Professional Documents
Culture Documents
6, 2000 1387
FOOD CHEMICAL CONTAMINANTS
arious analytical approaches have been used for analysis of patulin in apple juice with essentially similar
(fairly simplistic) extraction and cleanup but with different determinative steps. Thin-layer chromatography
(TLC; 1), gas chromatography (GC; 2) with derivatization,
gas chromatography/mass spectrometry (GC/MS) with
derivatization (3), and liquid chromatography (LC; 47) have
all been successfully used for patulin analysis. For TLC with
3-methyl-2-benzothiazolinone hydrazoneHCl as spray reagent, a full collaborative study was undertaken in 1974 (1),
resulting in a First Action AOAC Method. For that study at
patulin levels of 50, 120, and 340 ng/g, the relative standard
deviations for repeatability (RSDr) and reproducibility
(RSDR) ranged from 31 to 38%, and 21 to 30%, respectively.
An ISO standard for determination of patulin in apple juice,
apple juice concentrates, and drinks containing apple juice
was also based on a TLC approach (8).
More recent studies have used LC rather than TLC or GC.
A collaborative study by IUPAC in 1988 (9) compared 2 LC
methods, one extracting into ethyl acetate and back-extracting
into sodium carbonate, and the other extracting into ethyl acetate with cleanup on a silica gel column before LC. Both approaches were acceptable. The first gave RSDr values of
about 7% and RSDR values of about 8%; the second gave
RSDr values of 1218% and RSDR values of 1417%. Apple
juice test samples with patulin contents ranging from 40 to
200 ng/g were used in the 2 trials, and led to an ISO standard
for patulin (10) based on the first of the 2 approaches (without
silica gel cleanup).
The most recent collaborative study was undertaken by
Brause et al. (11) and led to the AOAC First Action LC
Method (12). That study involved 22 participants from
10 countries who analyzed blind duplicates of apple juice containing patulin at 4 spiked levels (20, 50, 100, and 200 ng/g)
and one naturally contaminated apple juice containing an average level of 35 ng/g. The method used by Brause et al. (11)
was essentially identical to that used in the present study for
1388 MACDONALD ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 83, NO. 6, 2000
clear apple juice. The study reported here, however, was extended to cloudy apple juice and apple puree, which have presented problems in the past in terms of recoveries and are
more prone to interference. The method has been modified
and includes pectinase enzyme treatment to improve the clarity of juices and puree before analyte isolation. The collaborative trial results for clear juices in the present study may be
compared directly with those of Brause et al. (11), but there
are no comparable studies for cloudy juices or apple puree.
Regulations control patulin contamination in about
11 countries worldwide (13) with limits in most cases set at
50 ng/g for fruit (apple) juices, although in the Czech Republic
a limit of 30 ng/g is applied to infant foods. At present, there
are no European Union (EU) limits for patulin, although
5 member states of the EU have limits in their national regulations and others (e.g., the UK) have voluntary guideline limits
and there are some moves to harmonize limits in the EU. The
European Standardization Organization CEN has indicated a
need for a validated method for patulin as a European standard, which would cover not only clear juice but also the increasingly popular organic cloudy juices and apple puree,
which may be produced from damaged and molded apples.
Test Materials for the Collaborative Trial
blank juice were added as necessary to reach the required concentration of patulin.
MACDONALD ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 83, NO. 6, 2000 1389
Table 2000.02. Interlaboratory study results for patulin in clear and cloudy apple juices and apple puree
0, ng/g
ID
No. of labsa(b)
Sr
RSDr, %
SR
RSDR, %
HORRAT
Rec., %
67
12 (0)
8.4
13
15.3
23
0.95
89
26
12 (0)
3.7
14
8.4
33
1.18
nc (c)
54
12 (0)
11
13.6
25
1.02
nc (d)
128
10 (2)
9.9
14
11
0.50
12.5
21
0.85
80
75 ng/g (a)
d
nc (b)
8
Cloudy Apple Juice
75 ng/g (a)
60
11 (1)
7.8
13
nc (b)
26
12 (0)
8.9
35
nc (c)
69
9 (2)
4.3
nc (d)
106
10 (2)
10.2
10
35
1.25
10
8.9
14
0.61
12.9
12
0.54
Apple Puree
75 ng/g (a)
69
9 (1)
7.5
11
9.2
13
0.56
92
nc (b)
23
8 (1)
6.4
27
8.5
36
1.23
nc (c)
38
9 (2)
3.8
10
12.6
33
1.27
nc (d)
121
10 (0)
23.6
19
34.8
29
1.31
a(b)
c
d
a = number of labs retained after eliminating outliers; (b) = number of labs removed as outliers.
Laboratories received 4 sets of duplicate test samples labeled a, b, c, and d.
nc = naturally contaminated.
1390 MACDONALD ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 83, NO. 6, 2000
A 1000
C
3160
A MW 1000 CF
MACDONALD ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 83, NO. 6, 2000 1391
range of standard curve, dilute extract with pH 4 water, reinject, and re-analyze diluted extract solution.
Calculate concentration of patulin in test sample (ng/g) as
follows:
ng/g Patulin =
C T 1000
d
10
where CT = concentration of patulin in extract (ng/g), 10 = ratio of test portion in test solution (5 g apple juice or apple puree is represented by 0.5 mL test solution); d = dilution factor,
which = 1 for undiluted test portion.
Ref.: J. AOAC Int. 83, 13891391(2000)
1392 MACDONALD ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 83, NO. 6, 2000
Table 1. Collaborative trial results of liquid chromatographic determination of patulin in clear apple juice
Patulin concentration, ng/g
Lab ID
75
aa
75
68.8
68
67.4
68.6
<5
<5
80.2
78.1
11.8
50.4
48.2
76
76.4
20.1
71.9
86.8
5.5b
27.2
26.7
52.6
53.9
109.6
113.9
28
27.6
59.2
58.6
120
115.6
8.3
40.8
34.8
64.7
68.8
136.7
124.7
1.9b
18
26.8
38.2
44.7
91b
149b
17.5
29.3
28.8
52.4
53.7
111.2
145
11.6
21.1
62.8
44.7
140.3
133.6
50.5
59.1
106.7b
78
78.6
147.8
141.6
49.8
47
3.3
6.8
11.7
3.9
82.4
73.2
7.8
8.6
32.4
29.7
10
68.5
72.5
<5
<5
24.4
25.1
40.1
50.8
116.1
120.6
11
51.6
28.2
23.5b
<4.2b
28.1
21
38.9
37.1
129.2
151.3
13
69.4
97.8
18.7
15.8
26.6
24.1
27.2
43.8
116.9
109.8
14
61.8
65.4
6.5
8.6
34.2
36.2
70
69.1
139.3
138.9
a
b
Table 2. Collaborative trial results of liquid chromatographic determination of patulin in cloudy apple juice
Patulin concentration, ng/g
Lab ID
1
aa
75
75
60.6
54.7
<3
a
<3
21.6
26.4
69.3
65.8
95
96
109.4
62.2
60.8
nd
nd
24.3
24.2
64
64.4
107
73.4
72.9
8.9
8.1
29.9
23.7
70.2
82.1
103.6
8.4
6.8
nd
nd
26.5
73.6
72.4
21.5
7
8
65.6
43.2
43.4
49.6
6.5
4.7
5.4
108.7
2.8c
6.4
nd
12.8
29.6
73.1
84
116.1
128
51.8
23.6
57.5
60
113.4
117.8
32.2
8.6
24.7c
14.4c
30.6
29.3
84.3
82
116.6
123.1
29
31
79.6
75.6
78.6
113.6
0c
2.8c
73.1
66.9
10
66.8
81
<5
<5
11
51.1
44.5
<4.2
<4.2
20.8
21.8
59.6
54.3
87.1
96.2
13
47.8
67
64.8c
0c
28.8
25.3
0c
64.6c
93.3
116.6
14
50.2
39
2.7
3.4
25.6
28.7
62.4
62.1
99.9
95.7
a
b
c
d
5.4
MACDONALD ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 83, NO. 6, 2000 1393
Table 3. Collaborative trial results of liquid chromatographic determination of patulin in apple puree
Patulin concentration, ng/g
75
75
aa
62.2
61.9
<3
<3
19.9
23.7
37.4
39.4
127.4
67.4
65.6
ndb
nd
22.6
25.6
42.2
38.3
135
120.2
75.1
85.6
79.3c
25.9c
89.6c
19.2c
161.2
151.6
3.7
nd
Lab ID
d
96.9
17.5d
6.8
21.4
nd
nd
9.3
nd
10.4
69.4
77.9
24.5
39.1
50.5
46.2
140
78.4
79.5
0c
29.5c
103.7
0e
0e
0e
0e
0e
0e
0e
0e
81.6
62.2
4.2
2.4
29.7
23.7
44.5
35.1
128.9
127.3
19
0e
59.8
104
0e
10
57.1
78.5
<5
19
14.3
<5
42.8
48.3
153.7
152.2
11
67.7d
xd,f
<2.3
<2.3
18.5
19.3
34
41.5
132.7
78.6
13
61.9
64.2
21.5
24.4
35.2
37.9
129.3
151.9
14
57.8
59.6
17.4
2.9
30.4
32.8
46.4
47.4
137.2
129.5
a
b
c
d
e
f
RSDr values were <20% and all RSDR values <25%, except
apple puree containing 121 ng/g patulin for which an RSDR of
29% was obtained. Although LC chromatograms showed no
evidence of interferences, any co-extractive compounds from
juices that interfered with the patulin determination would have
been more in evidence at lower patulin levels. There was no obvious correlation between laboratories that found higher than
average patulin levels in the control juices and their results for
positive samples. Thus, Laboratories 6 and 13 found mean levels of 18.8 and 17.2 ng/g, respectively, in the control clear apple
juice, but mean levels of 29.0 and 25.3 ng/g, respectively, in the
lowest naturally contaminated juice as compared with 26 ng/g
patulin, which was the mean result from14 laboratories.
Where it was possible to make direct comparisons e.g., for
clear apple juice, the values for precision characteristics for
this trial compared favorably with those for First Action
AOAC Method (12). Thus, for First Action AOAC Method,
RSDr values of 11 and 13%, and RSDR values of 22 and 21%
were obtained for spiked apple juice containing 50 and
100 ng/g patulin, respectively, compared with RSDr values of
11 and 8% and RSDR values of 25 and 11% for naturally contaminated apple juice containing average levels of 54 and
128 ng/g patulin. Recoveries of 91108% were reported for
the AOAC First Action Method compared with 89% for this
study; the AOAC Method covered a wider range
(20200 ng/g) of patulin spike concentrations.
The principal difference between this study and others was
the application of the collaborative trial to cloudy apple juice
and apple puree, which were previously more difficult matrixes to analyze. The values for recoveries of patulin derived
1394 MACDONALD ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 83, NO. 6, 2000
from the spiked samples ranged from 80 to 92% (Table 2000.02). There were negligible differences in recoveries
for all 3 matrixes, which were acceptable for determining
patulin in all 3 product types.
Interpretation of Results
Acceptability of the precision characteristics of the method
were assessed on the basis of the HORRAT values (15) which
compare the RSDR obtained for a particular level and matrix
with the value statistically predicted on the basis of collaborative trial studies taken from the published literature. When
outliers were excluded, the HORRAT values for patulin
ranged from 0.5 to 1.3. Because all HORRAT values were
<2.0, which indicates an acceptable precision, and were better
than or comparable to values reported in the AOAC-IUPAC
Official First Action Method (12), this method is clearly acceptable as an AOAC Official Method.
Recommendation
It is recommended that this LC method for patulin at
>50 ng/g be adopted Official First Action for determination of
patulin in cloudy apple juice and apple puree. The method is a
minor modification of the existing First Action Method for
clear juice, but has an additional step required for matrix extension to cloudy juice and apple puree. This is the first time
validation of a method has been carried out for these matrixes.
Acknowledgments
This project was financially supported by the European
Commission, Standards Measurement and Testing Program
(Brussels, Belgium) SMT Project CT96-2045. There have
been many who contributed to the success of this study,
including A. Boenke (EC SMT-Program Scientific Officer for
the project); K. Mathieson (CSL Food Science Laboratory,
York, UK), who performed the statistical analysis; A. Williams (formerly of Leatherhead Food RA, Leatherhead, UK),
who prepared calibrant standards; H.P. van Egmond (RIVM,
Bilthoven, Netherlands), who took responsibility for the pretrial workshop for participants; and A.E. Buckle (CSL, York,
UK), who assisted at the workshop. The authors also express
their appreciation to the following collaborators for their participation in the study:
Carlo Brera, Istituto Superiore di Sanita, Rome, Italy
Kevin Jrgensen, Danish Veterinary and Food Administration, Srborg, Denmark
Maria Lgia Martins, Laboatorio Nacional De Veterinria,
Lisbon, Portugal
Ma Luz Macho, Laboratrio de Salud Pblica, Bilbao,
Spain
Paul Majerus, Chemiches Untersuchungsamt, Trier, Germany
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
(15)