International Journal of Food Microbiology 58 (2000) 1–37 www.elsevier.nl / locate / ijfoodmicro

Review

Histamine fish poisoning revisited a, b Leigh Lehane *, June Olley a

National Office of Animal and Plant Health, Agriculture, Fisheries and Forestry – Australia, GPO Box 858, Canberra, ACT 2601, Australia b School of Agricultural Science, University of Tasmania, GPO Box 252 – 54, Hobart, Tasmania 7001, Australia Received 29 February 2000; received in revised form 11 March 2000; accepted 16 March 2000

Abstract Histamine (or scombroid) fish poisoning (HFP) is reviewed in a risk-assessment framework in an attempt to arrive at an informed characterisation of risk. Histamine is the main toxin involved in HFP, but the disease is not uncomplicated histamine poisoning. Although it is generally associated with high levels of histamine ( $ 50 mg / 100 g) in bacterially contaminated fish of particular species, the pathogenesis of HFP has not been clearly elucidated. Various hypotheses have been put forward to explain why histamine consumed in spoiled fish is more toxic than pure histamine taken orally, but none has proved totally satisfactory. Urocanic acid, like histamine, an imidazole compound derived from histidine in spoiling fish, may be the ‘‘missing factor’’ in HFP. cis-Urocanic acid has recently been recognised as a mast cell degranulator, and endogenous histamine from mast cell degranulation may augment the exogenous histamine consumed in spoiled fish. HFP is a mild disease, but is important in relation to food safety and international trade. Consumers are becoming more demanding, and litigation following food poisoning incidents is becoming more common. Producers, distributors and restaurants are increasingly held liable for the quality of the products they handle and sell. Many countries have set guidelines for maximum permitted levels of histamine in fish. However, histamine concentrations within a spoiled fish are extremely variable, as is the threshold toxic dose. Until the identity, levels and potency of possible potentiators and / or mast-cell-degranulating factors are elucidated, it is difficult to establish regulatory limits for histamine in foods on the basis of potential health hazard. Histidine decarboxylating bacteria produce histamine from free histidine in spoiling fish. Although some are present in the normal microbial flora of live fish, most seem to be derived from post-catching contamination on board fishing vessels, at the processing plant or in the distribution system, or in restaurants or homes. The key to keeping bacterial numbers and histamine levels low is the rapid cooling of fish after catching and the maintenance of adequate refrigeration during handling and storage. Despite the huge expansion in trade in recent years, great progress has been made in ensuring the quality and safety of fish products. This is largely the result of the introduction of international standards of food hygiene and the application of risk analysis and hazard analysis and critical control point (HACCP) principles.  2000 Elsevier Science B.V. All rights reserved. Keywords: Histamine; Scombroid; Cadaverine; Mast cells; Urocanic acid; Fish poisoning; Fish spoilage

*Corresponding author. Tel.: 161-2-6272-4697; fax: 161-2-6272-4533. E-mail address: [email protected] (L. Lehane) 0168-1605 / 00 / $ – see front matter  2000 Elsevier Science B.V. All rights reserved. PII: S0168-1605( 00 )00296-8

2

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

1. Introduction Histamine (or scombroid) fish poisoning (HFP) is a foodborne chemical intoxication caused by eating spoiled, or bacterially contaminated, fish. The fish are harmless when fresh, and after they have become toxic they may still have a normal appearance and odour (Sapin-Jaloustre and Sapin-Jaloustre, 1957). No available method of preparation, including freezing, canning and smoking, will destroy the causative toxin(s) (Etkind et al., 1987). The disease is not uncomplicated histamine poisoning, but is generally associated with high levels of histamine ( $ 50 mg / 100 g) in the spoiled fish. Histamine is a naturally occurring substance in mammalian physiology. It is contained in mast cells and basophils, and its biological effects are usually seen only when it is released in large amounts in the course of allergic and other reactions. Histamine exerts its effects by binding to receptors on cellular membranes in the respiratory, cardiovascular, gastrointestinal and haematological / immunological systems and the skin (Cavanah and Casale, 1993). HFP is a common form of fish poisoning, but many incidents go unreported because of the mildness of the disease, lack of required reporting, and misdiagnosis. Symptoms can be confused with those of ‘‘Salmonella infection’’ (Russell and Maretic, 1986) and food allergy (Taylor, 1985). While the syndrome is that of histamine toxicity, there is individual variation in susceptibility, and clinical signs are more severe in people taking medications such as isoniazid, which inhibit enzymes that normally detoxify histamine in the intestine (Stratton et al., 1991). HFP is a significant public health and safety ´ concern (Sanchez-Guerrero et al., 1997; Wu et al., 1997) and a trade issue (Anonymous, 1998a). The earliest record of the disease was in 1828 (Henderson, 1830). Since then, it has been described in many countries, and is now the most prevalent form of seafood-borne disease in the United States (Lipp and Rose, 1997). The worldwide network for harvesting, processing and distributing fish products has made HFP a global problem. However, because it is a consequence of improper handling or storage of fish and there are effective testing methods to identify

likely toxic fish, control and prevention are possible (Institute of Medicine, 1991). Scientific information on HFP is reviewed here in a risk-assessment framework (Kindred, 1996; Buchanan, 1998). An attempt has been made to address, as accurately as possible, the questions: ‘‘What causes outbreaks of HFP?’’, ‘‘What are the underlying factors contributing to outbreaks?’’ and ‘‘What are the consequences of outbreaks?’’

2. Hazard identification The involvement of histamine as the main hazard in HFP is supported by: (i) symptoms identical to those of intravenous histamine administration or allergic reaction; (ii) the efficacy of antihistamine therapy; and (iii) the presence of increased levels of histamine in spoiled fish that cause the syndrome. Morrow et al. (1991) claimed that there was definitive evidence that histamine is the toxic agent in HFP after analysing the urine of poisoned subjects. They found that urinary levels of histamine in patients with HFP were higher than those of subjects who had been injected intravenously with histamine to produce toxic symptoms. However, consuming spoiled fish containing histamine is more likely to cause toxic effects than taking the same amount of pure histamine by mouth. Pure histamine taken orally is substantially metabolised in crossing the intestinal wall or in the liver, and produces only mild symptoms at relatively high doses (Taylor et al., 1984). Thus, while elevated blood histamine levels are undoubtedly characteristic of HFP, there is uncertainty as to how these occur. This has led to speculation that there are other ‘‘scombroid toxins’’ acting with histamine. Taylor (1986) said that HFP appears to be caused by exogenous histamine (i.e., from spoiled fish), potentiated by other biogenic amines and possibly other substances. Clifford et al. (1991) and Ijomah et al. (1991), on the other hand, postulated that a toxin or toxins other than histamine in the spoiled fish cause(s) release of endogenous histamine from mast cells in the human body. The indication that there may be multiple toxins, perhaps acting synergistically, may account for the clinical variability encountered.

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

2.1. Production of histamine and other biogenic amines and imidazole compounds in spoiling fish Histamine production in fish is related to the histidine content of the fish, the presence of bacterial histidine decarboxylase (HD), and environmental conditions (Ienistea, 1973). During spoilage, certain bacteria produce decarboxylase enzymes, which act on free histidine and other amino acids in the fish muscle to form histamine and other biogenic amines. Chemically, histamine (from histidine), putrescine (from ornithine), cadaverine (from lysine), and spermidine and spermine (from arginine), which are produced post-mortem in fish muscle, are low-molecular-mass, aliphatic, alicyclic or heterocyclic organic bases (Eitenmiller and De Souza, 1984; Rawles et al., 1996). Scombroid fish belonging to the families Scombridae (e.g., tuna and mackerel) and Scomberesocidae (e.g., saury) are most commonly associated with HFP, but non-scombroid species (e.g., mahi-mahi, sardines, pilchards, anchovies, herring, marlin and bluefish) can also be involved. These species are characterised by having relatively high levels of histidine in their flesh (Taylor, 1986). Histidine levels vary from 1 g / kg in herring to as much as 15 g / kg in tuna (Ijomah et al., 1992). Fresh fish contains negligible quantities of histamine, usually , 0.1 mg / 100 g (Frank et al., 1981). Histamine can be produced rapidly by bacterial decarboxylases in scombroid and other fish that have relatively high free histidine levels in their muscles when alive (Love, 1980). This occurs before postmortem proteolysis liberates additional histidine from muscle protein, and explains why histamine can reach high concentrations without the formation of organoleptic (sensory) spoilage indicators (SapinJaloustre and Sapin-Jaloustre, 1957). After investigating HD production by Morganella morganii in mackerel, Eitenmiller et al. (1982) concluded that the ready availability of free histidine in the muscle to act as both an inducer and substrate makes it an ideal environment for histamine formation. The main bacteria responsible for histidine decarboxylation and HFP are members of the family Enterobacteriaceae (Frank et al., 1985; Taylor and Sumner, 1986). Endogenous production of decarboxylase enzymes in fish muscle is insignificant when

3

compared with the exogenous (bacterial) pathway (Rawles et al., 1996). Spoilage, ammonia production and biogenic amine production by these bacteria are enhanced at elevated storage temperatures (Arnold et al., 1980). Once a large bacterial population has been established, residual enzyme activity continues slowly at refrigeration temperatures, although bacterial growth ceases (Klausen and Huss, 1987b; Stratton and Taylor, 1991). Histamine is also produced, but to a lesser extent, by bacteria that can grow at refrigeration temperatures (Okuzumi et al., 1981). Only free histidine can be decarboxylated (Geiger, 1944b; Arnold and Brown, 1978). However, the decarboxylation of histidine to form histamine is only one of two routes of histidine metabolism, and the occurrence of this pathway in fish spoilage is quite limited. The pathway favoured by most bacteria is a catabolic one in which glutamate is the ultimate product formed. The first step in this pathway is the loss of ammonia from histidine by the action of L-histidine ammonia lysase (HAL), or histidase, resulting in the formation of urocanic acid (isomer not stated) (Baranowski, 1985). HAL has a wide distribution among bacteria (Shibatani et al., 1974; Baranowski, 1985) and, unlike HD, is also found as an endogenous component of fish muscle (Kawai and Sakaguchi, 1968; ´ ´ Mackie and Fernandez-Salgeuro, 1977). Some bacteria also possess urocanase, which catabolises urocanic acid, but Shibatani et al. (1974) found that no strain of 106 tested had urocanase activity equal to its HAL activity. Urocanic acid (isomer not stated) was found at a much higher concentration than histamine (4.74 versus 0.19 mg / 100 ml) in mackerel after 18 days’ storage at 08C (Mackie and ´ ´ ´ ´ Fernandez-Salgeuro, 1977; Fernandez-Salgeuro and Mackie, 1979), although in this experiment it was attributed to endogenous rather than bacterial HAL. Shewan (1955) investigated the distribution of free bases and amino acids in muscle from fresh and spoiled mackerel by fractionation by displacement chromatography on ion-exchange resins of aqueous– alcoholic extracts. Following spoilage, he found large amounts of histamine and moderate amounts of an unidentified component (compound 8) not found in fresh muscle. Later, Hughes (1959) demonstrated a major imidazole spoilage product (compound X) from spoiled herring. It gave a cherry red colour with

4

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

Pauly’s reagent, suggesting the imidazole portion of the molecule was still intact. However, his efforts to identify compound X as urocanic acid (imidazolylacrylic acid), imidazolylcarboxylic acid or imidazolylacetic acid by comparison with ‘‘authentic samples’’ (of unspecified origin) were unsuccessful.

2.2. Metabolism of histidine and histamine in mammals Humans metabolise histidine to urocanic acid through the activity of HAL, to form glutamate and then a-ketoglutarate, which enters the citric acid cycle (Stryer, 1981; Furuta et al., 1996), or to histamine through the activity of HD (Stryer, 1981). Histamine can be catabolised by several routes. It can be oxidatively deaminated by diamine oxidase (DAO, or histaminase) to imidazole acetaldehyde and imidazoleacetic acid, methylated by histamine methyl transferase (HMT) to form methylhistamine, or its side chain can be methylated or acetylated (Rice et al., 1976; Taylor, 1986). The fact that oral histamine alone is not toxic to adults except at doses $ 100 mg appears to be due mainly to the presence of DAO and HMT in the intestinal tract, and HMT in the liver, which detoxify histamine. Histamine is also converted to inactive acetylhistamine in the intestine, presumably by bacterial enzymes. All these enzymes decrease the amount of unmetabolised histamine available for absorption into the systemic circulation. The relative importance of the DAO and HMT pathways varies among species. The oxidative deamination pathway predominates in rats and guinea pigs, and the methylation pathway is of prime importance in humans, mice, cats, pigs and hamsters (Taylor, 1986). When [ 14 C]histamine was administered orally to humans, 68–80% of the administered radioactivity was recovered in the urine (Sjaastad and Sjaastad, 1974). Some histamine remained unchanged in the faeces, and additional amounts were catabolised by intestinal bacteria and radioactivity was exhaled as 14 CO 2 from the lungs. Hesterberg et al. (1984) demonstrated that HMT activity is widespread in human tissues, with the order of activity being liver . . colon . spleen . lung . small intestine . stomach, and that DAO is mainly localised in the intestine. Helander et al. (1965) had shown earlier that the human kidney has a considerable capacity for removing histamine from the blood. When heal-

thy individuals were infused intravenously with histamine, a large proportion was methylated by the kidney and excreted in the urine and a smaller proportion was excreted unchanged in the urine. HMT is very selective for histamine, and requires S-adenosylmethionine as a methyl donor. It is subject to substrate inhibition by high concentrations of histamine and is inhibited by analogues of methylmethionine such as adenosyl-homocysteine, antimalarial drugs, and numerous agonists and antagonists of histamine receptors. The type of inhibition (competitive, uncompetitive, or non-competitive) varies depending on the inhibitor (Taylor, 1986). DAO oxidises other diamines such as putrescine, as well as histamine. It is subject to substrate inhibition when certain diamines including histamine are used as substrates. Many inhibitors of DAO have been identified, such as aminoguanidine and some antihistaminic drugs. Foodborne inhibitors of DAO include carnosine, thiamine, cadaverine and tyramine. Monoamine oxidase (MAO) is also important in histamine metabolism (Taylor, 1986).

2.3. Possible mechanisms of toxicity 2.3.1. Toxicity of histamine and other biogenic amines Histamine exerts its toxicity by interacting with receptors on cellular membranes. There are three types of histamine receptors, H 1 , H 2 and H 3 (Cavanah and Casale, 1993). The most common symptoms result from action on the cardiovascular system. Histamine causes dilatation of peripheral blood vessels, causing urticaria, hypotension, flushing and headache. Histamine-induced contraction of intestinal smooth muscle causes abdominal cramps, diarrhoea and vomiting. Pain and itching associated with the urticarial lesions may be due to sensory and motor neuron stimulation (Taylor, 1986). Although there is compelling evidence to implicate histamine as the causative agent in HFP, there is not a straightforward dose–response relationship, as spoiled fish containing histamine tends to be more toxic than the equivalent amount of pure histamine dosed orally. Studies on HFP have been confused by interacting variables affecting both fish samples and the consumer, and most investigations have been limited in their ability to disentangle them (Ijomah et al., 1992). Complicating factors in sampling can include the wrong sample analysed, variable his-

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

tamine levels within the sample, and the presence of microbial toxins, or other toxins or contaminants or metabolites. Variables affecting the consumer include misdiagnosis, innate individual variation, bodyweight differences, gender differences in metabolism, concomitant medication, idiosyncratic intolerance, and the presence of true allergy. Weiss et al. (1932) found that 180 mg histamine base (given as 500 mg histamine phosphate) administered orally was without noticeable effect in humans, while 7 mg administered intravenously caused vasodilatation and increased heart rate. Granerus (1968) gave humans up to 67.5 mg histamine orally without toxic effect. Motil and Scrimshaw (1979) found that administration of 100–180 mg histamine orally mixed in grapefruit juice or in 100 g high-quality tuna caused characteristic symptoms of mild histamine poisoning (mild-to-severe headache and obvious flushing) in some people (1 / 4 and 4 / 8, respectively). Subsequently, Clifford et al. (1989) gave volunteers 50 g of fresh mackerel to which 300 mg histamine had been added (a dose of about 5 mg / kg bodyweight), and recorded only mild symptoms of histamine poisoning (oral tingling, headache and flushing in some subjects). However, 50 g of spoiled mackerel with 300 mg of added histamine was no stronger in its effect. A survey of scombrotoxic fish poisoning in Britain by Bartholomew et al. (1987) suggested that most cases were uncomplicated histamine poisoning, but that other toxins may have been involved when suspect fish contained little histamine, or when symptoms were not typical. These authors reported 258 incidents of suspected scombrotoxic fish poisoning. Of 240 fish samples from these incidents, 101 contained . 5 mg histamine / 100 g fish. The symptoms most consistently reported were rash, diarrhoea, flushing and headache. In any one incident, the symptoms of all patients were similar, although each patient did not experience all symptoms. Of fish with . 20 mg histamine / 100 g, 94% were from incidents in which scombrotoxic symptoms were characteristic. Where fish contained only 5–20 mg histamine / 100 g fish, only 38% of incidents were ‘‘clinically distinctive’’, with rash, flushing and burning of the mouth. Although gastrointestinal symptoms especially diarrhoea were frequently experienced, in contrast with some previous reports (Arnold and Brown, 1978), gastrointestinal symptoms alone were not regarded as indicative of scombrotoxic fish poison-

5

ing. Surprisingly, 36 / 93 fish samples (39%) with , 5 mg / 100 g histamine also gave rise to characteristic scombrotoxic (or HFP) symptoms. Analyses were not done for other biogenic amines, which may have been histamine potentiators or toxic in their own right, or urocanic acid. Til et al. (1997) examined the acute and subacute toxicity of five biogenic amines, cadaverine, putrescine, tyramine, spermidine and spermine, in rats. The doses of amines tested and tolerated were many times higher than doses expected in humans following the ingestion of a meal of spoiled fish. However, the amines were tested individually, not in combination with one another, or with histamine. Spermine and spermidine had the highest acute toxicity, with LD 50 values of 600 mg / kg, compared with . 2000 mg / kg for the other three amines. Spermine was the most toxic in a 5–6 week feeding study, causing liver changes, elevated activity of plasma enzymes associated with the liver, nephrotoxicity and myocardial degeneration at a dietary level of 500 ppm. However, Mietz and Karmas (1977) found that levels of spermine and spermidine tended to fall and sometimes reached zero during the decomposition process in tuna.

2.3.2. Inhibition of histamine detoxification by histamine potentiators A number of scientists have postulated that histamine is potentiated by some other component or components in toxic fish (Bjeldanes et al., 1978; Taylor and Lieber, 1979; Chu and Bjeldanes, 1981; Lyons et al., 1983; Taylor, 1986; Stratton et al., 1991). Such potentiators would act to decrease the threshold dose of histamine needed to provoke an adverse reaction in humans challenged orally. Certain drugs have definitely been implicated as contributing factors in cases of histamine poisoning (Chin et al., 1989; Stratton et al., 1991). Taylor (1986) stated that doses of pure histamine required to produce mild reactions were ‘‘several times higher than the doses producing more severe symptoms when consumed with spoiled fish’’. However, in the same paper he stated that ‘‘the variability in histamine levels in spoiled fish makes estimates of the toxic threshold difficult to obtain’’ and quoted an estimated threshold dose of histamine in fish at about 60 mg / 100 g flesh. Even after allowing for variability in human susceptibility and variable histamine

6

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

content in different parts of fish (Lerke et al., 1978), it seems that there is a difference between the relative lack of toxicity of pure histamine and the (often) apparent toxicity of histamine in spoiled fish. In support of the histamine-potentiator hypothesis, several in vivo and in vitro studies have suggested that the absorption, metabolism, and / or potency of one biogenic amine might be modified in the presence of a second amine (Bjeldanes et al., 1978; Taylor and Lieber, 1979; Lyons et al., 1983). The biogenic amines putrescine and cadaverine occur in appreciable quantities in toxic fish (Arnold and Brown, 1978) and at low levels in non-toxic fish (Mietz and Karmas, 1977). When given in higher ratios relative to histamine than those that usually occur in toxic fish, these amines potentiate the biological activity of histamine in laboratory animals. Uptake of unmetabolised histamine alone would not be sufficient to elicit some of the symptoms observed in HFP. To exert its full toxic effects, histamine must reach the peripheral tissues. The detoxification of histamine in extra-intestinal tissues must also be inhibited to achieve the full effects (Hui and Taylor, 1985). Particularly if hepatic first-pass clearance of histamine is normally substantial, potentiators must act beyond the intestinal lumen to have an effect in increasing the toxicity of histamine taken orally. Mongar (1957) observed that cadaverine and putrescine competitively inhibited DAO and potentiated histamine-induced contractions in guinea pig ileum. Taylor and Lieber (1979) tested the effects of 38 chemicals (mainly nitrogen-containing bases) likely to be consumed with tuna on the in vitro activity of rat jejunal mucosal HMT and DAO. The most potent inhibitors (at 10 mM concentrations) were monoamines, diamines and guanidines, including tyramine, b-phenylethylamine and tryptamine. Correlations between chemical structure and inhibitory activity were difficult to define. Many of the identified chemicals are found in spoiled tuna along with histamine, and their ability to inhibit histamine catabolism could magnify the oral toxicity of histamine, thus explaining its apparently greater toxicity when consumed with spoiled tuna. Cadaverine and aminoguanidine were both strong inhibitors of rat intestinal DAO and HMT in vitro. Putrescine was a weak inhibitor of HMT only (Taylor and Lieber, 1979). Lyons et al. (1983)

subsequently showed that cadaverine and aminoguanidine enhanced the absorption of unmetabolised histamine in perfused rat intestinal segments by inhibiting the conversion of histamine to less-toxic metabolites. Aminoguanidine is not known to occur in spoiled fish (Lyons et al., 1983). The oral toxicity of histamine in the guinea pig was increased 10 times when it was administered 40 min after oral administration of putrescine (Parrot and Nicot, 1966), and the oral toxicity of histamine in the guinea pig was potentiated by simultaneous administration of cadaverine (Bjeldanes et al., 1978). Other biogenic amines that may act as potentiators of histamine toxicity include tyramine (a MAO inhibitor that increases blood pressure), tryptamine (which inhibits DAO), and b-phenylethylamine (an inhibitor of both DAO and HMT) (Stratton et al., 1991). Bjeldanes et al. (1978) suggested that, to exhibit synergism as measured by LD 50 values in the guinea pig, Parrot and Nicot (1966) used an unrealistically high ratio of putrescine to histamine (5:1, compared with 1:100 in toxic fish). In their own LD 50 experiments on guinea pigs, Bjeldanes et al. (1978) found that cadaverine was synergistic only at a ratio to histamine of 1:5 or greater, whereas in toxic fish (tuna) the ratio is 1:10 (Mietz and Karmas, 1977; Kim, 1978). They also demonstrated that while optimal cadaverine synergism was obtained by simultaneous administration with histamine to guinea pigs, putrescine required a time lag before histamine was administered. Thus, it is unlikely that putrescine in spoiled fish acts as a histamine potentiator, and cadaverine may be involved only sometimes. Klausen and Lund (1986) found, when comparing the spoilage of mackerel and herring at low temperatures, that levels of cadaverine in mackerel may exceed those of histamine by 2–5-times, which may help explain why mackerel are often associated with HFP. Hui and Taylor (1985) determined the effects of enzyme inhibitors of histamine in vivo and their possible role in potentiating histamine toxicity by studying the effects of known inhibitors of DAO, HMT and MAO, both foodborne and pharmacological, on the urinary excretion pattern of histamine and its metabolites in rats. When they administered [ 14 C]histamine orally to rats, an average of 80% of the administered radioactivity was recovered in the urine after 24 h. About 10% of the total dose was excreted via the faeces. Analysis of 4-h urine sam-

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

ples showed imidazoleacetic acid was the predominant metabolite (60.6%), and N-methyl-imidazoleacetic acid (8.6%), N-methylhistamine (7.3%) and N-acetylhistamine (4.5%) were minor metabolites. While the administration of inhibitors resulted in an increased amount of unmetabolised histamine and a decreased amount of metabolites in the urine, the total rate of excretion of histamine and its metabolites was similar in the presence and absence of potentiators. This indicated that the total amount of absorption and excretion of histamine and its metabolites was equivalent, with or without potentiators. Hui and Taylor (1985) found that pharmacological inhibitors were more potent and had a longer duration of action in rats than foodborne ones. Cadaverine dihydrochloride, putrescine dihydrochloride, tyramine hydrochloride and b-phenylethylamine hydrochloride were weak inhibitors, effective at doses 4–5-times higher than that of simultaneously administered histamine dihydrochloride. Inhibitors of DAO (aminoguanidine, isoniazid, cadaverine, bphenylethylamine and tyramine) had a more profound effect on excretion of unmetabolised histamine than inhibitors of only HMT. To test the possibility that more than one potentiator might be involved in foods implicated in histamine poisoning incidents, the authors tested a mixture of histamine (0.5 mmol / kg) and the foodborne inhibitors cadaverine, putrescine, tyramine, b-phenylethylamine and tryptamine (each dosed at 0.5 mmol / kg). The mixture increased the amount of unmetabolised histamine in 4-h urine samples 100% over that present when histamine was dosed alone. The significance of potentiators of histamine intoxication in humans has not been established. However, experiments in rats and guinea pigs indicate that cadaverine and other amines could possibly potentiate histamine toxicity of spoiled fish if present alone (in the case of cadaverine) or collectively. It is also possible that they could act in an additive or synergistic fashion with other, as yet undiscovered, potentiators.

2.3.3. Barrier disruption hypothesis The ‘‘barrier disruption hypothesis’’ hypothesis was first proposed by Parrot and Nicot (1966), who suggested that potentiators might interfere with the protective actions of intestinal mucin. Intestinal mucin is known to bind histamine, and it has been

7

suggested that this binding is essential to prevent the intestinal absorption of histamine. Potentiation would occur by disruption of the bonding and enhancement of absorption (Jung and Bjeldanes, 1979; Chu and Bjeldanes, 1981). Jung and Bjeldanes (1979) found that cadaverine exhibited a marked influence on the rate of transport of 14 C-labelled histamine and metabolites across the gut wall in isolated gut sections of the guinea pig, but had a minor effect on histamine metabolism. They suggested that cadaverine potentiation of histamine toxicity may result from induction of increased rates of absorption of histamine and metabolites, and that the established antihistaminase activity of cadaverine does not appear to play a significant role. They believed their results were consistent with a proposed role of substances such as mucin in maintaining a barrier to histamine transport. However, experiments based on this theory have not so far provided a convincing rationale for the apparent toxicity of histamine in fish (Taylor, 1986; Mitchell, 1993).

2.3.4. Release of endogenous (mast cell) histamine by scombroid toxin(s) Mongar (1957) had reported that diamines in the series NH 2 (CH 2 ) n NH 2 release histamines from isolated tissues. His results were obtained using minced guinea pig lung and isolated rat diaphragm to measure histamine-releasing activity of diamines of chain length C 5 –C 15 . The activity increased with both chain length and concentration. At a concentration of 1 mM, cadaverine was inactive both on minced guinea-pig lung and rat diaphragm, but at 100 mM it released about 80% of the histamine content of the tissue. Histamine-releasing activity increased with pH and was attributed mainly to the non-ionised base. However, since the compounds were inactive at pH 7 and their activity rose sharply with pH to reach a high value at pH 8.5, they would be ionised at physiological pH and not active in producing endogenous histamine in the human body. Attention began to focus on mast cells, the granules of which contain a histamine–heparin complex (Riley, 1959). In 1972, Olley postulated that ‘‘some other basic substance produced during spoilage (of fish) may release histamine from the histamine–heparin complex of mast cells’’. Arnold and Brown (1978) also pointed out that release of histamine from mast cell granules could account for

8

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

some, if not all, of the symptoms associated with scombroid toxicity. Subsequently, Clifford et al. (1991) and Ijomah et al. (1991) of the UK Ministry of Agriculture, Fisheries and Food postulated that the scombroid toxin(s) is a mast cell degranulator, and antihistamine therapy is effective because it eliminates the effect of endogenous histamine. Ijomah et al. (1991, 1992) fed mackerel deliberately exposed to poor storage conditions to volunteers. They found that the volunteers fell into susceptible and non-susceptible subgroups, and that the level of histamine in fish did not correlate with toxicity exhibited. However, when subjects were dosed with either placebo or an H 1 antagonist, the antihistamine abolished vomiting and diarrhoea associated with the ingestion of scombrotoxic fish. The scientists concluded that dietary histamine played a minor role in the toxicity and that the major contributor was a postulated agent that released histamine by degranulation of mast cells in the gastrointestinal tract. Clifford et al. (1991) analysed mackerel fillets that had been associated with an outbreak of ‘‘scombrotoxicosis’’ for their contents of cadaverine, histamine, putrescine, spermidine, spermine and tyramine. The same fillets were fed to healthy volunteers. Susceptibility to the disease was quite variable. Of the 86 fillets examined, 30 rapidly induced nausea / vomiting and / or diarrhoea when 50g portions were consumed. The remaining fillets failed to provoke such symptoms, even though volunteers proven to be susceptible to scombrotoxicosis tested 17 of them. Statistical analysis failed to detect differences in amine content between fillets shown to be scombrotoxic and those failing to induce symptoms, and failed to establish significant relationships between the amine doses and volunteer responses, even after manipulations to simulate additive or synergistic interactions. The minimum doses associated with scombrotoxicosis and the maximum doses giving negative responses in susceptible volunteers were: histamine, 0.23 and 3.33; cadaverine, , 0.02 and 0.38; putrescine, 0.0008 and 0.01; tyramine, 0.009 and 0.75 mg / kg bodyweight, respectively. The scientists concluded that the contents of such amines in mackerel have little or no role in the aetiology of scombrotoxicosis, and that the primary scombrotoxin, which remained unknown, was responsible for mast cell degranulation.

Other clinical reports of HFP, for example those caused by the consumption of salmon, support the suggestion that exogenous histamine is not always responsible for intoxication, and that other toxins and / or endogenous histamine may be involved. Bartholomew et al. (1987) reported six incidents of scombrotoxic fish poisoning in canned salmon in Britain in 1983. In five of the incidents, the incriminated fish contained , 1 mg / 100 g histamine, despite characteristic symptoms of histamine poisoning. Salmon from the sixth episode contained only 17 mg / 100 g histamine. Gessner et al. (1996) also described a typical and severe scombrotoxic-like illness following the ingestion of smoked sockeye salmon that demonstrated low histamine levels and high toxicity on mouse bioassay. Symptoms included flushing, pruritus, nausea, sweating, vomiting, diarrhoea, dizziness, etc. The implicated salmon had histamine levels 25-fold less than the United States Food and Drug Administration (FDA) toxicity level for tuna of 50 mg / 100 g (FDA, 1998). The patient ingested an estimated 0.0006 mg of histamine / kg bodyweight – far less than the estimated 1 mg of histamine / kg bodyweight reported to cause human illness (Taylor, 1986). Estimated levels for putrescine and cadaverine were 0.67 mg / 100 g and 0.19 mg / 100 g. Some imidazole compounds are known to release histamine from mast cells by a non-immunological mechanism; for example, imidazole fungicides (Gietzen et al., 1996). It is interesting to note that urocanic acid, an imidazole compound and a histidine metabolite of spoiling fish (Kawai and ´ ´ Sakaguchi, 1968; Mackie and Fernandez-Salguero, 1977; Baranowski, 1985), has recently been described as an inducer of histamine in vivo in mice (Hart et al., 1997) and as a mast cell degranulator in human skin organ cultures (Wille et al., 1999) in the cis-isomer form. Urocanic acid may be one of the ‘‘toxic imidazole compounds’’ in spoiling fish mentioned by Olley, 1972 and the ‘‘scombroid toxin’’ that researchers have been seeking for decades. Some experimental evidence seems to reject the mast cell degranulation theory, but this may reflect the large number of variables involved in the pathogenesis of HFP. Mast cell degranulation may not occur in all cases of the disease, but may be an important feature of other cases. When Van Gelderen et al. (1992) gave eight healthy volunteers either

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

herring paste containing 90 mg histamine formed from induced spoilage with photobacteria or fresh fish to which 90 mg histamine had been added, the maximum histamine concentration in plasma did not differ significantly between the two dose regimes. However, herring may not have been the most suitable fish to use in this experiment. Klausen and Lund (1986) found that herring contained 4–5-times less free histidine and free lysine (cadaverine precursor) than fresh mackerel. In addition, the photobacteria used to induce spoilage may not have possessed HAL, which is required to produce urocanic acid from histidine. Morrow et al. (1991) measured the urinary excretion of histamine and its metabolite, N-methylhistamine, in three people who had HFP after eating marlin that contained high levels of histamine. They also measured 9a,11b-dihydroxy-15-oxo-2,3,18,19tetranorprost-5-ene-1,20-dioic acid (PGD-M), the principal metabolite of prostaglandin D 2 , a mast cell secretory product, to assess whether mast cells had been activated to release histamine. The levels of histamine and N-methylhistamine were 9–20- and 15–20-times, respectively the normal means in urine samples collected 1–4 h after the meal. During the subsequent 24 h, the levels fell to 4–15- and 4–11times the normal values, and 14 days later they had returned to normal. PGD-M excretion was not increased at any time. Thus, HFP was associated with urinary excretion of histamine in quantities far exceeding those associated with toxicity. The high histamine levels in the incriminated fish, together with the failure to find evidence of increased endogenous release of prostaglandin D 2 , indicated that the histamine was derived from the fish. Analyses were not done for other biogenic amines, which could have been histamine potentiators. Increases in the urinary excretion of both histamine and N-methylhistamine of similar magnitude in the poisoned people suggested that the spoiled fish did not contain a substance or substances that potentiated histamine toxicity by inhibiting its inactivation by HMT. ´ Like Morrow et al. (1991), Sanchez-Guerrero et al. (1997) rejected the proposal that endogenous histamine is involved in HFP. The latter authors used immunoassay to measure tryptase, another indicator of mast cell degranulation, in seven patients intoxicated after eating tuna. The quantification of tryptase release in vivo allows a precise valuation of mast cell

9

activation, as the enzyme is located in secretory granules of mast cells and released at the same time as histamine during degranulation. The low levels obtained caused the authors to reject an anaphylactictype reaction and favour poisoning by an exogenous source of histamine (fish). The role of endogenous histamine release, if any, in histamine / scombroid fish poisoning remains uncertain and unproved. The syndrome described by Ijomah et al. (1991, 1992) and Clifford et al. (1991) was not what would be considered typical of HFP. Many toxins produce nausea, vomiting and diarrhoea. Mitchell (1993) wrote that a possible factor in explaining the apparent contradiction of this work with previous observations on HFP is the definition of poisoning that was applied. Ijomah et al. (1991) considered that vomiting and / or diarrhoea were essential factors in poisoning. Flushing and tingling in the mouth were considered ‘‘minor’’ and possibly caused by dietary histamine. The syndrome described by Ijomah, Clifford and colleagues may well have been complicated by a toxin or toxins additional to those commonly associated with HFP. The complexity of fish poisoning makes it difficult to do clinical trials on HFP. Unless there is consensus of opinion on the clinical signs and symptoms of HFP, such trials are of little value.

2.3.5. Are paralytic and diarrhetic shellfish poisons involved? Clifford et al. (1993) found low levels of paralytic shellfish poisons (saxitoxins) in toxic mackerel (0.02–1.30 mg saxitoxin equivalent / kg) by enzymelinked immunosorbent assay (ELISA) and suggested that, possibly in combination with diarrhetic shellfish poisons (saxitoxins), they may be responsible for scombroid fish poisoning. These concentrations of saxitoxin equivalents corresponded to doses in the range 0.04–1.0 ng / kg bodyweight and were orders of magnitude less than the intraperitoneal and oral LD 50 values in mice (10 and 263mg / kg, respectively). Calculations made following an outbreak of mussel poisoning in England suggested that doses as low as 9 mg saxitoxin equivalent / kg bodyweight may cause illness in humans, but doses as high as 86 mg / kg bodyweight are not necessarily fatal (McCollum et al., 1968). Nine of 23 mackerel samples associated with scombrotoxic incidents apparently

10

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

contained less saxitoxin than the negative control sample, and dose (in saxitoxin equivalents) could not be correlated with the volunteer response. Doses associated with nausea / vomiting and / or diarrhoea ranged from 0.11 to 1.0 ng / kg bodyweight, whereas doses not producing these symptoms ranged up to 0.5 ng / kg bodyweight. Clifford et al. (1993) concluded that paralytic shellfish poisons are at least partly responsible for ‘‘scombrotoxicosis’’. However, the observation during volunteer testing of nausea and vomiting at doses orders of magnitude less than previously reported for saxitoxin, coupled with the presence of diarrhoea, which is not characteristic of paralytic shellfish poisoning, suggest that this is unlikely.

2.3.6. Absorption of histamine from mouth and throat An unusual hypothesis (cited by Lange, 1988) is that when affected fish are eaten the histamine may be absorbed through the mucous membranes of the mouth and throat, thus bypassing the digestive process that destroys it. It is difficult to propose a mechanism whereby a fish substrate would enhance the absorption of histamine through mucous membranes. 2.4. Clinical characteristics and treatment 2.4.1. Clinical signs and symptoms Signs and symptoms of HFP occur from several minutes to several hours after ingestion of the toxic fish. The illness typically lasts a few hours, but may continue for several days. The primary signs / symptoms are those of histamine poisoning – cutaneous (rash, urticaria, oedema and localised inflammation), gastrointestinal (nausea, vomiting, diarrhoea), haemodynamic (hypotension) and neurological (headache, palpitations, tingling, burning, itching) (Taylor, 1985; Wu et al., 1997). In severe cases there may be bronchospasm and respiratory distress (Shalaby, 1996). The most consistent clinical sign reported by Arnold and Brown (1978) was flushing of the skin of the face and neck, which caused a feeling of intense heat and general discomfort. This predominantly involves exposed areas, resembling sunburn (Kim, 1979). Gastrointestinal symptoms (nausea and diarrhoea) were experienced by fewer than 25% of

victims described by Arnold and Brown (1978). However, when Gilbert et al. (1980) reported on 150 patients affected in 30 separate outbreaks of ‘‘scombrotoxic fish poisoning’’ in Britain, in each outbreak the symptoms were similar and diarrhoea was predominant, occurring in 24 / 30 outbreaks. Diarrhoea was also a prominent clinical sign in 77% of patients (second in frequency only to skin rash, which occurred in 82%) in 10 incidents of HFP involving ¨ 22 patients in South Africa (Muller et al., 1992). Wu et al. (1997) described two outbreaks of HFP in Taiwan in 1996. In the first, caused by a nonscombroid fish, the main symptoms were facial flush, dizziness, headache, conjunctival hyperaemia and hypotension. In the second outbreak, in which a scombroid fish was involved, dizziness, diarrhoea, flushing and headache were the most common symptoms. Levels of histamine and cadaverine in fish samples taken from the two outbreaks were 84 and 8.5 mg / 100 g (first outbreak) and 272 and 23 mg / 100 g (second outbreak), respectively. Anxiety may be a prominent symptom of HFP (e.g., Russell and Maretic, 1986; Sabroe and Kobza Black, 1998; Specht, 1998). A case of transient loss of vision in association with atrial tachycardia was also reported. Other signs and symptoms in the latter patient were typical of HFP, and response to antihistamine therapy, including return of vision, was rapid (McInerney et al., 1996; Clark, 1997). Serious complications are rarely encountered in cases of HFP. However, on rare occasions, cardiac and respiratory complications occur in individuals ´ with pre-existing conditions (Russell and Maretic, 1986; Taylor et al., 1989; Ascione et al., 1997).

2.4.2. Diagnosis and treatment A tentative diagnosis of HFP can be made if the clinical signs / symptoms are typical of histamine poisoning, the onset time is short, and the patient has eaten a type of fish that has previously been implicated in cases of histamine poisoning. The diagnosis can be confirmed by detecting high levels of histamine in the contaminated food, meal remnants or in a similar fish obtained from the same source (Taylor, 1986). Apart from being somewhat inconsistent, the symptoms of HFP are not definitive. Many can occur with other illnesses, both foodborne and non-foodborne. When diarrhoea is the predominant symptom,

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

histamine may not be the main toxin involved. In addition, histamine poisoning is often confused diagnostically with food allergies. Identical symptoms occur and antihistamines are equally effective in treating both illnesses. However, histamine poisoning can easily be distinguished from food allergy on the basis of the lack of a previous history of allergic reactions to the incriminated food, the high attack rate in group outbreaks, and the detection of high levels of histamine in contaminated food. With food allergy, it is unusual for more than one person in a group to experience symptoms caused by a specific food. Further, allergic reactions to some of the fish commonly incriminated in HFP outbreaks, such as tuna and mahi-mahi, are quite rare. IgEmediated allergic reactions could be detected by using skin prick tests with extracts of similar fish with known low histamine content (Taylor, 1986). A clear definition of histamine / scombroid fish poisoning is required, based on the clinical signs / symptoms described by Taylor (1985) and Wu et al. (1997) above. Many instances of fish poisoning are probably the result of a number of unrelated toxins acting in concert. As such, they are not typical of histamine / scombroid fish poisoning, and should not be given that classification. Much research is needed to sort out the many toxins in spoiling fish and the syndromes they cause. Symptoms of HFP rapidly subside after antihistamine treatment. H 1 antagonists such as diphenhydramine or chlorpheniramine are usually selected, but H 2 antagonists such as cimetidine may also be effective (Guss, 1998). Induced emesis has also been recommended as a treatment (Downs, 1997). Since the disease is self-limited, running a short course, pharmacological intervention is not always necessary (Taylor, 1986).

2.5. Detection of histamine and other biogenic amines in fish 2.5.1. Detection of histidine-decarboxylating bacteria Many methods have been developed to detect histidine-decarboxylating bacteria (HDB) that produce histamine in foods such as fish. Most of these detection systems are based on specific media that are selective and / or differential for HDB (Omura et al., 1978; Niven et al., 1981; Taylor and

11

Woychik, 1982; Stratton and Taylor, 1991). In these procedures, the histamine is usually extracted and measured by fluorometric assay. However, there have been problems with identification of HDB. Taylor et al. (1978a) reviewed earlier work and found that evidence linking some bacterial species to histamine production was lacking. They attributed discrepancies to differences between strains within a particular species, growth media, or analytical methodology used in the detection of histamine or HD. For example, Niven’s medium (Niven et al., 1981) has been associated with false positive results. Garland (1985) attributed this to the non-specific alkaline products generated through the metabolism of tryptone and yeast extract by bacteria. McMeekin (1986) suggested that the incorporation of 0.1% glucose could balance non-specific alkaline products, but this medium still gave problems with M. morganii. When 0.1% maltose was incorporated by Anggawati (1986), most false positive results were eliminated (Leung, 1987). According to Taylor (1986), many early studies of bacterial histamine production were not comparative, making it difficult to determine whether histamine production by bacteria was prolific or inconsequential. A reason for measuring histamine production rather than HD activity is that several bacterial species also produce histaminase, which may limit histamine accumulation by these species in foods (Ienistea, 1971). Klausen and Huss (1987a) developed a method based on automated conductance measurements in a histidine-containing medium incubated at 258C. The method is rapid with HDB being detected within 24 h, and could be used in quality assurance programs to screen large numbers of fish samples that may be contaminated with HDB. However, it has the disadvantage of requiring expensive instrumentation (Stratton and Taylor, 1991). An improved test for screening for HDB is required, ideally encompassing and distinguishing cadaverine- and putrescineproducing bacteria.

2.5.2. Analysis of histamine Although canned tuna with high levels of histamine imparts a ‘‘peppery’’ feel to the mouth when chewed, tasting on a routine basis is not a feasible means of quality assurance (Etkind et al., 1987).

12

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

Several other measurements of fish deterioration have been associated with histamine levels (e.g., Cantoni et al., 1976), but no objective measurement other than determining histamine concentration has emerged as an effective indicator of histamine levels. The classical method for determination of histamine, based on the fact that it causes contraction of guinea pig ileum, was used by Geiger (1944a) for histamine analysis in fish. He and fellow workers had earlier identified a biologically active substance in marine fish as histamine (Geiger et al., 1944). Geiger (1944a) reported findings on histamine levels in raw and canned sardines and mackerel and pointed out that canning did not interfere with subsequent analysis for histamine. The validity of chemical testing for histamine in fish muscle depends on the design of the sampling plan. Because histamine is generally not uniformly distributed in a decomposed fish, a guidance level of 50 ppm has been set (FDA, 1998). Additionally, recent studies suggest that, if levels of HD are high, histamine formation can continue even in frozen storage (Price, 1999). Other problems associated with chemical testing for histamine are the lack of standardisation of histamine detection methodology and the fact that the presence of histamine alone is not necessarily a reliable indicator of fish likely to cause HFP. Many diverse analytical procedures have been published for histamine in unprocessed and canned fish since the guinea pig ileum method of Geiger (1944a). Some of the more important include tests for histamine in: • Fish products: thin-layer chromatographic method (Shultz et al., 1976) • Canned tuna: fluorometric method (Lerke and Bell, 1976) • Fish: fluorometric method (Taylor et al., 1978b) • Fish: enzyme-based screening test (Lerke et al., 1983) • Seafood: flow-injection method (Hungerford et al., 1990) • Canned fish: high-performance liquid chromatography (HPLC) (Yen and Hsieh, 1991) • Fish: capillary electrophoresis (Mopper and Sciacchitano, 1994) • Tuna: copper chelation method (Bateman et al., 1994)

• Fish: oxygen-sensor-based method (Ohashi et al., 1994) • Seafood: biological method (Association of Official Analytical Chemists, AOAC, 1995a) • Seafood: chemical method (AOAC, 1995b) • Seafood: fluorometric method (AOAC, 1995c) • Seafood: modification of AOAC method (Rogers and Staruszkiewicz, 1997) • Fish: capillary zone electrophoresis (Trenerry et al., 1998) Rapid testing methods are also available for histamine detection (Hungerford et al., 1997; Miller et al., 1997; Price, 1999). For example, commercial competitive ELISA kits ALERT (sensitivity 5–50 ppm, qualitative, 2 h) and Veratox (sensitivity , 5 ppm, quantitative from 0 to 50 ppm, 1 h) are sold by Neogen, USA and Elisa Systems, Australia. There are also several enzyme immunoassay test kits, including Histamarine Test Kit (sensitivity 0.5 ppm, quantitative from 1 to 500 ppm, 1 h) from Immunotech, France, which is approved by the AOAC (Price, 1999). The ELISA kits provide good reproducibility and ‘‘low’’ false negative and false positive rates when tested against the official AOAC fluorometric method for histamine (Anonymous, 1999). They are still too expensive for routine testing of numerous samples on a quality-assurance or preventive-testing basis, but are suitable for disease outbreak investigatory studies.

2.5.3. Analysis of other biogenic amines and chemical quality index The assessment of the biogenic amine content in foods has received much attention for many years due to the possibility of using amine concentrations as an index of food quality, and has included various HPLC methods besides those for histamine (Draisci et al., 1998). Because histamine alone is not always useful as an indicator of fish quality, Mietz and Karmas (1977) established a chemical quality index of canned tuna for estimating the extent of decomposition in fresh tuna prior to canning. They used qualitative and quantitative HPLC to examine the relationship of dansyl derivatives of five amines (histamine, putrescine, cadaverine, spermine and spermidine) extracted from canned fish. They prepared standards of pure

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

putrescine, cadaverine, spermine and spermidine at concentrations of 1 mg / 100 ml, and the level of each of the five compounds in canned fish samples was calculated on a ppm basis to generate an index of tuna decomposition. The resulting chemical index scores, based on the formula below, compared favourably to organoleptic value scores and ‘‘authentic pack’’ samples collected and decomposed under controlled conditions: ppm histamine 1 ppm putrescine 1 ppm cadaverine Index 5 ]]]]]]]]] 1 1 ppm spermine 1 ppm spermidine

Since it measures several different compounds resulting from several different processes of decomposition, this method could be used as a chemical indicator for decomposition of tuna. However, it is too complex for routine screening. Wagener (1984) prepared dansyl derivatives of extracted amines from fresh and spoiled fish using the method of Hui and Taylor (1983) and examined them by HPLC. He confirmed the earlier findings, that levels of histamine, putrescine and cadaverine rise rapidly during the initial stages of fish spoilage. He also identified tryptamine in tuna, hake and mackerel and tyramine in mackerel, but found that levels of these did not change much during storage. Interestingly, there were unidentified peaks on the chromatograms, some increasing and some decreasing during spoilage. An unknown peak marked ‘‘X’’ showed an increase in both tuna and hake, and was prominent in canned mackerel. Wagener (1984) speculated that the peak could be the substance ‘‘saurine’’ noted by Japanese workers in spoiling saury (Kawabata et al., 1955) (Foo, 1976, identified saurine as the phosphate salt of histamine). The unidentified compound was apparently not tested by Pauly’s reagent, and may have been the same as Shewan’s compound 8 (Shewan, 1955) or Hughes’ compound X (Hughes, 1959). Mietz and Karmas (1977) also found a dansylated peak (shoulder) between cadaverine and histamine in decomposed canned tuna that remained unidentified. Rogers and Staruszkiewicz (1997) of the United States FDA published details of a gas chromatographic method for putrescine and cadaverine in seafood. When 14 laboratories analysed 14 samples of canned tuna and raw mahi-mahi (including blind duplicates and a spike) containing 0.2–2.6 ppm putrescine and 0.6–9.1 ppm cadaverine using the

13

method, recoveries ranged from 71 to 102% for putrescine and from 77 to 112% for cadaverine. AOAC International has adopted the method for determination of putrescine in canned tuna and cadaverine in canned tuna and mahi-mahi. More recently, an Italian group has developed an improved method for the simultaneous determination of underivatised biogenic amines (cadaverine, putrescine, spermidine, histamine and tyramine) and amino acid precursors (histidine and tyrosine) in food products. The method, which is based on ion-exchange chromatography (IC) with integrated pulsed amperometric detection (IPAD), was used successfully for the analysis of biogenic amines and amino acids in food of vegetable and animal origin (pilchards) and in fermented foods such as cheese and salami. The main advantages of this method over a previous HPLC method coupled with IPAD are its application to a larger number of analytes and matrices, a simpler extraction and clean-up procedure, and an improved chromatographic separation and a lower detection limit. The IC–IPAD method is suitable for the detection of biogenic amines in a large number of samples and is particularly useful for routine checks in repetitive analyses (Draisci et al., 1998).

2.5.4. Analysis of urocanic acid Although he did not postulate that urocanic may be a candidate scombroid toxin, Baranowski (1985) suggested that urocanic acid may be a useful alternative to histamine as a spoilage index in scombroid and other fish rich in endogenous histidine. He used thin-layer chromatography to analyse for urocanic acid using standards in 95% alcohol. Morrison et al. (1980) and Caron et al. (1982) have used HPLC methods.

3. Dose–response assessment This section has been approached by looking at the incidence of HFP; fish characteristics that affect the clinical response; the nature and amount of bacterial contamination in relation to histamine content; histamine levels and toxic dose; human factors that affect the clinical response; and morbidity and mortality rates.

14

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

3.1. Incidence of histamine fish poisoning HFP occurs throughout the world and is perhaps the most common form of toxicity caused by the ingestion of fish (Mines et al., 1997). However, there are no reliable statistics on its incidence. Many countries do not have adequate systems for reporting foodborne diseases. In countries that do, cases are often missed because of the mild nature of the illness, which causes patients not to seek medical attention, or ignorance by medical personnel, who may misdiagnose HFP as a food allergy. HFP is not a reportable disease, even in those countries that keep records (Taylor, 1986) and its under-reporting is a worldwide problem (FDA, 1992). Since 1970, the countries with the most reported incidents of HFP are Japan, the United States and the UK, although this most likely represents better reporting by these countries. Less frequent outbreaks have been reported in many other countries.

3.1.1. Africa Although no information was found on the incidence of HFP in Africa, Ababouch et al. (1986) discussed histamine levels in commercially processed fish in Morocco, indicating that the problem had existed there. Morocco has a large fisheries industry based in the Atlantic Ocean, with sardines the predominant fish landed. Fishmeal is the main product, followed by canned sardines, tuna and mackerel, and fresh or frozen fish. In 1986, worldwide sales of Moroccan fish products were falling, owing to complaints about the quality of canned products and several outbreaks of food poisoning in Europe implicating Moroccan products. Morocco has since developed a progressive strategy to meet European standards for fish products, including the use of mandatory in-plant hazard analysis and critical control point (HACCP)-based quality control systems (Ababouch, 1997). 3.1.2. Asia HFP probably occurs frequently in Asia. High levels of histamine in fish on sale in Asian countries have been reported in various FAO Fisheries reports (Mitchell, 1993). There are also reports of extremely high levels of histamine in some salted, and dried or salted, fermented products in Asia. Histamine has

also been a problem in canned products from Asia (James and Olley, 1985). HFP was recognised as a major cause of illness in Japan in the early 1950s, and remains a major foodborne disease in that country. Forty-two outbreaks involving 4122 cases were reported by the Ministry of Health and Welfare, Japan, in 1970– 1980. Incriminated fish included mackerel, tuna, anchovies, sardines and marlin. The largest outbreak yet recorded in the world, involving 2656 cases, occurred in Japan in 1973 from the consumption of dried horse mackerel (Trachurus japonicus). Given the Japanese preference for raw fish, it is surprising that cooked fish have been involved in more incidents than raw fish. The utilisation of only the highest quality fish in the raw fish market of Japan is probably the reason for this (Taylor, 1986).

3.1.3. Australia and New Zealand There is little in the scientific literature on outbreaks of HFP in Australia. Incidents have been reported associated with the consumption of tailor (Pomatomus saltatrix) (Taylor, 1985), juvenile Western Australian salmon (Arripis truttaceus) (Smart, 1992), and tuna of unknown source (Brown, 1993). From 1973 to 1975, several incidents of histamine poisoning occurred in New Zealand associated with consumption of canned mackerel, smoked kahawai, kingfish and trumpeter fish. In addition, New Zealand authorities assisted in the investigation of an outbreak in 1974 involving canned skipjack tuna in the Solomon Islands (Taylor, 1985). Incidents involving smoked kahawai, kingfish and canned mackerel in the early 1970s were reported and discussed by Foo (1975a,b, 1976, 1977). Other reports involved kingfish and smoked kahawai (Mitchell, 1984) and smoked mackerel (Mitchell and O’Brien, 1992). In a report on HFP prepared in 1993 for the New Zealand Ministry of Health, survey data from 1992 to 1993 on histamine levels in (mostly smoked) fish were presented, together with a summary of recent (1990–1993) HFP outbreaks in New Zealand (Mitchell, 1993). Nineteen outbreaks were reported between March 1990 and June 1993, all resulting from the consumption of smoked fish, including kahawai, mackerel, marlin and trevally. Mitchell’s (1993) report concluded that, because of the number of poisoning incidents and the number of samples exceeding the acceptable level of histamine

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

in retail smoked fish, further monitoring and investigation of the causes of high levels of histamine in fish were warranted.

3.1.4. Canada In Canada, HFP was first confirmed in 1975, although there were probably earlier incidents. In the 1970s and 1980s, canned tuna was the most likely vehicle of contamination, but between 1990 and 1995 vehicles were fresh tuna, mahi-mahi or marlin. The three largest outbreaks were caused by the ingestion of smoked mackerel (October 1987: 14 cases), fresh marlin (July 1991: 12 cases) and fresh tuna (August 1994: 12 cases). Most incidents in Canada have been associated with imported fish. Histamine levels of 28–710 mg / 100 g were implicated in seven separate incidents involving tuna imported from Sri Lanka (Todd et al., 1992; Todd, 1997). 3.1.5. Europe An early outbreak of HFP in Great Britain occurred among a group of British sailors aboard the Triton of Leith in 1828 (Henderson, 1830). Five of the crew became ill following the consumption of bonito, a scombroid fish and a likely vehicle for HFP. After this early incident, reports of HFP in Great Britain were scarce or non-existent until 1976. From 1976, numerous outbreaks have been reported in England, Wales and Scotland associated with mackerel, tuna, sardines, herring and pilchards (e.g., Gilbert et al., 1980; Bartholomew et al., 1987). Scoging (1991) described incidents of ‘‘scombrotoxic fish poisoning’’ in the United Kingdom between 1976 and 1990. There were 441 suspected incidents involving 962 cases. Scoging (1998) investigated 405 incidents between 1987 and 1996: 243 sporadic incidents (60%), 105 general outbreaks (26%) and 56 family outbreaks (14%). Tuna (fresh / frozen and canned) and mackerel were most commonly implicated in both suspected and confirmed (elevated histamine) incidents. Salmon was involved in 30 suspected incidents, but remnants from only one incident contained slightly elevated histamine levels (5.1 mg / 100 g). Bacterial contamination occurred both during harvesting and processing and during food retail and preparation, for example after canned tuna was opened (Scoging, 1998). Sockett (1991) reported on food poisoning out-

15

breaks associated with manufactured foods in England and Wales between 1980 and 1989. Thirty-five outbreaks associated with processed fish and shellfish were reported during this period, including two outbreaks caused by imported products. ‘‘Scombrotoxin poisoning’’ was the most common cause of illness, and gave rise to 20 outbreaks affecting 59 people. All outbreaks were associated with mackerel products, including smoked mackerel (18), mackerel pate´ (1) and filleted mackerel (1). The symptoms were those of histamine poisoning. Molinari et al. (1989) reviewed the hygiene and health importance of histamine toxicity in Italy and concluded that the problem was greatly underestimated. Tuna and mackerel were most commonly associated with toxicity. An outbreak of HFP occurred in Palermo in 1979 that affected 250 people. In 25 years (1966–1991), 76 poisoning incidents after intake of tuna were reported to the Swiss Toxicological Information Centre (Maire et al., 1992). Twenty-seven reports came from physicians and, of these, 18 fulfilled the criteria of ‘‘scombroid fish poisoning’’.

3.1.6. United States It is difficult to determine the true incidence of HFP in the United States, because of incomplete record keeping. Local health departments recommend, but do not require, that cases be reported. Statistics on the disease are also inaccurate because of misdiagnosis (Lange, 1988). Data compiled by the CDC in the mid-1970s indicated no particular season for HFP and named Hawaii and California as the states with largest number of outbreaks (Hughes et al., 1977). In 1980, histamine poisoning was one of the most prevalent forms of foodborne disease of chemical aetiology in the United States, ranking only behind ciguatera fish poisoning (Sours and Smith, 1980). By 1997, scombroid toxicity or HFP was the most prevalent form of seafood-borne disease in the United States. Lipp and Rose (1997) reported outbreaks of seafood-borne disease associated with fish, 1983–1992, and listed the aetiological agents as ‘‘scombroid’’ (57% of outbreaks), ciguatoxin (19%), bacteria (14%), unknown (9%) and chemicals (1%). In contrast to the Japanese situation, most outbreaks in the United States have involved a small number of people, typically fewer than five individuals.

16

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

Between 1970 and 1974, ‘‘scombroid fish poisoning’’ was responsible for 29 (43%) of the 68 outbreaks caused by fish and shellfish toxins. In 45% of outbreaks, laboratory tests demonstrated elevated histamine levels in the incriminated fish. Tuna and mahi-mahi were most frequently implicated (Hughes et al., 1977). In 1970, some 40 children in a school lunch program became ill after eating imported canned tuna. A larger outbreak of HFP occurred in 1973, involving commercially canned tuna (Merson et al., 1974). There were 254 clinical cases in eight states. Nine assays for histamine produced values ranging from 68 to 280 mg / 100 g, with controls at 3 mg / 100 g. FDA recalled the incriminated lots, which showed evidence of honeycombing, indicating advanced decomposition (Merson et al., 1974; Arnold and Brown, 1978). This outbreak is probably still the largest on record in the United States, and it signalled an increasing awareness of histamine poisoning in that country (Taylor, 1986).

3.2. Fish characteristics that affect the clinical response Fish that commonly cause HFP include mackerel (Scomber spp.), tuna (Thunnus spp.), skipjack tuna (Katsuwonus pelamis), saury (Cololabis saira) and bonito (Sarda spp.) (scombroid fish); and mahi-mahi (Coryphaena spp.), sardines (Sardinella spp.), pilchards (Sardina pilchardus), anchovies (Engraulis spp.), herring (Clupea spp.), marlin (Makaira spp.) and bluefish (Pomatomus spp.) (non-scombroid fish) (Taylor, 1986). Non-scombroid species such as Western Australian salmon (Arripis truttaceus), sockeye salmon (Oncorhynchus nerka), amberjack (Seriola spp.) and Cape yellowtail (Seriola lalandii) have also been implicated in HFP (Lange, 1988; ¨ Muller et al., 1992; Smart, 1992; Gessner et al., 1996). As non-scombroid fish can cause HFP, the term scombroid fish poisoning is a misnomer (Taylor, 1988). Since free histidine in fish muscle is the substrate for microbial decarboxylation to produce histamine, species difference in histidine content has a large effect on the potential hazards of poor handling practices (Pan, 1985b). Although the muscle of fatty, red-meat, active and migratory species contains much more free histidine than that of white-meat, slower species, the latter still contain combined

histidine as a component of muscle protein. In the active fish, the free amino acid may act as a buffer, protecting the tissues from the effects of sudden increases in lactic acid. Histidine deaminase (HAL) and urocanase, which mediate the first two steps of the degradative process, regulate the level of histidine in the live fish (Love, 1980). Arnold and Brown (1978) discussed factors such as the effect of pH and temperature on the amount of histamine produced in ‘‘red-meat fish’’ such as mackerel and ‘‘white-meat fish’’ such as rockfish. The terms ‘‘red-meat fish’’ and ‘‘white-meat fish’’ are based on the general surface appearance of the fish. The terms ‘‘red muscle’’ and ‘‘white muscle’’ refer to muscle types within individual fish, red muscles being those containing predominantly red fibre types and white muscles being those containing predominantly white fibre types. All scombroid fish contain both muscle types in varying amounts, depending on the species. Takagi et al. (1969) examined the amounts of histidine and histamine in 21 species of aquatic animals during spoilage and found that more histamine was produced in red-meat fish such as mackerel than in white-meat fish such as rockfish. Suyama and Yoshizawa (1973) found relatively high free histidine contents in 13 species of migratory red-meat fish (286–1460 mg / 100 g), whereas whitemeat fish were low in free histidine (0–38.2 mg / 100 g) and did not produce histamine during spoilage. However, within the same species of fish, more histidine was found in white than in red muscle, and the resulting histamine formation followed this pattern with regard to muscle type; that is, more histamine in white muscle (Takagi et al., 1969; Pan, 1985b). Klausen and Lund (1986) demonstrated differences in amine formation between a scombroid and a non-scombroid fish. They chose mackerel (Scomber scombrus) and herring (Clupea harengus), because in Denmark the methods of catching, handling and manufacturing of the two species are very much alike, but only mackerel causes HFP. The contents of free histidine and free lysine were 4–5-times higher in fresh mackerel than in fresh herring. When stored in vacuum packs at 2 or 108C, both species accumulated similar small amounts of histamine at the time of rejection by sensory means. However, larger amounts of cadaverine were formed in mackerel than

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

in herring, and cadaverine values for mackerel corresponded to about 2–5-times the histamine values, before rejection. When exposed to higher temperatures (20–308C) and contamination with histidine-decarboxylating bacteria, mackerel accumulated much higher levels of both histamine and cadaverine than herring. Only moderately toxic levels of histamine (50 mg / 100 g) could be formed from the smaller amounts of free histidine present in herring (Klausen and Lund, 1986). Cadaverine and putrescine are probably formed in fish by enzymes from bacteria not involved in histidine decarboxylation, and cadaverine occurs in spoiled fish more frequently than histamine. The amount varies, but levels of 10–60 mg / 100 g have often been found. Putrescine levels in spoiled fish are usually much lower than those of cadaverine, usually ,10 mg / 100 g, probably because of the limited quantities of ornithine in fish tissues. Spoiling fish that accumulate cadaverine and / or putrescine, but not histamine, are not known to be toxic (Taylor and Sumner, 1986). Whereas data compiled by the CDC in the mid1970s indicated no particular season for HFP (cited by Lange, 1988), there seems to be a definite connection between season, temperature, histidine content and histamine formation in fish products (Pan, 1985b). Free histidine content in herring varied with seasons from 260 to 1600 mg / kg, being highest in summer (Hughes, 1959). Skovgaard and Ellemann (1978) reported that confirmed histamine poisoning outbreaks caused by smoked mackerel occurred during late summer, when both fish and water have the highest temperature, and the mackerel are at their fattest. Clifford et al. (1989) noted that mackerel show quite marked fish-to-fish and seasonal variations in chemical composition and susceptibility to spoilage by various microbiological, enzymic and auto-oxidative processes. The authors concluded that it is possible that only certain fish in an apparently homogenous batch will develop all factors necessary to induce HFP. Concentrations of histamine in fish tend to vary widely from one part of the fish to another. Histamine in raw fish is usually present at higher levels in tissue adjacent to the gills or the intestines, the main reservoirs of HDB (Lerke et al., 1978; Taylor, 1986). Lerke et al. (1978) showed that the dis-

17

tribution of histamine in spoiled tuna is quite uneven, varying more than four-fold over 3 cm, and being considerably higher near the gut cavity than elsewhere. Frank et al. (1981) observed a similar histamine gradient in spoiled skipjack tuna. Yoshinaga and Frank (1982) attributed this to the non-uniform distribution of spoilage organisms throughout the fish, and noted that muscle levels of histidine were essentially uniform in fresh skipjack tuna. Hardy and Smith (1976) showed that the histamine content of ungutted mackerel was 10-times more than that of gutted fish after storage at ambient temperature for 140 h.

3.3. Nature and amount of bacterial contamination in spoiling fish 3.3.1. The common histidine-decarboxylating bacteria Although only about 1% of the surface microflora of live fish represents histamine producers (Kimata, 1961), HDB form a greater proportion of the microbial population as a fish spoils. Omura et al. (1978) reported that 31% of isolates from spoiled skipjack tuna and jack mackerel growing at warm temperatures were HDB, and Yoshinaga and Frank (1982) found 13.4% in decomposing skipjack tuna. Frank and Baranowski (1984) found with mahi-mahi that 7% of isolates growing at warm temperatures and 9% growing at refrigeration temperatures were HDB. Taylor et al. (1978a) identified 112 species of bacteria that are known to possess HD. Most were members of the family Enterobacteriaceae, or the genera Clostridium and Lactobacillus. Enteric bacteria, specifically Morganella morganii, certain strains of Klebsiella pneumoniae and a few strains of Hafnia alvei are the most prolific histamine producers in fish when they are maintained at temperatures greater than 48C (Stratton and Taylor, 1991) and are the most commonly associated with scombrotoxic fish (Taylor, 1983). Certain strains of Lactobacillus that are also prolific histamine producers are probably of importance only in fermented fish (Taylor, 1986). Niven et al. (1981) identified Vibrio spp. as HDB, and histamine-producing isolates of V. alginolyticus have been found in decomposing skipjack tuna and mahi-mahi (Frank et al., 1985). V. harvei, V. fisheri and Photobacterium leiognathi are capable of

18

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

producing histamine at warm temperatures (Ramesh and Venugopalan, 1986). Other Photobacterium spp. and Vibrio spp. may be primarily responsible for histamine production at lower temperatures (Morii et al., 1986; Van Spreekens, 1987). Van Spreekens (1987) maintained that because photobacteria are thermolabile and have special requirements for growth they have often been overlooked as HDB and causative agents in HFP.

3.3.2. Temperature–growth relationships and histamine production Doe et al. (1998) pointed out that the traditional classification of bacteria into psychrophiles, psychrotrophs, mesophiles and thermophiles was only arbitrary and that there was a gradation from one category to the other. Because of the overlapping in growth curve characteristics and inconsistencies in the use of this terminology in the literature, the traditional classification will not be used here. Rather, the reader is asked to utilise the concept displayed in Fig. 1 below from McMeekin et al. (1993), where the square root of the bacterial growth

rate is plotted against temperature for organisms typical of four thermal classes, groups A, B, C and D. For histamine production at refrigeration temperatures, bacteria in group A are nearing their maximum growth rate and those in group B are growing slowly. At temperate ambient temperature and during cold smoking in temperate areas (Anonymous, 1988a), the growth of group A bacteria has slowed markedly, that of group B is close to the maximum, and group C are growing slowly. At tropical ambient temperature and temperatures used for cold smoking in tropical areas (Chng and Kuang, 1987), group C bacteria are almost at their optimum growth rate, group B are almost at their upper limit of temperature, and group D are beginning to grow well. In temperatures of the hot smoking kiln (Anonymous, 1988b), bacteria in groups A, B and C have been eliminated, but those in group D, if present, may not be destroyed. From these data, it is evident that there are no clear-cut distinctions between the temperature groups, but rather a continuum of values. The references Anonymous (1988a,b) above are

Fig. 1. Growth characteristics for four categories of bacteria at temperatures used in fish processing and storage (adapted from Fig. 10.7 of McMeekin et al., 1993). Histamine producers can be found across most of the temperature–growth spectrum: Photobacterium phosphoreum (group A); several Vibrio species (group B); Klebsiella pneumoniae, Morganella morganii and Hafnia alvei (group C); and group D bacteria, if present.

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

extracted from extensive tables on the preparation and composition of dried and smoked fishery products in 12 countries. The lower limits for production of toxicologically significant levels of histamine in tuna fish infusion broth for some common HDB were 78C for K. pneumoniae, 158C for two M. morganii strains, and 308C for H. alvei, Citrobacter freundii and E. coli (Behling and Taylor, 1982). Because of the importance of temperature in the production of histamine, several attempts have been made to predict histamine formation in spoiling fish at different temperatures. For example, Frank (1985) constructed normographs to predict histamine production in skipjack tuna, and Pan (1985a) used Arrhenius plots for such estimations in mackerel and bonito. However, the methods involved too many assumptions and were inaccurate in their predictions. A further development that took temperature fluctuations into account, temperature function integrators, was also considered unsuitable for predicting the complicated process of histamine production in spoiling fish (Olley and McMeekin, 1985). It is clearly not possible to relate the amount of bacterial contamination to histamine production by such methods. Numerous factors introduce bias. First, a reasonable estimate of the initial numbers of significant bacteria needs to be made. Then histamine concentration represents the amount produced by potentially diverse population of bacteria with different capacities for histamine production and / or destruction in association with bacteria that are not histamine producers or destroyers. Even when one group or strain of HDB predominates, histamine production is unlikely to follow a growth curve fitting that group / strain, not least because metabolism of histamine by histaminase may be taking place simultaneously. Further, the amount of histamine production varies from site to site within a fish, and substrate changes such as proteolysis at high temperatures affect the amount of histidine available for histamine formation. Subsequent cooking or processing of a spoiled fish will further alter the relationship between bacterial numbers and histamine production by reducing or removing the microbial population without affecting the histamine content significantly. For example, Fletcher et al. (1995) reported that, when whole fresh kahawai (Arripis trutta) was stored under a

19

variety of temperature regimes, elevated levels of histamine (defined as $50 mg / kg) occurred only when aerobic plate counts at 20 or 358C exceeded 10 6 colony forming units (CFU) / g. There was only one sample with elevated histamine where the aerobic plate count was ,10 7 CFU / g. When such fish were smoked, bacterial numbers, but not histamine concentration and toxicity, were considerably reduced (Fletcher et al., 1998).

3.3.3. Studies on variability in species /strain composition of microflora Scientists have found that factors affecting the dominating species / strains of HDB in the microflora of a particular spoiling fish are many and varied. The diversity can be attributed to differences in the species of fish, handling procedures, holding times and temperatures. In addition, the character of the microflora can be influenced by the fish’s feeding habits, geographical location, the season, ocean temperature, etc. (Kimata, 1961; Shewan, 1977; Yoshinaga and Frank, 1982). It is not surprising that results have been variable in many studies on the effect of storage temperature of histamine formation in fish (Behling and Taylor, 1982). HDB that grow at refrigeration temperatures usually produce histamine in smaller quantities than species that grow at warmer temperatures, and toxic levels may not be reached. Okuzumi et al. (1984a) looked at the occurrence of various histamine-forming bacteria on / in fresh fish purchased from fish markets in Japan. For the summer samples, M. morganii was found in the greatest numbers, followed by the ‘‘N-group’’ bacteria, later identified as Photobacterium phosphoreum by Fujii et al. (1997). For the winter samples, only N-group bacteria were detected. This is consistent with the observation of Simidu and Hibiki (1954) of less histamine in sardines and bonito during winter than in other seasons under the same handling conditions. In a subsequent study, Okuzumi et al. (1984b) studied microbial populations at various storage temperatures in mackerel bought at fish markets, and correlated these with histamine production. In fish samples stored at 58C and 108C, the main histamine formers were N-group bacteria, which reached levels of 10 7 –10 8 / g after 7 days. In samples stored at 158C, the formation of histamine was still attributed mainly to N-group bacteria, particularly in the first 3

20

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

days, although numbers of M. morganii increased throughout. In samples stored at 208C, a great part of histamine at the early part of storage (0–4 days) was attributed to N-group bacteria, and at a later stage (6–8 days) to M. morganii. At the final stage (11–14 days), histamine content decreased from .700 mg / 100 g, and remained constant at about 500 mg / 100 g. In samples stored at 258C and 308C, M. morganii was the main histamine producer. Ryser et al. (1984) obtained 60 isolates of indigenous bacteria from raw tuna, and identified them as Pseudomonas spp. Fewer than half (35%) the isolates produced histamine during incubation at 218C for 48 h, and the maximum amount produced by a single isolate was only 3.2 mg / 100 ml. This would suggest that bacteria that grow at lower temperatures are not significant toxicologically. However, Okuzumi et al. (1981) reported significant histamine production by halophilic (salt-loving) N-group bacteria growing at 58C. Husain (1996) estimated the growth rates of different HDB such as Vibrio spp., Bacillus spp. and Morganella spp. incubated at 288C in seawater broth of different salt concentrations. At concentrations ranging from 1% to 10%, a visible growth was seen in all cultures. Maximum growth occurred at 4%, 3% and 7% salt for Vibrio spp., Bacillus spp. and Morganella spp., respectively. Husain (1996) also found that traditional Indian preservatives garlic, turmeric, ginger and black pepper all had a profound inhibitory effect on the growth of HDB isolates at concentrations of 1 to 5%. There was a large decrease in bacterial growth rate as concentrations increased. Garlic was the most effective, completely inhibiting growth of Vibrio spp. at 4% concentration and Bacillus spp. and Morganella spp. at 5%. Earlier, workers in Japan (Wendakoon and Sakaguchi, 1992) had tested the effects of powdered spices and their ethanol extracts on growth and biogenic amine formation of Enterobacter aerogenes and M. morganii. Clove and cinnamon were the best inhibitors of the spices tested (allspice, cardamom, chilli, cinnamon, clove, cumin, black pepper, nutmeg, sage and thyme).

3.3.4. Species /strain variation in histidine decarboxylase activity Different bacteria vary significantly in the quantity of HD they produce and / or the specific activity of

their decarboxylases. According to Ferencik (1970), M. morganii (then Proteus morganii) required histidine levels of 100–200 mg / 100 g to induce the production of HD. By contrast, H. alvei decarboxylated histidine when it was present at ,50 mg / 100 g. This suggests that bacteria differ according to their relative importance as histamine producers on different species of fish. Eitenmiller et al. (1981,1982) investigated factors influencing HD production by M. morganii. Examination of 22 strains of M. morganii revealed that each possessed HD activity, although at variable levels. With one high-activity strain, maximal HD activity occurred at 378C and pH 6.5. Normal muscle pH of fresh yellowfin tuna (6.5) thus corresponds closely to the pH required for optimal activity of HD, and is low enough to permit rapid enzyme synthesis by M. morganii. Minimal enzyme activity was present when the culture was grown at pH 8.5. HD activity decreased as the age of the culture increased. Rapid enzyme and histamine formation occurred in yellowfin tuna fillets inoculated with M. morganii and stored at 24 and 308C, and histamine soon reached toxic levels at both temperatures. Little enzyme activity was present in inoculated fillets stored at 158C, and in control fillets that were not inoculated. Olley and Baranowski (1985) pointed out that low-temperature enzyme activity by microorganisms that grow at warm temperatures is important in histamine formation, provided sufficient bacterial numbers have been reached before cold storage. Klausen and Huss (1987b) studied growth and histamine production by M. morganii in histidinecontaining broth and in mackerel. Following storage at higher temperatures (10–258C), large amounts of histamine were formed at low temperatures (0–58C), when no growth took place. Fujii et al. (1994) found that the specific activity of HD of halophilic histamine-forming bacteria Photobacterium phosphoreum and P. histaminum sp. nov. remained at 27% and 53% of the initial value after storage in cell suspensions for 7 days at 48C and 2208C, respectively, while viable cell numbers decreased. This indicates that outbreaks of HFP may be caused by the ingestion of frozen–thawed fish even when the viable count of histamine forming bacteria is low. A common feature of studies on HDB is that when multiple strains of the same species are isolated, only

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

one or a few of these strains are prolific histamine producers (Gale, 1946; Voigt and Eitenmiller, 1977; Taylor et al., 1978a; Taylor et al., 1979; Yoshinaga and Frank, 1982). Most strains of M. morganii are able to produce histamine, but only certain strains of K. pneumoniae and H. alvei are described as potent histamine producers (Havelka, 1967; Taylor et al., 1978a; Taylor et al., 1979). Taylor et al. (1979) isolated a high-histamine-producing strain of K. pneumoniae from a sample of tuna sashimi implicated in an outbreak of HFP. Only 12 / 50 other K. pneumoniae strains isolated from foods, representing five distinct biochemical types, were equally efficient histamine producers. Behling and Taylor (1982) divided HDB into those species capable of producing large quantities of histamine (.100 mg / 100 ml) in tuna fish infusion broth during a short incubation period (,24 h) at temperatures .158C and those capable of producing somewhat lesser quantities (,25 mg / 100 ml) after prolonged incubation (.48 h) at 308C or above. M. morganii, K. pneumoniae and E. aerogenes were prolific histamine producers, and tested strains of H. alvei, C. freundii and E. coli were slow producers (Taylor et al., 1978a; Behling and Taylor, 1982). Taylor et al. (1978a) suggested that most of the other bacteria identified as HDB in the scientific literature are in the slow-producer category. However, in reporting that M. morganii, K. pneumoniae and H. alvei were the only bacteria that had been isolated from fish implicated in fish poisoning incidents, Taylor (1985) postulated that certain other bacteria may also fall into the prolific-histamine-producer classification. He cited as an example the isolation of Clostridium perfringens from decomposing skipjack tuna by Yoshinaga and Frank (1982). Leung (1987) found that of 36 bacterial isolates from spoiled chub mackerel skin washings all the Morganella strains (10) produced large quantities of histamine. However, of two strains of Acinetobacter calcoaceticus isolated, one was the strongest producer of all strains tested whereas the other gave a totally negative result. Similarly, only 1 / 3 K. pneumoniae strains and 1 / 4 Aeromonas hydrophila strains were histamine producers. In attempting to explain her results, Leung (1987) postulated that HD may be controlled by a plasmid and that the plasmid may be transferred from strain to strain, species to species or genus to genus. Barancin et al. (1998)

21

have recently found that the fish pathogen Vibrio anguillarum harbours a plasmid-encoded HD gene. Many bacteria (especially of the Enterobacteriaceae family) possess either ornithine decarboxylase or lysine decarboxylase, the enzymes necessary to produce potential potentiators of histamine, putrescine and cadaverine, respectively (Taylor and Sumner, 1986). Since only a few bacteria possess HD responsible for histamine production, it is likely that the bacteria forming histamine differ from those forming putrescine and cadaverine. Leung (1987) found that of 36 bacterial isolates from spoiled chub mackerel skin washings none had the ability to decarboxylate all three amino acids, histidine, lysine and ornithine.

3.3.5. Bacterial destruction of histamine The histamine levels in a toxic fish depend on free histidine levels in the muscle and the balance between histamine production and histamine destruction by the contaminating microflora (Okuzumi et al., 1984b). Histaminase (DAO) activity has been detected in several types of bacteria including M. morganii, Vibrio spp. and Klebsiella spp. (Gale, 1942; Satake et al., 1953; Ienistea, 1971). Gale (1942) found that bacterial histaminase is best produced under somewhat alkaline conditions (pH 7.5–8), but that moderate activity also occurs under slightly acidic conditions. Yamanaka (1984) showed that in saury, mackerel, yellowtail and big-eyed tuna the overall rate of histamine production was greater at 208C than at 358C, while in skipjack tuna more histamine was produced at 358C. The lower levels of histamine production at high temperatures in the former species was attributed either to histaminases or to the presence of HDB with a growth optimum of ,358C (Olley and McMeekin, 1985). Ferencik (1970) reported that a strain of M. morganii, after being inoculated into a sterile tuna flesh homogenate, produced large amounts of histamine. However, a significant amount of histamine was soon decomposed by the organism. A subsequent sterile addition of histidine to the inoculated homogenate again resulted in histamine formation followed by histamine destruction. He concluded that the amount of histamine production and destruction by the M. morganii strain was determined by the concentration of free histidine in the fish homogenate.

22

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

3.4. Histamine levels and toxic dose While the presence of histamine in fish muscle is a good indication that decomposition has taken place, its occurrence is extremely variable. As is evident from the above, its production is a function of the species of fish and the individual fish, the part of the fish sampled, time, temperature, and the types and numbers of bacteria present (Rawles et al., 1996). The threshold toxic dose for histamine in foods is not precisely known (Taylor, 1986). The variation in histamine levels within a spoiled fish (Lerke et al., 1978; Frank et al., 1981) makes it difficult to estimate. Simidu and Hibiki (1955) estimated the threshold toxic dose for histamine in fish at about 60 mg / 100 g, but their methods were not precise. Shalaby (1996) reviewed the oral toxicity to humans of histamine and other biogenic amines in fish muscle. He considered that histamine-induced poisoning is, in general, slight at 8–40 mg / 100 g fish, moderate at .40 mg / 100 g and severe at .100 mg / 100 g. Based on an analysis of recent poisoning episodes, he suggested the following guideline levels for histamine content of fish: (i) ,5 mg / 100 g (safe for consumption); (ii) 5–20 mg / 100 g (possibly toxic), (iii) 20–100 mg (probably toxic), and (iv) .100 mg / 100 g (toxic and unsafe for human consumption). This assessment correlates with those of Simidu and Hibiki (1955) and Mitchell (1993) and emphasises that poisoning does occur at lower histamine concentrations than 100 mg / 100 g fish. There is uncertainty regarding the threshold toxic concentration because potentiators of toxicity may be present in fish and lower the effective dosage compared with pure histamine (Institute of Medicine, 1991). Different fish could contain different potentiators, and the levels of potentiators could also vary considerably from one fish to another. The types and levels of potentiators in a fish would depend on a variety of factors, including the natural constituents of the fish, the contaminating bacteria and their metabolic capabilities, and the environmental conditions (mainly temperature). Until the identity, levels and potency of possible potentiators and / or mast cell degranulating factors are elucidated, it is difficult to establish regulatory limits for histamine in foods on the basis of potential health hazard. FDA guidelines for tuna, mahi-mahi and related

fish specify 500 ppm (50 mg / 100 g) as the toxicity level, and 50 ppm (5 mg / 100 g) as the defect action level because histamine is not uniformly distributed in a decomposed fish. Therefore, if 50 ppm is found in one section, there is a possibility that other units may exceed 500 ppm (FDA, 1998). These levels, based on years of investigative experience, allow for the non-uniform distribution of histamine in a spoiled fish, but ignore the existence of suspected potentiators or other toxins. The acceptable level set as a manufacturing standard by legislation in the United Kingdom in 1992 was 10 mg / 100 g (Anonymous, 1998b). For control of histamine in fish belonging to the Scombridae and Clupeidae families, European Union Directive No. 91 / 493 stipulates that nine independent samples from each batch should correspond to: (1) An average histamine concentration lower than 100 ppm (10 mg / 100 g). (2) No more than two samples out of the nine with a concentration of between 100 and 200 ppm. (3) No sample with a histamine content higher than 200 ppm. In Australia and New Zealand, the Australian Food Standards Code (ANZFA, 1998a) states that histamine must not exceed 100 mg / kg in a composite sample of fish or fish products, other than crustaceans and molluscs. However, this is currently under review, with a proposal by the Australia New Zealand Food Authority (ANZFA) to increase the maximum allowable level of histamine in fish and fish products to 200 mg / kg (ANZFA, 1998b). A level of 10 mg histamine / 100 g of fish seems conservative and appropriate to protect public health and safety. If this were regarded as the highest level that can be consumed safely, this figure must be related to the amount of fish eaten and the weight of the person to calculate a likely safe dose. If a 60 kg person eats, say, 300 g (wet weight) of this fish, the safe dose would be 0.5 mg / kg bodyweight. Such a calculation is of limited value, however, considering the variable nature of HFP and the lack of understanding of its pathogenesis.

3.5. Human factors that affect the clinical response 3.5.1. Variation in individual susceptibility The severity of the clinical response will depend

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

on the amount of toxin(s) ingested and variation in individual susceptibility. The literature would suggest that there is a big variation in the latter. This was illustrated in the experiment of Motil and Scrimshaw (1979), where oral administration of 100–180 mg histamine in grapefruit juice or in 100 g high-quality tuna caused characteristic symptoms of mild histamine poisoning in only 1 / 4 and 4 / 8 volunteers, respectively. According to Taylor et al. (1989), attack rates in group outbreaks of histamine poisoning vary from 50 to 100%. In the human volunteer trial of Van Gelderen et al. (1992), only 2 / 8 subjects (both female) exhibited symptoms after eating portions of spoiled fish containing 90 mg histamine, and these two volunteers did not have higher concentrations of histamine in plasma than volunteers who did not show symptoms. In the four cases of ‘‘classic scombrotoxism’’ described by Blakesley (1983), amounts of histamine ingested were not known. The nature of the symptoms varied between the subjects. Most (3 / 4) exhibited burning skin, headache and nausea, 2 / 4 had abdominal cramps and pruritus, and 1 / 4 had hive-like skin lesions and palpitations.

3.5.2. Influence of diet Geiger (1955) suggested that alterations in the intestinal tract caused by seasoned hot dishes prepared from spoiled fish or the simultaneous consumption of alcoholic beverages might cause histamine to be absorbed at an increased rate, such that its detoxification could not keep up with its entry into the circulation. Zee et al. (1981) also suggested that histamine may be absorbed more rapidly in the presence of alcohol and exert a more-marked biological effect on the circulatory system. However, in rats ethanol enhanced DAO and HMT activity to 123 and 111% of normal, respectively (Taylor and Lieber, 1979). Mitchell and Code (1954) reported that histamine taken with a meal (bread, butter and milk) was absorbed to a greater extent than histamine consumed by itself. Granerus (1968) also studied the effect of diet on histamine metabolism in humans. His studies indicated the variability in urinary excretion of histamine and its main metabolite, N-methyl-imidazoleacetic acid. He quoted a study by Irvine et al. (1959) that concluded that some intestinal bacteria could in fact contribute to urinary histamine by decarboxylating

23

histidine in food. While most proteins differ by only 2–3% in their histidine content, these proteolytic bacteria are favoured by a diet rich in animal protein and inhibited by a diet that favours the acidophilic group of bacteria (e.g., vegetables, grains, lactose and dextrin). Thus, dietary differences may explain the highly varying amounts of histamine metabolite found in the urine of some subjects. Johnson and Overholt (1967) observed a significant increase in the concentration of histamine in the gastric venous blood of dogs whose stomachs had been filled with dilute acetic acid solutions. He suggested that the lemon or vinegar commonly used to enhance the flavour of fish might also enhance the absorption of histamine from toxic fish. Acids in lemon or vinegar may affect the pH of the intestinal contents and inhibit the activity of histaminemetabolising enzymes. A number of other foodborne substances are known to inhibit histamine metabolism (Taylor, 1986). Amines such as histamine, cadaverine, putrescine and tyramine are found in meat and meat products, cheeses and other fermented foods such as sauerkraut, beverages such as wine and beer, and fruits and vegetables (Shalaby, 1996) and in Chinese foods such as tamari and soy sauce (Chin et al., 1989). Thus, dietary items other than fish containing histamine and other amines might exacerbate HFP if consumed at the same time as spoiled fish. In addition, the increasing popularity of mixed fish dishes such as seafood marinara may promote HFP by providing histamine potentiators from different fish species eaten in the same meal.

3.5.3. Influence of medication Some drugs inhibit histamine-metabolising enzymes and potentiate histamine activity when taken in conjunction with food containing high concentrations of histamine (Taylor, 1986; Chin et al., 1989). On the other hand, antihistamines taken for an unrelated reason may protect individuals from HFP to some extent (Taylor et al., 1989). HMT is inhibited by analogues of methylmethionine such as adenosyl-homocysteine, antimalarial drugs, and numerous agonists and antagonists of histamine receptors. Some antihistaminic drugs and aminoguanidine are inhibitors of DAO. Several compounds once classified as specific MAO inhibitors are known to inhibit DAO as well, includ-

24

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

ing isoniazid, an anti-tuberculosis medication that has often been incriminated as an exacerbating factor in HFP (Uragoda and Kottegoda, 1977; Senanayake and Vyravanathan, 1981; Taylor, 1986; Stratton et al., 1991).

3.5.4. Disease states and age Histidine metabolism in humans can be altered in certain disease states. Histidinaemia, an inborn error of metabolism resulting from the absence of HAL, is associated with increased excretion of histamine and its metabolites. Patients with this disease or with altered histamine metabolism may be more susceptible to HFP. Exogenous histamine would also exacerbate diseases such as allergies in which elevated endogenous histamine levels play a role, and mastocytosis (Taylor, 1986). Activities of histidine- and histamine-metabolising enzymes may also be diminished with age.

3.6. Morbidity and mortality rates HFP is usually a rather mild illness (Taylor, 1986). In rare cases it has proved fatal (Eitenmiller et al., 1981). Arnold and Brown (1978) reported that morbidity values vary from 0.07 to 100% in the literature, but said that such figures are misleading, as different lots of canned fish may differ greatly in their histamine content. The distribution of histamine in individual fish is also highly variable, as is individual susceptibility to histamine poisoning. In group outbreaks the attack rate may approach 100%, but when an outbreak involves raw fish it is sometimes substantially below 100% (Taylor, 1986; Wu et al., 1997). The lack of any visual sign of spoilage in scombrotoxic fish probably accounts for the high morbidity rate.

4. Exposure assessment In discussing exposure assessment in relation to HFP, we will examine, in a non-quantitative way, factors affecting the probability of HFP occurring, amounts of fish consumed and at-risk population groups, and future exposure trends.

4.1. Factors affecting the probability of histamine fish poisoning occurring 4.1.1. Post-catching contamination It appears that HDB, which have been isolated from the skin, gills, intestines and muscle tissues of spoiling fish (Lerke et al., 1978), can originate from the marine environment or from post-catching contamination (Arnold and Brown, 1978). Kimata (1961) and Yoshinaga and Frank (1982) estimated that histamine formers occupied about 1% of the regular surface microflora of fresh fish and suggested that, with extended storage at elevated temperatures, the bacteria invade the muscle and convert free histidine to histamine. However, in their study of HDB contaminating fish in retail markets in Spain, ´ Lopez-Sabater et al. (1996) found that Plesiomonas shigelloides was the only HDB frequently associated with the marine environment. Most scientists believe that post-harvesting contamination is the main source of histamine formers. Taylor and Speckhard (1983) devised a method for the recovery of HDB from frozen skipjack tuna obtained from a major tuna packer in the United States. Gills, intestines and muscles were sampled. HDB (M. morganii and C. freundii) were isolated from 3 / 10 fish, in each case from the gills only. The evidence suggested that Enterobacteriaceae are not part of the normal microflora of tuna and the isolation of these bacteria from the gills, which are usually used as ‘‘hand-holds’’ in handling, was attributed to post-harvesting contamination. Taylor et al. (1989) went further to say that most of the histamine formers found in fish are common enteric bacteria of humans and animals. However, the status of C. perfringens and N-group bacteria that grow at refrigeration temperatures cannot be confirmed, since early recovery procedures used did not cater for their detection. Scoging (1991, 1998) also attributed outbreaks of HFP to fish becoming contaminated after harvesting. In the United States, most outbreaks of HFP are associated with improperly handled fish from private catches (Lange, 1988). HFP has also been linked to restaurant-prepared meals (Yamani et al., 1981) and to canned tuna, both of which involve commercial fishing operations (Lange, 1988). Post-catching contamination with HDB may occur at several levels – aboard the fishing vessel, at the processing plant, in

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

the distribution system (fresh and frozen fish), and at the level of the user. Restaurant contamination could be particularly important with raw tuna (Taylor, 1986). To understand the sources of contamination of fish with HDB, Subburaj et al. (1984) investigated the fish market environment of Mangalore, India, for counts of these bacteria. They cultured samples from carrying baskets, ice, the market floor, and water for wetting fish. They also recorded total plate counts, HDB counts and histamine levels in samples of seerfish and mackerel. The results demonstrated that HDB were widely distributed in the market environment and water. The generic composition of HDB in a mackerel sample containing 20 mg histamine / 100 g revealed Morganella (52%), Pseudomonas (21%), Plesiomonas (12%), Providencia (8%), Flavobacteria (6%) and Aeromonas (1%). In multiple fish samples, there was no direct correlation between HDB count and histamine level, confirming that the generic composition of the microflora is more important in histamine formation than numbers. If fish are stored ungutted, the gut itself may be a source of contamination, particularly if chilling is delayed. For example, if anaerobic bacteria of the gut such as C. perfringens (a prolific histamine producer) are allowed to proliferate at high temperatures, they will produce enzymes that are highly active at lower temperatures (Olley et al., 1985).

4.1.2. Temperature abuse on fishing vessels At any time between catching and consumption, exposure of certain fish to elevated temperatures can cause formation of histamine from histidine by HDB, which are inevitably present. Quality assurance on board fishing vessels is the first link in the quality chain that should run from the start of production to final consumption. The key to the reduction of histamine production is the rapid cooling of the fish after catching (Ritchie and Mackie, 1979). Formation of histamine can be induced by shortterm temperature abuse early in the chilling step and detected later (Andersen, 1997). Initial cooling is important in reducing the rate of histamine production, even when temperature rises at a later stage (Mitchell, 1993). Many fish species commonly implicated in outbreaks of HFP are caught in warm water, and fish temperatures at capture frequently exceed 208C. Fish may remain in purse–seiner nets

25

or on longlines at this elevated temperature for several hours after capture, and fish may die before they leave the water. Tuna have higher body temperatures than most other fish. In the western Pacific, where sea surface temperatures are 27–308C, tuna often come on board with internal temperatures of $328C. Spoilage at this temperature is about 30-times faster than at 08C. Because of elevated temperatures, the length of time from death until tuna are chilled is the most critical phase of shipboard handling (Bartram, 1997a,b). Once on the fishing vessel, fish may or may not be cooled, and methods of cooling employed vary widely in their efficiency. American purse seiners cool fish in holds filled with refrigerated seawater. Once holds are full, the fish can be frozen. The rate of cooling depends on the size of the catch and on the size of individual fish. The fish may be held at elevated temperatures for some time, and are not gutted until they reach the processing plant. Any time a fish is held at greater than about 48C significantly reduces the expected safe shelf life, but fish that have been handled particularly well on board the harvest vessel may safely withstand somewhat more exposure to elevated temperatures during post-harvest handling (Taylor, 1986; Price, 1999).

4.1.3. Inadequate chill-storage procedures Taylor (1986) cited a number of studies on the effect of storage temperature on histamine formation in various types of fish. While all the studies agreed that histamine formation is negligible in fish stored at 08C or below, other results were variable. Widely varying data exist for both the lower temperature limits for safe storage and for the optimal temperature for producing histamine. This is not surprising, given the variability of the nature of the microbial populations on fish. While low-temperature storage effectively controls the growth of most HDB, bacteria that grow at refrigeration temperatures can produce smaller amounts of histamine in fish stored at temperatures between 0 and 108C (Ritchie and Mackie, 1979; Klausen and Huss, 1987b; Stratton and Taylor, 1991). Regardless of the species involved, bacteria must grow to a large enough population for significant production of histamine to occur. Geiger (1944a) found that histamine concentration in scombroid fish

26

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

increased from 0.09 mg / 100 g when fresh to 95 mg / 100 g when kept at room temperature for 10 h. In fish subjected to elevated temperatures even for short periods, a large population of bacteria is soon established. During subsequent refrigeration, although bacterial growth ceases, residual enzyme activity continues slowly and histamine levels continue to increase (Klausen and Huss, 1987b; Stratton and Taylor, 1991). Ritchie and Mackie (1979) monitored the formation of histamine, putrescine, cadaverine, spermine and spermidine in freshly caught mackerel and herring held ungutted on ice for 28 days at 18C, in an incubator at 108C, or in an insulated box at ambient temperature (258C). In both mackerel and herring, the concentrations of histamine and other amines were quite low after prolonged storage at 18C, even when the fish were putrid. As expected, the amines were produced in relatively larger amounts at the two higher temperatures. In an experiment involving nine lots of sardines with undocumented post-harvest history bought at a fish market in Morocco, Ababouch et al. (1986) demonstrated the effect of refrigeration on histamine development. Sardines stored at 88C had a 12-h longer shelf life than those held at 178C. A combination of salting and refrigeration was more effective. Fish held at 88C and salted at a level of 5% or 8% had a shelf life 35 h longer than fish held at 178C with no salt. Chen and Malison (1987) reported typical scombroid symptoms caused by mackerel stored on ice for 2 days and then kept at room temperature of 308C for 3–4 h before cooking. Toxin production apparently occurred at a rapid rate, because people who had lunch from 12.30 to 1 pm had a much higher attack rate than those who ate 1–2 h earlier. Other fish from the same catch did not cause illness after standing at room temperature for only 1 h. Shipments of unfrozen fish packed in refrigerated containers have posed a significant problem in the United States because of inadequate temperature control (FDA, 1999).

4.1.4. Inadequate freezing and thawing procedures A study by Ben-Gigirey et al. (1998) demonstrated the importance of good handling practices during the thawing of fish, since the absence of certain bacteria potentially involved in spoilage is

not guaranteed, even after prolonged frozen storage, and the levels of some biogenic amines may increase. Ben-Gigirey et al. (1998) investigated the changes in biogenic amines and numbers of bacteria reported to have decarboxylase activity in albacore (white tuna) muscle during frozen storage at 2188C or 2258C. The initial contents of all the amines, except for spermidine (11.9 mg / 100 g), were ,0.8 mg / 100 g. Putrescine decreased after 6 months, but increased after 9 months, of frozen storage to show the largest increase of all the amines (up to 942%). Histamine decreased after 3 months, but increased to 103% of the initial level after 9 months, storage at 2188C. There was no increase in histamine in samples stored at 2258C. Cadaverine contents tripled or doubled after 3 months of storage at 218 or 2258C, respectively, and decreased thereafter. Spermine increased slightly as frozen storage progressed. Spermidine persisted at the initial range during the first 6 months of frozen storage, but after 9 months the final concentrations of this amine were significantly lower than the initial level. Aerobic bacteria that grow at ambient temperatures survived frozen storage at 2188C (39%) and 2258C (92%) for 9 months. The survival rate at 2258C of bacteria that grow at refrigeration temperatures was 4.6% after 9 months.

4.1.5. Temperature abuse in the preparation of dried and /or smoked products Toxic levels of histamine have been found in dried and / or smoked products of mackerel, horse mackerel and sardines (Taylor, 1983), as well as in the fishmeal made from these fish. Exposure of raw fish to high ambient temperatures accelerates this reaction. Histamine content increases to a maximum, then decreases with prolonged drying time. The drying of sardines previously brined in 5–15% sodium chloride for 2 h caused pronounced increases in histamine (Pan, 1988). Trinidad et al. (1986) reported on histamine production in Spanish mackerel smoked for 8 h at 458C and stored at refrigeration temperature (0–58C) as 3-mm thick slices packed in flexible plastic films under vacuum or not. Histamine increased in both kinds of packs, but not to a significant extent because the fish used for smoking were very fresh, kept chilled, and contamination during handling and processing was kept to a minimum. Vacuum-packed

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

samples had lower microbial counts at the middle of storage, but at the end (day 30) the difference between the two packs was negligible. Van Spreekens (1987) reported high levels of histamine in vacuum-packed, lightly salted herring fillets stored at refrigeration temperatures, particularly after low-level contamination with histamineproducing photobacteria. The salt concentration of the product was a vital factor in their growth (0.5– 4%). In the same paper, histamine production in salted or hot-smoked mackerel was attributed to Photobacterium spp., which grow at refrigeration temperatures. The fish were exposed to temperature abuse as a result of delayed cooling on board ship and / or during processing. Hot smoking ‘‘practically sterilises’’ the product and denatures enzymes, thus imparting some degree of preservation, but does not destroy histamine already formed (Poulter, 1988). Bremer et al. (1998) carried out thermal death trials using a H. alvei strain isolated from a portion of hot-smoked kahawai with a histamine level of 166 mg / 100 g. Results of the trials, carried out in 0.1% peptone suspensions and in kahawai at 54–588C, indicated that hot smoking has the potential to eliminate H. alvei from seafood products. Fletcher et al. (1998) reported the results of a retail survey of the levels of histamine in hot-smoked fish products available in New Zealand, where most cases of HFP are associated with smoked fish. They purchased 107 samples from Auckland retail markets between July 1995 and March 1996 and determined their histamine and bacterial levels. Eight samples (from 5 / 9 retail markets) contained .50 mg / kg histamine. In two samples, histamine levels were .200 mg / kg (346 and 682 mg / kg). Within or between fish species there were no consistent relationships between levels of histamine in the samples and either the total aerobic plate counts or the numbers of HDB. Some of the samples with elevated levels of histamine had relatively low microbial counts. For example, the sample with the highest histamine level had microbial counts of ,6?10 3 CFU / g. Histamine had been formed prior to smoking, and HDB were eliminated during smoking.

4.1.6. Poor canning procedures In the past, incidents of HFP caused by high levels of histamine in commercially canned scombroid

27

products occurred all over the world (Taylor, 1983). This was largely the result of temperature abuse before canning. However, improvements in handling and processing associated with the establishment of HACCP procedures are becoming widespread. These offer the advantage of in-process control, while avoiding the need for large-scale terminal analyses of finished product, which is time consuming and beyond the capabilities of many small canneries (Warne, 1985).

4.1.7. Low-quality fermented products Fermented fish containing high levels of histamine could conceivably present a hazard. In Finland, sugar–salted herring in barrels is more likely to contain high levels of histamine than canned herring, but has not been implicated as causing illness. More research is needed to define suitable limits for histamine in fermented products (Taylor, 1985). 4.1.8. Temperature abuse of raw tuna for the sashimi market The principal hazards in the lucrative raw tuna trade are decomposition and histamine formation. Both are linked to temperature abuse. Raw tuna buyers are presented with so much quality variation that each fish is inspected and graded prior to purchasing. The highest quality raw tuna is sold in sushi bars. Grading fulfils the objectives of HACCP, because raw tuna core temperatures are checked and fish are organoleptically examined at critical control points (Bartram, 1997a,b). 4.2. Amounts and types of fish consumed, and atrisk population groups Figures were not available for consumption of various fish species in different countries. Even if they were found, their value would be limited in assessing possible exposure to HFP, because details of handling and storage conditions would not be known. All population groups are susceptible to HFP, but there are regional differences in the amounts and types of fish consumed and the way they are processed and stored. HFP still occurs frequently in developed countries such as the United States, Japan and the United Kingdom (Taylor, 1986). The disease is probably also common in developing countries where fish

28

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

preserved by traditional methods is an important part of the diet. Such diets remain extremely popular and are a major source of inexpensive dietary protein. Their popularity is largely due to desirable flavours produced by chemical changes that occur during the smoking and drying processes. Traditional African products are rarely salted, but some may be sundried before smoking starts. In the course of sun and air drying, bacteria can multiply in the moist interior of the fish, which has high water activity, causing the inside to spoil while the outside appears satisfactory. The problem of HFP would be expected to be greater in tropical countries where higher ambient temperatures promote the growth of prolific HDB. In tropical countries, fish are cold smoked at 30–408C, depending on species, to prevent coagulation of protein, which occurs with cooking. This process is designed to give a desired flavour to the product rather than a significant degree of preservation (Poulter, 1988). It will not destroy HDB or deactivate HD. Even if hot smoking follows cold smoking, histamine concentrations may have already become high enough to become hazardous. Products cured in the traditional way have become less popular in most developed countries. However, in some non-tropical countries such as the United Kingdom, fish are cold smoked at low temperatures (,308C) (Anonymous, 1988a). The modern, mildly smoked and less extensively dried products are less stable than the traditional ones, and the trend towards these products has been made possible by chilling and freezing. Neither cold- nor hot-smoked fish will keep for long periods unless supplementary techniques such as drying, freezing or chilling are used (Poulter, 1988). The increasing popularity of raw fish probably presents less risk of HFP because only the highest quality fish are directed to this trade. Also, as the product is a delicacy, smaller quantities of fish would be likely to be consumed per meal. Fermented fish products are also consumed in smaller amounts than fresh or smoked fish, for example, and so higher concentrations of histamine could probably be tolerated in these products.

seafood quality are improving and exposure to spoiled seafood and thus HFP is likely to become less prevalent. Developing countries have increased their net income from fish and fishery products from about US$3 billion to some US$18 billion over the past 10 years (FAO, 1999). By 1997, they had equalled the production of the developed world, contributing almost 50% of global fishery exports (Karnicki, 1997). Great progress has been made in the quality of fish products at the same time as the huge expansion of international trade. This is the result of the introduction of international standards in food hygiene and the application of risk analysis and HACCP principles as part of a common, global approach for maximising the quality and safety of all food products (FAO, 1999). HFP is now much less likely to occur as a result of commercial canning or smoking as HACCP quality assurance measures are increasingly being adopted. In Australia, for example, the Australian Quarantine and Inspection Service, as part of Agriculture, Fisheries and Forestry – Australia, has introduced HACCP and International Standards Organization 9000 quality-management inspection systems. These initiatives, together with those of other government and industry agencies, have encouraged the Australian export seafood industry and other sectors of the processed food industry to implement qualitymanagement systems that address issues such as good handling practices and temperature control (Aitken, 1997). Predictive microbiology is an emerging area of food microbiology in which microbial responses to environmental factors are measured under defined and controlled conditions. The potential applications of predictive modelling in HACCP systems are numerous (Ross and McMeekin, 1995). Although predictive modelling cannot be applied to histamine production directly, it has an important future role in the potential to control microbiological spoilage of fish and thus will assist indirectly in the control of HFP.

4.3. Future exposure trends 5. Risk characterisation In the past 10 years, authorities and producers around the world have become more aware of the need for quality assurance in relation to food (Barker and McKenzie, 1997). As a result, standards for

Although much is known about HFP, there are still many gaps in our knowledge. Sufficient data are not available at this stage for an adequate risk characteri-

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

sation and thus risk assessment. Further strategies for risk management and risk communication can be put in place only after additional data are generated and analysed.

5.1. Nature and magnitude of risk 5.1.1. Impact on human health HFP does not have a major impact on human health. It is generally a mild but unpleasant disease, likely to cause discomfort for only a few hours. The morbidity rate may approach 100% following the consumption of toxic fish, but mortalities are rare. However, the disease may be more serious in people taking certain medicines or with pre-existing pulmonary, cardiac or renal problems. Because HFP is generally mild and of short duration, and responsive to antihistamine therapy, small outbreaks are unlikely to be reported, even in the daily press. However, HFP is important from the food-safety aspect. Even in countries where random monitoring is carried out, it is possible that products will escape this safety net from time to time. Consumers are becoming more demanding in regard to food safety. Litigation following food poisoning incidents is becoming more common, and producers, distributors and restaurants will increasingly be held liable for the quality of the products they handle and sell. 5.1.2. Impact on fishing industries Major outbreaks of HFP attract media attention, which affects fish consumption and has a negative impact on the marketing of seafood. Outbreaks caused by imported fish products seriously affect the trade of the exporting country. Although such events are becoming increasingly less likely because of the widespread adoption of HACCP analysis and quality assurance, they will still occur occasionally owing to incidents such as those caused by equipment failure, human error or negligence. 5.2. Uncertainties and problem areas in risk characterisation 5.2.1. Defining histamine fish poisoning and elucidating its pathogenesis The mechanism of toxicity of HFP is still not clearly understood, which is unsatisfactory. More research is needed to determine the role that potentiators or other toxins may play in causing the

29

disease. It is particularly important that HFP be clearly differentiated from syndromes caused by endogenous toxins such as ciguatoxins that may be present in finfish from time to time (Lehane, 2000). As suggested by Mitchell (1993), there may be potential in fractionation of the chemical constituents of toxic fish and assessment of the toxicity of different single and combined fractions. In addition, it is the view of the authors that little attention has been given to the pKa values of the carboxyl, imidazole and amine groups of the compounds discussed in this review, although pKa greatly influences biological properties. A list of pKa values can be found in any handbook of data for biochemical research (e.g., Dawson et al., 1969), and is the pH at which half of the ionisable group is in the dissociated and half is in the unionised form. It is significant that the pKa of the imidazole group is in the normal physiological range. Similarly, in separating these compounds found in spoiling fish chromatographically, little consideration has been given to the pH of the solvent or to the stabilising effects of solvent counter-ions. Future regulatory action may need to take into account the impact of potentiators, and the identification of such substances in various foods. When analysing suspect fish for histamine, the simultaneous detection of other common putrefactive amines such as cadaverine and putrescine could be advantageous. The safety of urocanic acid has been investigated only in relation to its use in cosmetic products and sunscreens (Cosmetic Ingredient Review Expert Panel, 1995). Animal toxicity studies have involved only topical administration, so its toxicity profile is incomplete. While animal data indicated that the substance has immunosuppressant properties, clinical data were inconclusive. The Cosmetic Ingredient Review Expert Panel (1995) reported that ‘‘it cannot be concluded that urocanic acid is safe in cosmetic formulations’’. The possible role of urocanic acid in HFP as a potential mast cell degranulator may be a suitable research project, beginning with oral toxicity studies in laboratory animals. When the pathogenesis of the disease is understood, clinical HFP can be defined more accurately and making a differential diagnosis will be easier. There is a need to assess the actual incidence of fish allergy and to determine what percentage of cases diagnosed as fish allergy represent misdiagnosis of

30

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

HFP. The latter is differentiated from allergy by the occurrence of clusters of affected people rather than single cases, and by the application of skin tests using non-toxic fish extracts.

5.2.2. Investigating and managing post-harvesting contamination More research still needs to be done on postharvesting contamination in order to improve quality-control procedures. Identification of points at which temperature abuse and / or bacterial contamination occur would include the gathering of information, where this has not been done already, on current fish harvesting, transporting, storage, processing and retailing practices for the species identified as those most likely to cause HFP. For example, temperature abuse may occur on recreational boats that lack adequate refrigeration; and Gellert et al. (1992) suggested that rinsing fish in seawater during handling and cleaning could promote contamination with bacteria capable of metabolising histidine to histamine. Regardless of the origin of the spoilage bacteria, histamine serves as an indicator of microbial spoilage in histidine-containing fish, and many countries have defined maximum histamine levels for fish products. Histamine analysis as a means of monitoring spoilage is associated with a number of problems. It is time consuming and / or expensive, and samples need to be taken from multiple sites to compensate for variations in histamine content throughout the spoiled fish. Perhaps more important is the lack of standardisation of histamine detection methods employed around the world, and the multitude of tests available. There is a need for global standardisation of histamine detection methods, and laboratory accreditation and proficiency testing. A rapid and cheap assay for detecting histamine in fish would also be of value, particularly to recreational fishers. The current ELISA tests are rapid, but not cheap. For the above reasons, together with the fact that monitoring fish histamine levels may not always ensure protection from HFP, a method other than direct measurement of histamine may be preferable for quantitative measurement of quality deterioration. Baranowski (1985) suggested that urocanic acid may be a useful alternative to histamine as a spoilage index in scombroid and other fish that are rich in

endogenous histidine. This idea should be investigated, especially in the light of new knowledge that cis-urocanic acid is a mast cell degranulator (Wille et al., 1999). A study of the correlation between the amine content and bacterial counts needs to be done. Ideally, each amine would be correlated with its respective amine-producing bacteria. The detection of histamine, cadaverine and putrescine can be achieved satisfactorily by TLC or HPLC. A solid medium that allows direct enumeration of histidine, lysine and ornithine decarboxylase activities does not appear to be available. Government regulatory agencies require the removal of contaminated fish from the marketplace when human illness occurs, which should lead to prevention of additional cases. However, mechanisms need to be in place to allow efficient and complete traceback of incriminated fish to point of origin, in order to rectify problems leading to spoilage. At present in Australia, tracing fresh or frozen fish from commercial outlets is difficult, because of a lack of a ticketing system, especially in the middle (wholesaler) stages of distribution (Rawlin and Herfort, 1999). This is a problem area that could be improved considerably, given the necessary resources and cooperation between stakeholders. Quality control programs in the tuna-canning and other fish-processing industries have largely eliminated HFP as a concern in these products (Taylor et al., 1989).

Acknowledgements This review was derived from a larger review written for, and available from, the National Office of Animal and Plant Health, Agriculture, Fisheries and Forestry – Australia, Canberra: Histamine (scombroid) fish poisoning: a review in a risk-assessment framework, L. Lehane and J. Olley, 1999. The authors are grateful to many personnel from the following organisations for their assistance: Agriculture, Fisheries and Forestry – Australia, the University of Tasmania, Queensland Department of Primary Industries, Australia New Zealand Food Authority, Australian Government Analytical Laboratories, Port Lincoln Tuna Processors Pty Ltd., and Elisa Systems.

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

References Ababouch, L., 1997. Application of the HACCP principles in the fish-canning industry in Morocco. In: Martin, R.E., Collette, R.L., Slavin, J.W. (Eds.), Fish Inspection, Quality Control, and HACCP: A Global Focus, Proceedings of the Conference held 19–24 May 1996, Arlington, VA, Technomic, Lancaster, Basel, pp. 302–310. Ababouch, L., Alaoui, M.M., Busta, F.F., 1986. Histamine levels in commercially processed fish in Morocco. J. Food Prot. 49 (11), 904–908. Aitken, L., 1997. A regulatory approach to HACCP – the Australian perspective. In: Martin, R.E., Collette, R.L., Slavin, J.W. (Eds.), Fish Inspection, Quality Control, and HACCP: A Global Focus, Proceedings of the Conference held 19–24 May 1996, Arlington, VA, Technomic, Lancaster, Basel, pp. 613– 618. Andersen, E.M., 1997. Quality assurance on board fishing vessels. In: Martin, R.E., Collette, R.L., Slavin, J.W. (Eds.), Fish Inspection, Quality Control, and HACCP: A Global Focus, Proceedings of the Conference held 19–24 May 1996, Arlington, VA, Technomic, Lancaster, Basel, pp. 427–436. Anggawati, A., 1986. ACIAR Report on Visit by RIFT Personnel Australia in October–December 1986, unpublished results. Anonymous, 1988a. Dried and smoked fishery products: preparation and composition, Table 10, cold smoking. In: Burt, J.R. (Ed.), Fish Smoking and Drying: The Effect of Smoking and Drying on the Nutritional Properties of Fish. Elsevier Applied Science, London, New York, Chapter 9, p. 156. Anonymous, 1988b. Dried and smoked fishery products: preparation and composition, Table 8, hot smoking. In: Burt, J.R. (Ed.), Fish Smoking and Drying: The Effect of Smoking and Drying on the Nutritional Properties of Fish. Elsevier Applied Science, London, New York, Chapter 9, p. 148. Anonymous, 1998a. New system for seafood safety. Environ. Health Perspect. 106 (10), A475–A478. Anonymous, 1998b. The Food Safety (Fishery Products and Live Shellfish) (Hygiene) Regulations 1998. SI No. 994. HMSO, London. Anonymous, 1999. Rapid tests for seafood quality: histamine detection. Food Aust. 51 (3), 102. ANZFA, 1998a. Standard D1 – fish. In: Australian Food Standards Code (Issue 41, 18 November 1998), Australia New Zealand Food Authority, Canberra. ANZFA, 1998b. Development of Joint Australia New Zealand Food Standards as part of the process of the Review of the Food Standards Code. Fish and Fish Products. Explanatory notes, proposal / full assessment, P183. Australia New Zealand Food Authority, Canberra, August 1998. AOAC, 1995a. Histamine in seafood: biological method. Sec. 35.1.30, method 954.04. In: Cunniff, P.A. (Ed.), Official Methods of Analysis of AOAC International, 16th Edition, Association of Official Analytical Chemists (AOAC) International, Gaithersburg, MD, pp. 14–15. AOAC, 1995b. Histamine in seafood: chemical method. Sec 35.5.31, method 957.07. In: Cunniff, P.A. (Ed.), Official Methods of Analysis of AOAC International, 16th Edition, Association of Official Analytical Chemists (AOAC) International, Gaithersburg, MD, pp. 15–16.

31

AOAC, 1995c. Histamine in seafood: fluorometric method. Method 35.1.32, method 977.13. In: Cunniff, P.A. (Ed.), Official Methods of Analysis of AOAC International, 16th Edition, Association of Official Analytical Chemists (AOAC) International, Gaithersburg, MD, pp. 16–17. Arnold, S.H., Brown, W.D., 1978. Histamine toxicity from fish products. Adv. Food Res. 34, 113–154. Arnold, S.H., Price, R.J., Brown, W.D., 1980. Histamine formation by bacteria isolated from skipjack tuna, Katsuwonas pelamis. Bull. Jpn. Soc. Sci. Fish. 46, 991–995. Ascione, A., Barresi, L.S., Sarullo, F.M., De Silvestre, G., 1997. [Two cases of ‘‘scombroid syndrome’’ with severe cardiovascular compromise]. Cardiologia 42 (12), 1285–1288, In Italian, with English abstract. Barancin, C.E., Smoot, J.C., Findlay, R.H., Actis, L.A., 1998. Plasmid-mediated histamine biosynthesis in the bacterial fish pathogen Vibrio anguillarum. Plasmid 39 (3), 235–244. Baranowski, J.D., 1985. Low-temperature production of urocanic acid by spoilage bacteria isolated from mahimahi (Coryphaena hippurus). Appl. Environ. Microbiol. 50 (2), 546–547. Barker, J., McKenzie, A., 1997. Review of HACCP and HACCPbased food control systems. In: Martin, R.E., Collette, R.L., Slavin, J.W. (Eds.), Fish Inspection, Quality Control, and HACCP: A Global Focus, Proceedings of the Conference held 19–24 May 1996, Arlington, VA, Technomic, Lancaster, Basel, pp. 73–81. Bartholomew, B.A., Berry, P.R., Rodhouse, J.C., Gilbert, R.J., 1987. Scombrotoxic fish poisoning in Britain: Features of over 250 suspected incidents from 1976 to 1986. Epidemiol. Infect. 99, 775–782. Bartram, P.K., 1997a. Proper processing and handling of raw tuna. In: Martin, R.E., Collette, R.L., Slavin, J.W. (Eds.), Fish Inspection, Quality Control, and HACCP: A Global Focus, Proceedings of the Conference held 19–24 May 1996, Arlington, VA, Technomic, Lancaster, Basel, pp. 233–237. Bartram, P.K., 1997b. Proper shipboard handling of fresh tuna for the sashimi market. In: Martin, R.E., Collette, R.L., Slavin, J.W. (Eds.), Fish Inspection, Quality Control, and HACCP: A Global Focus, Proceedings of the Conference held 19–24 May 1996, Arlington, VA, Technomic, Lancaster, Basel, pp. 448– 457. Bateman, R.C., Eldrige, D.B., Wade, S., McCoy, M.J., Jester, E.L.E., Mowdy, D.E., 1994. Copper chelation assay for histamine in tuna. J. Food Sci. 59 (3), 517–518. Behling, A.R., Taylor, S.L., 1982. Bacterial histamine production as a function of temperature and time of incubation. J. Food Sci. 47, 1311–1317. Ben-Gigirey, B., Vieites Baptista de Sousa, J.M., Villa, T.G., Barros-Velaquez, J., 1998. Changes in biogenic amines and microbiological analysis in albacore (Thunnus alalunga) muscle during frozen storage. J. Food Prot. 61 (5), 608–615. Bjeldanes, L.F., Schutz, D.E., Morris, M.M., 1978. On the aetiology of scombroid poisoning: cadaverine potentiation of histamine toxicity in the guinea pig. Food Cosmetic Toxicol. 16 (2), 157–159. Blakesley, M.L., 1983. Scombroid poisoning: prompt resolution of symptoms with cimitedine. Ann. Emerg. Med. 12, 104. Bremer, P.J., Osborne, C.M., Kemp, R.A., van Veghel, P.W.C., Fletcher, G.C., 1998. Thermal death times of Hafnia alvei cells

32

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

in a model suspension and in artificially contaminated hotsmoked kahawai (Arripis trutta). J. Food Prot. 61 (8), 1047– 1051. Brown, C., 1993. Scombroid poisoning (Letter). Med. J. Aust. 158 (6), 435–436. Buchanan, R., 1998. Principles of risk assessment for illness caused by foodborne biological agents. J. Food Prot. 61 (8), 1071–1074, representing the National Advisory Committee on Microbiological Criteria for Foods. Cantoni, C., Bianchi, M.A., Beretta, G., Persiani, G., 1976. [Tuna flesh deterioration and histamine levels]. Arch. Vet. Ital. 27, 60–63, In Italian, English abstract. Caron, J.C., Martin, B., Shroot, B., 1982. High-performance liquid chromatographic determination of urocanic acid isomers in biological samples. J. Chromatogr. 230, 125–131. Cavanah, D.K., Casale, T.B., 1993. Histamine. In: Kaliner, M.A., Metcalfe, D.D. (Eds.), The Mast Cell in Health and Disease, Marcel Dekker, New York, Basel, Hong Kong, pp. 321–342, Chapter 11. Chen, K.-T., Malison, M.D., 1987. Outbreak of scombroid fish poisoning. Taiwan. Am. J. Public Health 77 (10), 1335–1336. Chin, K.W., Garriga, M.M., Metcalfe, D.D., 1989. The histamine content of oriental foods. Food Chem. Toxicol. 27 (5), 283– 287. 1987. Smoked products. In: Chng, N.M., Kuang, H.K. (Eds.), Southeast Asian Fish Products Marine Fisheries Research and Development, Southeast Asian Fisheries Development Center, Singapore, p. 28. Chu, C.H., Bjeldanes, L.F., 1981. Effect of diamines, polyamines and tuna fish extracts on the binding of histamine to mucin in vitro. J. Food Sci. 47, 79–88. Clark, R.B., 1997. Scombroid poisoning revisited (Letter). Ann. Emerg. Med. 29 (3), 426–427. Clifford, M.N., Walker, R., Wright, J., Hardy, R., Murray, C.K., 1989. Studies with volunteers on the role of histamine in suspected scombrotoxicosis. J. Sci. Food Agric. 47 (3), 365– 375. Clifford, M.N., Walker, R., Ijomah, P., Wright, J., Murray, C.K., Hardy, R., 1991. Is there a role for amines other than histamines in the aetiology of scombrotoxicosis? Food Addit. Contam. 8 (5), 641–652. Clifford, M.N., Walker, R., Ijomah, P., Wright, J., Murray, C.K., ¨ Hardy, R., Martlbauer, E.P., Usleber, E., Terplan, G., 1993. Do saxitoxin-like substances have a role in scombrotoxicosis? Food Addit. Contam. 6, 657–667. Cosmetic Ingredient Review Expert Panel, 1995. Final report on the safety assessment of urocanic acid. J. Am. Coll. Toxicol. 14 (5), 386–423. Dawson, R.M.C., Elliot, D.C., Elliot, W.H., Jones, R.M., 1969. Data for Biochemical Research, 2nd Edition, Clarendon Press, Oxford. Doe, P.E., Olley, J., Hard, N., Gopakumar, K., 1998. Methodology for quality measurements. In: Doe, P.E. (Ed.), Fish Drying and Smoking: Production and Quality, Technomic, Lancaster, Basel, pp. 117–135, Chapter 5. Downs, M., 1997. Scombroid poisoning revisited. Ann. Emerg. Med. 29 (3), 426–427. Draisci, R., Giannetti, L., Boria, P., Lucentini, L., Palleschi, L., Cavalli, S., 1998. Improved ion chromatography–integrated

pulsed amperometric detection method for the evaluation of biogenic amines in food of vegetable or animal origin and in fermented foods. J. Chromatogr. A 798 (1–2), 109–116. Eitenmiller, R.R., De Souza, S.C., 1984. Enzymic mechanisms for amine formation in fish. In: Ragelis, P. (Ed.), Seafood Toxins, American Chemical Society, Washington, DC, pp. 431–442, Chapter 6. Eitenmiller, R.R., Wallis, J.W., Orr, J.H., Phillips, R.D., 1981. Production of histidine decarboxylase and histamine by Proteus morganii. J. Food Prot. 44 (11), 815–820. Eitenmiller, R.R., Orr, J.H., Wallis, W.W., 1982. Histamine formation in fish: microbial and biochemical conditions. In: Martin, R.E., Flick, G.J., Hebard, C.E., Ward, D.R. (Eds.), Chemistry and Biochemistry of Marine Food Products, Avi, Westport, CT, pp. 39–50. Etkind, P., Wilson, M.E., Gallagher, K., Cournoyer, J., 1987. Bluefish-associated scombroid poisoning. An example of the expanding spectrum of food poisoning from seafood. J. Am. Med. Assoc. 258 (23), 3409–3410. FAO, 1999. FAO Expert Consultation on the Trade Impact of Listeria in Fish Products. University of Massachusetts, Amhurst, MA, 17–20 May 1999. FDA, 1992. Foodborne Pathogenic Microorganisms and Natural Toxins, Food and Drug Administration, Center for Food Safety and Applied Nutrition, Rockville, MD. FDA, 1998. FDA and EPA guidance levels. In: Fish and Fishery Products Hazards and Controls Guide, 2nd Edition, Department of Health and Human Services, Public Health Service, Food and Drug Administration, Center for Food Safety and Applied Nutrition, Office of Seafood, Washington, DC, pp. 245–248, Appendix 5. FDA, 1999. Foodborne Pathogenic Microorganisms and Natural Toxins Handbook. Food and Drug Administration, Center for Food Safety and Applied Nutrition, Rockville, MD (http: / / www-micro.msb.le.ac.uk / others / FDA / |mow / chap38.html, updated 19 / 6 / 99). Ferencik, M., 1970. Formation of histamine during bacterial decarboxylation of histidine in the flesh of some marine fishes. J. Hyg. Epidemiol. Microbiol. Immunol. 14, 52–60. ´ ´ J., Mackie, I.M., 1979. Histidine metabolism Fernandez-Salguero, in mackerel (Scomber scombrus). Studies on histidine decarboxylase activity and histamine formation during storage of flesh and liver under sterile and non-sterile conditions. J. Food Technol. 14, 131–139. Fletcher, G.C., Summers, G., Winchester, R.V., Wong, R.J., 1995. Histamine and histidine in New Zealand marine fish and shellfish species, particularly kahawai (Arripis trutta). J. Aquatic Food Prod. Technol. 4, 53–74. Fletcher, G.C., Summers, G., van Veghel, P.W.C., 1998. Levels of histamine and histamine-producing bacteria in smoked fish from New Zealand markets. J. Food Prot. 61 (8), 1064–1070. Foo, L.Y., 1975a. Scombroid-type poisoning induced by the ingestion of smoked kahawai. NZ Med. J. 81, 476–477. Foo, L.Y., 1975b. The content of histamine and fish poisoning. NZ Med. J. 82, 381–383. Foo, L.Y., 1976. Scombroid poisoning. Isolation and identification of ‘‘saurine’’. J. Sci. Food Agric. 27, 807–810. Foo, L.Y., 1977. Scombroid poisoning – recapitulation on the role of histamine. NZ Med. J. 85, 425–427.

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37 Frank, H.A., 1985. VI.1. Use of normographs to estimate histamine formation in tuna. In: Histamine Formation in Marine Products: Production by Bacteria, Measurement and Prediction of Formation, FAO Fisheries Technical Paper No. 252, Food and Agriculture Organization of the United Nations, Rome, pp. 18–20. Frank, H.A., Baranowski, J.D., 1984. Authentic fishery product samples, collected and decomposed under controlled conditions. Contract 223-80-2295. US Department of Health and Human Services, Washington, DC. Frank, H.A., Yoshinaga, D.H., Nip, W.-K., 1981. Histamine formation and honeycombing during decomposition of skipjack tuna, Katsuwonus pelamis, at elevated temperatures. Mar. Fish Rev. 43 (10), 9–14. Frank, H.A., Baranowski, J.D., Chongiriwatana, M., Brust, P.A., Premaratne, R.J., 1985. Identification and decarboxylase activities of bacteria isolated from decomposed mahimahi (Coryphaena hippurus) after incubation at 0 and 328C. Int. J. Food Microbiol. 2, 331–340. Fujii, T., Kurihara, K., Okuzumi, M., 1994. Viability and histidine decarboxylase activity of halophilic histamine-forming bacteria during frozen storage. J. Food Prot. 57, 611–613. Fujii, T., Hiraishi, A., Kobayashi, T., Yoguchi, R., Okuzumi, M., 1997. Identification of the psychrophilic histamine-producing marine bacteria previously referred to as the N-group bacteria. Fish Sci. 63, 807–810. Furuta, T., Okamiya, K., Shibasaki, H., Kasuya, Y., 1996. Pharmacokinetics of stable isotopically labelled L-histidine in humans and the assessment of in vivo histidine ammonia lysase activities. Drug Metab. Dispos. 24 (1), 49–54. Gale, E.F., 1942. The oxidation of amines by bacteria. Biochem. J. 36, 54. Gale, E.F., 1946. Bacterial amino acid decarboxylases. Adv. Enzymol. 6, 1–32. Garland, C.D., 1985. Unpublished results (cited by Leung, 1987). Geiger, E., 1944a. Histamine content of processed and uncanned fish. A tentative method of quantitative determination of spoilage. Food Res. 9, 293–297. Geiger, E., 1944b. On the specificity of bacterium decarboxylase. Proc. Soc. Exp. Biol. Med. 55, 11–13. Geiger, E., 1955. Role of histamine in poisoning with spoiled fish. Science 121, 865–866. Geiger, E., Courtney, G., Schnakenberg, G., 1944. The content and formation of histamine in fish muscle. Arch. Biochem. 3, 311–319. Gellert, G.A., Ralls, J., Brown, C., Huston, J., Merryman, R., 1992. Scombroid fish poisoning. Underreporting and prevention among noncommercial recreational fishers. West. J. Med. 157 (6), 645–647. Gessner, B., Hokama, Y., Isto, S., 1996. Scombrotoxicosis-like illness following the ingestion of smoked salmon that demonstrated low histamine levels and high toxicity on mouse bioassay. Clin. Infect. Dis. 23, 1316–1318. Gietzen, K., Penka, L., Eisenburger, R., 1996. Imidazole fungicides provoke histamine release from mast cells and induce airway contraction. Exp. Toxicol. Pathol. 48 (6), 529–531. Gilbert, R.J., Hobbs, G., Murray, C.K., Cruickshank, J.G., Young, S.E.J., 1980. Scombrotoxic fish poisoning: Features of the first

33

50 incidents to be reported in Britain (1976–1979). Br. Med. J. 281, 71–72. Granerus, G., 1968. Effects of oral histamine, histidine and diet on urinary excretion of histamine, methylhistamine, and 1-methyl4-imidazoleacetic acid in man. Scand. J. Clin. Lab. Invest. 22 (Suppl. 104), 49–58. Guss, D.A., 1998. Scombroid fish poisoning: successful treatment with cimetidine. Undersea Hyperb. Med. 25 (2), 123–125. Hardy, R., Smith, J.G.M., 1976. The storage of mackerel (Scomber scombrus). Development of histamine and rancidity. J. Sci. Food Agric. 27, 595–599. Hart, P.H., Jaksic, A., Swift, G., Norval, M., El-Ghorr, A.A., Findlay-Jones, J.J., 1997. Histamine involvement in UVB- and cis-urocanic acid-induced systemic suppression of contact hypersensitivity responses. Immunology 91 (4), 601–608. Havelka, B., 1967. Role of Hafnia bacteria in the rise of histamine in tuna fish meat. Cesk. Hyg. 12, 343–352. Helander, C.G., Lindell, S.-E., Westling, H., 1965. The renal 14 removal of C-labelled histamine from the blood in man. Scand. J. Clin. Lab. Invest. 17, 524–528. Henderson, P.B., 1830. Case of poisoning from the bonito (Scomber pelamis). Edinburgh Med. J. 34, 317–318. Hesterberg, R., Sattler, J., Lorenz, W., Stahlknecht, C.-D., Barth, H., Crombach, M., Weber, D., 1984. Histamine content, diamine oxidase activity and histamine methyltransferase activity in human tissues: fact or fiction? Agents Actions 14 (3–4), 325–334. Hughes, J.M., Horwitz, M.A., Merson, M.H., Barker, W.H., Gangarosa, E.J., 1977. Foodborne disease outbreaks of chemical aetiology in the United States, 1970–1974. Am. J. Epidemiol. 105 (3), 233–244. Hughes, R.B., 1959. Chemical studies on the herring (Culpea harengus). 2. The free amino acids of herring flesh and their behaviour during post-mortem spoilage. J. Sci. Food. Agric. 10, 558–564. Hui, J.Y., Taylor, S.L., 1983. High pressure liquid chromatographic determination of putrefactive amines in foods. J. AOAC Int. 66 (4), 853–857. Hui, J.Y., Taylor, S.L., 1985. Inhibition of in vivo histamine metabolism in rats by foodborne and pharmacologic inhibitors of diamine oxidase, histamine N-methyl transferase, and monoamine oxidase. Toxicol. Appl. Pharmacol. 8, 241–249. Hungerford, J.M., Walker, K.D., Wekell, M.M., LaRose, J.E., Throm, H.R., 1990. Selective determination of histamine by flow injection analysis. Anal. Chem. 62, 1971–1976. Hungerford, J.M., Manger, R., Wekell, M., Hollingworth, T., 1997. Rapid chemical tests: Laboratory and field screening methods for seafood analysis. In: Martin, R.E., Collette, R.L., Slavin, J.W. (Eds.), Fish Inspection, Quality Control, and HACCP: A Global Focus, Proceedings of the Conference held 19–24 May 1996, Arlington, VA, Technomic, Lancaster, Basel, pp. 538–550. Husain, S.S., 1996. Effect of traditional preservatives on the growth of histamine formers. Dissertation submitted in partial fulfilment of the requirements for the degree of Master of Science in Marine Biology and Oceanology. Centre of Advanced Study in Marine Biology, Annamalai University, Parangipettai. Ienistea, C., 1971. Bacterial production and destruction of his-

34

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

tamine in foods, and food poisoning caused by histamine. Nahrung 15, 109–113. Ienistea, C., 1973. Significance and detection of histamine in food. In: Hobbs, B.C., Christian, J.H.B. (Eds.), The Microbiological Safety of Foods, Academic Press, New York. Ijomah, P., Clifford, M.N., Walker, R., Wright, J., Hardy, R., Murray, C.K., 1991. The importance of endogenous histamine relative to dietary histamine in the aetiology of scombrotoxicosis. Food Addit. Contam. 8 (4), 531–542. Ijomah, P., Clifford, M.N., Walker, R., Wright, J., Hardy, R., Murray, C.K., 1992. Further volunteer studies on scombrotoxicosis. In: Burt, J.R., Hardy, R., Whittle, K.J. (Eds.), Pelagic Fish: The Resource and its Exploitation, Fishing News Books, Oxford, pp. 194–199, Chapter 17. Institute of Medicine, 1991. Seafood Safety, National Academy Press, Washington, DC. Irvine, W.T., Duthie, H.L., Ritchie, H.D., Waton, N.G., 1959. Urinary output of histamine after a meat meal. Lancet 1, 1061. James, D., Olley, J., 1985. IX. Summary and future research needs. In: Histamine in Marine Products: Production by Bacteria, Measurement and Prediction of Formation, FAO Fisheries Technical Paper No. 252, Food and Agriculture Organization of the United Nations, Rome, pp. 47–51. Johnson, L.R., Overholt, B.F., 1967. Release of histamine into gastric venous blood flow following injury by acetic or salicyclic acid. Gastroenterology 52, 505–509. Jung, H.P.K., Bjeldanes, L.F., 1979. Effects of cadaverine on histamine transport and metabolism in isolated gut sections of the guinea pig. Food Cosmetic Toxicol. 17, 629–632. Karnicki, Z.S., 1997. The impact of international trade and trade agreements on fish inspection and quality control. In: Martin, R.E., Collette, R.L., Slavin, J.W. (Eds.), Fish Inspection, Quality Control, and HACCP: A Global Focus, Proceedings of the Conference held 19–24 May 1996, Arlington, VA, Technomic, Lancaster, Basel, pp. 3–22. Kawabata, T., Ishizaka, K., Miura, T., 1955. Studies on the allergy-like food poisoning associated with putrefaction of marine products. II. Separation of causative substance and some of its chemical characteristics. Jpn. J. Med. Sci. Biol. 8, 503–519. Kawai, A., Sakaguchi, M., 1968. Histidine metabolism in fish. 2. Formation of urocanic, formimoglutamic, and glutamic acids from histidine in livers of carp and mackerel. Bull. Jpn. Soc. Sci. Fish. 34, 507–511. Kim, I.-S., 1978. Simple analysis of amines in canned tuna fish. M.S. Thesis, University of California, Berkeley, CA (cited by Bjeldanes et al., 1978). Kim, R., 1979. Flushing syndrome due to mahimahi (scombroid fish) poisoning. Arch. Dermatol. 115, 963–965. Kimata, M., 1961. The histamine problem. In: Borgstrom, G. (Ed.), Fish as Food: Production, Biochemistry and Microbiology, Vol. 1, Academic Press, New York, pp. 329–352. Kindred, T.P., 1996. Risk analysis and its application in FSIS. J. Food Prot. Suppl., 24–30. Klausen, N.K., Huss, H.H., 1987a. A rapid method for detection of histamine-producing bacteria. Int. J. Food Microbiol. 5, 137–146. Klausen, N.K., Huss, H.H., 1987b. Growth and histamine pro-

duction by Morganella morganii under various temperature conditions. Int. J. Food Microbiol. 5, 147–156. Klausen, N.K., Lund, E., 1986. Formation of biogenic amines in herring and mackerel. Z. Lebensm. Unters Forsch. 182 (6), 459–463. Lange, W.R., 1988. Scombroid poisoning. Am. Fam. Physician 37, 163–168. Lehane, L., 2000. Ciguatera update. Med. J. Aust. 172 (4), 176–179. Lerke, P.A., Bell, L.D., 1976. A rapid fluorometric method for the determination of histamine in canned tuna. J. Food Sci. 41, 1282–1284. Lerke, P.A., Werner, S.B., Taylor, S.L., Guthertz, L.S., 1978. Scombroid poisoning. Report of an outbreak. West. J. Med. 129, 381–386. Lerke, P.A., Porcuna, M.N., Chin, H.B., 1983. Screening test for histamine in fish. J. Food Sci. 48, 155–157. Leung, E., 1987. Amine producing bacteria isolated from chub mackerel (Rastrelliger sp.). A thesis in partial fulfilment of the requirements for the degree of Bachelor of Agricultural Science with Honours, University of Tasmania, Hobart. Lipp, E.K., Rose, J.B., 1997. The role of seafood in foodborne diseases in the United States of America. Rev. Sci. Technol. 16 (2), 620–640. ´ ´ ´ Lopez-Sabater, E.I., Rodrıguez-Jerez, J.J., Hernandez-Herrero, M., Mora-Ventura, M.T., 1996. Incidence of histamine-forming bacteria and histamine content in Scombroid fish species from retail markets in the Barcelona area. Int. J. Food Microbiol. 28 (3), 411–418. Love, R.M., 1980. Histidine. In: The Chemical Biology of Fishes, Advances 1968–1977, Vol. 2, Academic Press, London, New York, Toronto, Sydney, San Francisco, p. 427. Lyons, D.E., Beery, J.T., Lyons, S.A., Taylor, S.L., 1983. Cadaverine and aminoguanidine potentiate the uptake of histamine in vitro in perfused intestinal segments of rats. Toxicol. Appl. Pharmacol. 70, 445–458. McInerney, J., Sahgal, P., Vogel, M., Rahn, E., Jonas, E., 1996. Scombroid poisoning. Ann. Emerg. Med. 28 (2), 235–238. ´ ´ J., 1977. Histidine metabolism Mackie, I.M., Fernandez-Salguero, in fish. Urocanic acid in mackerel (Scomber scombrus). J. Sci. Food Agric. 28, 935–940. McCollum, J.P.K., Pearson, R.C.M., Ingham, H.R., Wood, P.C., Dewar, H.T., 1968. An epidemic of mussel poisoning in north-east England. Lancet ii, 767–770. McMeekin, T.A., 1986. Personal observation (cited by Anggawati, 1986). McMeekin, T.A., Olley, J., Ross, T., Ratkowsky, D.A., 1993. A thermodynamic approach to bacterial growth. In: Predictive Microbiology: Theory and Application, Research Studies Press–Wiley, Taunton, New York, pp. 287–309, Chapter 10. Maire, R., Dreiding, K., Wyss, P.A., 1992. [Incidence and clinical aspects of scombroid fish poisoning]. Schweiz Med. Wochenschr. 122 (50), 1933–1935, In German, English abstract. Merson, M.H., Baine, W.B., Gangarosa, E.J., Swanson, R.C., 1974. Scombroid fish poisoning. Outbreak traced to commercially canned tuna fish. J. Am. Med. Assoc. 228 (10), 1268– 1269. Mietz, J.L., Karmas, E., 1977. Chemical quality index of canned tuna as determined by HPLC. J. Food Sci. 42, 155–158.

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37 Miller, B.M., Johnson, C.G., Ledden, D.J., 1997. A rapid, easily used test kit to determine histamine concentrations in fish. In: Martin, R.E., Collette, R.L., Slavin, J.W. (Eds.), Fish Inspection, Quality Control, and HACCP: A Global Focus, Proceedings of the Conference held 19–24 May 1996, Arlington, VA, Technomic, Lancaster, Basel, pp. 551–559. Mines, D., Stahmer, S., Shepherd, S.M., 1997. Poisonings: food, fish, shellfish. Emerg. Med. Clin. North Am. 15 (1), 157–177. Mitchell, J.W., 1984. Scombroid food poisoning from fish. NZ Med. J. 97, 127. Mitchell, J.W., 1993. Scombrotoxic Fish Poisoning. A Report Prepared For the Ministry of Health, Institute of Environmental Health and Forensic Sciences Limited, Mt. Albert Science Centre, New Zealand. Mitchell, J.W., O’Brien, G., 1992. Food poisoning – and mackerel. NZ Med. J. 105, 929. Mitchell, R.G., Code, C.F., 1954. Effect of diet on urinary excretion of histamine. J. Appl. Physiol. 6, 387–391. Molinari, G., Montagnoli, G., Pellegrini, G., Caroli, G., 1989. [Hygiene and health importance of histamine as an unhealthy factor in several food products]. Ann. Ig. 1 (3–4), 637–646, In Italian, English abstract. Mongar, J.L., 1957. Effect of chain length of aliphatic amines on histamine potentiation and release. Br J. Pharmacol. 12, 140– 148. Morii, H., Cann, D.C., Taylor, L.Y., Murray, C.K., 1986. Formation of histamine by luminous bacteria isolated from scombroid fish. Bull. Jpn. Soc. Sci. Fish. 52, 2135–2141. Morrison, H., Avnir, D., Zarrella, T., 1980. Analysis of Z and E isomers of urocanic acid by high-performance liquid chromatography. J. Chromatogr. 183, 83–86. Mopper, B., Sciacchitano, C.J., 1994. Capillary zone electrophoretic determination of histamine in fish. J. AOAC Int. 77 (4), 881–884. Morrow, J.D., Margoles, G.R., Rowland, J., Roberts, L.J., 1991. Evidence that histamine is the causative toxin of scombroidfish poisoning. New Engl. J. Med. 324 (11), 716–720. Motil, K.J., Scrimshaw, N.S., 1979. The role of exogenous histamine in scombroid poisoning. Toxicol. Lett. 3, 219–223. ¨ Muller, G.J., Lamprecht, J.H., Barnes, J.M., de Villiers, R.V.P., Honeth, B.R., Hoffman, B.A., 1992. Scombroid poisoning. Case series of 10 incidents involving 22 patients. South Af. Med. J. 81 (8), 427–430. Niven, C.F., Jeffrey, M.B., Corlett, D.A., 1981. Differential plating medium for quantitative detection of histamine-producing bacteria. Appl. Environ. Microbiol. 41 (1), 321–322. Ohashi, M., Nomura, F., Suzuki, M., Otsuka, M., Adachi, O., Arakawa, N., 1994. Oxygen-sensor-based simple assay of histamine in fish using purified amine oxidase. J. Food Sci. 59 (3), 19–22. Okuzumi, M., Okuda, S., Awano, M., 1981. Isolation of psychrophilic and halophilic histamine-forming bacteria from Scomber japonicus. Bull. Jpn. Soc. Sci. Fish. 47, 1591–1598. Okuzumi, M., Yamanaka, H., Kubozuka, T., 1984a. Occurrence of various histamine-forming bacteria on / in fresh fishes. Bull. Jpn. Soc. Sci. Fish. 50 (1), 161–167. Okuzumi, M., Yamanaka, H., Kubozuka, T., Ozaki, H., Matsubara, K., 1984b. Changes in numbers of histamine-forming bacteria

35

on / in common mackerel stored at various temperatures. Bull. Jpn. Soc. Sci. Fish. 50 (4), 653–657. Olley, J., 1972. Unconventional sources of fish protein. CSIRO Food Res. Quarterly 32 (2), 27–32. Olley, J., Baranowski, J., 1985. V. Temperature effects on histamine formation. In: Histamine Formation in Marine Products: Production by Bacteria, Measurement and Prediction of Formation, FAO Fisheries Technical Paper No. 252, Food and Agriculture Organization of the United Nations, Rome, pp. 14–17. Olley, J., McMeekin, T.A., 1985. VI.3. Use of temperature function integrators. In: Histamine Formation in Marine Products: Production by Bacteria, Measurement and Prediction of Formation, FAO Fisheries Technical Paper No. 252, Food and Agriculture Organization of the United Nations, Rome, pp. 24–29. Olley, J., Orejana, F., Pan, B.S., 1985. VII.4. Effect of gutting. In: Histamine Formation in Marine Products: Production by Bacteria, Measurement and Prediction of Formation, FAO Fisheries Technical Paper No. 252, Food and Agriculture Organization of the United Nations, Rome, p. 37. Omura, Y., Price, R.J., Olcott, H.S., 1978. Histamine-forming bacteria isolated from spoiled skipjack tuna and jack mackerel. J. Food Sci. 43, 1779–1781. Pan, B.S., 1985a. VI.2. Use of Arrhenius plot to estimate histamine formation in mackerel and bonito. In: Histamine Formation in Marine Products: Production by Bacteria, Measurement and Prediction of Formation, FAO Fisheries Technical Paper No. 252, Food and Agriculture Organization of the United Nations, Rome, pp. 21–23. Pan, B.S., 1985b. VIII.2. Histamine formation in canning processes. In: Histamine Formation in Marine Products: Production by Bacteria, Measurement and Prediction of Formation, FAO Fisheries Technical Paper No. 252, Food and Agriculture Organization of the United Nations, Rome, pp. 41–44. Pan, B.S., 1988. Formation of histamine toxicity. Undesirable factors in dried fish products. In: Burt, J.R. (Ed.), Fish Smoking and Drying: The Effect of Smoking and Drying on the Nutritional Properties of Fish, Elsevier Applied Science, London, New York, pp. 66–67, Chapter 5. Parrot, J., Nicot, G., 1966. Pharmacology of histamine. In: Eichler, O., Farah, S. (Eds.), Handbook of Experimental Pharmacology, Springer-Verlag, New York, pp. 148–161. Poulter, R.G., 1988. Processing and storage of traditional dried and smoked fish products. In: Burt, J.R. (Ed.), Fish Smoking and Drying: The Effect of Smoking and Drying on the Nutritional Properties of Fish, Elsevier Applied Science, London, New York, pp. 85–90, Chapter 7. Price, R.J., 1999. Compendium of Fish and Fishery Product Processes, Hazards, and Controls. National Seafood HACCP Alliance for Training and Education (http: / / seafood.ucdavis.edu / haccp / compendium / compend.htm, updated 8 / 9 / 99). Ramesh, A., Venugopalan, V.K., 1986. Densities and characteristics of histamine-forming luminous bacteria of marine fish. Food Microbiol. 3, 103–105. Rawles, D.D., Flick, G.J., Martin, R.E., 1996. Biogenic amines in fish and shellfish. Adv. Food Nutr. Res. 39, 329–364.

36

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37

Rawlin, G., Herfort, A., 1999. Ciguatera detection in fish flesh and tracking of possible ciguatera-tainted fish in the marketplace. Internal advice note, Agriculture, Fisheries and Forestry – Australia. Rice, S.L., Eitenmiller, R.R., Koehler, P.E., 1976. Biologically active amines in food: a review. J. Milk Food Technol. 39 (5), 353–358. Riley, J.F., 1959. In: The Mast Cells, Livingstone, Edinburgh, London, p. 130. Ritchie, A.H., Mackie, I.M., 1979. The formation of diamines and polyamines during storage of mackerel (Scomber scombrus). In: Connell, J.J. (Ed.), Advances in Fish Science and Technology, Fishing News Books, Surrey, pp. 489–494. Rogers, P.L., Staruszkiewicz, W., 1997. Gas chromatographic method for putrescine and cadaverine in canned tuna and mahimahi and fluorometric method for histamine (minor modification of AOAC Official Method 977.13): collaborative study. J. AOAC Int. 80 (3), 591–602. Ross, T., McMeekin, T.A., 1995. Predictive microbiology and HACCP. In: Pearson, A.M., Dutson, T.R. (Eds.), HACCP in Meat, Poultry and Fish Processing, Blackie, pp. 330–357, Chapter 13. ´ Z., 1986. Scombroid poisoning: miniRussell, F.E., Maretic, review with case histories. Toxicon 24 (10), 967–973. Ryser, E.T., Marth, E.H., Taylor, S.L., 1984. Histamine production by psychrotrophic pseudomonads isolated from tuna fish. J. Food Prot. 47, 8–80. Sabroe, R.A., Kobza Black, A., 1998. Scombrotoxic fish poisoning. Clin. Exp. Dermatol. 23 (6), 258–259. ´ Sanchez-Guerrero, I.M., Vidal, J.B., Escudero, A.I., 1997. Scombroid fish poisoning: a potentially life-threatening allergic-like reaction. J. Allergy Clin. Immunol. 100 (3), 433–434. Sapin-Jaloustre, H., Sapin-Jaloustre, J., 1957. [A little-known food poisoning: histamine poisoning from tuna]. Concours Med. Paris 79, 2705–2708, In French, English abstract. Satake, K., Ando, S., Fujita, H., 1953. Bacterial oxidation of some primary amines. J. Biochem. 40, 299–315. Scoging, A.C., 1991. Illness associated with seafood. CDR (Lond. Eng. Rev.) 1 (11), R117–R122. Scoging, A.C., 1998. Scombrotoxic (histamine) fish poisoning in the United Kingdom. Commun. Dis. Public Health 1 (3), 204–205. Senanayake, N., Vyravanathan, S., 1981. Histamine reactions due to ingestion of tuna fish (Thunnus argentivittatus) in patients on anti-tuberculosis therapy. Toxicon 19, 184–185. Shalaby, A.R., 1996. Significance of biogenic amines to food safety and human health. Food Res. Int. 29 (7), 675–690. Shewan, J.M., 1955. The nitrogenous extractives from fresh fish muscle. III. Comparison of several flat fishes and members of the herring–mackerel group. J. Sci. Food Agric. 6, 99–104. Shewan, J.M., 1977. The bacteriology of fresh and spoiling fish and the biochemical changes induced by bacterial action. In: Sutcliffe, P., Disney, J. (Eds.), Proceedings of Conference on the Handling, Processing and Marketing of Tropical Fish, Tropical Products Institute, London, pp. 51–66. Shibatani, T., Nishimura, N., Nabe, K., Kakimoto, T., Chibata, I., 1974. Enzymatic production of urocanic acid by Achromobacter liquidum. Appl. Microbiol. 27, 688–694. Shultz, D.E., Chang, G.W., Bjeldanes, L.F., 1976. Rapid thin layer

chromatographic method for the determination of histamine in fish products. J. AOAC Int. 59 (6), 1224–1225. Simidu, W., Hibiki, S., 1954. Studies on putrefaction of aquatic products. 14. Comparison on putrefaction of different kinds of fish (2). Bull. Jpn. Soc. Sci. Fish. 20 (4), 302–304. Simidu, W., Hibiki, S., 1955. Studies on putrefaction of aquatic products. 23. On the critical concentration of poisoning for histamine. Bull. Jpn. Soc. Sci. Fish. 21, 365–367. Sjaastad, O., Sjaastad, O.V., 1974. Catabolism of orally administered 14 C-histamine in man. Acta Pharmacol. Toxicol. 34, 33. Skovgaard, N., Ellemann, G., 1978. Levendsmiddelbaren histaminforgiftning. Ugeskr. Laeg. 140, 313–316, Cited by Pan, 1985b. Smart, D.R., 1992. Scombroid poisoning. A report of seven cases involving the Western Australian salmon, Arripid truttaceus. Med. J. Aust. 157 (11–12), 748–751. Sockett, P.N., 1991. Food poisoning outbreaks associated with manufactured foods in England and Wales. Commun. Dis. Rep. 1 (10), R105–R109. Sours, H.E., Smith, D.G., 1980. Outbreaks of foodborne disease in the United States. J. Infect. Dis. 142, 122–125. Specht, D., 1998. Scombroid fish poisoning. J. Emerg. Nursing 24 (2), 118–119. Stratton, J.E., Taylor, S.L., 1991. Scombroid poisoning. In: Ward, D., Hackney, C. (Eds.), Microbiology of Marine Food Products, Spectrum, New York, pp. 331–351, Chapter 13. Stratton, J.E., Hutkins, R.W., Taylor, S.L., 1991. Biogenic amines in cheese and other fermented foods. A review. J. Food Prot. 54, 460–470. Stryer, L., 1981. Amino acid degradation and the urea cycle, and biosynthesis of amino acids and heme. In: Biochemistry, W.H. Freeman, San Francisco, CA, p. 417, Chapter 18 and p. 504, Chapter 21. Subburaj, M., Karunasagar, I., Karunasagar, I., 1984. Incidence of histidine decarboxylating bacteria in fish and market environs. Food Microbiol. 1, 263–267. Suyama, M., Yoshizawa, Y., 1973. Free amino acid composition of the skeletal muscle of migratory fish. Bull. Jpn. Soc. Sci. Fish. 39, 1339–1343. Takagi, M., Iida, A., Murayama, H., Soma, S., 1969. On the formation of histamine during loss of freshness and putrefaction of various marine products. Hokkaido Daigaku Suisan Gakubu Kenkyu Iho 20, 227–234. Taylor, S.L., 1983. Monograph on histamine poisoning. In: Codex Alimentarius Commission, FAO / WHO, Rome. 19th Session of the Codex Committee on Food Hygiene, Washington, DC, 26–30 September 1983. Taylor, S.L., 1985. VPH / FOS / 85.1. Histamine Poisoning Associated With Fish, Cheese, and Other Foods, World Health Organization, Geneva. Taylor, S.L., 1986. Histamine food poisoning: toxicology and clinical aspects. Crit. Rev. Toxicol. 17 (2), 91–128. Taylor, S.L., 1988. Marine toxins of microbial origin. Food Technol. 42 (3), 94–98. Taylor, S.L., Lieber, E.R., 1979. In vivo inhibition of rat intestinal histamine-metabolizing enzymes. Food Cosmetic Toxicol. 17, 237–240. Taylor, S.L., Speckhard, M.W., 1983. Isolation of histamine-

L. Lehane, J. Olley / International Journal of Food Microbiology 58 (2000) 1 – 37 producing bacteria from frozen tuna. Mar. Fish Rev. 45 (4–6), 35–39. Taylor, S.L., Sumner, S.S., 1986. Determination of histamine, putrescine, and cadaverine. In: Kramer, D.E., Liston, J. (Eds.), Seafood Quality Determination, Elsevier, Amsterdam, pp. 235– 245. Taylor, S.L., Woychik, N.A., 1982. Simple medium for assessing quantitative production of histamine by Enterobacteriaceae. J. Food Prot. 45 (8), 747–751. Taylor, S.L., Guthertz, L.S., Leatherwood, M., Tillman, F., Lieber, E.R., 1978a. Histamine production by food-borne bacterial species. J. Food Safety 1, 173–187. Taylor, S.L., Lieber, E.R., Leatherwood, M., 1978b. A simplified method for histamine analysis of foods. J. Food Sci. 43, 247–250. Taylor, S.L., Guthertz, L.S., Leatherwood, M., Lieber, E.R., 1979. Histamine production by Klebsiella pneumoniae and an incident of scombroid fish poisoning. Appl. Environ. Microbiol. 37 (2), 274–278. Taylor, S.L., Hui, J.Y., Lyons, D.E., 1984. Toxicology of scombroid poisoning. In: Ragelis, E.P. (Ed.), Seafood Toxins, American Chemical Society, Washington, DC, pp. 417–430. Taylor, S.L., Stratton, J.E., Nordlee, J.A., 1989. Histamine poisoning (scombroid fish poisoning): an allergy-like intoxication. J. Toxicol. Clin. Toxicol. 27 (4–5), 225–240. Til, H.P., Falke, H.E., Prinsen, M.K., Willems, M.I., 1997. Acute and subacute toxicity of tyramine, spermidine, spermine, putrescine and cadaverine in rats. Food Chem. Toxicol. 35 (3–4), 337–348. Trenerry, V.C., Marshall, P.A., Windahl, K., 1998. Method optimisation: determination of histamine in fish by capillary zone electrophoresis. J. Cap. Electrophoresis 5 (1–2), 27–32. Trinidad, L.M., Narida, M.M., Reilly, A., 1986. Production and storage of smoked Spanish mackerel (Scomberomorus commerson). In: McLean, J.L., Dizon, L.B., Hosillos, L.V. (Eds.), The First Asian Fisheries Forum, Asian Fisheries Society, Manilla, pp. 473–475. Todd, E.C., 1997. Seafood-associated diseases and control in Canada. Rev. Sci. Technol. 16 (2), 661–672. Todd, E.C., Brown, D.J., Rutherford, M., Grolla, A., Saarkoppel, H., Chan, A., McGregor, G., Reffle, J., Ryan, S., Jerrett, N., 1992. Scombroid poisoning – an outbreak in two Ontario communities. Can. Commun. Dis. Rep. 18 (3), 17–19. Uragoda, C.G., Kottegoda, S.R., 1977. Adverse reactions to isoniazid on ingestion of fish with high histamine content. Tubercle 58, 83–89. Van Gelderen, C.E.M., Savelkoul, T.J.F., van Ginkel, L.A., van Dokkum, W., 1992. The effects of histamine administered in fish samples to healthy volunteers. Clin. Toxicol. 30, 585–596.

37

Van Spreekens, K., 1987. Histamine production by the psychrophilic flora. In: Kramer, D.E., Liston, J. (Eds.), Seafood Quality Determination, Elsevier, Amsterdam, pp. 309–318. Voigt, M.N., Eitenmiller, R.R., 1977. Production of tyrosine and histidine decarboxylase in dairy-related bacteria. J. Food Prot. 40, 241–245. Wagener, W.W.D., 1984. Biogenic amines in fresh and spoiled fish: high performance liquid chromatography of the dansyl derivatives. In: 38th Annual Report, Fishing Industry Research Institute, University of Capetown, Rondebosch, pp. 48–50. Warne, D., 1985. VIII.3. Application of hazard analysis critical control point concept. In: Histamine Formation in Marine Products: Production by Bacteria, Measurement and Prediction of Formation, FAO Fisheries Technical Paper No. 252, Food and Agriculture Organization of the United Nations, Rome, pp. 45–46. Weiss, S., Robb, G.P., Ellis, L.B., 1932. The systemic effects of histamine in man. Arch. Int. Med. 49, 360. Wendakoon, C.N., Sakaguchi, M., 1992. Effects of spices on growth and biogenic amine formation by bacteria in fish muscle. In: Huss, H.H., Jakobsen, M., Liston, J. (Eds.), Quality Assurance in the Fish Industry, Elsevier, Amsterdam, pp. 305–313. Wille, J.J., Kydonieus, A.F., Murphy, G.F., 1999. cis-Urocanic acid induces mast cell degranulation and release of preformed TNF-alpha: a possible mechanism linking UVB and cisurocanic acid to immunosuppression of contact hypersensitivity. Skin Pharmacol. Appl. Skin Physiol. 12 (1–2), 18–27. Wu, M.-L., Yang, C.-C., Yang, G.-Y., Ger, J., Deng, J.-F., 1997. Scombroid fish poisoning: an overlooked marine food poisoning. Vet. Hum. Toxicol. 39 (4), 236–241. Yamanaka, H., 1984. Changes in histamine contents in red meat fish during storage at different temperatures. Bull. Jpn. Soc. Sci. Fish. 50, 695–701. Yamani, M.I., Dickertmann, D., Untermann, F., 1981. Histamine poisoning by Proteus spp. from tuna fish. Zbl. Bakt. Hyg., I Abt. Orig. B 173, 478–487. Yen, G., Hsieh, C., 1991. Simultaneous analysis of biogenic amines in canned fish by HPLC. J. Food Sci. 56 (1), 158–160. Yoshinaga, D.H., Frank, H.A., 1982. Histamine-producing bacteria in decomposing skipjack tuna (Katsuwonas pelamis). Appl. Environ. Microbiol. 44 (2), 447–452. Zee, J.A., Simard, R.E., Vaillancourt, R., Boudreau, A., 1981. Effect of Lactobacillus brevis, Saccharomyces uvarum and grist composition on amine formation in beers. Can. Institute Food Sci. Technol. 14 (4), 321–325.