U. S. Food and Drug Administration
Center for Food Safety and Applied Nutrition
March 29, 2001

Potential Hazards in Cold-Smoked Fish: Biogenic Amines

Scope

1. Introduction

Cadaverine, putrescine, and histamine are diamines that may be produced post mortem from the decarboxylation of specific free amino acids (Table IV-1) in fish or shellfish tissue (Silla Santos 1996). The decarboxylation process can proceed through two biochemical pathways: endogenous decarboxylase enzymes naturally occurring in fish or shellfish tissue or exogenous enzymes released by the various microorganisms associated with the seafood product. Endogenous production of diamines is insignificant when compared to the exogenous pathway (Wendakoon and Sakaguchi 1992a). The nature of the microflora and the composition of the product affect the amount of decarboxylase a bacterial cell may release (Wendakoon and Sakaguchi 1992b; Suzuki and others 1990). In general, histamine, putrescine, cadaverine, tyramine, tryptamine, -phenylethylamine, spermine, and spermidine are considered the most important biogenic amines in foods (Shalaby 1996). However, -phenylethylamine, spermine, and spermidine are not end products of bacterial decomposition in fishery products.

Table IV-1. Biogenic amines and their chemical precursors (from Shalaby 1996)

Fish muscle is naturally rich in free amino acids and the content may increase even further post mortem. The high content of proteolytic enzymes in the intestinal tract is responsible for the rapid autolytic process (Gilberg 1978; Aksnes 1988) and the high free amino acid content in fishery products. Amino acid formation depends on the harvesting season and feeding activity prior to capture. For example, fish harvested in summer or feeding season quickly liberated large quantities of lysine and arginine (Aksnes and Brekken 1988).

The activity of amino acid decarboxylase depends on a range of factors, including fermentable sugars, pH, and redox potential (Gale 1946). The influence of environmental temperature, nature of microflora, decarboxylase activity, and intestinal tract content on biogenic amine formation may be major reasons for the discrepancies that have been reported in the literature concerning levels of biogenic amines in fresh and processed fish. Another reason for discrepancies may be poor experimental design. Regardless of the discrepancies, it is clear that a high amino acid content and bacterial activity could rapidly result in an elevated concentration of biogenic amines if the proper controls are not in place.

1.1. Safety aspects

Fish often associated with histamine poisoning are the scombroid fish belonging to the families Scomberesocidae and Scombridae. Fish included in these families are the tunas, bonito, mackerels, bluefish, and saury. Tuna and mackerel are the most common fish associated with the poisoning, but other fish are also associated with outbreaks of scombroid poisoning. Examples include mahimahi, sardines, anchovies, herrings, and marlin. The association of type of fish and biogenic amine poisoning may reflect the amount of consumption of a specific fish.

Research on the quantitative determination of histamine, cadaverine and putrescine in fishery products at FDA have resulted in the 2 only accepted Association of Official Analytical Chemists (AOAC) methods for regulatory purposes (Rogers and Staruszkiewicz 1997). The research was the basis for the establishment of the defect action levels used in FDA's regulatory programs. Recently, the Food and Drug Administration (FDA) (21CFR123) established a guidance level for histamine of 5 mg / 100 g (50 ppm) for assuring the safe consumption of scombroid or scombroid-like fish and recommended the use of other data to judge fish freshness, such as the presence of other biogenic amines associated with fish decomposition (FDA 1996). A maximum average histamine content of 10 mg / 100 g (100 ppm) has been established in the European Community (EC) for acceptance of tuna and other fish belonging to the Scombridae and Scomberesocidae families (Veciana-Nogues and others 1997). The EC has suggested that in the future a maximum of 300 ppm for total biogenic amines in fish and fish products may be an appropriate legal limit. It is important to note, however, that there may be a type of poisoning that does not arise from high levels of histamine. Thus a low histamine level may not be absolute assurance of a safe product. It may be more appropriate to say that the absence of decomposition in the fish renders it a safe product. As such, a safe product would have no evidence of spoilage including odors of decomposition, high histamine levels, and other amines such as cadaverine.

2. Toxicity

2.1. Histamine toxicity

The role of saurine (implicated in histamine poisonings in Japan) as a compound able to act synergistically with histamine was reviewed by Arnold and Brown (1978), but it was later concluded that the compound was in fact histamine.

2.2 Toxicity potentiators

Studies have shown that the levels of cadaverine in toxic or decomposed fish are generally several times greater than the levels of putrescine. When cadaverine was administered through stomach catheters simultaneously with histamine, peroral toxicity was observed in the guinea pigs (Bjeldanes and others 1978). Klausen and Lund (1986) reported that at 10 °C the high cadaverine contents of mackerel in comparison with herring could be responsible for mackerel often being implicated in scombroid poisoning and not herring, since histamine levels were similar in both. Cadaverine and putrescine, as well as other diamines, have been suggested to facilitate the transport of histamine through the intestinal wall and to increase its toxicity (Fernandez-Salguero and Mackie 1987b).

Arnold and Brown (1978) reported on the possibility that bacterial endotoxins, which are widespread, could result in hypersensitivity to histamine. These compounds are complex, heat-stable, lipopolysaccharide materials produced primarily by gram-negative bacteria. They also reported that endotoxin is known to be capable of inducing histamine release in animals (sometimes called endotoxin shock) similar to that seen in anaphylaxis. J. Baronowski (personal communication), however, reported extremely low levels of endotoxin in both good tuna and tuna known to have caused illness in humans.

From these discussions, it is clear that concentration of biogenic amines producing observable toxicity may differ significantly, depending on a variety of circumstances. Also, although a variety of histamine potentiators are known, there is not a clear understanding of the level and the manner by which synergism occurs.

3. Prevalence in fish

3.1 Muscle type

In the study by Gloria and others (1999), no difference was observed on amine levels of upper and lower loin light muscles, but dark muscles contained higher spermidine (Table IV-2). Intestine wall samples contained high amine levels.

Table IV-2. Levels of biogenic amines in light and dark muscles and intestine wall of albacore tuna (from Gloria and others 1999)

	Biogenic amines1 (ppm)
Samples	SPM	SPD	HIM	PUT	CAD	SER	Total
light muscle upper loin	0.68b (0.12)	0.26c (0.07)	0.00b	0.22ab (0.07)	0.13b (0.02)	0.00b	1.29c (0.17)
lower loin	1.21b (0.26)	0.25c (0.05)	0.00b	0.14b (0.05)	0.11b (0.06)	0.00b	1.77c (0.37)
dark muscle	2.50ab (0.97)	0.79b (0.18)	0.00b	0.06b (0.03)	0.07b (0.05)	0.00b	3.42b (0.72)
intestine wall	5.35a (2.46)	3.63a (1.18)	0.52a (0.25)	0.43a (0.16)	1.96a (0.59)	4.38a (1.33)	16.3a (4.59
1 Mean values (standard deviation) were calculated by using 0 for not detected levels (spermine -SPM, spermidine-SPD, histamine-HIM, putrescine-PUT, cadaverine-CAD æ 0.08; and serotonin-SER æ 0.18 mg/100g). Mean values with the same superscript in the same column do not differ significantly (p 0.05), Turkey test).

Takagi and others (1969) examined the amounts of histidine and histamine in 21 aquatic species during spoilage. Their conclusions were consistent with those of other researchers in that more histamine was produced in the red muscle fish such as tuna and mackerel than in white muscle species such as rockfish. Within a given fish species, more histidine and histamine were found in white than in red muscle.

Wendakoon and others (1990) reported that most of the bacteria that convert amino acids into non-volatile amines possess more than one decarboxylase. In contrast to results reported by Wendakoon and Sakaguchi (1992b), they also reported that in the dark muscle, the amine levels were always much higher and the amine production was more rapid than that in the white muscle.

3.1.3 Microflora

A variety of microorganisms are able to produce biogenic amines. The production of cadaverine and putrescine by microorganisms is not surprising since the covalent linking of cadaverine and putrescine to the peptidoglycan is necessary for normal microbial growth (Suzuki and others 1988). Several inoculation studies on both culture media and on fish have demonstrated that Morganella spp., Proteus morganii, Proteus spp., Hafnia alvei, and Klebsiella spp. are able to produce histamines and other biogenic amines. The majority of the studies also concurred that the potential of these microorganisms to produce toxic levels of biogenic amines is enhanced at abusive temperatures (see section 4.2 of this chapter).

The following tables summarize research on production of biogenic amines by microorganisms. Tables IV-3 and 4 list studies on production of biogenic amines by bacterial isolates inoculated on different culture media and on fish that may be cold-smoked, respectively. In addition, studies where isolates from fish have been incubated in media and histamine production monitored are listed in Table IV-5. It is noteworthy to point out that spoilage and toxin formation occur due to a variety of microorganisms, and therefore identical storage times for similar fish species may produce varying levels of scombrotoxin.

Okuzumi and others (1990) investigated the relationship between microflora on horse mackerel (Trachurus japonicus) and dominant spoilage bacteria. The results of their study showed that Pseudomonas I/II, Pseudomonas III/IV-NH, Vibrio, and Photobacterium were dominant when high levels of putrescine, cadaverine, and histamine were detected.

Table IV-3. Production of biogenic amine by bacteria growing on media culture

Table IV-4. Production of biogenic amines by bacteria isolates incubated on fish

Table IV-5. Production of biogenic amines on culture media by microorganism isolated from fish

The activity of decarboxylase can be an indirect measurement of potential for biogenic amine formation. A study by Middlebrooks and others (1988) showed that 14 bacterial isolates (Acinetobacter lowffi, Aeromonas hydrophila, Clostridium perfringens, Enterobacter aerogenes, Enterobacter spp., H. alvei, Morganella morganii, Proteus mirabilis, Proteus vulgaris, Proteus spp., Pseudomonas fluorescens/putida, Pseudomonas putrefaciens, Pseudomonas spp., and Vibrio alginolyticus) from mackerel tissue were capable of exhibiting decarboxylase activity (production of histamine, cadaverine, and putrescine) when incubated in Spanish mackerel at 0 °C (-32 °F), 15 °C (59 °F), and 30 °C (90 °F). Other bacteria strong histidine decarboxylase activities: Klebsiella pneumonia (Taylor and others 1979), Klebsiella planticola (Taylor and Lieber 1979), Alteromonas putrefaciens (Frank and others 1985), Photobacterium phosphoreum (Morii and others 1986); Staphylococcus xylosus (Rodriguez-Jerez and others 1994), Cedecea lapagei, Cedecea neteri, Plesiomonas shigelloides (Lopez-Sabater and others 1994a), Providencia spp. (Ababouch and others 1991b), Lactobacillus curvatus LTH 975 and Lactobacillus buchneri LTH 1388 (Leuschner and Hammes 1999), Serratia spp. (Lopez-Sabater and others 1996a), and Escherichia spp. (Gale 1946).

Okuzumi and others (1984) studied histamine-forming bacteria in addition to N-group (psychrophilic halophilic, histamine-forming) bacteria in and on fresh fish. The histamine-forming bacteria were N-group bacteria, P. morganii, P. vulgaris, H. alvei, Citrobacter spp., Vibrio spp., and Aeromonas spp. For the summer samples, P. morganii was found most frequently, followed by the N-group bacteria. On the other hand, for the winter samples, only the N-group bacteria were found, and other histamine bacteria were not detected.

Changes in the concentration of tyramine, agmatine, putrescine, cadaverine, spermidine, tryptamine, spermine, histamine, and trimethylamine were studied in parallel with the development of the microbial population and sensory scores during the storage of Mediterranean gilt-head sea bream (Sparus aurata) at three temperatures (0 °C [32 °F)] 8 °C [(46 °F] 15 °C [59 °F]) (Koutsoumanis and others 1999). Pseudomonads and H₂S-producing bacteria were dominant microorganisms. Enterobacteriaceae and lactic acid bacteria were also present in the fish microflora. Among the biogenic amines, putrescine and cadaverine were detected when pseudomonads exceeded 10⁶ to 10⁷ cfu / g. Histamine was produced only in samples stored at 15 °C and reached >50 ppm levels at 48 h. Putrescine and cadaverine reached high levels also at 15 °C after 120 h. Tyramine, tryptamine, agmatine, and trimethylamine were absent regardless of the storage temperature. The authors concluded that only putrescine and cadaverine could be used as an index of freshness. The role and significance of putrescine and cadaverine in food safety and biogenic amine poisoning is yet to be established. Furthermore, in prevalence studies, researchers are challenged regularly to select or obtain samples that are representative of the existing total population of fish, muscle, microorganisms, or whatever is being studied.

4. Effect of processing steps

4.1. Gutted versus ungutted fish

4.2 Effect of post-harvest handling

The effect of temperature on histamine formation has frequently been studied. Table IV-6 lists selected research studies where fish were kept either in ice, refrigerated, or at abusive temperatures for different storage times. Different studies reported that skipjack tuna that was allowed to spoil under similar conditions had 100-fold variations in histamine concentrations. Although there is great variability in the results within the same study, longer storage times and higher temperatures seem to induce histamine production.

Table IV-6. Levels of biogenic amines on fresh fish stored at different temperature and time combinations

Fish	Temperature/time	Histamine content (ppm)	Other biogenic amines	Sensory	Reference
Salmon (Oncorhynchus gorbuscha)	10 °C for 14 d	Not detected		Spoilage	Crapo and Himelbloom 1999
Tuna (Thunnus thunnus) Albacore tuna (Thunnus alalunga) Albacore tuna (whole fish)	4 °C 8 °C 20 °C iced for 33 d 15 - 23 °C for 24 h 15 - 23 °C for 4 d 25 °C for 7 d	Toxic levels Toxic levels Toxic levels 825 insignificant <50 1,000		Unacceptable Acceptable Unacceptable	Lopez Sabater and others (1996b) Price and others (1991)   Ben-Gigirey and others (1998b)
Rainbow trout (Salmo irideus)	0 °C for 24 d	<1	<1		Dawood and others (1988)
Sardines	24 h at ambient T 8 d in ice	2,350 2,350	>1,000 >1,000		Ababouch and others (1991b)
Mahimahi (Coryphaena hippurus)	21 °C for 2 d 32 °C for 12 h 32 °C for 24 h	1,540 18 2,920			Baranowski and others (1990)
Sardine, saury pike, mackerel, horse mackerel	5 °C for 6 - 9 d 20 °C for 2 d 35 °C for 2 d	Toxic levels Toxic levels Toxic levels			Yamanaka and others (1984)
Sardine (Sardina pilchardus), horse mackerel (Trachurus trachurus), chub mackerel (Scomber japonicus), and mackerel (Scomber scombrus)	iced for 7 d iced for 7 d iced for 7 d	<100 <100 <100			Mendes (1999)
Mackerel (Scomber scombrus)   Mackerel   Spanish mackerel (Scomberomorus maculatus Mackerel (Scomber scombrus)	0 °C for 25 d 2 °C for 12 d 10 °C for 120 h 23 °C for 36 h 6 °C for 150 - 200 h 17 °C for 75 h 23 °C for 46 h 35 °C for 20 h 24 °C for 2 d 0 °C for 10 d 10 °C for 4 d 10 °C for 2d and 8d at 0 °C	<1 14 1,820 50 500 - 700 3,540 8,260 140 238 0 1,000 200		Unacceptable Unacceptable Unacceptable Unacceptable	Fernandez Salguero and Mackie (1979)   Kimata and Kawai (1953)   Edmunds and Eitenmiller (1975)   Klausen and Huss (1987)
Kahawai	15 - 23 °C for 2 d	1,500 - 3,500			Ben-Gigirey and others (1998b)
Herring, Clupea harengus pallasi	10 °C for 14 d	55		Spoiled by 6 d	Crapo and Himelbloom (1999)

Post-harvest handling conditions have a significant effect on the presence and concentration of putrescine and cadaverine. Ababouch and others (1991b) reported that bacteria on the skin and gills of freshly harvested sardines (Sardina pilchardus) quickly invaded and grew within the muscle tissue, reaching 5 x 10⁸ cfu / g and 6 x 10⁸ cfu / g respectively after 24 h at ambient temperature and 8 d in ice. Histamine, cadaverine, and putrescine accumulated to levels of 2,350 ppm, 1,050 ppm, and 300 ppm respectively after 8 d of ice storage but only 24 h at ambient temperature.

Dawood and others (1988) showed that initial holding temperatures above 0 °C (32 °F) resulted in increased concentrations of non-volatile amines in freshly caught whole and eviscerated rainbow trout (Salmo irideus). In another study on skipjack tuna (Euthynnus lineatus), Mazorra-Manzano and others (2000) concluded that endogenous and microbial deterioration processes could be controlled at 0 °C, since, even after 24 d, there was <1 ppm of any of the biogenic amines analyzed (histamine, cadaverine, and putrescine) in hook- and line-caught fish that were immediately iced upon landing. Consistent with those results, Atlantic herring (Clupea harhengus) and Atlantic mackerel (Scomber scombrus) contained insignificant amounts (<3 ppm) of cadaverine and putrescine after 7 and 3 d of iced (1 °C, 34 °F) storage (Ritchie and Mackie 1980). These reports are in contrast to an earlier one, where the histamine content of albacore tuna was reported as 7.5 mg/100 g (75 ppm) of fresh fish during unloading from a fishing vessel (Leitao and others 1983) and 82.5 mg/100 g (825 ppm) after 33 d of ice storage (Price and others 1991). (These high values may be indicative of mishandling on the fishing vessel and the resulting fish decomposition). Also, Shewan and Liston (1955) reported that histidine was easily decarboxylated at 0 °C (32 °F). The variability of these findings reflects the challenges of representative sampling, species differences, quality of initial raw material, and other experimental conditions. Nevertheless, control of biogenic amine production by low temperatures (for example, 0 °C, 32 °F) is a constant observation.

Klausen and Huss (1987) found no histamine formation in mackerel stored in ice, whereas a rapid increase was noted at 10 °C (50 °F). Interestingly, storage at 10 °C for 2 d with no detectable histamine formation and subsequent storage at 0 °C (32 °F) led to formation of 200 ppm histamine after 8 d. These studies indicate that although the histamine-forming bacteria do not grow at 0 °C, decarboxylase formed during growth at 10 °C may be active at 0 °C.

Temperature-abuse potentiates histamine formation in fresh mahimahi (Coryphaena hippurus) at 32 °C (86 °F) increasing from 1.6 ppm to 2,920 ppm in 24 h (see Table IV-11) (Baranowski and others 1990). In a second study, Baranowski and others (1990) showed that heat penetration was much more rapid during incubation in seawater than in air, affecting the histamine content and quality score of the fish (Table IV-7).

Table IV-7. Effect of air incubation on decomposition of mahimahi held 18 h at 32 °C (from Baranowski and others 1990)

Storage Condition*	Histamine (ppm)	Quality Score**
Seawater	1,230	2.8
Air	101	6.0
*Four (4) fish per treatment **Decreasing 10-point scale where 10 - 9 = fresh, acceptable; 8 - 6 = slight decomposition; 5 - 3 = definite decomposition; and 2 - 1 = advanced decomposition

The histamine formation of big eye tuna (Thunnus obesus) and skipjack (Katsuwonus pelamis) tuna during storage at 4 °C (39 °F), 10 °C (50 °F), and 22 °C (72 °F) occurred very quickly at 22 °C, exceeding 50 mg / 100 g (500 ppm) in 1 d for skipjack and 2 d for big eye tuna (Silva and others 1998). The rise in histamine content was delayed at refrigerated temperatures (10 °C and 4 °C), but notable amounts were detected after 3 d at 10 °C and 6 d at 4 °C (Table IV-8).

Table IV-8. Changes in histamine (ppm) in big eye tuna and skipjack during storage at 4 °C, 10 °C, and 22 °C (from Silva and others 1998).

	Histamine in Skipjack			Histamine in Big Eye Tuna
Days	4 °C	10 °C	22 °C	4 °C	10 °C	22 °C
0	-	0	0	0	0	0
1	-	0	600	-	-	300
2	-	-	3,000	-	-	500
3	0	1,000	3,500	0	300	1500
6	1,000	5,500	-	250	1,500	-
9	2,500	8,750	-	1,000	4,250	-
12	4,000	7,000	-	1,200	1,750	-

The changes in histamine content during storage at 5 °C (44 °F), 20 °C (68 °F), and 35 °C (95 °F) were examined in the ordinary and dark meats of sardine, saury pike, mackerel, yellowtail, skipjack, big eye tuna, and horse mackerel (Yamanaka and others 1984). During storage at 20 °C and 35 °C, histamine was produced and accumulated >500 ppm levels at 2 - 6 d of storage, depending on the species. During 5 °C storage, however, the amounts of histamine gradually increased up to those levels in 9 d in sardine, saury pike, mackerel, and horse mackerel. Histamine formation in dark meat was less than that in white meat at the same temperature.

Recently, changes in histamine, cadaverine, putrescine, and agmantine contents were examined in sardine (Sardina pilchardus), Atlantic horse mackerel (Trachurus trachurus), chub mackerel (Scomber japonicus), and Atlantic mackerel (Scomber scombrus) during ice storage (2 - 3 °C, 35 - 37 °F) and storage at room temperature (20 - 23 °C, 68 - 73 °F) (Mendes 1999). At day 0, the initially high aerobic colony counts were 10⁵ - 10⁶ cfu / g. They reached a maximum within 48 - 55 h in fish stored at 22 - 23 °C, but only after prolonged times (10 - 16 d) in fish at 2 - 3 °C. Histamine formation, as well as other amines, varied greatly with species of fish and storage conditions. The levels of histamine, cadaverine, and putrescine increased gradually in all species as decomposition progressed, regardless of storage temperatures, and reached maximum limits for human consumption after 24 h of storage at room temperature. In contrast, amine production in iced fish was considerably reduced and histamine concentration increased slowly until day 7, after which a significant rise was detected, but generally was below 100 mg / Kg. No correlation was observed for histamine or other amine levels and the degree of fish decomposition. Consequently, the belief that decomposition protects consumers from hazardous biogenic amines seems disputable. Again in contrast, a recent report (Kaneko 2000) described the development of a Hazard Analysis Critical Control Point (HACCP) approach using Vessel Standard Operating Procedures for control of histamine on Hawaiian fishing vessels. They concluded that odors of decomposition were reliable indicators of histamine risk and that sensory evaluation is an effective HACCP control measure in the Hawaiian fishery. In this study, 583 mixed pelagic fish (fresh bigeye, yellowfin, albacore tuna, striped marlin, blue marlin, and mahimahi stored in ice) were sampled at the time of delivery from commercial fishing vessels. Fish were graded for quality, by using sensory indicators of decomposition, and analyzed for histamine concentration. A total of 119 fish were rejected because of decomposition. Only 14 fish exceeded 5mg / 100 mg (50 ppm) histamine defect action limit. All 14 fish were first rejected from the market because of odors of decomposition. None of the fish that passed the sensory evaluation exceeded the defect action limit. These conflicting results pose a challenge if biogenic amines are to be used as legal safety indices. Another challenge is to develop an acceptable definition for an odor of decomposition.

It is imperative to recognize that the fish species affect the production of biogenic amines and that many species are rarely, if ever, currently utilized in cold-smoked fish products. Variable and differing data are frequently reported. For instance, 34 albacore tuna (Thunnus alalunga) samples left on the deck (deck temperature 15.5 - 23.5 °C, 59 - 73 °F) for <12 h contained negligible histamine (<0.40 mg/100 g muscle [4 ppm]). Four samples left on the deck for up to 24 h did not contain any significant amounts of histamine (<0.18 mg / 100 g [1.8 ppm]). Among 9 fish left at up to 4 d, only 2 exhibited histamine levels higher than 5.0 mg / 100 g (50 ppm), that is, 9.31 and 6.19 mg/100 g (93.1 and 61.9 ppm) (Ben-Gigiery and others 1998b). Kahawai (Arripis trutta) incubated for 2 d, however, showed average histamine levels of 150 and 350 mg / 100 g (1,500 and 3,500 ppm) in two different trials (Fletcher and others 1995). Spanish mackerel (Scomberomorus maculatus) fillets contained histamine of 1.8 and 23.8 mg/100 g (18 and 238 ppm) when incubated at 24 °C (75 °F) for 1 and 2 d, respectively (Edmunds and Eitenmiller 1975). Mahimahi (Coryphaena hippurus) developed histamine of 154 mg/100 g (1,540 ppm) when incubated at 21 °C (7 °F) for 2 d (Baranowski and others 1990). The optimum temperature for histamine formation in albacore tuna (Thunnus alalunga) was 25 °C (77 °F) (100 mg / 100 g [1,000 ppm]) in whole fish stored for 7 d.

Fresh pink salmon (Oncorhynchus gorbuscha) fillets and whole Pacific herring (Clupea harengus pallasi) were stored for 2 wk at 10 °C (50 °F) to determine if significant amounts of histamine were produced prior to spoilage (Crapo and Himelbloom 1999). Spoilage odors in salmon were moderate by day 4 and intense by day 7, while herring had detectable spoilage by day 4 and became potent by day 6. Aerobic colony counts increased from 10² -10³cfu / g initially to 10⁷ - 10⁸ cfu / g by day 14. Histamine was not detected in salmon, while concentrations reached 55 ppm in herring at day 14. If spoilage were to be used as protection from histamine poisoning, according to this study, 10 °C would be an appropriate temperature to store salmon and herring, since toxic levels were not reached before spoilage occurred. Again caution must be taken because variability among many of the findings on decomposition-histamine relationships reflect the challenges of representative sampling, species differences, and other experimental conditions.

The effect of temperature has also been investigated through inoculation studies with histamine-producing bacteria. Biogenic amine concentrations and sensory changes in fresh and Morganella morganii inoculated blue fish (Pomatomum saltatrix) stored at 5, 10, and 15 °C (41, 50, and 59 °F) were reported by Gingerich and others (1999). Histamine content in fresh fish ranged from <1 to 99 ppm, with an average of 39 ppm. Putrescine and cadaverine were not present. Within 5 d of storage, high concentrations of histamine occurred while the fish were judged acceptable for consumption by the sensory panel. Kim and others (2000) isolated histamine-producing bacteria from albacore tuna stored at 0, 25, 30, and 37 °C (32, 77, 90, 98 °F). The optimum temperature for growth of histamine-producing bacteria was 25 °C. The bacterium producing the highest level of histamine isolated from fish abused at 25 °C was identified as M. morganii. The M. morganii isolate was inoculated into tuna fish infusion broth medium, and the effect of temperature was determined for microbial growth and formation of histamine and other biogenic amines. The isolate produced the highest level of histamine, 5,253 ppm, at 25 °C in the stationary phase. At 15 °C, histamine production was reduced to 2,769 ppm. Neither microbial growth nor histamine formation was detected at 4 °C. Cadaverine, putrescine, and phenylethylamine were also detected. The optimum temperature for histamine, cadaverine, putrescine, and phenylethylamine formation was 25 °C.

Histamine production by P. morganii, P. vulgaris, and H. alvei cultures isolated from skipjack tuna (Katsuwonus pelamis) was measured at storage temperatures of 1, 7, 19, and 30 °C (4, 44, 66, 90 °F) in skipjack infusion broth (Arnold and others 1980). The highest histamine concentrations were observed at 19 and 30 °C depending on the bacterial species (Table IV-9). No histamine was formed at 1 °C, indicating that rapid cooling of tuna flesh may adequately suppress histamine formation. At 19 °C and 30 °C, the Proteus organisms at first formed high levels of histamine, much of which was subsequently destroyed. It appears the histamine concentration may eventually depend on an equilibrium between histamine production and destruction. The authors noted that their conclusion was similar to that of other investigators who reported that tuna flesh homogenate incubated at 25 °C was able to produce approximately 600 mg / 100 g (6,000 ppm) histamine on day 1 but only 350 mg / 100 g (3,500 ppm) remained on day 3. However, this study was performed in an infusion broth and not in fish muscle, which may have changed the results.

Table IV-9. Histamine concentration in skipjack infusion broth (mg/100 ml histamine) (Arnold and others 1980)

	Temperature
	30 °C			19 °C			7 °C
Microorganism	1	3	7	1	3	7	1	3	7
Proteus morganii	370	120	120	365	200	180	130	140	180
Proteus vulgaris	340	220	180	345	235	200	130	192	180
Hafnia alvei	0	70	110	0	20	44	0	0	4

Post-harvest antimicrobial treatments did not show much promise in inhibiting histamine formation. Fish were incubated in seawater (off the coast of Hawaii) and in seawater containing 100 ppm of sodium hypochlorite or chlorine dioxide; however, neither histamine formation nor quality loss was inhibited (Table IV-10) (Baranowski and others 1990).

 

Table IV-10. Effect of chlorine disinfectants (100 ppm) on decomposition of mahimahi incubated 18 h at 32 °C (Baranowski and others 1990)

Incubation medium*	Histamine (ppm)	Quality Score**
Seawater (SW)	1,230	2.8
SW + sodium hypochlorite	2,340	3.0
SW + chlorine dioxide	2,360	2.5
*Four (4) fish per treatment. **Decreasing 10-point scale where 10 - 9 = fresh, acceptable; 8 - 6 = slight decomposition; 5 - 3 = definite decomposition; and 2 - 1 = advanced decomposition

4.3 Freezing

Albacore tuna (Thunnus alalunga) specimens of high quality were analyzed for their biogenic amine contents after 1, 3, 6, and 9 mo of frozen storage at -18 °C (-0.4 °F) or -25 °C (-13 °F) by Ben-Gigirey and others (1998a). Putrescine showed the greatest increase, reaching concentrations of 59 ppm (815% of the initial level) and 68 ppm (942% of the initial level) in the white muscle after 9 mo of storage at -18 °C and -25 °C, respectively. Cadaverine, histamine, and spermidine concentrations were below 3, 5, and 11 ppm respectively after 9 mo of frozen storage.

Fresh mackerel (Scomber scombrus) with no detectable histamine contained 3, 51, and 53 mg / kg (3, 51, and 53 ppm) when stored at -20 °C (-4 °F) for 11, 22, and 33 wk, respectively (Zotos and others 1995). This is in contrast to a report by Hardy and Smith (1976) who stored high quality mackerel (Scomber scombrus) at -14, -21, and -29 °C (7, 6, -20 °F) for 72 wk (1.5 years) and reported no measurable histamine formation.

Studies on the effect of incubation at 32 °C (86 °F) after frozen storage for 24 wk showed that histamine levels were greatly reduced (Table IV-11). Furthermore, fish frozen for 40 wk had almost no histamine formation during incubation suggesting that its microflora had undergone a greater reduction that occurred during the 24-wk storage period (Baranowski and others 1990).

Table IV-11. Effect of prior frozen storage at -20 °C on histamine formation and decomposition during incubation at 32 °C (from Baranowski and others 1990)

	H at 32 °C
	0	6	12	24
Frozen storage* (wk)	Histamine (ppm)
0	1.6	1.5	18.0	2,920
24	0.5	0.8	0.5	850
40	0.7	0.7	1.2	12
	Quality Score**
0	9.6	7.0	4.7	1.4
24	8.9	6.7	4.8	1.6
40	8.7	7.4	5.9	2.0
*Four (4) fish per treatment. **Decreasing 10-point scale where 10 - 9 = fresh, acceptable; 8 - 6 = slight decomposition; 5 - 3 = definite decomposition; and 2 - 1 = advanced decomposition

The findings of Baranowski and others (1990) may be explained by the data of Fujii and others (1994) and Mendes and others (1999). The specific activity of histidine-decarboxylase of halophilic histamine-forming bacteria, Photobacterium phosphoreum and Photobacterium histaminum, remained at 27 - 53% of the initial value after 7 d of storage at -20 °C (-4 °F) (Fujii and others 1994). During this time the viable cells decreased by more than 6 log cycles of the initial counts. Similarly, Mendes and others (1999) reported that freezing sardines (Sardina pilchardus) reduced the numbers of bacteria capable of forming biogenic amines; however, even when the post-freezing viable count of histamine-forming bacteria is low, earlier reports (Yamanaka and others 1982; Karolus and others 1985; Yamanaka and others 1987b) suggested the possibility that if the fish had been temperature-abused before freezing, histamines may still be present in toxic amounts. Therefore, it is extremely important to know the temperature history of the frozen fish, since outbreaks of scombroid fish poisoning can be caused by the ingestion of frozen-thawed fish containing degradation products if the fish were previously temperature-abused.

4.4 Salting

Table IV-12. Bacterial growth characteristics during storage of sardines (Ababouch and others 1991a)

A subsequent study by Leroi and others (2000) showed that the inhibition of bacteria in cold-smoked salmon stored for 5 wk at 5 °C (41 °F) and salt (5% wt/wt) and smoke was linearly proportional to the salt and smoke content (the higher the concentration, the greater the inhibition). No synergistic inhibition effect was observed between the two factors.

4.5 Smoked product

Gessner and others (1996) reported a scombrotoxicosis-like illness occurring in an individual within 10 min after eating a 25 g strip of home-smoked sockeye salmon. The meat came from 1 of 8 salmon caught and stored in a cooler for up to 12 h. Strips cut from the fish bellies had been placed in saltwater brine for 7 min, cooled with a fan for 6 h, and smoked for 2 d at a maximum temperature of 38 - 44 °C (100 - 111 °F) using untreated alder chips. A random sample of 6 strips was tested by the FDA and showed a mean histamine level of 0.19 mg / 100 g; a mean putrescine level of 0.67 mg / 100 g; and a mean cadaverine level of 0.19 mg / 100 g. Two fish strips had water phase salt concentrations of 2.7% and 2.1%. The patient ate an estimated 0.0006 mg of histamine/kg of body weight, well below the estimated 1 mg of histamine/kg of body weight reported to cause illness. The authors did not give a reason for this apparent high sensitivity to such a low concentration of histamine. This reported illness of a single individual is contrary to other published information on histamine toxicity. Also, the samples tested may not have represented the product consumed by the subject of the illness.

The production of biogenic amines during chill storage (5 °C, 41°F)) of cold-smoked salmon (Salmo salar) from 3 smoke houses over a 2-year period (1997 and 1998) was studied by Jorgensen and others (2000). Results of the study showed the production of biogenic amines is unlikely to result in histamine poisoning in humans as indicated by epidemiological data (Table IV-13). Some samples exceeded the defect action level of 50 ppm established by the FDA for Scombridae and 100 - 200 ppm by E.U. regulations for Scombridae and Clupeidae, but no samples reached toxic levels of 500 ppm, a value at which one would expect illness and that the FDA would use in legal proceedings (EEC 1991; FDA 1998).

Although the temperatures used for a hot-smoking process may inhibit histamine producers, cold-smoking does not expose the fish to temperatures high enough to inhibit the latter bacteria. The effect of hot-smoking previously frozen mackerel (Scomber scombrus) on histamine formation was reported by Zotos and others (1995) (Table IV-14). Smoking was done for a total of 7 h, at sequential temperatures of 30, 40 and 70 °C. From Table IV-14 it can be observed that a significant (p >0.05) increase in histamine formation in fresh, frozen (11 or 33 wk) mackerel was solely due to the smoking process. The histamine increase appeared to be independent of frozen storage time prior to smoking. Although this is a hot smoking process example, it demonstrates the importance of controlling the temperature and time of the smoking process.

Table IV-14. Effect of smoking previously frozen mackerel on histamine formation (mean ± SD)( Zotos and others 1995)

In many situations the production of biogenic amines is highly variable and difficult to predict. For example, cold-smoked, fermented rainbow trout (Oncorhynchus mykiss) were prepared with 3 different lactic acid bacteria (LAB) inocula plus staphylococci, with the control group being prepared without inoculum (Petaja and others 2000). The fish were cured by injecting brine (20% NaCl, 18% glucose, 0.5% ascorbic acid, and 0.625% KNO₃) at amounts corresponding to 5% of the weight of the fish fillet. The lactic acid bacteria inoculum was at 10⁷ cfu / g and staphylococci at 5 x 10⁶ cfu / g. The products were acceptable by sensory analysis, the LAB inoculum grew to >10⁸ cfu / g, the pH reduced to 5.0 - 5.3, and a_w to 0.927 and the pseudomonads, the predominate flora, disappeared. The fish raw material and products contained low amounts of biogenic amines with one exception: cadaverine, histamine, and tyramine increased in all product groups except in one experimental series (II) out of three (Table IV-15). This broad variability was again evident in this report.

Microbiological, chemical, and sensory changes in cold-smoked salmon were studied during 5 wk of vacuum storage at 5 °C (41 °F) (Leroi and others 1998). Total aerobic colony counts reached 3 x 10⁶ after 6 d; however, the shelf life was judged by a sensory panel to be acceptable for 2-3 wk. During the first 2 wk, gram-negative bacteria were dominant, mainly represented by Swanella putrefaciens immediately after the smoking process and then Photobacterium phosphoreum. Aeromonas spp. were present throughout the storage but in smaller amounts. Gram-negative bacteria then progressively decreased while gram-positive bacteria increased, dominated by LAB. A diversification was observed at the end of storage, with the appearance of Lactobacillus farciminis, Lactobacillus sake, and Lactobacillus alimentarius.

Table IV-15. Levels of biogenic amines in experimental fish fillets after 0 and 3 d of fermenting and 35 d after preparation (7 d of fermenting +28 d of storing) (Petaja and others 2000)