dehydro PAs and TAs with
the current European recommendations as well as the high
levels of dehydro PAs reported in the German and Swiss
markets.
Methods and materials
Reagents and standards
Methanol (LC-MS grade), formic acid (analytical grade),
acetic acid (analytical grade) and ammonia solution (32%,
analytical grade) were bought from Merck (Darmstadt,
Germany). Ammonium formate (LC-MS grade) was from
Fluka (Seelze, Germany). Analytical PA standards of mono-
crotaline, monocrotaline-N-oxide, echimidine, erucifoline,
erucifoline-N-oxide, europine, europine-N-oxide, helio-
trine-N-oxide, intermedine, intermedine-N-oxide, jacobine,
jacobine-N-oxide, lycopsamine, lycopsamine-N-oxide, sene-
cionine-N-oxide, seneciphylline-N-oxide, senecivernine-N-
oxide and trichodesmine were obtained from Phytolab
(Vestenbergsgreuth, Germany); heliotrine, lasiocarpine,
lasiocarpine-N-oxide, retrorsine, retrorsine-N-oxide, sene-
cionine, seneciphylline and senkirkine from AppliChem
(Darmstadt, Germany); senecivernine from Carl Roth
(Karlsruhe, Germany) and echimidine-N-oxide from Cfm
Oskar Tropitzsch (Marktredwitz, Germany). Analytical stan-
dards of atropine and (–)-scopolamine and their correspond-
ing N-oxides were obtained from Sigma-Aldrich
(Schnelldorf, Germany).
Samples
All tea and herbal tea samples analysed (n = 70) were
produced by four international brands commonly consumed
in Israel, Europe and the USA. The tea and herbal tea
samples were bought from Israeli supermarkets in the Tel
Aviv area, covering a wide range of popular herbal tea types,
namely mixed herbal tea, peppermint tea, chamomile tea,
fennel tea, black tea, green tea, rooibos tea, and melissa tea
 4 J.A. Shimshoni et al.
(n = 10 for each tea or herbal tea type, except for melissa and
fennel with n = 5). Each commercial product was sold in the
form of pre-packed teabags. The tea samples were purchased
during the period of January–March 2015 and were stored in
closed containers at room temperature until sampling.
Multi-residue sample preparation
A 1 g amount of sample was taken at random from each
package representing a different batch and extracted with
10 ml of aqueous acetic acid/methanol solution (1:2, v/v)
by 30 min vigorous shaking. The samples were centri-
fuged at 4000g for 10 min and the supernatants were
brought to pH 5–6 with diluted ammonia solution. An
aliquot of 2 mL of each supernatant sample was evapo-
rated under a stream of N2 at 40°C. The dried samples
were reconstituted in 4 ml water and filtered into a 2-ml
glass vial using 0.45 µm syringe filter.
Downloaded by [Arne Dübecke] at 05:37 20 October 2015
LC-MS/MS parameters
The LC-system consisted of a Shimadzu degasser (DGU-
20A3) and two Shimadzu LC-20AD pumps controlled by
a Shimadzu CBM-20A controller unit.
Injection of 10 µl of sample was done using a HTC
PAL autosampler of CTC Analytics AG. Column tempera-
ture was held steady at 25°C. The column was eluted
using a gradient flow (300 ml/min) of two solvents.
Solvent A consisted of 99.5% water plus 0.5% formic
acid; solvent B of 94.5% methanol, 5% water and 0.5%
formic acid. To both solvents 0.1 mM oxalic acid and
2.0 mM ammonium acetate were added. The mobile
phase was maintained for the first 30 s at 97:3 (A:B) and
then changed with a linear gradient to 100% solvent B
over 6 min, which was kept for 1 min before a re-equili-
bration phase over 2.5 min restored the initial mobile
phase of 97:3. Separation was achieved using a Thermo
Hypersil Gold C18 column (50 x 2.1 mm, 1.9 mm particle
size). The short column allowed for a reduced analysis
time of 10 min per sample. An Applied Biosystems API
4000 QTRAP triple quadrupole mass spectrometer was
used to detect the PAs, PA-N-oxides, TAs and TA-N-
oxides (Table 1). The instrument was set to multi-reaction
monitoring (MRM) mode (source temperature 650°C,
cone voltage 5500 V, collision gas on “high”, curtain gas
25 psi, ion source gas 1 and 2 at 35 and 45 psi, respec-
tively), using one MRM transition as quantifier and
another two as qualifiers for each analyte. Quantification
was done using external calibrations. The calibrations
were prepared with the reference materials in blank
mixed herbal tea. Finally, concentrations were corrected
against the recovery of the internal standard resulting in
recoveries between 70% and 110% for the analysed
compounds. The limit of quantitation (LOQ) for each
analyte is depicted in Table 1.
Method validation
Methods for the determination of dehydro PAs, TAs and
their corresponding oxides in tea and herbal teas were
validated in-house with regard to recovery, precision,
LOD and LOQ. LODs and LOQs were determined by
the calibration curve method, whereby the LOQ was
threefold the value of LOD. The resulting LOQs in tea
and herbal teas are shown in Table 1. Recovery rates were
determined by repeated analysis of fortified blank material
at concentrations of 20, 150 or 200 µg/kg.
Quantification of dehydro PAs, TAs and their corre-
sponding N-oxides was carried out by means of matrix-
matched calibration from blank extracts of the respective
type of tea.
Results and discussion
Validation
Method validation for the determination of 30 different
dehydro PAs and PAs-N-oxides and for four different
TAs (atropine, scopolamine and their corresponding
N-oxides) was performed according to the guidelines
defined by the EU and the results are summarised in
Table 1. The recoveries obtained from blank matrix
spiked with dehydro PAs and TAs were in the range of
80–95%. The calibration curves for dehydro PAs and
TAs displayed excellent correlation between analyte
concentration and peak area (R2 for all curves were
>0.995). The LOD and LOQ values for the dehydro
PAs and TAs were in the range of 0.001–0.05 mg/kg
and (Table 1). The percent coefficients of variation (%
CV) of within- and between-days for dehydro PAs and
TAs as well as for their corresponding N-oxides were
1.0%–8.5% and 0.6%–3.8% respectively. The accuracy
values for all analytes tested were in the range of 90%–
110%. Hence, in the examined range of concentrations
analysed (1–500 µg/kg), the optimised analytical method
was considered accurate and precise.
PAs in tea and herbal tea samples
In the present study, 70 tea and herbal tea samples of 8
different popular tea (green and black tea) and herbal tea
types (chamomile, fennel, melissa, mixed herbal teas,
peppermint and rooibos) purchased in various Israeli
supermarkets during January–March 2015 were analysed
for their dehydro PAs (sum of 30 individual dehydro
PAs and dehydro PA-N-oxides) and TAs (atropine,
(–)-scopolamine and their corresponding N-oxides) levels.
 Food Additives & Contaminants: Part A 5

Table 1. Mass spectrometric conditions used, indicative retention times and limit of quantification of utilised
pyrrolizidine and tropane alkaloids determined in the study.

Standard used for Precursor ion Fragment 1 Fragment 2 Retention Limit of quantification
quantification (m/z) (m/z) (m/z) time (min) (µg/kg)

Lycopsamine 300.2 94.0 156.0 4.71 10

Lycopsamine-N-oxide 316.2 94.0 172.0 5.27 10
Heliotrine 314.2 138.0 156.0 5.81 10
Heliotrine-N-oxide 330.2 111.0 172.0 6.14 10
Europine 330.2 94.0 138.0 4.63 10
Europine-N-oxide 346.2 172.0 256.0 4.91 10
Erucifoline 350.0 94.0 138.0 4.19 50
Erucifoline-N-oxide 366.0 118.0 164.0 4.66 20
Intermedine 300.1 120.0 156.0 4.57 10
Intermedine-N-oxide 316.0 111.1 172.0 5.15 10
Echimidine 398.2 120.0 220.0 7.28 10
Echimidine-N-oxide 414.2 254.0 352.0 7.12 10
Lasiocarpine 412.2 120.0 220.0 8.30 10
Lasiocarpine-N-oxide 428.2 138.0 254.0 8.62 10
Jacobine 352.0 120.1 155.0 4.52 50
Jacobine-N-oxide 368.0 120.0 296.0 4.79 20
Downloaded by [Arne Dübecke] at 05:37 20 October 2015

Monocrotaline 326.0 194.0 120.0 3.48 20

Monocrotaline-N- 342.0 120.0 138.0 5.53 20
oxide
Retrorsine 352.0 138.0 324.0 5.50 20
Retrorsine-N-oxide 368.0 136.0 120.0 5.58 20
Riddelliine 350.0 120.0 138.1 4.93 20
Riddelliine-N-oxide 366.0 94.0 136.0 5.04 20
Senecionine 336.0 94.0 308.0 6.48 20
Senecionine-N-oxide 352.0 118.0 220.0 6.60 10
Seneciphylline 334.0 138.0 306.0 5.73 20
Seneciphylline-N- 350.0 95.0 118.0 5.92 10
oxide
Senecivernine 336.0 94.0 308.4 6.29 20
Senecivernine-N-oxide 352.0 94.1 118.2 6.41 10
Senkirkine 366.0 122.0 168.0 7.49 10
Trichodesmine 354.0 120.0 138.0 5.45 10
Atropine 290.2 93.0 124.1 6.27 1
Atropine-N-oxide 306.1 124.2 140.3 7.06 1
Scopolamine 303.9 102.9 156.1 5.03 1
Scopolamine-N-oxide 320.1 94.2 154.2 4.68 1

Moreover, the dehydro PAs profile of each tea-type was
analysed according to their necine base and esterification
type (mono-, di- or cyclic esters), in order to possibly
identify the plant families contributing to the contamina-
tion of the aforementioned teas and herbal teas.
Table 2 summarises the total dehydro PAs levels
obtained for the various tea/herbal tea types. Among the
herbal teas, dehydro PAs were below the LOD in all the
fennel and melissa tea samples analysed in this study. In
contrast, the survey results of the German market con-
ducted by the German Federal Institute for Risk
Assessment, showed that melissa teas were the most heav-
ily contaminated herbal teas with dehydro PAs levels
ranging from 5.0 to 2579.2 µg/kg, whereas the fennel
teas for babies in the German market were less heavily
contaminated with concentrations ranging from 1.8 to
96.5 µg/kg total dehydro PAs and from 11 to 198 µg/kg
total dehydro PAs for regular fennel teas (Bodi et al. 2014;
Mädge et al. 2015). In the Swiss market, however, dehy-
dro PAs were below the LOD in all fennel tea samples
analysed, in agreement with the result of our study.
In the Israeli market, chamomile, peppermint and rooi-
bos tea samples were the most heavily contaminated herbal
teas with dehydro PAs detected in all samples analysed with
mean values of 261 µg/kg (range: 20–507 µg/kg), 564 µg/
kg (range: 74–1729 µg/kg) and 314 µg/kg (range:
82–657 µg/kg), respectively (Table 2). The mean total
dehydro PAs values found for the aforementioned three
herbal teas are quite high. For instance, consumption of
the beverage prepared from one teabag (assuming 100%
transfer from the teabag to the water) containing 2 g
(equivalent amount of 1 whole teabag) chamomile, pepper-
mint, or rooibos herbs exceeds the recommended maximum
tolerable daily intake of 0.49 µg dehydro PAs for a 70 kg
 6 J.A. Shimshoni et al.
Table 2. Total pyrrolizidine alkaloids concentrations measured in different teas and herbal teas.a.
Mean dehydro PA ± STD
(µg/kg)
Tea (range) Samples above LOQ
Herbal tea
Fennel NDb
Peppermint 261 ± 179 (20–507)
Chamomile 564 ± 670 (74–1729)
Melissa ND
Mixed teas 315 ± 281 (48–723)
Rooibos 314 ± 211 (82–657)
Teas
Black 107 ± 84 (10–160)
Green 81 ± 30.5 (60–116)
Notes: aEach tea type was randomly selected for analysis in various supermarkets during January–March 2015. The sample size for green, black, rooibos,
mixed herbal tea, chamomile and peppermint was 10, whereas for melissa and fennel teas the sample size was 5. Mean value were calculated only for
samples with dehydro PAs levels ≥ LOQ.
b
ND, not detected (level below LOD).
Downloaded by [Arne Dübecke] at 05:37 20 October 2015
adult and is more than 13-fold the recommended safe daily
dose (0.04 µg) for infants weighing 5 kg (Table 2; EFSA
2011; Mädge et al. 2015). In agreement with the results of
the present study, the German survey also found exception-
ally high levels of dehydro PAs in rooibos (mean level of
1856.4 µg/kg) and chamomile (mean level of 439.7 µg/kg)
teas, occurring in 90% of the chamomile and 100% of the
rooibos samples. Furthermore, the peppermint teas in the
German market displayed a lower mean level of dehydro
PAs of 134.2 µg/kg, occurring in 87% of the samples (Bodi
et al. 2014). In the Swiss market, on the other hand, the
total dehydro PAs burden for chamomile, peppermint and
rooibos was lower, with mean values ranging from 101 to
146.5 µg/kg and occurrence frequencies ranging between
50% (peppermint) and 100% (rooibos; Mathon et al. 2014).
It is reasonable to assume that the lower dehydro PAs levels
reported for the Swiss market may be an underestimation
due to the fact that only nine different dehydro PAs were
monitored, whereas in the Israeli and German market,
respectively, 30 and 17 different dehydro PAs were mon-
itored. Because several chamomile and peppermint samples
analysed form the German and Swiss markets had no
detectable levels of dehydro PAs and their corresponding
pure herbal drugs were shown by several studies to be free
of dehydro PAs, it can be concluded that both plant species
are non-dehydro PAs producers and that the appearance of
dehydro PAs in their tea samples is the result of contamina-
tion with PA-producing plants (Bodi et al. 2014; Mathon
et al. 2014).
Interestingly, dehydro PAs were found in all of the
rooibos tea samples obtained from the Israeli, German
and Swiss markets, in contrast to all other tea types
which displayed variable occurrences of dehydro PAs.
The rooibos plant (Aspalathus linearis) is a member of
the Crotalarieae tribe, which includes numerous dehydro
% of tea bag samples exceeding Average amount of
the maximal recommended levels dehydro PA consumed
of 210 µg/kg for 60 kg adult per 2 g teabag (µg)
ND ND –
10 60 0.52
10 60 1.13
ND ND –
5 30 0.63
10 80 0.63
6 0 0.21
6 0 0.16
PAs-producing Crotolaria weed species (Van Wyk &
Verdoorn 1990). Because there are no available pub-
lished data regarding the alkaloid content of the rooibos
plants, it is of the utmost importance to determine
whether rooibos plants are potential producers of dehy-
dro PAs.
In herbal mixed teas obtained from the Israeli market,
dehydro PAs occurred in 50% of the analysed samples with
levels ranging from 48 to 723 µg/kg and with a mean value
of 315 µg/kg (Table 2). The composition of the mixed
herbal teas analysed in the present study included the
following major ingredients: green tea, black tea, lemon-
grass, peppermint, chamomile and melissa. High levels of
dehydro PAs of 455–723 µg/kg occurred in all mixed
herbal tea samples containing chamomile as a common
ingredient, resulting in an exposure of 0.91–1.4 µg per
teabag. The aforementioned levels of dehydro PAs are
more than twice the recommended maximum tolerable
daily intake of 0.42 µg for a 60 kg person (EFSA 2011).
In comparison, dehydro PAs occurred in 91% of the mixed
herbal tea samples analysed from the German market with a
mean level of 151.4 µg/kg, whereas in the Swiss mixed
herbal tea products, dehydro PAs occurred only in 13% of
the samples with a mean value of 2 µg/kg (Bodi et al. 2014;
Mathon et al. 2014).
Mean levels of dehydro PAs in green and black tea
samples in the Israeli market were 81 and 107 µg/kg,
respectively, occurring in 60% of the samples (Table 2).
The total levels of dehydro PAs per 2 g teabag in all
analysed green and black tea samples were below the
recommended daily intake for a 60 kg person (Table 2).
However, for 2 black tea and 1 green tea samples, the
total level of dehydro PAs exceeded 105 µg/kg per tea-
bag, potentially exposing long-term tea consumers con-
suming more than 1–2 tea cups daily to significant levels

of dehydro PAs, possibly leading to increased risk of
health impairment. According to the recent dehydro
PAs survey in green and black tea samples in the
German market, higher occurrence rates and mean levels
of dehydro PAs were found in both tea types, namely in
87% of the green tea and 95% of the black tea samples
with mean values of 109 and 255.9 µg/kg, respectively
(Bodi et al. 2014). In contrast, no dehydro PAs were
found in green and black tea samples analysed from the
Swiss market (Mathon et al. 2014). The widely seen
variability in dehydro PAs content in different tea types
within and between the various retail markets clearly
demonstrate the possibility to produce pre-packed super-
market teabag products below the recommended maxi-
mum daily intake of dehydro PAs, thereby minimising
potential health risks associated with exposure to dehy-
dro PAs. Recently, Mädge and co-workers analysed 44
herbal tea samples for infants, pregnant or lactating
Downloaded by [Arne Dübecke] at 05:37 20 October 2015
women purchased across Germany, with 86% of the
samples testing positive for dehydro PAs with alarmingly
high concentrations of up to 821.1 µg/kg and a mean
value of 105 µg/kg total dehydro PAs (Mädge et al.
2015). These findings are of major concern to public
health, mainly due to the facts that dehydro PAs are
well known to be partially excreted into mothers’ milk
and the known high susceptibility of infants and children
to the hepatotoxic effects of dehydro PAs.
Furthermore, the dehydro PAs profile for each tea
and herbal tea type was determined (Figure 2). Each
tea type analysed in this study displayed a different
composition of dehydro PAs comprising 3–10 dehydro
PAs, with 19 different dehydro PAs being detected in all
the samples analysed (Figure 2). Except for black tea
samples, which displayed only 3 dehydro PAs of the
retrorsine-mono-ester type, the remaining tea and herbal
teas comprised a mixture of mainly retrorsine-type 12-
membered macrocyclic, mono-ester and open chain-die-
ster dehydro PAs, with only peppermint tea and the
Figure 2. (colour online) Mean total dehydro PAs concentration profile of each analysed tea variety. The mean values were calculated
only for samples with concentration values above the LOQ.
Food Additives & Contaminants: Part A 7
mixed herbal teas containing dehydro PAs of the helio-
tridine-type (Figures 1 and 2). Because at present no
dehydro PAs-producing weed species is known to pro-
duce a mixture of mono-, open chain-diesters and macro-
cyclic-esters, it is reasonable to assume that chamomile,
green, peppermint and rooibos teas were contaminated
with several dehydro PAs-producing plant species of
unknown genera. The lack of macrocyclic dehydro PAs
in black tea indicates that the source of contamination
might be a weed species of the Boraginaceae family.
Furthermore, the presence of only 12-membered macro-
cyclic dehydro PAs found in herbal teas as well as in
green teas indicate that the sources of contamination
were mainly plant species of the Asteraceae family. In
agreement with the survey results in the Berliner area, no
11-membered macrocyclic dehydro PAs such as mono-
crotaline were found in the present study, which implies
the possibility of ruling out Crotalaria species as the
potential source of contamination.
Senecionine-N-oxide and retrorsine-N-oxide were the
only dehydro PAs commonly found in all positive-tested
herbal tea types as well as in green tea (Figure 2). The
most frequently detected dehydro PAs occurring in more
than 30% of all tested samples were: senecionine-N-oxide
followed by lycopsamine and intermedine-N-oxide
(Figure 2). Dehydro PAs found at concentrations above
100 µg/kg in some of the tea samples analysed in the
present study were intermedine, senecionine-N-oxide,
lycopsamine-N-oxide, lycopsamine, intermedine-N-oxide
and heliotrine-N-oxide. Comparable results regarding the
frequent occurrences and high concentrations of various
dehydro PAs were obtained for the German market survey
(Bodi et al. 2014).
Although presently the source of contamination of
various tea and herbal tea products with dehydro PAs is
unknown, it is reasonable to focus on two major key
factors during the manufacturing process possibly respon-
sible for the contamination, namely the harvest technique
 8 J.A. Shimshoni et al.

being used in the field and the processing (withering,
rolling, fermentation, drying, sorting and packing) of tea/
herbal tea leaves yielding the final product. The utilisation
of harvesting machinery on the field in contrast to hand-
picking techniques commonly practiced on tea plantations
might present a major risk factor for dehydro PAs con-
tamination as dehydro PAs-producing weeds growing on
the same field can be easily co-harvested by modern
harvesting machinery. The relatively low dehydro PAs
burden of green and black tea as compared to chamomile
and peppermint herbal teas might be due to the fact that
teas in general are commonly handpicked in contrast to
many herbal teas (Hicks 2009). Furthermore, airborne
dehydro PAs-producing weed particles such as pollen
can easily contaminate nearby crops and tea plantations
and therefore their removal from the teas and herbal tea
leaves by various physical means and/or the eradication of
dehydro PAs producing weeds in nearby fields needs to be
Downloaded by [Arne Dübecke] at 05:37 20 October 2015
of tropane alkaloids (TAs) such as atropine, scopolamine
and their corresponding N-oxides in commercially avail-
able teas and herbal teas. The present study reports for the
first time the occurrence of the aforementioned alkaloids
in eight different tea and herbal tea types. Because the
naturally occurring plant alkaloids (−)-hyoscyamine and
(−)-scopolamine can racemise to atropine ((±)-hyoscyamine)
and (±)-scopolamine, respectively) during the extraction and
purification procedure, we have determined the levels of the
racemic mixtures as well as their corresponding N-oxides as
a measure of total TAs burden in the tea and herbal tea
samples. Table 3 summarises the TAs concentrations and
occurrences in the various teas and herbal teas analysed. No
TAs were detected in all black tea, green tea and melissa tea
samples analysed. However, low levels of atropine and
scopolamine occurred in 80% of the peppermint samples
analysed with mean values of 171 µg/kg (range:
20–208 µg/kg) and 81 µg/kg (range: 14–171 µg/kg),

considered. The utilisation of the same processing equip-
ment for tea and herbal tea leaves during the manufactur-
ing process might present an additional major cause for
cross-contamination and therefore a vigorous cleaning
protocol needs to be implemented after each batch
processing.
All tea and herbal tea samples of the global brands
analysed herein revealed dehydro PAs to various
levels, thereby strengthening our assumption that the
contamination occurs during harvest and possibly as a
result of cross-contamination during the manufacturing
process.
TAs in tea and herbal tea samples
To the best of our knowledge, no published data in peer
reviewed literature are available regarding the occurrence
respectively (Table 3). Their corresponding N-oxides were
found in trace levels (1–2.8 µg/kg) in only 40% of the
samples. The highest total TAs level (sum of atropine,
scopolamine and their corresponding N-oxides) detected
in the peppermint samples (379 µg/kg) yielded a maximal
amount of 0.76 µg TAs per 2 g teabag (equivalent to the
amount per one cup of tea), which is below the acute daily
reference dose (0.96 µg) for a person weighing 60 kg
(EFSA 2013). However, the consumption of 2 or more
cups of peppermint herbal tea containing more than
220 µg/kg TAs per teabag by a person weighing 60 kg for
a longer period of time might result in hazardous chronic
exposure according to the recommended guidelines (EFSA
2013). Because the health risks associated with chronic
exposure to low levels of TAs in humans and animals are
currently unknown, toxicity studies are necessary for a
rational risk evaluation.

Table 3. Total tropane alkaloid concentrations measured in different teas and herbal teas.a.

Mean total N-oxides

of atropine and Mean total atropine Average amount
Mean atropine ± Mean scopolamine scopolamine ± and scopolamine ± Samples of TA consumed per
STD (µg/kg) ± STD (µg/kg) STD (µg/kg) STD (µg/kg) above tea bag (of 2 g)
Tea (range) (range) (range) (range) LOQ (µg)

Herbal tea
Fennel 83 11 1 94 1 0.19
Peppermint 171 ± 53 (20–208) 81 ± 69 (20–208) 3.3 ± 0.8 (1–2.8) 181 ± 154 (34–379) 8 0.36
Chamomile NDb 2.1 ND 2.1 1 0.004
Melissa ND ND ND ND ND –
Rooibos ND 2 ND 2 1 0.004
Teas
Black ND ND ND ND ND –
Green ND ND ND ND ND –
Notes: aEach tea type was randomly selected for analysis in various supermarkets during January–March 2015. The sample size for green, black, rooibos,
mixed herbal tea, chamomile and peppermint was 10, whereas for melissa and fennel teas the sample size was 5. Mean value were calculated only for
samples with TAs levels ≥ LOQ.
b
ND, not detected (level below LOQ).

Chamomile and rooibos tea samples were free of
TAs contamination except for one sample of each
herbal tea type containing non-significant residual
levels of scopolamine (2 µg/kg; Table 3).
Furthermore, fennel herbal tea samples were also
mostly free of TAs except for one herbal tea sample
containing a total TAs of 94 µg/kg, a value 5 times
lower than the recommended acute daily reference
concentration for a 60 kg adult.
Conclusion
A preliminary survey of 70 tea and herbal tea samples
bought in the Israeli market showed that herbal teas
(chamomile, peppermint and rooibos) contained high
levels of dehydro PAs in almost all samples.
Comparatively lower amounts were detected in black
Downloaded by [Arne Dübecke] at 05:37 20 October 2015
and green teas, while none were found in fennel and
melissa herbal teas. Total dehydro PAs concentrations in
chamomile, peppermint and rooibos teas ranged from 20
to 1729 μg/kg. Except for black tea displaying only
mono-ester retrorsine-type dehydro PAs, all other teas
and herbal teas revealed mixed patterns of dehydro PA
ester types, indicating that the dehydro PAs originated
from contamination by various weed species during
harvesting and/or production.
This is the first publication reporting the presence
of TAs in herbal teas. Although the concentrations per
teabag were below the recommended acute reference
dose, consumption of more than two tea cups for a
longer period of time of the most contaminated pep-
permint teas might expose humans to hazardous
adverse effects. The positive findings of TAs in pep-
permint tea samples warrant a more extensive survey
to determine the extent of TAs contamination in var-
ious herbal teas.
Numerous in vitro studies as well as fatal case
reports in humans clearly indicate the health hazards
associated with long-term exposure to low levels of
dehydro PAs, which may lead to veno-occlusive disease
and liver cirrhosis. Hence, it is of the utmost importance
to minimise dehydro PAs content in food products and
to implement maximum residual levels of the regulatory
authorities. Furthermore, research activities to uncover
the source of dehydro PAs contamination in teas and
herbal teas are essential for establishing prevention
measurements.
Acknowledgements
The authors thank Professor Alan Shlosberg for his excellent
technical assistance.
Food Additives & Contaminants: Part A 9
Disclosure statement
No potential conflict of interest was reported by the authors.
References
Adamse P, Van Egmond HP, Noordam MY, Mulder PPJ, De Nijs
WCM. 2014. Tropane alkaloids in food: poisoning incidents.
Qual Assur Saf Crops Foods. 6:15–24.
Bodi D, Ronczka S, Gottschalk C, Behr N, Skibba A, Wagner M,
Lahrssen-Wiederholt M, Preiss-Weigert A, These A. 2014.
Determination of pyrrolizidine alkaloids in tea. Herbal Drugs
Honey Food Addit Contam: Part A. 31:348–358.
Bryson PD. 1978. Burdock root tea poisoning. JAMA: J Am
Med Assoc. 240:1586–1586.
Bull LB, Culvenor CCJ, Dick AT. 1968. The pyrrolizidine alka-
loids: their chemistry, pathogenicity and other biological
properties. In: Neuberger A, Tatum EL, editors. Frontiers
of biology series. Amsterdam: North- Holland Publishing
Company.
Chen Z, Huo JR. 2010. Hepatic veno-occlusive disease asso-
ciated with toxicity of pyrrolizidine alkaloids in herbal pre-
parations. Neth J. 68:252–206.
Committee on Commodity Problems. 2012. Current situation and
medium term outlook for tea. Colombo, Sri Lanka; [cited
2015 Aug 20]. Available from http://www.fao.org/fileadmin/
templates/est/COMM_MARKETS_MONITORING/Tea/
Documents/IGG_20/12-CRS7-CurrentSit_01.pdf
Edgar JA, Molyneux RJ, Colegate SM. 2014. Pyrrolizidine alka-
loids: potential role in the etiology of cancers, pulmonary
hypertension, congenital anomalies and liver disease. Chem
Res Toxicol. 28:4–20.
EFSA Panel on Contaminants in the Food Chain (CONTAM).
2011. Scientific opinion on pyrrolizidine alkaloids in food
and feed. Efsa J. 9:134. Available from: http://www.efsa.
europa.eu/en/efsajournal/doc/2406.pdf
EFSA Panel on Contaminants in the Food Chain (CONTAM).
2013. Scientific opinion on tropane alkaloids in food and
feed. Efsa J. 11:113. Available from: http://www.efsa.europa.
eu/de/efsajournal/doc/3386.pdf
El-Shazly A, Wink M. 2014. Diversity of Pyrrolizidine alkaloids
in the Boraginaceae structures, distribution and biological
properties. Diversity. 6:188–282.
Euromonitor International. 2015. Tea in Israel. [cited 2015 Aug
20]. Available from: http://www.euromonitor.com/tea-in-
israel/report
Fu P, Xia Q, Lin G, Chou MW. 2004. Pyrrolizidine alkaloids -
genotoxicity, metabolism enzymes, metabolic activation and
mechanisms. Drug Metab Rev. 36:1–55.
Galizia E. 1983. Clinical curio hallucinations in elderly tea
drinkers. Br Med J. 287:979.
Ghanem J, Hershko C. 1981. Veno-occlusive disease and pri-
mary hepatic vein thrombosis in Israeli arabs. Isr J Med Sci.
17:339–347.
Griffin CT, Gosetto F, Danaher M, Sabatini S, Furey A. 2014.
Investigation of targeted pyrrolizidine alkaloids in traditional
Chinese medicines and selected herbal teas sourced in
Ireland using LC-ESI-MS/MS. Food Addit Contam Part.
31:940–961.
Harper PAW, Walker KH, Krahenbuhl RE, Christie BM. 1985.
Pyrrolizidine alkaloid poisoning in calves due to contamina-
tion of straw by Heliotropium europaeum. Aust Vet J.
62:382–383.
 10 J.A. Shimshoni et al.
Hayat K, Iqbal H, Malik U, Bilal U, Mushtaq S. 2015. Tea and
its consumption: benefits and risks. Crit Rev Food Sci Nutr.
55:939–954.
Hicks A. 2009. Current status and future development of global
tea production and tea products. Au J T. 12:251–264.
Hill BD, Gaul KL, Noble JW. 1997. Poisoning of feedlot
cattle by seeds of Heliotropium europaeum. Aust Vet J.
75:360–361.
Hks S, Zingerle H. 1980. Atropin-vergiftung durch “gesundheits-
tee”. Fortschr Med. 98:1525.
Kakar F, Akbarian Z, Leslie T, Mustafa ML, Watson J, Van
Egmond HP, Omar MF, Mofleh J. 2010. An outbreak of
hepatic veno-occlusive disease in Western Afghanistan asso-
ciated with exposure to wheat flour contaminated with pyr-
rolizidine alkaloids. J Toxicol. 2010:1–7.
Kumana CR, Ng M, Lin HJ, Ko W, Wu PC, Todd D. 1985.
Herbal tea induced hepatic veno-occlusive disease: quan-
tification of toxic alkaloid exposure in adults. Gut.
26:101–104.
Mädge I, Cramer L, Rahaus I, Jerz G, Winterhalter P, Beuerle
T. 2015. Pyrrolizidine alkaloids in herbal teas for infants,
Downloaded by [Arne Dübecke] at 05:37 20 October 2015
pregnant or lactating women. Food Chem. 187:491–498.
Mathon C, Edder P, Bieri S, Christen P. 2014. Survey of pyrro-
lizidine alkaloids in teas and herbal teas on the Swiss market
using HPLC-MS/MS. Anal Bioanal Chem. 406:7345–7354.
Mattocks AR. 1986. Chemistry and toxicology of pyrrolizidine
alkaloids. London: Academic Press.
Molyneux RJ, Gardner DL, Colegate SM, Edgar JA. 2011.
Pyrrolizidine alkaloid toxicity in livestock: a paradigm for
human poisoning? Food Add Contam A. 28:293–307.
Mroczek T, Głowniak K, Kowalska J. 2006. Solid-liquid extrac-
tion and cation-exchange solid-phase extraction using a
mixed-mode polymeric sorbent of Datura and related alka-
loids. J Chromatogr A. 1107:9–18.
Phillipson JD, Handa SS. 1975. N-oxides of hyoscyamine and
hyoscine in the solanaceae. Phytochemistry. 14:999–1003.
Phillipson JD, Handa SS. 1976. Hyoscyamine N-Oxide in
Atropa-Belladonna. Phytochemistry. 15:605–608.
Rivlin VS. 2005. Wissotzky tea company has increased its mar-
ket share by 7.8%, thanks to its new Galilee fragrances
brand. Globes online Israel business news. [cited 2015 Aug
20]. Available from: http://www.globes.co.il/en/article-
877767
Rizk AM. 1991. Naturally occurring pyrrolizidine alkaloids.
Boca Raton (FL): CRC Press; p. 234.
Roulet M, Laurini R, Rivier L, Calame A. 1988. Hepatic veno-
occlusive disease in newborn infant of a woman drinking
herbal tea. J Pediatr. 112:433–436.
Routledge PA, Spriggs TLB. 1989. Atropine as possible con-
taminant of comfrey tea. Lancet. 333:963–964.
Ruan J, Yang M, Fu P, Ye Y, Lin G. 2014. Metabolic activation
of Pyrrolizidine alkaloids: insights into the structural and
enzymatic basis. Chem Res Toxicol. 27:1030–1039.
Schmeller T, El-Shazly A, Wink M. 1997. Allelochemical activ-
ities of pyrrolizidine alkaloids: interactions with neurorecep-
tors and acetylcholine related enzymes. J Chem Ecol.
23:399–416.
Shimshoni JA, Mulder PP, Bouznach A, Edery N, Pasval I, Barel
S, Abd-El Khaliq M, Perl S. 2015. Heliotropium europaeum
poisoning in cattle and analysis of its pyrrolizidine alkaloid
profile. J Agric Food Chem. 63:1664–1672.
Sperl W, Stuppner H, Gassner I, Judmaier W, Dietze O, Vogel W.
1995. Reversible hepatic veno-occlusive disease in an infant
after consumption of pyrrolizidine-containing herbal tea. Eur
J Pediatr. 154:112–116.
Van Wyk BE, Verdoorn GH. 1990. Alkaloids as taxonomic
characters in the tribe Crotalarieae (Fabaceae). Biochem
Syst Ecol. 18:503–515.
Xia Q, Zhao Y, Von Tungeln LS, Doerge DR, Lin G, Cai L, Fu
PP. 2013. Pyrrolizidine alkaloid-derived DNA adducts as a
common biological biomarker of pyrrolizidine alkaloid-
induced tumorigenicity. Chem Res Toxicol. 26:1384–1396.