implementation and verification of an analytical … olajos...3.3.3 hplc quantification of biogenic...
TRANSCRIPT
Implementation and verification of an analytical method
for the quantification of biogenic amines in seafood
products
Ildikó Olajos
Thesis for the degree of Master of Science
Faculty of Food Science and Nutrition
School of Health Sciences
University of Iceland
Implementation and verification of an analytical method
for the quantification of biogenic amines in seafood
products
Ildikó Olajos
Supervisors: Helga Gunnlaugsdóttir and Hrönn Olina Jörundsdóttir
Thesis for the degree of Master of Science in Food Science
Faculty of Food Sciences and Nutrition
School of Health Sciences
University of Iceland
Reykjavík, January 2015
This thesis is 60 ects credits and is for Master degree in Food Science and may not be copied in any form without the permission of the rights holder
© Ildikó Olajos 2015
Prentun: Háskólaprentun
Reykjavík, Ísland 2015
i
Ágrip
Rotamín (e. biogenic amines, BA) eru hitaþolin, lífræn basísk efni með hátt suðumark sem myndast í
matvælum af völdum örvera vegna ensímatískra efnabreytinga á náttúrulegum amínósýrum. BA eru
áhættuþáttur varðandi matvælaöryggi þar sem þau geta valdið ofnæmisviðbrögðum hjá mönnum. BA
má finna í ýmsum matvælum, sérstaklega fisk af scombroid tegund (t.d. makríl, túnfisk, síld) sem ekki
hefur verið rétt meðhöndlaður eða geymdur við réttar aðstæður. BA brotna ekki niður við eldun og
greinast ekki við skynmat, því hefur Evrópuráðið sett reglugerð um mælingu á histamíni með HPLC
aðferð ásamt hámarksgildum í fisk og fiskafurðum með reglugerð (EB nr 2073/2005).
Markmið meistaraverkefnis sem hér er kynnt var að setja upp og sannprófa HPLC aðferð til að
ákvarða magn histamíns og annarra BA í fiski og fiskimjöli. Aðferðin sem sett var upp er byggð á
opinberri þýskari aðferð til að ákvarða BA í fiskafurðum. Uppsetning aðferðarinnar fól í sér
aðferðarbestun á m.a. sýnaundirbúningi, HPLC greiningu og úrvinnslu niðurstaðna. Þar á eftir var
aðferðin metin til að sannreyna að niðurstöðurnar sem fengust væru samræmdar, réttar og
fullnægjandi fyrir mælingar á BA í fiski og fiskmjöli.
Aðskilnaður milli mismunandi BA náðist á C18 súlu með öfugum stöðufasa og ferðafasastigli með
hækkandi natríum-asetati og asetónítrili. Til að sannprófa greininguna voru breyturnar sértækni,
línuleiki, mælisviðs, næmnimörk, greiningarmörk, hittni (accuracy), nákvæmni (precision) og heimtur
ákvarðaðar með tölulegum gildum. Stýririt fyrir histamín var sett upp fyrir innri gæðastjórnun til að meta
gæði, árangur og nákvæmni mælinganna í framtíðinni.
Aðferðin reyndist vera sértæk, BA voru aðskilin frá hverju öðru, upplausn toppa var góð og línuleiki
aðferðarinnar var mjög góður á fyrir styrk á bilinu 2,5-100ppm fyrir öll fjögur BA sem greind voru, þ.e.
r2≥0.994. Magngreiningarmörk aðferðarinnar voru ákvörðuð sem LOQ = 0,16-0,48 ppm, og því verður
jafnvel mögulegt að greina BA í mjög lágum styrk , heimtur á BA reyndust vera á bilinu 80-110%.
Aðferðin var nákvæm og endurteknar niðurstöður samkvæmar hverri annarri, sömuleiðis reyndist
enginn marktækur munur á niðurstöðum milli rannsóknarstofu Matís og annarra rannsóknarstofa.
Nákvæmni aðferðarinnar sýndi að hlutfallslegt staðalfrávik (RSD) af flatarmáli toppa í endurteknum
sýnum voru á bilinu RSD = 3-4,2% fyrir fiskhold og RSD = 4,2-5,9 % fyrir fiskimjöl, með hlutfallslegri
mælióvissu 3-4% fyrir fiskhold og 5,86-8,30% fyrir fiskimjöl. Samanburðarsýni leiddi einnig í ljós litla
hnikun (variation) í niðurstöðum, hlutfallslegt staðalfrávik var á bilinu 5-9% með hlutfallslega
mælióvissu milli 5-7,8%.
Þátttaka í samanburðarrannsóknum sýndi góðan árangur rannsóknarstofu Matís samanborið við
aðrar rannsóknarstofur. Útaukin staðalóvissa var ákvörðuð vera 7,4% fyrir fisk og 11,8% fyrir fiskmjöl
með hliðsjón að bjögun (bias) milli rannsóknarstofa.
Rannsóknin staðfestir að niðurstöður mæliaðferðarinnar voru samanburðarhæfar, réttar og
fullnægjandi fyrir greiningu á BA í fiski og fiskimjöli og mun því leiða til bætts matvælaöryggis.
Ávinningar rannsóknarverkefnisins eru m.a. að Ísland getur uppfyllt matvælalöggjöf ESB (EB tilskipun
nr 2073/2005) ásamt íslensku matvælalöggjöfinni og gerir framleiðendum sjávarafurða kleift að
staðfesta gæði og öryggi afurðanna.
ii
Abstract
Biogenic amines (BAs) are non-volatile, heat stabile, organic basis formed in food by microorganisms
through enzymatic decarboxylation of amino acid. BAs are a food safety hazard because they can
trigger an allergic response in humans, they can be found in various food such as inappropriately
handled and/or stored scombroid fish (e.g. mackerel, tuna, herring). Since these amines cannot be
destroyed by cooking or detected with organoleptic evaluation, the European Council requires the
determination of histamine in fish and fish products with HPLC and regulates the maximum levels of
histamine according to the Commission Regulation (EC) No 2073/2005.
The aim of this master thesis was the implementation and verification of an HPLC method for the
determination of histamine and other biogenic amines in fish and fish meal. The method applied was
based on an official German food testing method for the determination of BAs in fish based products.
The implementation of the method included optimization of: sample preparation, HPLC analysis and
data evaluation. Thereafter, the method performance was verified to demonstrate that the results
obtained were consistent, correct and satisfactory for analysis of BAs in fish and fish meal.
Separation of biogenic amines was achieved on a C18 reverse-phase column with gradient elution
separation with a binary mixture with increasing sodium-acetate and acetonitrile. For the verification
analytical parameters such as selectivity, linearity, working range, limit of detection and limit of
quantitation, accuracy, precision, recovery and were determined. For internal quality control, control
chart for histamine was prepared to monitor the future measurement performance and accuracy.
The method was selective, the biogenic amines separated from each other with a good resolution.
The analytical method demonstrated a very good linearity in the range of 2,5-100ppm for all four
biogenic amines, with r2≥0.994. The limit of quantification of the method was determined as LOQ=
0,16-0,48 ppm, giving a good opportunity to measure biogenic amines even in a very low
concentration. The recovery of biogenic amines ranged from 80-110%. The method was accurate,
repeatable and reproducible; there were no significant differences in inter-laboratory measurements.
The precision of the method showed that the relative standard deviation (RSD) of the peak areas of
the replicates were in the range of RSD= 3-4,2% in the case of fish flesh and RSD= 4,2-5,9% in the
case of fish meal, with a relative measurement uncertainty of 3-4% for fish flesh and 5,86-8,30% for
fish meal. The reproducibility also revealed a low variations in the results obtained, the relative
standard deviation was in the range of 5-9% with a relative uncertainty between 5-7,8% .
Participation in proficiency testing showed that the performance of the Matís laboratory was very
good compared to other laboratories. Extended relative uncertainty was determined, taking into
consideration the inter- laboratory bias, to be 11,8% and 7,4% for fish meal and fish flesh,
respectively.
This study verified that the results obtained were consistent, correct and satisfactory for analysis of
BAs in fish and fish meal will improve food safety. As a result of the thesis, Iceland will be able to
comply to the EU legislation (EC directive No 2073/2005) and Icelandic regulations and enable
seafood producers to confirm the quality and safety of their products
iii
Table of Contents
Introduction ........................................................................................................................................ 1 1.
Literature review ................................................................................................................................ 3 2.
2.1 Chemical and physical characteristics of biogenic amines........................................................ 3
2.2 Biogenic amines in food ............................................................................................................. 4
2.2.1 Formation and control of BAs in food and fish ................................................................ 5
2.2.2 Regulatory and dietary limits of BAs in fish and fish products ........................................ 6
2.2.3 Histamine metabolism and histamine intolerance ........................................................... 7
2.2.4 Symptoms of Scombroid Fish Poisoning: ....................................................................... 9
2.2.5 Biogenic amines in fish .................................................................................................. 10
2.3 Principles of analyzing BAs ..................................................................................................... 11
2.4 Importance of method validation in analytical chemistry ......................................................... 16
Methods and materials .................................................................................................................... 19 3.
3.1 Sample origin ........................................................................................................................... 19
3.2 Preparation of the sample ........................................................................................................ 20
3.3 Standards and standard stock solutions .................................................................................. 21
3.3.1 Chemicals mobile phase solvents and derivatization solution ...................................... 21
3.3.2 HPLC quantification of Biogenic Amines using post-column derivatization with OPA
(new-Matís method) .................................................................................................... 22
3.3.3 HPLC quantification of Biogenic Amines using pre-column derivatization with OPA
(old-Matís method) ...................................................................................................... 26
3.3.4 HPLC quantification of Biogenic Amines using post-column derivatization with OPA
in Nofima accredited laboratory .................................................................................. 28
3.3.5 Calculation of the results ............................................................................................... 29
3.3.6 Quality assurance (QA) ................................................................................................. 29
Results and discussion .................................................................................................................... 31 4.
4.1 Operational verification ............................................................................................................ 32
4.2 Confirmation of identity and selectivity/specificity .................................................................... 32
4.3 Calculation of limit of detection (LOD) and limit of quantitation (LOQ) .................................... 38
4.4 Linearity, working range ........................................................................................................... 39
4.5 Accuracy .................................................................................................................................. 41
4.5.1 Repeatability precision .................................................................................................. 41
iv
4.5.2 Reproducibility precision ............................................................................................... 45
4.6 Robustness .............................................................................................................................. 49
4.7 Analytical method efficiency (recovery) ................................................................................... 50
4.8 Internal quality control (IQC) .................................................................................................... 51
4.9 Participation in proficiency testing ........................................................................................... 52
4.10 Inter-laboratory uncertainty ...................................................................................................... 56
Conclusion ....................................................................................................................................... 57 5.
Future aspects ................................................................................................................................. 59 6.
References ...................................................................................................................................... 61 7.
Appendix .......................................................................................................................................... 69 8.
v
List of figures
Figure 1: Histamine degradation in the human body adapted from Maintz L. 2007 ............................... 8
Figure 2: Schematic of HPLC ................................................................................................................ 13
Figure 3: The principles of fluorescence detection ................................................................................ 14
Figure 4: Modified silica particles of non-polar stationer phase ............................................................ 16
Figure 5: The principle of RP-chromatography with gradient elution .................................................... 16
Figure 6: ISO/IEC 17025 requirements for testing laboratories ............................................................ 17
Figure 7: A schematic diagram of the sample preparation and detection ............................................. 20
Figure 8: Vacuum filtration of the solvents used in gradient elution ...................................................... 22
Figure 9: Shimadzu HPLC system used in BAs analyses ..................................................................... 23
Figure 10: Gradient elution applied for the separation of BAs (new-Matís method) ............................. 24
Figure 11: Schematic flow chart of post-column derivatization used in current verification study (new-Matís method) ........................................................................................................... 25
Figure 12: Derivatization reaction of BAs with OPA .............................................................................. 25
Figure 13: Flow chart of pre-column derivatization (old-Matís method) ................................................ 26
Figure 14: Gradient elution used in pre-column derivatization method (old-Matís-method) ................. 27
Figure 15: The sequence used in the measurements of BAs ............................................................... 29
Figure 16: Parameters determined in verification study (verification plan) ........................................... 31
Figure 17: Identification of BAs on different columns (new-Matís method); A) represents the Nucleosil-120-5 column, B) represents the Zorbax Eclipse Plus column .......................... 33
Figure 18: Identification the elution row and retention time (RT) of BAs by measuring standard dilutions at the concentration of 100mg/kg (on product weight basis); A): standard mixture of BAs, B): Tyramine standard, C): Putrescine standard, D): Cadaverine standard, E): Histamine standard solution ......................................................................... 34
Figure 19: Graphic comparison of gradient elution of pre (old-Matís), and post-column (new-Matís) derivatization ........................................................................................................... 36
Figure 20: Comparison of concentration data in fish meal measured with old-Matís and new-Matís methods and obtained from accredited laboratory (Nofima) .................................... 36
Figure 21: Investigation of selectivity in sardine-matrices, A: sample No.1. and B: sample No.2. ....... 37
Figure 22: Comparison of the BAs concentrations (mg/kg) in sardine sample No. 1. measured in LAVES and Matís (new-Matís method) laboratories .......................................................... 38
Figure 23: Comparison of BAs concentration (mg/kg) in sardine sample No. 2. measured in LAVES and Matís (new-Matís method) laboratories .......................................................... 38
Figure 24: Calibration curve of the standards ....................................................................................... 40
Figure 25: Components of accuracy ..................................................................................................... 41
Figure 26: Investigation of robustness: the influence of the age of the derivatization solution on detected BAs concentration; A) HPLC profile of the standard mixture of BAs derivatized with one day old OPA derivatization solution; B) HPLC profile of standard
vi
mixture of BAs derivatized using a three days old derivatization solution; Elution row of BAs are Tyr, Putr, Cad, His on Figure A and B ............................................................. 49
Figure 27: Recovery study A: unspiked Icelandic cod sample, B: Icelandic cod sample spiked with 1µg histamine standard .............................................................................................. 51
Figure 28: Control chart for internal quality control; monitoring measurement performance with the measurement of histamine concentration of LVU RM ................................................. 52
Figure 29: A) Proficiency testing: Measurement of histamine in fish flesh B) Proficiency testing: Measurement of histamine in fish meal ............................................................................. 53
Figure 30: Overview of Z scores of different laboratories from inter-laboratory ring test in fish flesh. Numbers are representing the participating laboratories. Matís laboratory is marked as no.3 (Z-score=0,7), the Z-score for laboratory no.1 and 2 was determined as 0. ................................................................................................................................... 55
Figure 31: Overview of Z-scores of different laboratories from inter-laboratory ring test in fish meal. Numbers are representing laboratories, Matís laboratory is marked as no.3. (Z-score=0,2) For laboratory no.6 the Z-score was determined to be equal to 0. .................. 55
vii
ListofTables
Table 1: Biogenic amines and their precursor amino acids .................................................................... 3
Table 2: The chemical structure and characteristics of biogenic amines analyzed in the study ............. 3
Table 3: BA content of raw and fermented products ............................................................................... 4
Table 4: Regulated histamine levels in fish products by the European Committee ................................ 7
Table 5: Histamine toxicity levels referring to 100g of food ..................................................................... 7
Table 6: Type of fish and fish sources involved in scombroid fish poisoning (outbreaks), with locations and number of cases between 1970-2008 ........................................................... 9
Table 7: Symptoms caused by scombroid fish poisoning according to Poison Management Manual ............................................................................................................................... 10
Table 8: Concentration of BAs in mg/kg in histidine-poor and histidine-rich fish stored at different conditions ........................................................................................................................... 10
Table 9 : BAs content and their concentration changes through time in raw fish ................................. 11
Table 10: Qualitative methods used for the screening of histamine in Quality Control ........................ 12
Table 11: Overview of different extraction and derivatization possibilities in the determination of BAs with HPLC ................................................................................................................... 15
Table 12: Preparation of the calibration curve....................................................................................... 21
Table 13: Details of the HPLC instrument used for analysis of Bas (new-Matís method) .................... 23
Table 14: Time Program applied for the separation of BAs (new-Matís method) ................................. 24
Table 15: The analytical conditions in pre-column derivatization method (old –Matís method) ........... 27
Table 16: Eluents and solutions applied in the method of Nofima accredited laboratory ..................... 28
Table 17: Comparison of different columns used in the identification of BAs ....................................... 33
Table 18: Determined retention times of BAs on Zorbax Eclipse Plus (250x4,6mm; 5µm) with the flow rate 0,9 ml/min ............................................................................................................ 34
Table 19: Results of the system suitability tests: SD and RSD of areas and retention times of BAs ... 35
Table 20: Limit of detection (LOD) and limit of quantitaion (LOQ) values of each BAs ........................ 39
Table 21: Correlation coefficients of calibration curves of BAs ............................................................. 40
Table 22: Correlation coefficient (R2) values between calibration points to investigate linearity .......... 41
Table 23: SDr and RSD values of relevant BAs under repeatability conditions in fish flesh (LVU RM) .................................................................................................................................... 42
Table 24: Uncertainty (U), trueness (H) values of relevant BAs in fish flesh (LVU RM) under repeatability conditions ...................................................................................................... 42
Table 25: Welchs’ test to investigate differences between the mean values of results (Matís) and reference material (LVU) under repeatability conditions.................................................... 44
Table 26: Overview of measurements repetition (10 times) for the investigation of repeatability in fish-meal ............................................................................................................................. 44
Table 27: Standard deviation (SDr), relative standard deviation (RSD) and uncertainty (U, Ur) values calculated under repeatability conditions in fish meal. ........................................... 45
viii
Table 28: SDR and %RSD values of relevant BAs in fish flesh (LVU RM) investigated under reproducibility conditions .................................................................................................... 45
Table 29: Uncertainty (U), relative uncertainty (Ur) trueness (H) values in fish flesh (LVU RM) under reproducibility conditions ......................................................................................... 46
Table 30: Welchs’ test to investigate differences between the mean values of results (Matís) and reference material (LVU) under reproducibility conditions ................................................. 46
Table 31: Reproducibility (R) and repeatability (r) limits calculated in fish flesh (LVU RM) .................. 47
Table 32: Overview of repeatability (r) and reproducibility (R) limit values in different matrices LVU fish meat homogenizate measured in Matís, compared to lax, tuna and herring matrices measured in the laboratory, which carried out the official validation of the method ............................................................................................................................... 48
Table 33: Comparison of the derivatization ability of OPA derivatization solution over a three day period ................................................................................................................................. 50
Table 34: Recovery values in % in spiked white fish (Icelandic cod) .................................................... 51
Table 35: Data obtained from proficiency testing .................................................................................. 54
Table 36: Operation qualification tests and their acceptances .............................................................. 69
ix
List of abbreviations
ACN: Acetonitrile
BAs: Biogenic amines
BAI: Biogenic amine index
BMEL: Federal Ministry of Food and Agriculture
BfR: Federal Institute for Risk Assessment
CAD: Cadaverine
CI: Confidence interval
DAO: Diamino oxidase
EC: Commission regulation of Europian Union
ELISA: enzyme-linked immunosorbent assay
Em: Emission
Ex. Excitation
FAO: United Nations Food and Agriculture Organization
FDA: Food and Drug Administration of the United States
FIA: Flow injection analysis
GC: Gas chromatography
GLP: Good laboratory pracices
HDC: L-Histidine decarboxylase
HFP: Histamine fish poisoning
HIS: Histamine
HNMT: Histamine N-methyltransferase
HPLC: High performance liquid chromatography
IR: Infrared
LOD: Limit of detection
LOQ: Limit of quantitation
MAO: monoamine oxidase
ND: not detected
OPA: o-Phthalaldehyde
PUT: Putrescine
OQ; Operational qualification
QA: Quality assurance
QC: Quality control
RP: Reversed phase
x
RPC: Reversed phase chromatography
RSD: Relative standard deviation
RT: Retention time
SD: Standard deviation
TCA: Trichloric acid
TI: Tolerance interval
TLC: Thin layer chromatography
TYR: Tyramine
UV: Ultraviolet
WHO: World Health Organization
LVU: Laborvergleichsuntersuchung,
LAVES: Lower Saxony State Office for Consumer Protection and Food Safety
MAST: Icelandic Food and Veterinary Authority
xi
Acknowledgements
I would like to express my special thanks of gratitude to my supervisors: Helga Gunnlaugsdóttir and
Hrönn Ólina Jörundsdóttir and also to Heiða Pálmadóttir for their scientific guidance and patience
through this study and all for their help enabling me finalizing my master thesis. I am very grateful for
their precious time they spent with me.
I would like to thank Matís, providing me the opportunity, facilities and equipment for the
measurements, also providing training possibility lead by external German experts of whom I thank the
most Dorothea Erika Ella Majohr, Stefan Effkemann and Roland Gerhard Körber, guiding and
educating me through the HPLC method verification. I am obliged to become a trained analyst not just
in HPLC, but thanks for the opportunity provided through the Icelandic German bilateral project also in
HPLC-MS/MS and GC-MS/MS.
I express my warm thank to Mr Sean Scully for his scientific and technical guidance and support
any time I have needed.
I am using this opportunity to express my gratitude to everyone who supported me throughout this
project especially Guðjón Þorkelsson, Ingibjörg Rósa Þorvaldsdóttir, Ásta Heiðrún E. Petursdóttir,
Natasa Desnica and Paulina E. Romotowska.
Secondly I would like to thank my beloved parents and my love Guðmundur Hilmar who stood next to
me all along this time.
1
Introduction 1.
It is not documented when humans first started to capture fish, however fishing was for sure one of
their earliest activities to obtain food. Fish has always been an important part of human diet not only
because it is tasty and easily digested, but also because it is nutritious as it high protein content (15-
20%), essential good source of essential amino acids, and excellent source of vitamins (A, D, B),
minerals (Ca, Fe, Cu, Se) as well as and polyunsaturated fatty acids (omega-3 and omega-6) [1, 2].
The demand for fish has increased in the last decades in line with the growth of the global population
and economic well-being. Also as people are getting more health conscious and educated the demand
for food that promotes health increases. In this respect fish and fish products stand out among other
food commodities [3]. Consumers today expect that food is convenient and palatable; however they
also demand that their food is safe to eat. Food businesses that participate in the global food trade are
responsible to ensure that their food products are safe and competent authorities must implement food
safety measures according to international standards in order to ensure consumer protection.
According to the World Health Organization (WHO), foodborne diseases are the most widespread
health problems in the world. Between 10-25% of the outbreaks related to food safety are caused by
seafood and thereof 86% are linked to fish that contains bio-toxins and histamine. In Iceland four
incident was reported as histamine poisoning between 2004 and 2005. In these cases raw and
canned tuna caused the intoxication (see Table 6) [4] . Histamine fish poisoning (HFP) is caused by
the consumption of scombroid fish such as tuna and other pelagic fish such as sardines and mackerel
naturally containing high level of free histidine that has been transformed to histamine by natural
bacteria. If fish is mistreated in relations to temperature during or after the catch, enzymatic bacterial
decarboxylation can cause histidine degradation into histamine. The consumption such as spoiled fish
leads to the development of HFP, where the severity of symptoms differs, depending on the digested
amount of biogenic amines (BAs) and also on the sensitivity of the individuals to relevant chemicals.
As hazardous level of histamine cannot be detected by organoleptic examination, chemical analysis
of foods for traces of this potential food allergen is necessary and the most widely used quantitative
analytical measurement is based on high performance liquid chromatography (HPLC) [5, 6].
In order to be able to evaluate food safety, the necessary laboratory capacities need to be in place
such as laboratory equipment and training of the responsible staff in operating the laboratory
equipment and carrying out official analytical testing procedures. The present work was carried out as
a part of a bilateral project between Iceland and Germany. The aim of this project was to strengthening
laboratory capacities in Iceland in order to improve food controls and product safety. The project was
supported by Matís, the Icelandic Food and Veterinary Authority (MAST) and the Ministry of Industries
and Innovations in Iceland as well as the Federal Ministry of Food and Agriculture (BMEL), Federal
Institute for Risk Assessment (BfR) and Lower Saxony State Office for Consumer Protection and Food
Safety (LAVES) in Germany. The objective of the present work is to implement an official validated
method for the quantification of biogenic amines in fish and fishmeal using HPLC and to verify this
method and prepare the method for official accreditation at the Matís laboratory. This will enable
official authorities and seafood producers to monitor the occurrence of BAs in Icelandic products and
lead to increased food and feed safety.
3
Literature review 2.
2.1 Chemical and physical characteristics of biogenic amines
Biogenic amines (BAs) are non-volatile, heat stabile, low molecular weight organic bases with
biological activity and have aliphatic, aromatic, or heterocyclic structures [6-8]. They are formed as a
result of either the transamination of ketones or aldehydes catalyzed by amino acid transaminases or
as a product of microbial decarboxylation of free amino acids [7]. The BAs and the amino acids they
are originating from presented in Table 1.
Table 1: Biogenic amines and their precursor amino acids
Precursor amino acid Biogenic amine
Histidine Histamine
Tyrosine Tyramine
Lysine Cadaverine
Glutamine
Arginine Putrescine
Agmatine
BAs can be classified according to the number of amine groups they possess, thus tyramine and
histamine belong to monoamines, cadaverine and putrescine to diamines, spermine and spermidine to
polyamines [9]. Table 2 shows the chemical structures of the aliphatic BAs: putrescine, cadaverine,
the heterocyclic histamine and the aromatic tyramine [10] that are the main focus of the present thesis.
Table 2: The chemical structure and characteristics of biogenic amines analyzed in the study
Name AbbreviationMolecular
formulaStructure formula pK
Molecular
weight
Tyramine TYR C8H11NO
OH
CH2CH2NH2
pK = 9.6 137.2
Putrescine PUT C4H12N2 NH2H2N
pK1 = 10.8
pK2 = 9.4 88.2
Cadaverine CAD C5H14N2 H2N NH2 pK1 = 11.0
pK2 = 9.9 202.2
Histamine HIS C5H10N3 N
NH
CH2CH2NH2
pK1 = 9.8
pK2 = 6.0 111.1
4
The human body naturally produces BAs; BAs are synthetized through cellular metabolism and
have diverse physiological functions. While some (histamine and tyramine) are hormonal mediators, or
neurotransmitters (dopamine, serotonin), others are precursors of hormones, proteins, alkaloids and
nucleic acids [11]. Another physiological significance of BAs is their regulatory influence on body
temperature and blood pressure [12, 13].
2.2 Biogenic amines in food
BAs are the products of normal metabolic processes in vegetable-, animal-, and microbial cells. BAs
are normally present in various foodstuffs in low concentrations including fish, meat, dairy products,
beverages, vegetables, fruits, nuts and even in chocolate [10, 14]. They play a role in the formation of
certain aroma compounds resulting the typical taste of matured vegetables and fruits [10, 15].
The most common biogenic amines found in foods are histamine, tyramine, cadaverine, 2-
phenylethylamine, spermine, spermidine, putrescine, tyramine, and agmatine [16]. High concentration
of BAs in food, feed and beverages is attributed to bacterial decarboxylation [14], during storage,
aging, spoilage or fermentation [10, 17-19].
Various levels of BAs can be detected in beer and wine as well as in raw and fermented food as
the result of metabolic activities of food-associated microorganisms (see Table 3). Consequently, BAs
can be used either as food spoilage indicators [10, 19, 20] or as the markers of the microbiological
contamination of the food, fish and fish products. However, the quantity of BAs is not a reliable
representation of the real levels of bacterial contamination as the amines can be substrates of
enzymatic degradation or a product of the fermentation process itself [21-23].
Table 3: BA content of raw and fermented products
Product Amount of BAs
Cheese 5-4500 mg/kg
Pickled cabbage 110-300 mg/kg
Inappropriately stored fish 2400-5000 mg/kg
Beer 2,8-13 mg/L
Processed meat 10-700 mg/kg
Wine 5-130 mg/L
According to Veseli et al.
5
2.2.1 Formation and control of BAs in food and fish
Several extrinsic and intrinsic factors play an important role in the formation of BAs in food. These
factors are the concentration of free amino acids (direct precursors for BAs), the presence of the
suitable biochemical milieu for the bacterial maturation, multiplication, and decarboxylase activity.
Additionally, factors such as a suitable amount of carbon and other nutrition sources, growing factors,
pH, temperature as well as the availability of oxygen [24-26], and the redox potential of the media [27].
The formation of BAs is also depends on the type of the bacteria strain and species. The bacteria
that are considered to be involved in BAs formation including the genera of Enterococcus;
Staphylococcus; Pseudomonas; Aeromonas; Campylobacter; Arcobacter and Firmicutes including
Lactic acid bacteria [28-30]. The temperature optimum for many members of these genera between
20°C and 37°C [27]. The decarboxylase can be usually characterized as having acidic pH optimum
and being inactive in aerobic conditions [27, 31].
Since the optimal temperature range of the organisms involved (-20-10°C), improper refrigeration
of the captured fish allows their proliferation, thus the formation of histamine. To avoid this, rapid
chilling, preferably below 10°C within 4 hours of capture is recommended. For longer storage, the
applied temperature should be below 0°C because the decarboxylase generated due to elevated
temperatures (10-25°C) for a short time, continues to produce the biogenic amines even though
temperature is decreased below 5°C, as lower temperature only hinder the bacterial growth not the
enzyme activity [27, 32, 33]. In 2002, Du and coworkers verified that histamine production was 31
times higher at 10°C, and 4 times higher at 4°C compared to 0°C in tuna. They also showed that icing
temperature retarded the histamine formation [34]. Thus storage temperature is the key factor in
keeping histamine level and bacterial number low [35]. Once histamine is formed it is almost
impossible to eliminate it by freezing, cooking or smoking [36], due to its heat stability [32]. As the
presence of histamine does not change the organoleptic characteristic of the food it is almost
undetectable even by highly trained panelist, which increases the risk for general consumers [6].
Beside temperature control sanitary conditions also have to be maintained during handling, processing
and distribution considering that most of the histamine production occurs near the intestines and
diffuses from there to the flesh. [33]
Many studies have investigated food processing/preserving methods, focusing on their capability in
the degradation of already formed biogenic amines such as gamma irradiation, in a dose-dependent
manner [37], the application of diamine oxidase bacteria as a starter culture in fermented food [38],
salting or modified atmosphere packaging. The application of ionizing radiation is beneficial on two
levels. Primarily, the technique is capable of eliminating viable bacterial cells by damaging their nucleic
acids and secondly to induce the radiolytic degradation of BAs decreasing their concentration with an
increase in radiation doseage [37]. Although irradiation seems to be good method for controlling the
formation of Bas, it may cause some adverse effects in terms of organoleptic properties and nutrition
content [38].
Salting appears to be successful method in reducing BA levels. A higher final NaCl content of
product (6-10%) leads to a lower Enterococci and Enterobacteriaceae count both in fermented food
6
such as sausages and feta cheese, and in salt dried fish. Reaching 10-16% salt content in salt-dried
sardine, histamine forming strains disappear [39].
Modified atmosphere packaging where the major filling-gas is carbon-dioxide with fungistatic and
bacteriostatic properties, compared with vacuum packaging inhibits the growth of the microbes with
amino acid decarboxylase activity much better, not to mention its ability to prolong the bacterial lag
phase. Antimicrobial chitosan film packaging was found having even a better inhibitory effect,
providing the lowest histamine concentration and Enterobacteriacea count outranking the modified
packaging [30, 39, 40].
2.2.2 Regulatory and dietary limits of BAs in fish and fish products
Although BAs toxicity is well established, it is very difficult to define the exact threshold of their toxicity.
Due to the fact that the toxicity is the result of the interaction of the quantitative and qualitative factors
of the food. Furthermore the individual characteristics and the presence of BAs and the state of the
health of the consumer [20, 25, 41, 42]. However a toxicological level, as 750-900mg/kg for the
maximum total level of the BAs has been proposed [43]. Although other BAs such as cadaverine,
putrescine and tyramine enhance the toxicity of histamine due to their competition with the histamine-
metabolising enzymes, still histamine is the only amine with established legal limits for human
consumption in fish and fish products in the Europian Union [20, 44, 45]. The Europian Council of
Directive regulates the maximum level of histamine in fish belonging to the Scombroidea, Coryfenidae,
Engraulidae, Clupeidae, Pomatomidae and Scrombresosidae families. According to the Commission
Regulation (EC) No 2073/2005, the maximum limits are given both in raw (100mg/kg) and salted fish
products (200mg/kg) associated with high amount of histidine as shown in Table 4 [46]. The maximum
level of histamine in fermented fish products, such as in fish sauce is established as 400mg/kg, which
is in compliance with the Codex Alimentarius standards and in the line with the consumer exposure
data report of the Food Safety research agency of the European Union [47]. Although there is a wide
range of histamine and other BAs in fish sauce, due to the present of the fermenting microorganisms
contributing histamine accumulation, however the health risk following the consumption is excluded
due to the small average uptake [21, 22, 48]. The same EC regulation also specified the sampling
activities both according to the fishery products and fish sauce. In the case of fish products, only one
sample has to be taken at the retail level. If the level of histamine exceeds the regulated limit (>100
mg/kg or >200 mg/kg), nine new samples from the same batch have be taken and analyzed. As the
distribution of histamine is expected to be more even in fish sauce, only one sample has to be taken
according to the Commission regulation [46].
7
Table 4: Regulated histamine levels in fish products by the European Committee
Regulated limits in mg/kg of food Designation of products
>100 mg/kg in raw fish
>200 mg/kg
in salted fish for species belonging to the
Scomboidae and Clupeidae families
400mg/kg fishery products undergoing enzyme
maturation treatment in brine/ fish sauce
In the USA, the regulated limit for histamine, established by the Food and Drug Administration of
the United States (FDA) is more strict compared to the Europian Union and was set as: 50mg/kg
which is in alignment with the histamine toxicity-level suggested by Shalaby and coworkers [44] (see
Table 5).
Table 5: Histamine toxicity levels referring to 100g of food
Histamine level Toxicity
<5 mg/100 g safe for consumption
5-20 mg/100 g possibly toxic
20-100 mg/100 g probably toxic
>100 mg/100 g toxic and unsafe for human consumption *Suggested by Shalaby et al
The same directive recommended to determine the other BAs levels, associated with fish
decomposition, but up to now only tyramine dietary levels were recommended. Suggesting an
acceptance level between 100-800mg/kg, and a toxicity level over the amount of 1080mg/kg tyramine
intake [32].
2.2.3 Histamine metabolism and histamine intolerance
Histamine (2-[4-imidazolyl] ethylamine) is a biogenic amine that occurs in various degree in many
foods. It was discovered in 1910 by Dale and Laidlaw and later identified as an anaphylactic reaction
mediator which is synthetized by the decarboxylation of histidine (free amino acid) by L-histidine
decarboxylase (HDC) and can be degraded in the human body in two ways as presented in Figure 1
[49].
8
Figure 1: Histamine degradation in the human body adapted from Maintz L. 2007
Histamine levels in food vary depending on the maturation process and the degree of freshness.
The longer food is stored or left to mature, the greater its histamine content and the more problematic
it can be for individuals with food sensitivities and intolerance. The reason for this intolerance is the
histamine accumulation and the shortage of its degradation in the human body. The impaired
degradation is caused by the reduced level of secretory protein called diamino oxidase (DAO) which is
responsible of the inactivation of extracellular histamine along with cytosolic (intracellular) protein:
histamine N-methyltransferase (HNMT). As the result of the reduced DAO activity the non-degraded
histamine-excess into cells becomes enlarged causing numerous allergy-like symptoms [50].
The most common health impact of biogenic amines is histamine-poisoning also referred as
scombroid fish poisoning. The poisoning which occurs throughout the world as a food-borne chemical
intoxication caused by mostly the ingestion of fish that contains high levels of histamine (≥50 mg/100g
of food) as shown in Table 6 [35]. According to the United Nations Food and Agriculture Organization
(FAO), fish species with high free histidine tissue-level like sardines, anchovy, or particularly (dark
flesh) scombroids such mackerel, bonito and tuna are more likely to be involved in scombroid fish
poisoning.
Even though histamine is the main mediator in scombroid fish poisoning, as well as tyramine, other
biogenic amines like agmatine, putrescine, cadaverine, anserine, spermine and spermidine enhances
9
the toxic effect of histamine and tyramine by either inducing the release of histamine from mast cells or
blocking the histamine degrading enzymes [51]. These synergic effect result in that histamine at lower
levels can be poisonous, combined with the presence of other biogenic amines in 5 times higher
concentration [45, 52, 53].
Table 6: Type of fish and fish sources involved in scombroid fish poisoning (outbreaks), with locations and number of cases between 1970-2008
Species involved/ source Time/location No. of cases Ref
Mackerel 2008, Romania 3 [54]
Fish cube 2007, Taiwan 347 [55]
Tuna 2005-2007, Israel 46 [56]
Fried fish cubes 2007, Taiwan 347 [55]
Tuna USA, 2006 11 [56]
Tuna 2006, Taiwan 7 [55]
Tuna 2004-2005 Iceland 4 [4]
VP-cold smoked tuna 2004, Denmark 10 [57]
Yellowtail 2004, South Africa 19 [58]
Billfish 2004, Taiwan 59 [59]
Garfish 2001, Denmark 13 [60]
Fish 1998-2002, USA 463 [58]
Fish 1993-1997, USA 297 [61]
Tuna, mackerel 1987-1996, United
Kingdom
404 [62]
Tuna, mackerel 1979, Italy 250 [63]
Dried horse mackerel 1973, Japan 2656 [64]
Mackerel, tuna, anchovies, sardines, marlin 1970-80, Japan 4122 [64]
2.2.4 Symptoms of Scombroid Fish Poisoning:
Under normal conditions exogenous BAs digested with food are quickly detoxified by the amino
oxidase enzymes (MAO and DAO) of the human gut epithelia, but if one of these enzymes is inhibited
for example in allergic individuals or the intake is too high (˃50mg/100g), the BAs accumulate in the
body causing different allergic-like symptoms [20]. The main symptoms of histamine poisoning are
summarised in Table 7. According to the Poison Management Manual the evolution of the initial
symptoms occur within 10-90 minutes, while advanced symptoms occur at later stages [14, 65].
10
Table 7: Symptoms caused by scombroid fish poisoning according to Poison Management Manual
Initial symptoms Advanced symptoms Sever cases
Facial flushing Facial rush Respiratory stress
Sweating Hives Swelling of tongue
Burning taste sensation in the
Mouth and throat
Edema
Dizziness, Nausea Short term diarrhea
Palpitations, Headaches Abdominal cramps
Normally, treatment is not necessary if the symptoms are mild, but antihistamine medication can be
applied if it is necessary, which leads to rapid relive of the symptoms and recovery.
2.2.5 Biogenic amines in fish
Naturally, fish contains different level of histamine and other BAs. The amount of histamine in freshly
caught fish can also be low as 0,1mg/100g [58, 66]. Bacteria associated with histamine and BAs
formation are present in the aquatic environment of the fish and naturally present in the normal
microflora of live fish. Appearing both on the external surface and in the gut of the fish, with no harm to
it. Due to death the defence mechanism is no longer capable to inhibit the bacterial growth, histamine
forming bacteria may grow in the muscle tissues [67]. These bacteria can contaminate fish flesh during
catching and processing, leading to the elevated formation of BAs if fish is subjected to temperature
abuse [35, 68]. BA content of fish is dependent of the species, those with a higher free histidine
content such as scombroids including tuna (8% of globally traded fish), bonito, mackerel, or sardine
are more likely to contain a higher BAs level after contamination and temperature abuse as Table 8
shows [68].
Table 8: Concentration of BAs in mg/kg in histidine-poor and histidine-rich fish stored at different conditions
Histidine
content
Fish type Temperature/time Histamine
(mg/kg)
Putrescine
(mg/kg)
Cadaverine
(mg/kg)
Tyramine
(mg/kg)
Ref
Histidine -
poor fish
Carp 3°C/8days ND* 11,8 8,1 0,2 [69]
Rainbow
trout
Ice/12 days 0,4±0,0 6,6±0,6 3,3±0,8 ND [70]
Histidine
rich fish
Sardine 4°C/15days 203±13 114±26 100±49 16±17 [30]
Indian
Anchovy
35°C/16h 2007,0 259,9 863,4 273,0 [71]
herring 0°C/16days 237,2 39,7 4,2 271,4 [72]
*ND=not detected
11
Further BAs can be used to evaluate the hygienic quality, decomposition of marine, freshwater fish
and crustacean[73] . Although several BAs can be found in fish, only cadaverine, putrescine, and
histamine are used in quality determination [74].The reason why tyramine is not included in this quality
determination is that the primarily low tyramine content of fish (0-40,3mg/kg) does not change as
significantly over time on low temperature storage (0°/4°C), such as the other three BAs as shown in
Table 9. Generally the formation of other BAs is considerably lower in histamine rich (dark muscle)
fish than histamine, thus histamine is used as a quality indicator in these species. In contrast the
formation of histamine is much lower in white flesh fish (histidine-poor) than the legal limit in
scombroids and changes in histamine content occurs over a longer period of time thus histamine is
less suitable as a quality indicator in white fish. In these species putrescine, and cadaverine are the
mainly forming BAs in, thus their level indicates the quality of shellfish, fermented seafood products
and white muscle fish (histidine-poor) [73]. Cadaverine also used as spoilage indicator on the initial
stage of fish decomposition [74].
Table 9 : BAs content and their concentration changes through time in raw fish
Biogenic amine Concentration (mg/kg)
Fish Temperature/time Cadaverine Putrescine Tyramine Histamine Reference
herring 0°C/0days 8,5 0,0 0,0 0,0 [72]
herring 0°C/16days 237,2 39,7 4,2 271,4 [72]
sardine 4°C/0days 3,9 13,4 0,0 19,5 [30]
sardine 4°C/15days 100,4 114,0 16,3 203,0 [30]
2.3 Principles of analyzing BAs
There are numerous methods available for the determination of histamine, however, some of them
such as the routine analysis methods listed in can only be used for screening and are not suitable for
quantification of histamine. The screening techniques are simple, inexpensive and do not require
expensive equipment or skilled technicians. They can be carried out on the spot (commercial kits) and
are therefore a useful parts of quality control. Despite of all these advantages they are limited and only
capable to supply qualitative, at best semi-quantitative, values. For further confirmation, more specific
quantitative analytical determinations are necessary since these procedures are more sensitive,
reliable and reproducible [75].
12
Table 10: Qualitative methods used for the screening of histamine in Quality Control
Methods/ routine analysis
Technical basis Detection Advantages LOD,LOQ Disadvantages
Colorimetric assay by
Saline extraction of histamine /centrifugation /extraction by n-butanol /evaporation/visual evaluation of color intensity
Spectrophoto-meter
Rapid (45min)
Inexpensive
Requires unskilled technician
LOQ:10mg/kg
Thin layer chromatography
Extraction by methanol or TCA
Separation by chromatography
Reveal by ninhydrin
Quantitative: densitometer
Rapid (2hours), simultaneous analysis of several samples
inexpensive
LOQ: 50mg/kg
semi quantitative
qualitative
Enzymatic Enzymatic conversion of histamine into imidazole acetaldehyde and hydrogen peroxide
The intensity of oxidation of leucocristal violet into purple crystal violet is proportional to histamine amount
Qualitative: Visual
Quantitative: spectrometry
Rapid (20min-2 hours), simultaneous analysis of several samples
inexpensive
LOQ:1,5mg/kg
LOD:0,5mg/kg
Overestimates histamine level
Immuno-enzymatic
(ELISA)
Extraction by water or acidic solution
Sample and enzyme labelled-antigen competition for binding-sites of antibodies coated wells.
Determination directly from standard curve
Quantitative: spectrometer
Rapid (15min-2 hours), simultaneous analysis of several samples,
Specific and sensitive
By the usage of commercial kits no equipment is necessary.
LOQ approx.= 50mg/kg
Semi-quantitative
*LOQ. Limit of quantitation, LOD: Limit of detection, TCA: Trichloric acid;
The most commonly used analytical methods for the separation and quantification of BAs are
chromatographic methods such as: thin layer chromatography (TLC), gas chromatography (GC) and
high-performance liquid chromatography (HPLC) [75].
13
Figure 2: Schematic of HPLC
HPLC systems consist of solvent reservoirs,pump(s) for mobile phase delivery (high or low
pressure), a degasser which removes the dissolved gasses from solvents preventing band spreading
and bad detector performance, injector, column, a detector, and integrator. Furthermore, thermostated
components (autosamplers, guard and analytical column , and detector are needed to keep
temperature consistent to ensure reproducable separations. A generic HPLC setup is given in Figure
2 [76].
For the separation either isocratic or gradient elution can be used. While the first one employs a
single solvent of constant composition, the second one uses at least two different solvents with
different polarity. In gradient elution the ratio of the solvents varry in a programmed way in a series of
steps. With the usage of gradient elution separation time can be reduced.
In HPLC the sample is introduced by an autosampler or manual injection onto the column, where
separation takes place. These analytical columns range in length from 10 to 30cm with an internal
diameter of 1 to 10mm. The particles of the column, which impact the separation and pressure can
also differ in the size (3-10µm). The particles can be made of pellicular or porous particles. Pellicular
particles, used currently rather for pre-columns (30µm to 40µm) consist of nonporous polymer or glass
beads, covered by thin pourous layer of silica, alumine or ion-exchange resin. The porous particle
packiging consists of pourous microparticles with 3-10µm sized diameters, made of alumina,
polystyrene-divinyl benzene, ion-exchange, or the the most commonly used silica resin. Silica resin is
coated by physically or chemically bonded organic films. To obtain better chromatograms close control
of temperature is recommended which is implemented by the column thermostat.
For detection different kind of detectors are used such as refractive index-, ultraviolet/visible light
(UV) absorption, Infrared light absorption, fluorescence emission and electrochemical detectors, are
the most commonly used-although mass spectroscopy is increasingly prominant [77, 78].
Fluorescence is detected by a photoelectric detector which in most cases utilizes a xenon lamp as a
light source to excite the fluorescence chemicals in the sample. These kind of lamps emit light in the
14
range of ultraviolet and infrared spectra. The excitation monochromator separates the optimal
wavelength from the emitted light to be used for the excitation of the analyte. Sample absorbs this
electromagnetic radiation and the electrons of the sample molecules move to higher excited states
from their ground state. Following this in a very short time, in 10-5-10-10 sec, without spin changing
fluorescence emission happens (∏→∏*) as the result of the movement of the electrons back to a
lower energy state. The wavelength of this emitted light (visible) is longer than the exiceted light (UV)
and has a lower energy. Emitted light is also specific to the compound itself and can be detected by
the fluorescent detector.
Figure 3: The principles of fluorescence detection
As BAs are non-UV absorbing analytes and have no native fluorescence characteristics
necessitating chemical derivatization to form detectable substances (fluorescent derivatves) [48, 79].
Introducing fluorofores, analytes sensitivity to UV absorption and fluorescence detection can be
increased. The most useful and intense fluorescence compounds contain fused aromatic funtional
groups with low energy ∏→∏* transmission levels. A higher number of aromatic group-content leads
to better quantum efficiency [80]. The most commonly used chemical for the derivatization of BAs is o-
phthalaldehyde (OPA) a fluorogenic agent which is able to react with primary amines in the presence
of 2-mercaptoethanol forming a fluorescent isoindol derivative under basic conditions (pH 9-11) [81].
The derivatization of the BAs is possible both before and after their separation. If the derivatization
agent is mixed to the sample prior to the separation than pre-column separation takes place, otherwise
post-column derivatization takes place. The derivatization agents are different in the case of these two
kind of labelling technique. Ninhydrin and OPA are commonly used in post-column derivatization,
while dansyl and dabsyl chloride, benzoyl chloride, fluoresceine, and 9-fluorenylmethyl chloroformate
are commonly used in pre-column derivatization [82-84]. Table 11 shows an overview of different
extraction and derivatization methods for HPLC analyses of BAs selected from the literature.
15
Table 11: Overview of different extraction and derivatization possibilities in the determination of BAs with HPLC
In liquid-solid (adsorption) chromatography the solid stationary phase involves high-surface-area
particles and the solute sample moves along with the mobile phase, sample components interacting
with the surface of the solid phase particles causing retention-migration pattern allowing separation of
individual components. The retention and the migration of the various compounds differs on the basis
of the interaction with the solid phase. This different migration depends on the equilibrium distribution
of the compounds between the mobile-, and stationary phases, determined by the separation
temperature and the composition of both the stationary-, and mobile phase [76]. In liquid-solid
separation normal or reversed phase columns can be applied. In reversed phase chromatography
(RPC), the stationary phase has a non-polar (hydrophobic) surface, such as silica particles modified
with C18 (Figure 4 [76]) while the mobile phase is polar, unlike the conditions in normal phase
chromatography. Thus, the hydrophobic solid phases used in RPC allows the separation of molecules
with hydrophobic characters as these non-polar components spend more time on the column due to
their interactions with that solid phase and elute later than more polar compounds [90].
Amines Sample-type
Sample-treatment
Derivatization Column/
stationary phase
Mobile phase Detection
λ
Ref
Various amines
Fish, cheese, meat products
Extraction with 0.1 M-hydrochloric acid
Pre-column derivatization with Dansyl chloride (heating for 60min to 40°C)
SPHERISORB 3S T6
(150mm x 1.6mm, 3µm)
Mixture of water and ACN
UV
250 nm
[85]
Various amines
Fish and fish products
Extraction with 0,6 M-perchloric acid
Post-column derivatization with OPA
NOVAPAK C18
150mm x
3,9 mm), 4µm
Gradient elution with A:0,1 M-sodium acetate and 10mM octanesulfonic acid (pH 5.2)
B:ACN, 0,2M-sodium acetate and 10mM-octanesulfonic acid (pH 4.5)
Fluorometric(excitation 340nm, emission 495 nm)
[86]
Putrescine, Cadaverine, Histamine
Fish Extraction with 5% trichloroacetic acid
Pre-column derivatization with fluorescein
Phenomenex IB-SIL 100mm x 4mm
0,02M-phosphate buffer (pH 7,2) and
ACN
Fluorometric(ex=390nm; Em=475nm)
[87]
Various amines
Fish Extraction with 5% TCA (heating 60°C for 15 min)
Pre-column derivatization with dansyl chloride
ALTEX ULTRASHERE-Si 250mm x 4,6mm,
Hexane-ethyl acetate(40:60) with addition of 0,01% amino ethanol
Fluorometric(ex=333nm; Em=470nm)
[88]
Putrescine, Cadaverine, Histamine, Spermidine, Spermine
Fish Extraction with 5% TCA
Pre-column derivatization with dansyl chloride (60 min, 55°C)
LICHROSORB
250mm x 3mm, 10µm RP-8
Gradient elution with methanol, ACN and 0,02M-acetic acid
UV 254nm [89]
Fish, sour cabbage wine
Extraction with 6% perchloric acid
Post-column derivatization with OPA
INERSIL ODS 2
250mm x 4,6mm, 5µm
Phosphate buffer (pH 7) and ACN
Fluorometric(ex=340nm; Em=455nm)
[84]
16
Figure 4: Modified silica particles of non-polar stationer phase
The hydrophobic binding interactions between the sample and the solid phase can be changed by
the alteration of the composition/polarity of the mobile phase. Thus adsorption and desorption of
solute can be controlled. By gradient elution separation, where in the initial conditions use a highly
polar solvent (containing a large % of water), the adsorption of the solute to the non-polar surface of
the stationary phase is promoted. This is followed by a stepwise or linear alteration of mobile-phase
polarity by increasing the percentage of the organic modifier which makes the solute to favor
desorption from the surface until that point that an extreme equilibrium is reached, namely solute will
be distributed 100% in mobile phase, subsequently eluted from the column as illustrated in Figure 5
below [90].
Figure 5: The principle of RP-chromatography with gradient elution
2.4 Importance of method validation in analytical chemistry
During method development, the method should be validated to that point that it will satisfy validation
requirements, so if given method is well developed, stabilized and optimized, and the initial
performance results are available from the laboratories developing the method and the application
laboratory does not have to perform full validation [91]. Method validation is essential in analytical
chemistry to ensure the quality and the reproducibility of the measurements [92]. Through validation
the developer and later the user have to test whether the method is fit for the intended purpose,
accurate, reliable and precise, capable to identify the analyte in question by demonstrating and
confirming the performance characteristics, represented by the functional characteristics and statistical
17
analysis. These characteristics show the degree of the reliability of the method under the investigated
operating conditions [91].
Validation and verification are essential requirements of accreditation according to ISO/IEC 17025
and ISO 1518 standards. The ISO/IEC 17025 standard requires fully documented procedures
including the performance check of the instrument, validation of the method, qualification of the
analysts and the outcomes of the tests. The implementation of ISO/IEC17025 which consists of both
Management and Technical Requirements improves the national and international reputation of the
accredited laboratory and is the basis for Good Laboratory Practices (GLP). In order to be accepted by
accreditation bodies, analysis have to be executed in alignment with protocols and the validation
process should be carried out according to the requirements of ISO/IEC17025 illustrated in Figure 6
[93].
Figure 6: ISO/IEC 17025 requirements for testing laboratories
Method validation should be performed routinely while the method being used [94]. It is important
to stress that there are no universal agreements neither regarding the definitions used in validation nor
the executions itself [92].
19
Methods and materials 3.
3.1 Sample origin
Different sample matrices (fish flesh and fish meal) were tested in the verification process. The reason
why these matrices were chosen was according to the expectation of future sample types that would
be analyzed by the laboratory where the newly implemented method (new-Matís metod) would be
executed. Samples originated from different sources.
Three samples of reference materials (RMs) containing BAs were obtained from ring tests. Thereof
one fish flesh homogenizate (LVU RM) was purchased from German proficiency test performed in
2013 with the participation of 19 laboratories (LVU: Laborvergleichsuntersuchung, „Biogene Amine
(2013) [95].
The other two RMs (one fish flesh homogenizate and one fish meal powder) were obtained from
AQUACULTURE Ring test (2014), which ran under the system established by ISO/IEC 17043
organized by Masterlab analytical services, where Matís laboratory participated.
A commercial fish meal sample was supplied from Icelandic local producers. Fish meal was
produced by the utilization of capelin (Mallotus villosus), herring (Clupea harengus) and blue whiting
(Micromesistius poutassou), caught mostly in the sea north and the east of the country.
Quality control material originated from sardine (Sardine No.1. and No. 2.) was received from
LAVES (Niedersӓchsisches Landesamt für Verbraucherscutz und Lebensmittelsicherheit, Institute für
Fische und Fischereierzeugnisse, Cuxhaven, Germany). These sardine samples were extracted by
well-trained analysts and sample-extracts were sent to Matis laboratory for measurements.
In addition, Icelandic cod was used for recovery studies, this fish flesh-extract was spiked with
standard mixture of BAs.
20
3.2 Preparation of the sample
In this study the preparation of the sample was carried out as illustrated on Figure 7.
Figure 7: A schematic diagram of the sample preparation and detection
For samples of fresh fish, whole fish from received batch were selected for the measurement.
Head, tail and guts were removed and fish were sliced with clean, stainless steel knife. Edible parts of
the fish were randomly chosen and homogenized with a food processor and transferred into plastic
container. In the next step 5g of the homogenate was weighted to the accuracy of 0,05g on an
analytical scale in a beaker. Sample was again homogenized, now with 45mL extraction solvent
(0,6mol/L perchloric acid (HClO4)) with a metallic staff homogenizer (T-25 digital Ultra-Turrax from
IKA®-Werke GmbH&Co., Germany) for about 2 min. The extraction solvent was removed and filtered
with Whatman filter papers (40, Ashless, 125mm, England), followed by filtration through a membrane
syringe filter (Milllipore Millex-HV, hydrophilic PVDF; 0,45µm, Ireland) into an Erlenmeyer flask. The
sample was then transferred into 2mL screw cap vials (Agilent Technologies) and placed in the HPLC
auto-sampler system for analysis or stored refrigerated for later measurement. In the case of fish meal
samples the extraction was carried out on the same way except for grounding.
SAMPLE
GROUNDING, HOMOGENIZATION
WITH FOOD PROCESSOR
SCALING
(5 0,05g g of sample)
EXTRACTION WITH PERCHLORIC ACID (homogenization)
PAPERFILTRATIONMEMBRANE FILTRATION
SEPARATION on reverse phase HPLC
column
POSTCOLUMN DERIVATZATION
FLUORESCENCE‐DETECTION
21
3.3 Standards and standard stock solutions
BAs standards were: Tyramine hydrochloride (C8H12ClNO), Cadaverine dihydrochloride (C5H16Cl2N2),
Histamine dihydrochloride (C5H11Cl2N3), Putrescine dihydrochloride (C4H14Cl2N2) purchased from
Sigma-Aldrich (Germany) and tandard material was HPLC grade. For the measurement of the target
BAs: tyramine, cadaverine, histamine and putrescine individual stock solutions were prepared with the
concentration of 100mg/mL: 182,8mg of Putrescine dihydrochloride, 126,6 mg of Tyramine
hydrochloride, 170,3 mg of Cadaverine Dihydrochloride, and 165,6 mg of Histamine dihydrochloride
were measured into 100 mL volumetric flask individually and dissolved in 0,6 mol / L perchloric acid
(HClO4). 0,6 M Perchloric acid was prepared from w=60% 6M HClO4, also obtained from Sigma-
Aldrich: 24 mL of HClO4 was diluted in 400 mL of deinozied water (˃ 18MΩ cm-1 ), produced in house
for chromatographic purposes with Millipore Milli-Q Academic Q-Gard® 1 deionizer (Millipore, Ireland).
Standard mixture was prepared from the stock solutions. From each stock solution, 1mL was
measured into 10mL volumetric flask and filled up with perchloric acid (concentration of
mixture=0,1mg/mL). Calibration curve of the standard mixture was prepared as shown in Table 12.
Table 12: Preparation of the calibration curve
Dilution Factor Amount taken out
from the mixture
Diluted
into (mL)
Concentration
( mg/mL)
BA-concentration
(mg/kg sample)
D50* 2mL 10 0.02 200
D100 1mL 10 0.01 100
D200 0.5 (500µl) 10 0.005 50
D500 200µl 10 0,002 20
D1000 100µl 10 0.001 10
D2000 50 µl 10 0,0005 5
D4000 25 µl 10 0,00025 2,5
D50*= standard mixture of BAs is diluted 50x times
3.3.1 Chemicals mobile phase solvents and derivatization solution
Chemicals used for the preparation of mobile phase solvents were all purchased from Sigma-Aldrich
(Germany) and were LC-grade or better, except for potassium hydroxide pellets (KOH) which was
obtained from MERCK (Germany). For the gradient elution separation two solvent mixtures, eluent A
and B, were prepared. In the preparation of eluent A, 8,03g of sodium acetate (CH3COONa) was
dissolved in 800mL of purified water. The solution was pH-adjusted with acetic acid (C2H4O2, w=
22
100%) to 4,5 ± 0,1. Next, 2,16g of sodium-1-octane sulfonate (CH3(CH2)7SO3Na*H2O) was added to
the solution and the solution diluted up to 1000 mL with purified water. To prepare eluent B, 12,73g of
sodium acetate and 600mL of deionized water was mixed. The pH was adjusted to 4,5±0,1 with acetic
acid (C2H4O2, w= ca. 100%) followed by addition of 2,16g of sodium-1- octane sulfonate and 230mL of
acetonitrile (CH3CN). In the final step, the solution was diluted to 1000 mL with deionized water in a
volumetric flask. Both eluent A and B were stirred with a magnetic stirrer (IKAMAG KMO-1 from IKA®-
Werke GmbH&Co., Germany), vacuum filtered with 0,4µm HTTP, Isopore Membrane filter (Millipore,
Ireland) shown in Figure 8 and sonicated for 30 min prior to use as mobile phase on the HPLC.
Figure 8: Vacuum filtration of the solvents used in gradient elution
The BAs derivatization solution was prepared by mixing the following substances, 3g BRIJ® L23
detergent (polyoxyethylenlaurylether) with 1g of o-phthaldialdehyde reagent that was dissolved in
10mL of methanol (CH3OH), followed by addition of 1L of borate buffer as well as 3mL of 2-
mercaptoethanol.
Borate buffer was prepared from 61,8g of boric acid and 40g of potassium hydroxide diluted to
1000mL with purified water.
3.3.2 HPLC quantification of Biogenic Amines using post-column derivatization with OPA (new-Matís method)
The separation of biogenic amines was carried out on Shimadzu HPLC instrument (Shimadzu
Corporation, Kyoto, Japan) (see Figure 9), the details regarding the instrument used are listed in
Table 13, using Zorbax Eclipse C18 reversed-phase chromatography column purchased from Agilent
Technologies (Canada, United States). The stationary phase consists of an ultra-high purity silica
support (SiO2), packed by a dense monolayer of dimethyl-n-octadecylsilane. The column is compatible
with water and all organic solvent. For the protection of the column, a guard column was used with the
23
same stationer phase to ensure a longer lifetime of the analytical column, purchased from the same
company as the analytical column.
Figure 9: Shimadzu HPLC system used in BAs analyses
Table 13: Details of the HPLC instrument used for analysis of Bas (new-Matís method)
MODULE TYPE OF MODULE
Fluorometric detector DGU-20A38/20A5R
Degassing unit DGU-20A38/20A5R
Auto sampler SIL-30AC
Column oven CTO-20A/20AC
Solvent delivery system (pump A) LC-30AD
Solvent delivery system (pump B) for post column
derivatization
LC-20AD
System controller CBM-20A
Analytical column ZORBAX Eclipse Plus C18 (4.6x250mm 5 –
micron) (Agilent)
24
Samples were prepared according to the description section 3.2 and 10µl of the final extract was
injected by the autosampler of the HPLC onto the Zorbax Eclipse column. A gradient elution
separation was applied to decrease the retention of later-eluting components and to provide a better
and sharper peak shape without tailing effect (common problem in isocratic elution). The gradient
consisted of a binary mixture of eluent A and B prepared as explained in section 3.3.1. The mixture
which contained ion-paring reagent to control the retention strength was pumped with pump A with the
flowrate of 0,926mL/min. Gradient program in which acetate buffer, and increasing proportion of
acetonitrile were used is shown in Figure 10 and Table 14. Column temperature was set to 50°C.
Figure 10: Gradient elution applied for the separation of BAs (new-Matís method)
Table 14: Time Program applied for the separation of BAs (new-Matís method)
0
10
20
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60
Solvent %
Time in min
A
B
Time
(min)
Ratio of
eluent B
Event
0,01 15% Start/Injec
tion
0,02 Zero
20,0 40%
30,0 65%
34,0 65%
34,1 100%
50,0 100%
51,0 15% Rinse
60,0 End
25
Figure 11: Schematic flow chart of post-column derivatization used in current verification study (new-Matís method)
Prior to fluorescent detection (exciation: 330nm, emission: 465nm), online post-column
derivatization took place: separated BAs were converted into fluorescent OPA-derivatives (see Figure
11 and Figure 12 for details). The prepared derivatizing solution was supplied with a flow-rate of
0.5mL/min by pump B and pumped through a 50 cm long reaction loop, where temperature control
was not necessary.
Figure 12: Derivatization reaction of BAs with OPA
26
After analysis of BAs, the analytical column was flushed with water:methanol in the ratio of 50:50%
5-10x of the column volumes to avoid any precipitation. The column volume was calculated by means
of the following equation: V=πr2L, where V=column volume in mL; r= column radius in cm; L=column
length in cm. During cleaning, the flowrate was 0,5mL/min, approximate the 1/2-1/5 of the typical
flowrate (here 0,926mL/min).
3.3.3 HPLC quantification of Biogenic Amines using pre-column derivatization with OPA (old-Matís method)
In the formerly used method (old-Matís method [96]) sample preparation and homogenization was
carried out in the same way as in the case of post- column derivatization (new Matís method) as
detailed in section 3.2. For extraction 5 g of sample was weighted and extracted in 50mL 10% TCA.
The extraction solvent was removed and filtered with Whatman 542 filter paper (Whatman
International Ltd Maidstone, England) and made up to 100mL. Then filtered through a membrane
syringe filter (Milllipore Millex-HV, hydrophilic PVDF; 0,45µm, Ireland).
Figure 13: Flow chart of pre-column derivatization (old-Matís method)
For pre-column derivatization OPA derivative was prepared; 90mg OPA was weighted into 10mL volumetric flask then 1mL methanol was added and soluted before 0,2mL 2-mercaptoethanol was added. Solution was made up to 10mL with boric acid buffer (pH 10,8). As Figure 13 shows the reagent addition and mixing settings used for the manual injection: 500µL of derivative was added to 250µL extraction, kept in dark for 3,5 min (derivatization interval) and then reaction was stopped with 2mL of ethyl acetate, vortexed (IKA minishaker, Sigma Aldrich, Germany) for a min and centrifuged (WIFUG-centrifuge; DJB Labcare, UK) for 0,5 min. After the phase-separation took place, supernatant was pipetted into 2mL vials and used for the HPLC measurements. The separation of the BAs was carried out on HP 1050 series (Hewlett-packard, Germany) HPLC instrument, using Hypersil BDS C18-RP column (Thermo Scientific, USA) as
27
Table 15 shows. For the protection of analytical column Hypersil BDS C18-RP guard column was
used. All chemicals used were purchased from Sigma-Aldrich (Germany)
Table 15: The analytical conditions in pre-column derivatization method (old –Matís method)
Type of derivatization
Pre-column OPA derivatization
Injection/injection volume
Manual
5 µl
Extraction 10% TCA
Calibration method Internal standard
Solvents A: Water
B: 10% ACN-90% 0,075M NaH2PO4
C: 60% ACN-40% 0,075M NaH2PO4
D: 100% ACN
Instrument HP 1050 series (Hewlett-packard, Germany)
Column Hypersil BDS C18-RP
250x4mm; 5µm (Thermo Scientific, USA)
Guard-column Lichrospher -18, 5µm 10x4mm (Merck, USA)
Detection Fluorescence: Varian 9070 detector
Ex: 336nm; Em:440nm
Datahandling system/calculation
Windows Chemstation
Excel 2010
In the gradient elution four different solvents were used with the gradient illustrated in Figure 14
Figure 14: Gradient elution used in pre-column derivatization method (old-Matís-method)
28
3.3.4 HPLC quantification of Biogenic Amines using post-column derivatization with OPA in Nofima accredited laboratory
For extraction 20g sample was weighted and extracted with 150ml 0,6M perchloric acid and then
250µl internal standard was added on dilution of the extract. Extract was filtered through a medicated
cotton then through a 2µm syringe filter. Pipetted into the autosampler vials and 20µl of sample was
injected on a Hypersil ODS (150cmX4,6mm) RP analytical column. Temperature was kept on constant
35°C. Gradient elution separation was applied with three solvents (A, B, C) containing ion-pairing
reagent as detailed in Table 16.
Table 16: Eluents and solutions applied in the method of Nofima accredited laboratory
Eluents and
solutions
Components of solutions
Eluent A Sodium acetate trihydrat (27,22g)
1-octanosulfonic acid (4,23g)
Distilled water (1800mL)
pH adjusted to 4,5 with acetic acid
Filled up to 2L with distilled water
Eluent B methanol
Eluent C Sodium acetate trihydrat (54,44g)
1-octanosulfonic acid (5,62g)
Distilled water (1800mL)
pH adjusted to 4,5 with acetic acid
Filled up to 2L with distilled water
Acetonitrile (mix solution: ACN=10:3)
Internal standard
solution
1,6 Diaminohexandihydrochloride (407,3mg)
Filled up to 250mL with 0,6M perchloric acid
OPA
derivatization
solution
OPA (1g)
Methanol (10mL)
Boric acid (1000mL, 1M)
Brij-35 (3mL)
2-merkaptoethanol (3mL)
All eluent gradients were linear. Gradient program applied increasing proportion of acetate buffer
and acetonitrile. Prior to fluorescent detection (ex: 365nm, em: 418nm) online post-column
derivatization took place applying OPA derivatization solution (see Table 16).
29
3.3.5 Calculation of the results
The identification of the amines was carried out by the comparison of the retention times (RT) with
those of standard substances. The quantitative determination was carried out according to the method
of external standards and integrating peak area in relation to the values of the standard substances.
The integration of the area was performed by LabSolution (Version 5.51; Shimadzu Corporation,
Kyoto, Japan) which is the Shimadzu HPLC system software. The evaluation of the concentrations
was done by using Microsoft Excel 2010. By means of linear equation of the regression line: y=ax+b
where y= area measurement in mAbs; x= concentration of the substance in mg/kg; a = slope of
regression curve and b = y-intercept. The amine contents were calculated in mg/kg on a wet weight
basis in fish flesh and on a product basis in fish meal.
3.3.6 Quality assurance (QA)
To ensure the quality of the measurement a reference material (further on LVU RM from former
proficiency testing) is used and measured along with the samples in each run. Concentration of the
RM is calculated and registered in a control chart to monitor the performance and accuracy of the
measurements. Action and warning limits for the method were determined from the control chart of in
the verification process and will be presented in section 4.9.
Figure 15: The sequence used in the measurements of BAs
The running sequence is presented in Figure 15 The sequence contains along with the RM and
samples four standard-samples on different concentration covering the range of 0,0005-0,01mg/mL
Two standard-samples are measured in the beginning and the end of the sequence. The first standard
sample also used to monitor the performance of the HPLC (function control). In the middle of the
sequence 1 blank sample is measured containing perchloric acid.
Standardc=0,01mg/mL
Standard
c=0,005mg/mLSamples
Blank
(perchloric acid)
LVU-RM))SamplesStandard c=0,0025mg/mL
Standard
c= 0,0005mg/mL
31
Results and discussion 4.
In this section the results of the verification of the installed analytical method for the quantification of
BAs with HPLC in fish and fish meal samples will be presented. The analytical method applied was an
official German reference method used by official control laboratories in Germany for the analysis of
BAs in fish and fish meal:§35 LFGB- 10.00-5; HPLC (1999-11) [97]. The investigated BAs were:
putrescine, cadaverine, tyramine and histamine.
The purpose was to demonstrate that the results obtained by the analytical method are consistent,
correct and satisfactory, reproducible despite of changes in technicians that perform the analysis or
slight changes in instrument performance or in chemicals used for the analysis. The following main
parameters that were investigated in the verification study were: identity, linearity, accuracy, trueness,
precision, repeatability [98]. The parameters of the verification (marked with purple) as well as the
procedures used to conform and demonstrate the different parameters are illustrated in Figure 16.
Figure 16: Parameters determined in verification study (verification plan)
Identity comparison of different colums
Linearity, working range
from calibration curves
Accuracy
Precision
Repeatability repetitions within-day by 1 operator
Reproducibility repetitions, between days by different operators
Trueness
Bias from RM, collaborative study, or spiked samples
Recovery from reproducibility
Robustness from derivatization ability
LOD,LOQ from signal/noise
Uncertainty from reproducibility and bias
32
Instrument verification (system suitability test), was also carried out as a part of the analytical
method verification. This included test that shows that the instrument was capable and reliable for
performing the measurements. System suitability test involved operational qualification and system
performance qualification and the results are presented in section 4.1 and 4.2.
4.1 Operational verification
Operational qualification (OQ) was carried out to demonstrate that the instrument functioned according
to its operational specification, and verified that the HPLC system complied with the key functional and
operational requirements as specified in the design qualification. A holistic testing was executed
following the verification plan presented in Table 36 in Appendix 1. This verification plan included the
accuracy of: flow rate, injection volume, temperature, wavelength, gradient concentration. Stability
tests and system reproducibility tests were also carried out according to the instructions of Shimadzu
Users manuals. The HPLC system passed the verification check and the investigated performance
characteristics were stable and fulfilled the acceptance criteria.
4.2 Confirmation of identity and selectivity/specificity
The first step of the new method installation included the identification of BAs. Standard mixture of
BAs were injected onto two different columns, Zorbax Eclipse Plus and Nucleosil‐120‐5, in the
concentration of 50mg/kg on wet weight basis and the results are illustrated in Figure 17. Two
different columns were tested as one was used and brought to Matís by LAVES expert (Nucleosil) and
the other had previously been purchased specifically for BAs analyses at Matís. The different
dimensions, physical and chemical properties of the two columns are presented in Table 17. Due to
the different column dimensions the appropriate flow rate was calculated for Zorbax Eclipse Plus
column and determined as 0,9mL/min, while in the case of the Nucleosil-120-5 column it was
0,7mL/min.
30,0 32,5 35,0 37,5 40,0 42,5
0
100
200
300
400
500
600
700
Ch1 Ex:330nm,Em:465nm
32,2
95 34,7
59 36,0
81
36,8
20
min
A
Inappropriate separation
33
Figure 17: Identification of BAs on different columns (new-Matís method); A) represents the Nucleosil-120-5 column, B) represents the Zorbax Eclipse Plus column
Table 17: Comparison of different columns used in the identification of BAs
Nucleosil‐120‐5 Agilent Zorbax Eclipse Plus
C18‐RP C18‐RP
Octanodecyl stationary phase Monolayer dimethyl N‐octadecyl silane stationer phase
Spherical silica support Porous silica support
pH 2‐8 stability pH 2‐9 stability
250x4 mm; 5µm 250x4,6mm; 5µm
The results showed that all four BAs were separated on both columns and the resolution of the
peaks were better on the Zorbax Eclipse Plus column (Figure 17).
For the identification of each peaks, elution order and retention times (RT) were determined.
Standard solutions in the concentration 100mg/kg on a product weight basis of each BAs were
injected separately on the Zorbax column and compared to the standard mixture of BAs with the same
concentration. The results from the identification study are shown in Figure 18. In Figure 18, A)
illustrates the chromatogram of the standard mixture, while B) illustrates the chromatogram of
tyramine, C) of cadaverine, D) of putrescine and E) of histamine. It was concluded that the elution
order is tyramine, putrescine, cadaverine and histamine, and their retention times are shown in Table
18.
30,0 32,5 35,0 37,5 40,0 42,5
0
250
500
750
1000
Ch1 Ex:330nm,Em:465nm
32,7
12
34,8
51
35,2
92
36,1
52
37,2
75
37,9
05
min
B
34
Figure 18: Identification the elution row and retention time (RT) of BAs by measuring standard dilutions at the concentration of 100mg/kg (on product weight basis); A): standard mixture of BAs, B): Tyramine standard, C): Putrescine standard, D): Cadaverine standard, E): Histamine standard solution
Table 18: Determined retention times of BAs on Zorbax Eclipse Plus (250x4,6mm; 5µm) with the flow rate 0,9 ml/min
Biogenic Amine Retention Time
(min)
Tyramine 32,1
Putrescine 35,7
Cadaverine 36,6
Histamine 37,4
30,0 32,5 35,0 37,5 40,0 min0
1000
2000Ch1 Ex:330nm,Em:465nm
32,1
44
35,6
09
36,6
77
37,4
25
30,0 32,5 35,0 37,5 40,0 min0
1000
2000Ch1 Ex:330nm,Em:465nm
32,1
25
30,0 32,5 35,0 37,5 40,0 min0
1000
2000Ch1 Ex:330nm,Em:465nm
35,6
65
30,0 32,5 35,0 37,5 40,0 min0
1000
2000Ch1 Ex:330nm,Em:465nm
36,6
54
30,0 32,5 35,0 37,5 40,0 min0
1000
2000Ch1 Ex:330nm,Em:465nm
37,4
50
A
B
C
D
E
35
A system suitability test was performed to investigate the repeatability/precision of the retention
times and areas of the BAs. For this test the mixture of standard solution was measured using the
same concentration level (100mg/kg on wet weight basis) in five replicates.
Standard deviation (SD) and relative standard deviation (RSD) of both area and retention time (RT)
were calculated according to equation (1) and equation (2).
SD=∑
equation (1)
Wherein is the average of the results obtained for RT/area, n= the number of the replicates
RSD= ∗ , equation (2)
Data obtained are summarized in Table 19.
Table 19: Results of the system suitability tests: SD and RSD of areas and retention times of BAs
Biogenic amine
Averages of Areas
SD of Area RSD of Area
(%)
Averages of RT
(min)
SD of RT
(min)
RSD of RT
(%)
Tyramine 14432712,8 1347845 9,34 32,31 0,33 1,02
Putrescine 36015789,8 241657 0,67 35,70 0,36 1,01
Cadaverine 31576063,2 1324896 4,20 36,75 0,32 0,86
Histamine 23747916,8 2014482 8,48 37,39 0,29 0,76
The results showed that the relative standard deviation for the RT of histamine was the lowest
(0,76%), while the results for the other three BAs were close to 1% and overall this repeatability of RT
is very good. The calculated RSD for the areas were different for each BAs and ranged from 0,67-
9,34%, the lowest RSD was obtained for putrescine and the largest for tyramine. In general RSD
values are acceptable if RSD is less than 10%.
Selectivity studies preferably include different type of comparisons such as comparing samples
analyzed by different methods. This was investigated by analyzing a sample of commercial fish meal
sample using three different methods for the quantification of BAs. The analytical methods used were
the post-column method (new method installed at Matís, here after called “new-Matís”), pre-column
method (previously used method at Matís, here after called “old-Matís”) and an accredited method
used by Nofima and performed in Norway. While the method description of Nofima is detailed in
section 0, then the post-column derivatization method described in section 3.3.2. Details of the pre-
column derivatization method can be found in section 3.3.3. The comparison of the gradient elution of
new-, and old-Matís method is shown in Figure 19.
36
Figure 19: Graphic comparison of gradient elution of pre (old-Matís), and post-column (new-Matís) derivatization
Commercial fish meal sample was analyzed in triplicate with new-Matís method and the average
concentrations were calculated. The results are presented in Figure 20. The comparison of old, new
Matís method and accredited method shows that comparable results were obtained for all three
methods. Concentration obtained from Nofima was higher in the case of putrescine, cadaverine and
histamine compared to the two other methods, and tyramine was not measured by this laboratory.
Results from new-Matís method gave lower concentrations in all cases of all four BAs, however, the
concentrations were closer to the concentrations obtained from accredited laboratory Nofima than to
the old-Matís method. The reason why the results from the new-Matís method and the method used
by Nofima are more comparable is probably because the Nofima laboratory also used OPA post-
column derivatization with 0,6M perchloric acid extraction and fluorescence detection (Ex: 365nm; Em:
418nm).
Figure 20: Comparison of concentration data in fish meal measured with old-Matís and new-Matís methods and obtained from accredited laboratory (Nofima)
0,24
0,575
0,897
0,107
0,64
0,91
0,290,21
0,54
0,74
0,18
0
0,2
0,4
0,6
0,8
1
Tyramine Putrescine Cadaverine Histamine
Concentration of BAs in m
g/kg
new Matís method NOFIMA old Matís method
37
An additional selectivity study was carried out, using quality control (QC) material received from
LAVES. This QC material originated from sardines (sample No. 1 and No. 2.) and was homogenized
and extracted by a trained technician at the LAVES laboratory in Cuxhaven, Germany, using the same
post-column derivatization method as installed at Matís (new-Matís).
Figure 21: Investigation of selectivity in sardine-matrices, A: sample No.1. and B: sample No.2.
At Matís, sardine samples were also analyzed in three replicates on three separate occasions. The
chromatograms from these analysis are shown in Figure 21 and show that all four BAs are present in
both samples and the elution row is the same as determined and presented in Figure 18. Same
samples were measured in both LAVES and Matís laboratories and then data were compared as
illustrated in Figure 22 and in Figure 23.
30,0 32,5 35,0 37,5 40,0 42,5 45,0 47,5 min0
500
1000
1500
2000 Ch1 Ex:330nm,Em:465nm
31
,83
2
35
,37
8
36
,43
7
37
,25
9
30,0 32,5 35,0 37,5 40,0 42,5 45,0 47,5 min0
500
1000
1500
2000 Ch1 Ex:330nm,Em:465nm
31
,83
3
35
,38
0
36
,43
9 37
,23
7
B
A
38
Figure 22: Comparison of the BAs concentrations (mg/kg) in sardine sample No. 1. measured in LAVES and Matís (new-Matís method) laboratories
Figure 23: Comparison of BAs concentration (mg/kg) in sardine sample No. 2. measured in LAVES and Matís (new-Matís method) laboratories
Results showed that LAVES and Matís laboratories detected all four BAs in similar concentrations
and method is selective for all BAs, no interference of other chemicals can be detected in the same
retention times in sardine matrix (see Figure 21).
4.3 Calculation of limit of detection (LOD) and limit of quantitation (LOQ)
As part of the verification study the limit of detection (LOD) and limit of quantitaion (LOQ) were
determined. While LOD states the lowest analyte concentration that can be detected in the sample,
then LOQ is the lowest concentration of the analyte can be quantified. For the determination of these
limits a blank sample containing perchloric acid was measured ten times. Since there was no peak
152
34
8 7
149
32,3
3,79 7,88
0
20
40
60
80
100
120
140
160
180
Histamine Cadaverine Putrescine Tyramine
BAs concentration in m
g/kg
Axis Title
LAVES Matís
162,00
71,00
5 6
147,98
71,14
6,09 11,32
0,00
20,00
40,00
60,00
80,00
100,00
120,00
140,00
160,00
180,00
Histamine Cadaverine Putrescine Tyramine
BAs concentration in m
g/kg
Axis Title
LAVES Matís
39
detectable in the blank samples at the same retention time of the BAs, thus these limits were
calculated from the baseline noise. The noise was determined by using the Labsolution software
program of the HPLC system. For the average of the baseline noise 543µV was calculated. From this
LOD was calculated as five times the baseline noise and LOQ as 10 times the baseline noise. LOD
was equal to 2,72mV and LOQ was equal to 5,43mV. To calculate the concentration values belonging
to the limits, signal value of histamine peak was used from the lowest measured concentration of
standard mixture of BAs (from calibration curve). This standard mixture contained 0,25µg/mL (2,5ppm)
histamine and the signal given by the histamine peak was 54mV. From this it can be seen that the
concentration of histamine that gives as much response as LOQ (5,43mV) is approximately
0,025µg/mL (0,25ppm). Thus the minimum concentration of histamine which can be quantified with an
acceptable precision and accuracy is 0,25ppm on product weight basis. On the other hand the
minimum concentration which can be reliably detected (LOD) is 0,125ppm on product weight basis.
With the same method LOD and LOQ were calculated for the other BAs and all values are presented
in Table 20 and values show that BAs can be detected even in a very low concentration.
Table 20: Limit of detection (LOD) and limit of quantitaion (LOQ) values of each BAs
BAs Signal (mV) LOD LOQ
Tyramine 28 0,024 µg/mL (0,24ppm) 0,048 µg/mL (0,48ppm)
Histamine 54 0,0125 µg/mL (0,125ppm) 0,025 µg/mL (0,25ppm)
Cadaverine 74 0,009 µg/mL (0,004ppm) 0,018 µg/mL (0,18ppm)
Putrescine 81 0,008 µg/mL (0,08ppm) 0,016 µg/mL (0,16ppm)
4.4 Linearity, working range
Linearity is the ability of the method to provide the test results that are proportional to the
concentration of the analyte in question [98]. The linear range of analytical methods are limited thus
calibration curves were applied and working range of these calibration curves were determined.
Calibrations were performed using standard stock solution containing all four BAs prepared as
described in section 3.3 in the concentrations of: 2,5; 5; 10; 20; 50; 100mg/kg on product weight basis.
Concentration points were equally distributed over the calibration range of interest. Calculation of the
regression line was carried out by the method of least squares. In the calculations, area was used
instead of peak height values because peak broadening is inevitable at higher concentrations. The
correlation coefficient (R2), y- intercept and slope of the regression line for each of the four BAs are
shown in Figure 24.
40
Figure 24: Calibration curve of the standards
The linearity was investigated both visually and from the correlation coefficients and regression
curves with correlation coefficient ≥ 0,0995 are considered to be linear. From this it is clear that curves
of all four BAs are linear (Figure 24 and Table 21). The calibration curves of putrescine, histamine
and tyramine have a high correlation coefficient i.e. higher than 0,994.
Table 21: Correlation coefficients of calibration curves of BAs
putrescine cadaverine histamine tyramine
0,9987 0,994 0,9986 0,9987
To determine at what concentration range the calibration curve starts to deviate from linearity the
correlation coefficients were investigated between individual calibration points to see when the
coefficients start to decrease.
Table 22 shows the coefficient values for histamine and the results show that the curve starts to
deviate from linearity at 100mg/kg concentration values, where R2 is equal to 0,093. Since
concentrations below 2,5mg/kg and over 100mg/kg were not tested it can be concluded that the linear
y = 361617x + 66958R² = 0,9987
y = 292408x - 21394R² = 0,9994
y = 199316x + 62149R² = 0,9986
y = 116669x + 204447R² = 0,9987
0
10000000
20000000
30000000
40000000
50000000
60000000
70000000
80000000
0 50 100 150 200 250
Are
a in
mA
bs
Concentration in mg/kg
Putrescine
Cadaverine
Histamine
Tyramine
41
portion of calibration curve is between 2,5-100mg/kg. This interval also represents the working range
of the new-Matis method. This deviation form linearity was also confirmed with visual inspection.
Table 22: Correlation coefficient (R2) values between calibration points to investigate linearity
Concentration range in mg/kg on product basis
R2
100-50 1
100-20 0,995
100-10 0,995
100-5 0,995
100-2,5 0,993
4.5 Accuracy
To investigate the accuracy of the method, the closeness between test results and”true” (accepted)
values were investigated; two components of accuracy: precision and trueness were determined as
illustrated in Figure 25.
Figure 25: Components of accuracy
Precision indicates the closeness of the test results to each other. In this study precision was
expressed as SD to evaluate repeatability and reproducibility. Trueness was determined as the
difference between measured and true (accepted) values. For the determination of all components of
accuracy a reference material with a certified amount of the analyte was used and this was obtained
from former proficiency testing, hereafter referred to as LVU RM.
4.5.1 Repeatability precision
According to Eurachem Guide repeatability is the smallest precision expected when the same analyst
executes the measurements with the same equipment over a short period of time [98]. Repeatability
was investigated in two different matrices: in fish flesh and in fish meal samples.
In repeatability studies both LVU RM (from former proficiency testing) and fish meal samples were
investigated, these measurements were carried out under the same experimental conditions. The LVU
RM was composed of fish flesh, six individual samples were prepared as described in section 3.2 and
all six samples were measured by the same analyst on the same day with the same HPLC system.
Standard deviation values for repeatability conditions (SDr) were calculated from the average
Accuracy
Precision
Repeatibility
Reproducibility
Trueness
42
concentrations using equation (1) (see section 4.2.). From these SDr values relative standard
deviation (RSD) was calculated according to equation (2) (see section 4.2.) and the results obtained
are shown in Table 23.
Table 23: SDr and RSD values of relevant BAs under repeatability conditions in fish flesh (LVU RM)
Tyramine Putrescine Cadaverine Histamine
Average (mg/kg) 130,5 99,58 298,3 53,45
Reference (LVU) (mg/kg) 124 112 289 60
SDr (mg/kg) 5,13 3,85 9,2 2,24
Repeatability RSD (%) 3,93 3,86 3 4,20
SDr values were the lowest in the case of histamine (2,24mg/kg) followed by putrescine (3,85mg
/kg), tyramine (5,13mg/kg) and cadaverine (9,2mg/kg). SDr values were low in the case of all four BAs
indicating that data points do not spread far from the mean, resulting a good precision. The low
percentages of RSD also indicate a low variability of measured data in the case of all four BAs. The
variability is lowest for cadaverine (3%) and highest for histamine (4,2%).
Table 24: Uncertainty (U), trueness (H) values of relevant BAs in fish flesh (LVU RM) under repeatability conditions
Tyramine Putrescine Cadaverine Histamine
Uncertainty (mg/kg) 5,3 4,08 9,7 2,6
Relative Uncertainty (%) 4 4 3 4
Trueness (bias) (mg/kg) 6,46 -12,42 8,7 -6,55
Uncertainty and relative uncertainty values were evaluated for this test according to equation (3)
and (4) and data obtained are presented in Table 24. Histamine showed the lowest measurement
uncertainty (2,6mg/kg). While, the deviation from the true value (uncertainty) was higher in the case of
putrescine (4,08mg/kg) and tyramine (5,3mg/kg) and was the highest in the case of cadaverine
(9,7mg/kg). Relative uncertainty was 4% for tyramine, putrescine and histamine and 3% in the case of
cadaverine and this relative uncertainty is considered to be very good.
The low uncertainty results indicate that the deviation from the (unknown) “true value” is very little
thus method is very reliable method repeatable.
Uncertainty: U= 1,96 equation (3)
Relative uncertainty: Urel = equation (4)
Furthermore, Table 24 also contains the data obtained from the calculation of, trueness calculated
with equations (5).
Trueness: H = -X ref equation (5)
Xref=reference mean (true value)
43
The trueness values indicated that the measurement bias was low. Thus, the difference between
the reference material (“true”) value and the values obtained in the current study was small. In the
case of histamine the measured concentration was 6,55mg/kg lower than the reference value. The
putrescine concentration analyzed was also lower, 12,42mg/kg lower than the concentration in the
reference material. On the other hand, the measured concentrations of the other two BAs, tyramine
and cadaverine was higher, 6,46mg/kg and 8,7mg/kg for tyramine and cadaverine, respectively,
compared to the reference value.
To investigate if results (Matís) have significantly differed from reference values (LVU) in the
repeatability study, Welsh’s test was carried out in all cases of BAs. For this standard deviation data of
former proficiency testing (2013), where LVU reference material was tested- were used. According to
the statistical test it is possible to compare individual measurements (here Matís results with reference
material from proficiency testing), with different number of repetitions (n=3, m=2) and evaluate whether
average values ( , ) of these different laboratories are significantly differ or not. Data used for the
calculations (such as averages and standard deviations of laboratories) and obtained results are
presented in Table 25. According to the Welch’s t- test first the t value and the degree of freedom (f)
were calculated according to equation (6) and (7) as presented below.
equation (6)
equation (7)
Where: is the average of results of Matís laboratory and is the average of former proficiency
testing
n is the number of measurements of Matis and equal to 3
m is the number of measurements of the former proficiency testing and equal to 2
Sx is the standard deviation of Matis triplicates
Sy is the standard deviation of duplicates of former proficiency testing (LVU)
44
Table 25: Welchs’ test to investigate differences between the mean values of results (Matís) and reference material (LVU) under repeatability conditions
Tyramine Putrescine Cadaverine Histamine
Average (mg/kg) ( ) 130,46 99,58 298,3 53,45
Reference from proficiency testing(mg/kg)( ) 124 112 289 60
SDr (mg/kg) Matis (Sx) 5,13 3,85 9,2 2,24
SD (mg/kg) of LVU proficiency testing (Sy) 12,9 12,4 50 11,9
Statistical ItI 0,6 1,9 0,3 0,7
Degree of freedom (f) 3 1 2 1
Theoretical t value at p=0,05 significance level 2,3 6,3 2,9 6,3
From results it is clear that there were no significant differences found between reference values
(LVU) and results of Matís on p=0,05 significance level because ItI< theoretical t value in all cases of
the four BAs. Which is considered to be excellent and shows that measurements were accurate.
In the second experiment, one fish meal sample (from local supplier) was prepared as described in
section 3.2 and measured ten times during the same day by the same analyst with the same HPLC
system. The results are presented in Table 26. Data obtained from the sequential measurements are
very close to each other as can be seen from the low SDr data: histamine (0,09 mg/kg), tyramine
(0,12mg/kg), putrescine (0,25mg/kg) and cadaverine (0,31mg/kg), which indicates an excellent
precision in all case of the four BAs (see Table 27)
Table 26: Overview of measurements repetition (10 times) for the investigation of repeatability in fish-meal
Number of measurements
Concentration in mg/kg
Tyramine Putrescine Cadaverine Histamine
1. 2,15 5,11 7,01 1,65
2. 2,08 5,40 7,47 1,76
3. 2,11 5,52 7,43 1,81
4. 2,21 5,79 7,82 1,87
5. 2,11 5,39 7,49 1,80
6. 2,01 5,45 7,32 1,76
7. 2,14 5,29 7,42 1,82
8. 2,10 5,50 7,57 1,81
9. 1,79 5,04 6,79 1,60
10. 2,22 5,80 7,74 1,88
min 1,79 5,04 6,79 1,60
max 2,22 5,80 7,82 1,88
45
Table 27: Standard deviation (SDr), relative standard deviation (RSD) and uncertainty (U, Ur) values calculated under repeatability conditions in fish meal.
Tyramine Putrescine Cadaverine Histamine
Average (mg/kg) 2,09 5,43 7,41 1,78
SDr (mg/kg) 0,12 0,25 0,31 0,09
RSD (%) 5,9 4,6 4,2 5,0
Uncertainty (mg/kg) 0,2 0,3 0,4 0,1
Rel. uncertainty (%) 8,30 6,38 5,86 7,06
From Table 27 it can be seen that the RSD values were low (between 4,2-5,9%) where the lowest
variability of measured data was found for cadaverine (4,2%), followed by putrescine (4,6%). Similar
data was obtained for histamine (5%) and tyramine (5,6%). Uncertainty values were very extremely
low, showing that the deviation from the true value was only 0,1mg/kg in the case of histamine;
0,2mg/kg in tyramine; 0,3mg/kg in putrescine and 0,4mg/kg in cadaverine. Relative uncertainty values
did not exceed 10%. The lowest RSD value was found for cadaverine (5,86%) and the highest for
tyramine (8,30%).
In both matrices, repeatability measurement showed that the method under same conditions is
repeatable,precise and data points cluster close around the mean, showed by the low SDr values. The
low RSD values indicate a low variability of data points. Further, the low uncertainty value represent
the low deviation of measured data from the true value.
4.5.2 Reproducibility precision
For the determination of intermediate precision, the LVU RM (fish flesh) was prepared as described in
section 3.2, while the analysis were carried out under different experimental conditions i.e. measured
during three consecutive days by different analysts using the same HPLC system. Standard deviation
for reproducibility conditions (SDR) were calculated from the average concentrations using equation
(1) as shown in section 4.2. From SDR values RSD% values were determined according to equation
(2) described in section 4.2 and results shown in Table 28.
Table 28: SDR and %RSD values of relevant BAs in fish flesh (LVU RM) investigated under reproducibility conditions
Tyramine Putrescine Cadaverine Histamine
Average (mg/kg) 136,1 119,3 300 52,8
Reference(median)(mg/kg) 124 112 289 60
SDR (mg/kg) 7,61 6,47 15 5
Reproducibility RSD (%) 5,6 5,4 5 9
The results reveal that the SDR values were low i.e. in the range of 5-15 mg/kg. The lowest SDR
was found for histamine (5 mg/kg). SDR values of putrescine (6,47mg/kg) and tyramine (7,61mg/kg)
46
were quite similar most likely due to their similar concentration. The highest SDR was found in the case
of cadaverine (15mg/kg) As expected lower concentrations yielded higher variation (RSD). Under
reproducibility conditions SDR and RSD data were higher than under repeatability conditions according
to expectations. These low SDR and RSD data indicate that the dispersion of measured data from the
mean are low and have a low variability under different experimental conditions, resulting a very good
and acceptable precision for the method.
Furthermore the SDR data was used to calculate uncertainty and relative uncertainty according to
equation (3) and (4) presented in section 4.5.1, and results obtained is shown in Table 29 below.
Table 29: Uncertainty (U), relative uncertainty (Ur) trueness (H) values in fish flesh (LVU RM) under reproducibility conditions
Tyramine Putrescine Cadaverine Histamine
Uncertainty (mg/kg) 10,7 9,1 15 3,1
Relative uncertainty (%) 7,8 7,6 5 5,8
Trueness (bias)(mg/kg) 12,1 7,3 11 -7,2
Histamine showed the lowest measurement uncertainty (3,1mg/kg). While, the deviation from the
true value (uncertainty) was higher in the case of putrescine (9,1mg/kg) tyramine (10,7mg/kg) and was
the highest in the case of cadaverine (15mg/kg). Relative uncertainty was in the range of 5-7,8% and
considered to be very good.
Trueness calculations (see Table 29) showed that measured concentration of histamine was
7,2mg/kg lower than the reference value. The highest difference (bias) was found in the case of
tyramine, where the measured concentration was 12,1mg/kg higher than the reference value.
Welchs’ test to investigate whether there were any significant differences between the
measurement of different laboratories was also carried out in all cases of BAs for the reproducibility
study, according to equation (6) and (7) (see section 4.5.1). Results are presented in Table 30.
Table 30: Welchs’ test to investigate differences between the mean values of results (Matís) and reference material (LVU) under reproducibility conditions
Tyramine Putrescine Cadaverine Histamine
Average (mg/kg) ( ) 136,1 119,3 300 52,8
Reference from proficiency testing(mg/kg)( ) 124 112 289 60
SDr (mg/kg) Matis (Sx) 7,61 6,47 15 5
SD (mg/kg) of LVU proficiency testing (Sy) 12,92 12,42 50 11,97
Statistical ItI 0,11 0,7 0,3 0,8
Degree of freedom (f) 2 1 1 1
Theoretical t value at p=0,05 significance level 2,9 6,3 6,3 6, 3
47
From results it is clear that there were no significant differences found between reference values
(LVU) and results of Matís on p=0,05 significance level because ItI< theoretical t value in all cases of
four Bas., Which is considered to be excellent and shows that measurements were accurate.
To investigate the acceptability of the test results obtained under repeatability and reproducibility
conditions repeatability limit (r) and reproducibility limit (R) were calculated from the obtained standard
deviation values (SDr, SDR). These limits are the comparison between two test results obtained under
repeatability and reproducibility conditions.
Table 31: Reproducibility (R) and repeatability (r) limits calculated in fish flesh (LVU RM)
Tyramine Putrescine Cadaverine Histamine
r (mg/kg) 14,7 10,7 36 8,7
R (mg/kg) 21,1 17,9 41,5 19,8
Repeatability limit (r) for all four BAs were calculated using the SDr values presented in Table 23
with the equation; r = t∞*√2*SD and the results are presented in Table 31. This means that the
difference between two sample results (replicas) measured the same day should not exceed the
repeatability limit (r) presented in Table 31 for the four BA determined in fish flesh, that is the
measured results for the different BAs will, with 95% probability, fall within the repeatability limit.
Reproducibility limit (R) on the other hand was calculated as R= t∞*√2*SDR (Table 31); wherein t∞ is
equal to 1,96 and represents the student factor at a 95% probability. This means that the difference
between two sample results (replicas) measured during different days and by different analysts should
not exceed the reproducibility limit (R) presented in Table 31 for BAs determined in fish flesh, that is
the measured results for the different BAs will, with 95% probability, fall within the reproducibility limit.
Results obtained are according to the expectations, between days measurements indicates a wider
variability limit.
Table 32 shows an overview of r and R limits in different matrices, such as lax, tuna and herring -
obtained from German laboratory (validated the official method [97]), compared to values of LVU
reference material (fish flesh) measured in Matís laboratory. Data show that R and r limit values differs
in a wide range in all cases of BAs in different matrices showing that these values are probably matrix
dependent.
48
Table 32: Overview of repeatability (r) and reproducibility (R) limit values in different matrices LVU fish meat homogenizate measured in Matís, compared to lax, tuna and herring matrices measured in the
laboratory, which carried out the official validation of the method
Lax Average r SDr R SDR
From official validation [97] All data in mg/kg
Cadaverine 296 26,2 9,3 51,7 18,3
Histamine 12 2,2 0,8 5,7 2,0
Putrescine 93 8,8 3,1 21,0 7,4
Tyramine 92 10,4 3,7 27,1 9,6
Tuna Average r SDr R SDR
From official validation All data in mg/kg
Cadaverine 13 3,6 1,3 9,9 3,5
Histamine 372 26,7 9,4 71,0 25,1
Herring Average r SDr R SDR
From official validation All data in mg/kg
Cadaverine 37 6,0 2,1 11,8 4,2
Histamine 18 3,4 1,2 8,8 3,1
Tyramine 167 17,9 6,3 25,5 9,0
LVU reference material Average r SDr R SDR
Measured at Matís All data in mg/kg
Cadaverine 300 36 9.2 41.5 15
Histamine 52,8 8,7 2,24 19,8 5
Putrescine 119,3 10,67 3,85 17,9 6,47
Tyramine 136,1 14,7 5,13 21,09 7,61
49
4.6 Robustness
Robustness was carried out to investigate how sensitive the method is to changes in the operating
condition and here the stability of the derivatization solution was investigated. For this investigation a
standard mixture of BAs was prepared as described in section 3.3 in the concentration of 50mg/kg (on
wet weight basis). The OPA-derivatization solution was prepared as described in section 3.3.1.and
used for sequential measurements of the standard mixture. Two measurements were carried out with
three day interval and the chromatograms obtained are illustrated in Figure 26.
Figure 26: Investigation of robustness: the influence of the age of the derivatization solution on detected BAs concentration; A) HPLC profile of the standard mixture of BAs derivatized with one day old OPA derivatization solution; B) HPLC profile of standard mixture of BAs derivatized using a three days old derivatization solution; Elution row of BAs are Tyr, Putr, Cad, His on Figure A and B
The concentration values of all four BAs were calculated on each occasion and the results obtained
are shown in Table 33.
0 10 20 30 40 50 min
0
250
500
750
1000
1250
1500
1750
Ch1 Ex:330nm,Em:465nm
31,22
7
34,484
35,645
36,308
0 10 20 30 40 50 min
-100
0
100
200
300
400
500
600
700
800
900
1000
1100
1200
1300 Ch1 Ex:330nm,Em:465nm
30,747
33,311
34,602
35,475
A
B
50
Table 33: Comparison of the derivatization ability of OPA derivatization solution over a three day period
Concentration of BAs in mg/kg
Tyramine Putrescine Cadaverine Histamine
1st day 0,59 0,60 0,60 0,57
2nd day 0,59 0,59 0,61 0,57
3rd day 0,40 0,41 0,39 0,39
The results show that there was a considerable difference between the two measurements as
approximately 32% lower concentration values were obtained on the 3rd day for all four BAs. These
results show that the measurements are influenced by the age of the derivatization solution and
therefore it is recommended that the solution should be used more than two days from preparation.
Concentration values obtained 2 days after preparation were very similar to the 1st day measurements
(Table 33). Further evaluations of robustness were not carried out as important analytical method
parameters such as pH and temperature were investigated as part of the official method validation
[97].
4.7 Analytical method efficiency (recovery)
The recovery study was carried out to investigate the efficiency of the method that is how effectively
the histamine was extracted from the matrix. For this investigation white fish (Icelandic cod) samples
were used, samples were prepared as detailed in section 3.2 and extracts were spiked with 1µg BAs
standard (0,001mg/mL). Spiked and unspiked samples were measured in triplicates and the
chromatograph for a spiked and unspiked cod samples are illustrated in Figure 27. The recovery was
calculated according to the equation Recovery%=
*100. Generally biological samples
contain a certain amount of the analyte of interest, therefore the unspiked sample concentration
(Cunspiked) has to be withdrawn from spiked concentration. The denominator contains the known
amount used of BAs for spiking.
35,0 37,5 40,0 min
0
50
100
150
200
250
300
350
400
450
500
550 Ch1 Ex:330nm,Em:465nm
38
,33
5
38
,99
9
A
51
Figure 27: Recovery study A: unspiked Icelandic cod sample, B: Icelandic cod sample spiked with 1µg histamine standard
Although this kind of determination of efficiency is recognized and accepted practice in chemical
analysis, it could provide unrealistic values because the artificially added analyte may not attach as
strongly to matrix as the natural analyte. The result obtained are shown in Table 34. Generally
acceptable recoveries are between 80-120%.
Table 34: Recovery values in % in spiked white fish (Icelandic cod)
Recovery in %
Sample matrices Tyramine Putrescine Cadaverine Histamine
Spiked Icelandic cod (fish flesh) 90 80 100 110
4.8 Internal quality control (IQC)
For internal quality control, control chart was established as a quality assurance tool (QA) as
represented in Figure 28, to be able to monitor whether the measurement performance is correct over
time. For this performance evaluation the LVU reference material was chosen. Since the focus was
mainly on histamine because of the EU legislations, thus histamine was used for monitoring by
creating the control chart. The histamine concentration was measured 20 times. For average
55,5mg/kg (wet weight basis) was calculated and 2,27mg/kg for SD. From the average and standard
deviation of measured data upper and lower action limits (UAL, LAL, respectively) for histamine were
determined according to the equation: ±3SD. Furthermore upper and lower warning limits (UWL,
UAL) were also calculated as ±2SD.
35,0 37,5 40,0 min
0
50
100
150
200
250
300
350
400
450
500
550 Ch1 Ex:330nm,Em:465nm
33
,94
2
38
,91
9
39
,54
2
39
,95
0
B
52
Figure 28: Control chart for internal quality control; monitoring measurement performance with the measurement of histamine concentration of LVU RM
Results show that upper and lower action limits for histamine were UAL=61,3mg/kg and
LAL=48,7mg/kg, which indicates that based on the normal distribution 99,7% of expected future
results will fall into the interval between 48,7-61,3mg/kg. The interval between the lower (LWL) and
upper warning limits (UWL) was equal to 50,8-59,2mg/kg and this interval is therefore expected to
contain the 95,5% of the future results for histamine. If concentration values of histamine measured in
the reference material will fall outside of the interval: 48,7-61,3mg/kg, or two successive values fall
outside 50,8-59,2mg/kg but remains inside the action limits, performance of measurement has to be
investigated.
4.9 Participation in proficiency testing
To evaluate the performance and the reliability of the measurements of BAs using the new analytical
method that had been set up and verified Matís took part in AQUACULTURE Ring test 2014
(http://masterlab.nl/), which is an inter laboratory proficiency testing that was established with the
standard ISO/IEC 17043 and organized by Masterlab analytical services. In this ring test different
laboratories compared their results, using their own analytical methods. These method can be slightly
different, but should be comparable and suitable for the measurement of the histamine content of the
reference material.
53
Figure 29: A) Proficiency testing: Measurement of histamine in fish flesh B) Proficiency testing:
Measurement of histamine in fish meal
Fish flesh and fish meal reference material was obtained from Masterlab analytical services.
Sample preparation was carried out as presented in section 3.2 and histamine was measured in both
matrices an example of the chromatographs obtained in the proficiency testing are shown in Figure
29. Samples were measured in duplicates in the same day and data obtained are shown in Table 35.
30,0 32,5 35,0 37,5 min0
250
500
750
1000
1250
1500
1750 Ch1 Ex:330nm,Em:465nm
31,9
06
34
,834
35
,44
2
36
,495
37,2
70
30,0 32,5 35,0 37,5 min0
250
500
750
1000
1250
1500
1750 Ch1 Ex:330nm,Em:465nm
31,0
19
31,8
94
33,8
59
35
,437
36,4
88
37
,264
37,8
63
B
A
54
Table 35: Data obtained from proficiency testing
Fish flesh Fish meal
Number of participants 7 11
Intra lab mean (Matís) (mg/kg) 450 140
Inter lab mean (Masterlab) (mg/kg) 430,78 128,74
Trueness (bias) (mg/kg) 19,22 11,26
Calculated Z score 0,7 0,2
SD of the ring test (mg/kg) 28,92 26,86
RSD (Rel. Reproducibility) % 6,71 20,05
Laboratory bias was calculated as the difference between the inter-lab and intra-lab mean values.
Bias was determined to be 11,26mg/kg in fish meal and 19,22mg/kg in fish flesh, therefore the bias
was higher for the fish flesh than in the fish meal. Results of Matís were considered good both in fish
flesh and in meal because results fall into the interval appointed by the interlab mean±SD of the ring
test. That is in the case of fish flesh measured 450 mg/kg falls into the interval of 401,86-478,92mg/kg
(430±28,92mg/kg) ; and in fish meal measured 140 mg/kg falls into the interval of 113,14-168,74mg/kg
(128±26,86mg/kg), thus there was no big difference between results of Matís and reference (intel ab
mean).
Mean value was calculated from participating laboratory-means values (robust-average), as well as
the Z-score. The Z score represents the laboratory performance compared to other participants and
was calculated as Z =
, where Xintralab is the result reported by a participant and Xinterlab
is the mean of all laboratories value. SD represents the standard deviation of the ring test.
Z-scores of each participants are illustrated on Figure 30 and Figure 31 and show that in both
matrices Matís laboratory showed a good laboratory performance and fitted into Z-score criteria as the
measured mean values fitted within the 2Z interval. If Z score would have fallen in the interval
between 2 and 3, the analysis should have been questionable and Z score over 3 would have
indicated a warning action according to the criteria of Z score and the analyses should be repeated if
no valid reason is found for the deviation.
55
Figure 30: Overview of Z scores of different laboratories from inter-laboratory ring test in fish flesh. Numbers are representing the participating laboratories. Matís laboratory is marked as no.3 (Z-score=0,7),
the Z-score for laboratory no.1 and 2 was determined as 0.
Figure 31: Overview of Z-scores of different laboratories from inter-laboratory ring test in fish meal. Numbers are representing laboratories, Matís laboratory is marked as no.3. (Z-score=0,2) For laboratory
no.6 the Z-score was determined to be equal to 0.
The obtained Z score values for Matís laboratory In fish flesh (0,7) and in fish meal (0,2) showed
that the deviation of the results from the reference “true” value compared with the ring test SD was
satisfactory as compared to the analyses performed by laboratories.
SD values of the ring test in both matrices was similarly quite high, RSD value was higher in fish
meal similarly as it was obtained in reproducibility study (see section 4.5.2.) The largest difference is
likely due to the fact that the fish meal reference material contained a lower amount of histamine thus,
resulting a much higher RSD value, or it was resulted by the different methods used by the
participants.
‐4
‐3
‐2
‐1
0
1
2
3
4
5
1 2 3 4 5 6 7
Z‐score
participants
‐4
‐3
‐2
‐1
0
1
2
3
4
5
6
1 2 3 4 5 6 7 8 9 10 11
Z‐score
participants
56
4.10 Inter-laboratory uncertainty
It is very necessary to estimate the measurement uncertainty of the analytical method to assess the
quality of the measurements carried out in the laboratory. This estimation is also important for the
laboratory customer, i.e. the fish producer, to be able to demonstrate that the goods produced and
distributed by the company are compliant with the legal concentration limits in respect to the
investigated analyte.
Uncertainty indicates how big the measurement error can be and characterizes the dispersion of
the results. The basis of this evaluation is statistical and uncertainty can originate from different
sources. In the first step of the evaluation both method and laboratory bias were taken into account.
Values were used from the AQUCULTURE proficiency testing (see section 4.9) and were investigated
both for the fish flesh and the fish meal. The calculations followed the instructions of Nordtest report
[99]. During the calculations bias and uncertainty between and within laboratory were taken into
account as presented below in the case of fish flesh.
Bias was calculated as: Xintralab-Xref=450-430,78=19,22mg/kg
1. Quantification of Method and Laboratory bias
RMSbias= = ,
= 5,13%
(n=14 because 7 laboratory measured duplicates)
U(cert)= √
=,
√ =5,43
RSD between laboratory SDR
2. Calculate standard uncertainty of bias
Ubias=√ =5,44%
3. Calculate combined standard uncertainty, where SDw is the intra-laboratory SD equal to
2,27mg/kg from control chart see section 4.8
Uc=√ = 5,88%
4. Expanded uncertainty
U=2*Uc=2*5,9%=11,8%
For the method 11,8% expanded uncertainty was obtained in fish flesh, which is considered very
good. Expanded uncertainty was also calculated for fish meal, using the same bias value (5,44%)-
used in the case of fish meal, calculated from the proficiency testing- and intra laboratory SDr
(5mg/kg), measured in the repeatability study see section 4.5.1. Expanded uncertainty for fish meal
was 7,4%, considered to be very good.
57
Conclusion 5.
The implementation of the new analytical method and the verification study was successful and proved
that HPLC system and method served its purposes. HPLC system passed operational qualification
tests and its performance characteristics were stable and suitable for the measurements. Verification
parameters were defined and the results confirmed the validity of BA measurement in fish and fish
products. The applied method was relatively simple to carry out, selective, accurate, sensitive,
repeatable, reproducible and robust for the quantification of histamine, cadaverine, tyramine and
putrescine.
The application of the new-Matís method (post-column) was much easier to carry out as compared
to the pre-column derivatization method formerly used by Matís. This was mainly due to the
application of post-column derivatization, which shortened the time for sample preparation, resulted in
stable derivatives of the analytes and enabled the possibility of continuous measurements applying
autosampler and online derivatization thus shortening the time for each analysis and improving
efficiency and sample turnover rate of the laboratory. Matrix effects were not investigated in this study
as the formal validation of the analytical method confirmed the applicability of the method for fish and
fish based product and it is only intended for these matrices.
59
Future aspects 6.
Due to the results of the thesis, Matís laboratory will be the first laboratory in Iceland that has applied
for accreditation for BAs analyses in fish and fish products. All the essential SOP have been
implemented in the Matís quality handbook and the method has recently been audited by an external
accreditation body. This will enable official authorities and seafood producers to monitor the
occurrence of BAs in Icelandic products and lead to increased food and feed safety. Furthermore,
Iceland will be able to comply to the EU legislation (EC directive No 2073/2005) and Icelandic
regulations and enable seafood producers to confirm the quality and safety of their products.
Since the European and Icelandic food legislations require measurements of histamine in fish and
fish based products the focus of this thesis was on histamine measurements. Future work should
focus on cadaverine analysis in fish meal as it is an important quality indicator of the meal and the
concentration will determine the price of the product. Therefore, the laboratory should participate in
further proficiency testing focusing on cadaverine measurements in fish meal.
61
References 7.
[1] Stansby M. Industrial Fishery Technology. London: Reinhold Publishing Corporation; 1963;41
[2] Favati F. Food and Agriculture Organization of the United Nations; Fisheries and Aquaculture Department:Nutritional elements of fish. Retrieved from: http://www.fao.org/fishery/topic/12319/en (15.07.2015)
[3] Sakaguchi M. More efficient utilization of Fish and Fisheries products; Proceedings of the International Symposium on the occasion of the 70th anniversary of the Japanese Society of Fisheries Science: Elsevier; 2004.
[4]. Retrieved from:http://www.laeknabladid.is/2005/03/nr/1905 (01.02.2015) Krílfiskieitrun á íslenskum veitingastað Læknablaðið 2005;91:251‐3.
[5] Valdimarson G, Cormier, R., Ababouch, L. Fish safety and quality in times of globalization (The fishery Industries Division. FAO. Rome. Italy First Joint Trans‐Atlantic Fisheries Technology Conference‐TAFT 2003, 33rd WEFTA and 48th AFTC meetings, Reykjavík‐Iceland; 11‐14 June 2003
[6] Tapingkae W, Tanasupawat S, Parkin KL, Benjakul S, Visessanguan W. Degradation of histamine by extremely halophilic archaea isolated from high salt‐fermented fishery products. Enz Microbial Tech 2010;46:92‐9.
[7] Zhai HL, Yang XQ, Li LH, Xia GB, Cen JW, Huang H, et al. Biogenic amines in commercial fish and fish products sold in southern China. Food Control 2012;25:303‐8.
[8] Lorenzo JM MS, Franco I, Carballo J. Biogenic amine content during the manufacture of dry‐cured lacón, a Spanish traditional meat product: Effect of some additives. Meat Sci ; 2007;77(2):287–93.
[9] Spano G RP, Lonvaud‐Funel A, Lucas P, Alexandre H, Grandvalet C, Coton E, Coton M, Barnavon L, Bach B, Rattray F, Bunte A, Magni C, Ladero V, Alvarez M, Fernández M LP, de Palencia PF, Corbi A, Trip H, Lolkema JS. Biogenic amines in fermented foods. Eur J Clin Nutr 2010 Nov;64 Suppl 3:S95‐100. doi: 10.1038/ejcn.2010.218.
[10] Silla Santos MH. Biogenic amines: their importance in foods. Int J Food Microbiol 1996;29:213‐31.
[11] Onal A. A review: Current analytical methods for the determination of biogenic amines in foods. Food Chem 2007;103:1475‐86.
[12] Bouchereau A, Guenot P, Lather F. Analysis of amines in plant materials. J Chrom B 2000;747:49‐67.
62
[13] Jansen SC, van Dusseldorp M, Bottema KC, Dubois AEJ. Intolerance to dietary biogenic amines: a review. Ann Allerg Asthma Immun 2003;91:233‐41.
[14] Suzzi G, Gardini F. Biogenic amines in dry fermented sausages: a review. Int J Food Microbiol 2003;88:41‐54.
[15] Moret S, Smela D, Populin T, Conte LS. A survey on free biogenic amine content of fresh and preserved vegetables. Food Chem 2005;89:355‐61.
[16] Naila A, Flint S, Fletcher G, Bremer P, Meerdink G. Control of Biogenic Amines in Food‐Existing and Emerging Approaches. J Food Sci 2010;75:R139‐R50.
[17] Bodmer S, Imark, C., Kneubuhl, M. Biogenic amines in foods: Histamine and food processing. Infl Res 1999;48:296‐300.
[18] Straub B, Schollenberger, M., Kicherer, M., Luckas, B., Hammes, W. P. Extraction and determination of biogenic‐amines in fermented sausages and other meat‐products using reversed‐phase‐HPLC. Z. Lebensm‐Unters‐Forsch 1993;197:230‐2.
[19] Vinci G, Antonelli, M. L. Biogenic amines: quality index of freshness in red and white meat. Food Control 2002;13:519‐24.
[20] Halász A. BA, Simon‐Sarkadi L., Holzapfel W. Biogenic amines and their production by microorganisms in food Trends Food Sci Tech 1994;5:42‐9.
[21] Bover‐Cid S, Miguelez‐Arrizado MJ, Moratalla LLL, Carou MCV. Freezing of meat raw materials affects tyramine and diamine accumulation in spontaneously fermented sausages. Meat Sci 2006;72:62‐8.
[22] Latorre‐Moratalla ML, Bover‐Cid S, Talon R, Garriga M, Zanardi E, Ianieri A, et al. Strategies to reduce biogenic amine accumulation in traditional sausage manufacturing. Lwt‐Food Sci Tech 2010;43:20‐5.
[23] Leuschner RG, Heidel, M. and Hammes, W. P. Histamine and tyramine degradation by food fermenting microorganisms. Int J Food Microbiol 1998;39(1‐2):992‐6.
[24] Cid SB, Miguelez‐Arrizado MJ, Becker B, Holzapfel WH, Vidal‐Carou MC. Amino acid decarboxylation by Lactobacillus curvatus CTC273 affected by the pH and glucose availability. Food Microbiol 2008;25:269‐77.
[25] Gardini F, Martuscelli M, Caruso MC, Galgano F, Crudele MA, Favati F, et al. Effects of pH, temperature and NaCl concentration on the growth kinetics, proteolytic activity and biogenic amine production of Enterococcus faecalis. Int J Food Microbiol 2001;64:105‐17.
63
[26] Greif G, Greifova M, Karovicova J. Effects of NaCl concentration and initial pH value on biogenic amine formation dynamics by Enterobacter spp. bacteria in model conditions. J Food Nutr Res 2006;45:21‐9.
[27] Karovicova J, Kohajdova Z. Biogenic amines in food. Chem Pap 2005;59:70‐9.
[28] Bover‐Cid S, Hernandez‐Jover T, Miguelez‐Arrizado MJ, Vidal‐Carou MC. Contribution of contaminant enterobacteria and lactic acid bacteria to biogenic amine accumulation in spontaneous fermentation of pork sausages. Eur Food Res Tech 2003;216:477‐82.
[29] Geornaras I, Dykes GA, Vonholy A. Biogenic‐amine formation by poultry‐associated spoilage and pathogenic bacteria. Lett Appl Microbiol 1995;21:164‐6.
[30] Ozogul F, Ozogul Y. Biogenic amine content and biogenic amine quality indices of sardines (Sardina pilchardus) stored in modified atmosphere packaging and vacuum packaging. Food Chem 2006;99:574‐8.
[31] Maijala R, Nurmi E, Fischer A. Influence of processing temperature on the formation of biogenic‐aminesin dry sausages. Meat Sci 1995;39:9‐22.
[32] ten Brink B, damink, C. Joosten, H. M. L. J. and Huis in 't Veld, J. H. J. Occurence and formation of biologically active amines in foods. Int J Food Microbiol 1990;11 (1):73‐84.
[33] Dean O. Cliver HR. Foodborne diseases. 2nd edition ed: Gulf Professional Publishing; 2002.
[34] Du WX, Lin CM, Phu AT, Cornell JA, Marshall MR, Wei CI. Development of biogenic amines in yellowfin tuna (Thunnus albacares): Effect of storage and correlation with decarboxylase‐positive bacterial flora. J Food Sci 2002;67:292‐301.
[35] Lehane L, Olley J. Histamine fish poisoning revisited. International Journal of Food Microbiology 2000;58:1‐37.
[36] Etkind P, Wilson ME, Gallagher K, Cournoyer J. Bluefish‐associated scromboid poisoning ‐An example of the expanding spectrum of food poisoning from seafood. Jama‐J Am Med Assoc 1987;258:3409‐10.
[37] Kim JH, Ahn HJ, Jo C, Park HJ, Chung YJ, Byun MW. Radiolysis of biogenic amines in model system by gamma irradiation. Food Control 2004;15:405‐8.
[38] Mbarki R, Sadok, S. and Barkallah, I. Influence of Gamma Irradiation on Microbiological, Biochemical, and Textural Properties of Bonito (Sarda sarda) During chilled Storege. Food Sci Tech Int 2008;14(4):367‐73.
[39] Chong CY, Abu Bakar, F., Russly, A. R., Jamilah, B. and Mahyudin, N. A. The effects of food processing on biogenic amines formation. International Food Res J 2011;18(3):867‐76.
64
[40] Alak G, Hisar SA, Hisar O, Genccelep H. Biogenic amines formation in Atlantic bonito (Sarda sarda) fillets packaged with modified atmosphere and vacuum, wrapped in chitosan and cling film at 4 degrees C. Eur Food Res Tech 2010;232:23‐8. [41] Eerola S, Sagues AXR, Lilleberg L, Aalto H. Biogenic amines in dry sausages during shelf‐life storage. Z. Lebensm‐Unters‐Forsch‐Food Res Tech 1997;205:351‐5.
[42] Ruiz‐Capillas C, Jimenez‐Colmenero F. Biogenic amines in meat and meat products. Crit Rev Food Sci Nutr 2004;44:489‐99.
[43] Ladero V, Calles‐Enriquez, M., Fernandez, M., Alvarez, M. A. Toxicological effects of dietary biogenic amines. Curr Nutr Food Sci 2010;6:145‐56.
[44] Shalaby AR. Significance of biogenic amines to food safety and human health. Food Res Int 1996;29(7):675‐90.
[45] Stratton JE, Hutkins RW, Taylor SL. Biogenic amines in cheese and other fermented foods‐ A review. J Food Prot 1991;54:460‐70.
[46] 1019/2013 CREN. amending Annex I to Regulation (EC) No 2073/2005 as rregards histamine in fishery products (text with EEA relevance). 13 October 2013.
[47] Retrieved from:(10):2393 EJ. http://www.efsa.europa.eu/de/efsajournal/pub/2393.htm) (19.05.2015)
[48] Stute R, Petridis K, Steinhart H, Biernoth G. Biogenic amines in fish and soy sauces. Eur Food Res Techy 2002;215:101‐7.
[49] Dale H. H. LPP. The physiological action of β‐iminazolyl‐ethylamine. J Phys 31. dec. 1910;41:318‐44.
[50] Maintz L. NN. Histamine and histamine intolerance 1' 2' 3'. Am J Clin Nutr 2007;85:1185‐96.
[51] Bartholomew BA, Berry PR, Rodhouse JC, Gilbert RJ, Murray CK. Scombrotoxic fish poisoning in Britain: features of over 250 suspected incidents from 1976 to 1986. Epidemiol Infect 1987;99:775‐82.
[52] Hernandez‐Jover T. I‐PM, Vecina‐Nogués M. T., Vidal‐Crou M. C. Biogenic amine and polyamine contents in meat and meat products. J Agr Food Chem 1997;45:2098‐102.
[53] Emborg J, Dalgaard P. Formation of histamine and biogenic amines in cold‐smoked tuna: An investigation of psychrotolerant bacteria from samples implicated in cases of histamine fish poisoning. J Food Prot 2006;69:897‐906.
[54] Lionte C. Un unusual case of hypertension and abnormal electrocardiogram (ECG)‐scromboid poisoning. Cent Eur J Med 2010;5(3):292‐7.
65
[55] Chen HC, Kung HF, Chen WC, Lin WF, Hwang DF, Lee YC, et al. Determination of histamine and histamine‐forming bacteria in tuna dumpling implicated in a food‐borne poisoning. Food Chem 2010;106:612‐8. [56] Lavon O LY, Bentur Y. Scrombroid fish poisoning in Israel, 2005‐2007. Israel Med Assoc J 2008;10:789‐92.
[57] Emborg J, Laursen BG, Dalgaard P. Significant histamine formation in tuna (Thunnus albacares) at 2 degrees C‐effect of vacuum‐and modified atmosphere‐packaging on psychrotolerant bacteria. Int J Food Microbiol 2005;101:263‐79.
[58] Auerswald L, Morren C, Lopata AL. Histamine levels in seventeen species of fresh and processed South African seafood. Food Chem 2006;98:231‐9.
[59] Tsai Y‐H, Hsieh H‐S, Chen H‐C, Cheng S‐H, Chai T‐j, Hwang D‐F. Histamine level and species identification of billfish meats implicated in two food‐borne poisonings. Food Chem 2007;104:1366‐71.
[60] Dalgaard P, Madsen HL, Samieian N, Emborg J. Biogenic amine formation and microbial spoilage in chilled garfish (Belone belone belone)‐effect of modified atmosphere packaging and previous frozen storage. J Appl Microbiol 2006;101:80‐95.
[61] Olsen SJ, Mackinon, L. C., Goulding, J. S., Bean, N. H. and Slutsker, L. Surveillance for foodborne diasease outbreaks, United States, 1993‐1997. NMWR CDC Surveill Sum 2000;49:1‐62. [62] Scoging AC. Scrombotoxic (histamine) fish poisoning in the United Kingdom. Commun Dis Public Health 1998;1:204‐5.
[63] Molinari G, Montagnoli G, Pellegrini G, Caroli G. Hygiene and health importance of histamine as an unhealthy factor in several food products. Annali di igiene : medicina preventiva e di comunita 1989;1:637‐46.
[64] Taylor SL. Histamine food poisoning: toxycology and clinical aspects. Crit Rev Toxicol 1986;17:91‐128.
[65] Kent WaL. Poison Management Manual, 4th Edition: BC Drug and Poison Information Centre; 1997.
[66]Retrieved from: http://histame.com/histamine‐rich‐foods‐substances (25.08.2015)
[67] (FDA) Fish and Fischery products Hazard and Controls Guidance, 4th Edition Washington DC: Department of Health and Human Services, Food and Drug Administration, Centre for Food Safety and Applied Nutrition. 2011.
[68] Joint FAO/WHO expert meeting on the Public health and other Biogenic Amines from fish and Fishery Products (meeting report), FAO Headquarters, Rome Italy, 23‐27 July 2012
66
[69] Krizek M, Pavlicek T, Vacha F. Formation of selected biogenic amines in carp meat. J Sci Food Agr 2002;82:1088‐93. [70] Rezaei M, Montazeri N, Langrudi HE, Mokhayer B, Parviz M, Nazarinia A. The biogenic amines and bacterial changes of farmed rainbow trout (Oncorhynchus mykiss) stored in ice. Food Chem 2007;103:150‐4.
[71] Yongsawatdigul J, Choi YJ, Udomporn S. Biogenic amines formation in fish sauce prepared from fresh and temperature‐abused Indian anchovy (Stolephorus indicus). J Food Sci 2004;69:C312‐C9.
[72] Ozogul F, Taylor KDA, Quantick P, Ozogul Y. Changes in biogenic amines in herring stored under modified atmosphere and vacuum pack. J Food Sci 2002;67:2497‐501.
[73] Prester L. Biogenic amines in fish, fish products and shellfish: a review. Food Additives and Contaminants Part a‐Chem Anal Contr Exp Risk Ass 2011;28:1547‐60.
[74] Al Bulushi I PS, Deeth HC, Dykes GA. Biogenic amines in fish: roles in intoxication, spoilage, and nitrosamine formation‐‐a review. Crit Rev Food Sci Nutr 2009 Apr;49(4):369‐77.
[75] Efienne M. I, Nantes. SEAFOODplus‐Traceability‐Valid‐Methods for chmical quality assessment‐Methodology for histamine and biogenic amines analysis. 2006, Febr.
[76] Retrieved from: http://lab‐training.com/wp‐content/uploads/2011/12/reverse‐phase‐HPLC‐separation.jp. (20.04.2015)
[77] Sagratini G, Fernandez‐Franzon M, De Berardinis F, Font G, Vittori S, Manes J. Simultaneous determination of eight underivatised biogenic amines in fish by solid phase extraction and liquid chromatography‐tandem mass spectrometry. Food Chem 2012;132:537‐43.
[78] Ito S, Nakata F. Simultaneous and Direct Analysis of Biogenic Amines in Food by LC‐MS‐MS Using Hydrophilic Chromatography. Lc Gc North Am 2011:36‐7.
[79] Loukou Z, Zotou A. Determination of biogenic amines as dansyl derivatives in alcoholic beverages by high‐performance liquid chromatography with fluorimetric detection and characterization of the dansylated amines by liquid chromatography‐atmospheric pressure chemical ionization mass spectrometry. J Chrom A 2003;996:103‐13.
[80] Skoog DA, HollerF. J., Nieman, T. A. Principles of instrumental analsis. USA: Harcourt Brace & Company; 1998.
[81] Roth M. Fluorescence reaction for amino acids. Anal Chem 1971;43:880‐&.
[82] Wei CI, Chen CM, Koburger JA, Otwell WS, Marshall MR. Bacterial growth and histamine production on vacuum packed Tuna. J Food Sci 1990;55:59‐63.
67
[83] Seiler N, Knodgen B. Determination of polyamines and related‐compounds by Reversed‐phase High‐Performance Liquid Chromatography. J Chrom 1985;339:45‐57.
[84] Beljaars PR, Van Dijk R, Jonker KM, Schout LJ. Liquid chromatographic determination of histamine in fish, sauerkraut, and wine: Interlaboratory study. J AOAC Int 1998;81:991‐8.
[85] Moret JaC, L. C. J Chrom 1996;729:369.
[86] Veciana‐Nogues MT. J AOAC International 1995;78:1050.
[87] Gingerich TM. J Food Prot 1999;62:1037.
[88] Lebiedzinska A, Lamparczyk, H., Ganowiak, Z., Eller, K. I. Differences in biogenic amine patterns in fish obtained from commercial sources. Z. Lebensm‐Unters‐Forsch 1991;192:240‐3.
[89] Rosier J, Van Peteghem, C., Z. Z Lebensm‐Unters‐Forsch‐Food Res Tech 1988;186:28.
[90] Mehta A. Principle of Reverse‐Phase Chromatography HPLC/Uplc.2012. Retrieved from: http://pharmaxchange.info/press/2012/12/principle‐of‐reversed‐phase‐chromatography‐hplcuplc‐with‐animation/(06.02.2015)
[91] AOAC Guidelines for Single Laboratory Validation of Chemical Methods for Dietary supplements and Botanicals. Retrieved from:http//www.aoacorg (10.03.2015)
[92] IUPAC. International Union of Pure and Applied Chemistry Harmonized Guideliness for International Quality Control in Analytical Chemistry Laboratories (Technical Report). Pure Appl Chem;67:649‐66.
[93] Retrieved from:http://www.labcompliance.com/tutorial/iso17025/default.aspx?sm=d_e (17.05.2015)
[94] (CDER) Center of Drug Evaluation and Research: Reveiwer Guidance, Validation of chromatographic Methods. Nov. 1994.:1‐2.
[95] LVU: Laborvergleichsuntersuchung BA. Biogene Amine und fluchtiger Basenstickstoff (2013) http://www.lvus.de/)
[96] Quality handbook of Matís Laboratory (AE‐9).
[97] §35 LMBG L 10.00‐5, HPLC (1999‐11) BVL L 10.00‐5 Untersuchung von Lebensmitteln ‐ Bestimmung des Gehaltes an biogenen Aminen in Fischen und Fischerzeugnissen‐Hochdruckflüssigkeitschromatographische Bestimmung; Referenzverfahren. 1999
68
[98] EURACHEM Guide: The fitness for purpose of Analytical Methods; A Laboratory Guide to Method Validation and Related Topics. 1998.
[99] Magnusson B. HH, Krysell M. Handbook for calculation of measurement uncertainty in environmental laboratories. 2004.
69
Appendix 8.
Table 36: Operation qualification tests and their acceptances
Solvent delivery system: LC-30AD
(pump)
Test
items/parameters
proced
ure
User limit (acceptance criteria)
Pumping stability
test
LC-30AD
3.5.4. Pressure fluctuation width≤0,15 MPa
Flow rate accuracy
test
LC-30AD
3.5.6 Measured flow rate
(setting flow rate 1.0mL/min)=1.00±0.02mL/min
Pump leak sensor
test
LC-30AD
3.5.8 [SENSOR GOOD] appears on the screen
Autosampler: SIL-30AC
Test
items/parameters
proced
ure
User limit (acceptance criteria)
Injection volume
accuracy test
SIL-30AC
3.6.5. Injection volume accuracy is less than ±2.0
If theck criteria is not passed, refer to the instruction manual
of SIL-30AC“ [ASP FACTOR]“ to calibrate
Leak sensor test
SIL-30AC
3.6.6. [SENSOR GOOD] appears on the screen
Temperature
accuracy test
SIL-30AC
3.6.7. Set temp. Accuracy: 4°C±3.0°C
Column oven: CTO-20A/20AC
70
Test
items/parameters
proced
ure
User limit (acceptance criteria)
Setting temp.
Accuracy and
regulation precision
CTO-20A/20AC
3.8.4 Temp. accuracy is:dT‹0.2°C.
[ACCURACY GOOD] appears
Leak sensor test
CTO-20A/20AC
3.8.5. [SENSOR GOOD] appears
Spectrometric detector: RF-20A/20AXS
Test
items/parameters
proced
ure
User limit (acceptance criteria)
Light source usage
time check
RF-20A/20AXS
3.10.4 D2 (deuterium) lamp maximum usage time: within 2,000
hours
Wavelength
accuracy check
RF-20A/20AXS
3.10.5. 254nm wavelength accuracy within:±1nm
656nm wavelength accuracy within:±1nm
Lamp intesity
check
RF-20A/20AXS
3.10.6. Reference intensity at 220nm≥400
Linearity check
RF-20A/20AXS
3.10.7. Deviation of no more then ±5.0% for each concentration
Displayed
absorbance value vs.
Output voltage check
3.10.8. For digital and analog signals : displayed absorbance
value / output signal value= 1.00±0.01
System validation
(to confirm the function of each component as well as the performance opf the entire system)
Test
items/parameters
proced
ure
User limit (acceptance criteria)
Drift noise check 3.11.2. Drift ≤ 1.0x10-3Au/h Noise ≤ 3.0x 10-4AU
Gradient
concentration
accuracy test
3.11.3. Within ± 1.0% of set value
System 3.11.4. Peak area CV% ≤ 1.0%
71
reproducibility test Retention time CV% ≤ 0.5%
Vertification of
gradient LC system
7.7.2. The RSD (C.V.)‘s obtained must satisfy
Peak area CV% ≤ 1.0%
Retention time CV% ≤ 0.5%
If vertification fails:
Check the service life of consumables, replace them if necessary or
Perform troubleshooting (for individual system components see their instruction manuals)