1.

Introduction

0740-0020/$ - see front matter 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0740-0020(03)00059-5

Several species of marine and freshwater fish including sea bream (Sparus aurata), sea bass (Dicentrarchus labrax) and rainbow trout (Onchorynchus mykiss) are being farmed in Greece and in other Mediterranean counties over the last decade in order to meet the increasing demand for fresh rather than frozen fish (Urch, 1994). Of the freshwater fish species, rainbow trout (O. mykiss) is being farmed mainly in the river waters of North Western Greece and is sold as either whole fresh fish in retail markets, or in fillet form in supermarkets and chain stores. Additionally, trout
"Corresponding author. Tel.: +30-265-10-98-342; fax: +30-265-1098-795. E-mail address: mkontomi@[Link] (M.G. Kontominas).

Freshness is the single most important attribute when assessing fish quality. Microbiological, biochemical and sensory changes are associated with deterioration of fish quality during handling and storage (Ehira and Uchiyama, 1986). Although a variety of biochemical, physical (Gill, 1992, 1997) and microbiological methods (Gram and Huss, 1996) have been used to assess fish freshness, sensory evaluation is still the most satisfactory method of achieving such a goal (Connell, 1975; Reineccius, 1990). Given that specific spoilage causing micro-organisms cannot be detected by organoleptic or chemical testing it is useful to conduct microbiological, chemical and organoleptic analyses when assessing the quality of fish (Ryder et al., 1993). Few studies have been published on the quality characteristics of freshwater fish during ice storage in contrast to a wealth of information available for marine fish species from tropical or cold waters (Shewan and Murray, 1979; Liston, 1980, 1992; Kyrana et al., 1997; Alasalvar et al., 2001; Masniyom et al., 2002; Gimenez et al., 2002; Kyrana and Lougovois, 2002; Tejada and Huidobro, 2002). The shelf-life of headed and gutted rainbow trout (Salmo gairdneri) was evaluated by Dawood et al. (1986) over a 14day period of storage in ice. The results indicated that when fish had been held at high ambient temperatures of 30C for 6 h before icing, there was a rapid deterioration in quality as shown by a linear increase in hypoxanthine values. Randell et al. (1997) studied the quality of filleted rainbow trout (S. gairdneri) in over-wrap packages, vacuum and gas packages stored at 2C. Results of microbiological and sensory analyses showed that the quality of trout fillets deteriorated faster in over-wrap and vacuum packages than in gas packages. The sensory characteristics and biochemical changes of freshwater rainbow trout (O. mykiss) during chilled storage were studied by Rodriguez et al. (1999). Trout samples were stored for up to 12 days in (a) ice without gutting, (b) in ice after gutting and (c) under refrigeration after gutting and vacuum packing. Results suggested that the chemical parameters: hypoxanthine and K value were useful as indicators of the freshness of trout stored in ice, regardless of whether fish were whole or gutted. However, these parameters were not useful as freshness indicators in vacuum-packaged trout under refrigeration. In another study of quality assessment of vacuum-packaged 'gravad' rainbow trout fillets, a shelf-life of 20 days (3C) and 18 days (8C) was reported based on microbiological and sensory analyses (Lyhs et al., 2001). Recently, the quality of aquacultured rainbow trout fillets packaged in over-wrap, vacuum and in modified atmospheres was studied by measurement of chemical (pH, TVB-N, hypoxanthine, TBARS (Thiobarbituric-reactive substances)), microbiological (aerobic psychro-trophic flora) and sensory parameters (Gimenez et al., 2002). Modified atmosphere packaging (MAP) gave a

fillets in the form of smoked and vacuum packaged products are being exported to various northern European countries and consumed with no further heating. The quality of fresh fish is a major concern to industry and consumers. Like marine fish, freshwater fish are extremely perishable food commodities. Deterioration of fish mainly occurs as a result of bacteriological activity leading to loss of quality and subsequent spoilage (Liston, 1980). Bacterial spoilage in refrigerated fish and fish products under aerobic storage conditions is caused by Gram-negative psychrotrophic organisms such as Pseudomonas, Alteromonas, Shewanella and Flavobacterium spp. (Hubbs, 1991). Faulty rearing, harvesting, and processing practices can result in crosscontamination of fish with foodborne pathogenic bacteria (ICMSF, 1998). significant extension of shelf-life of filleted trout as compared to vacuum and over-wrap packaging, due to the inhibitory effect of CO2 on microbial growth resulting in minor production of spoilage compounds (TVB-N hypoxanthine). Given the commercial significance of trout filleting and whole trout storage in ice, the present study was undertaken to evaluate the effect of filleting on the quality of fresh rainbow trout during ice storage.

2. Materials and methods

2.1. Preparation of the fish samples and storage conditions Aquacultured freshwater rainbow trout, O. mykiss (average weight and length: 400 g and 250 mm) was obtained from an aquaculture farm located on river Voidomatis in North Western (GIANNETAS, SA, Greece). The fish were harvested during the period of JulyAugust 2002. The fish were slaughtered by immersing in icecold water (hypothermia), and divided into two lots. One lot was headed, gutted and filleted; the second lot included whole trout. Both were delivered to the laboratory (whole and filleted) within 6 h of harvesting, packed in insulated polystyrene boxes containing ice. Six fish (3 whole and 3 filleted) were immediately sampled (day 0), while the rest were covered with a thin polyethylene film and an appropriate quantity of flaked ice on top. The polystyrene boxes, provided with outlets for water drainage, were stored in a refrigerator (2 + 0.5C). The ice/fish ratio (3:1) was maintained constant throughout the experiment. These methods of fish handling and storage conditions closely simulate normal commercial practice applied by the fish farming industry in Greece. After 0, 3, 6, 9, 12, 15 and 18 days, three randomly chosen fish were removed from ice, weighed, and their sensory attributes were determined as described in the sensory assessment section. On each sampling occasion three fish (whole and filleted) were analysed. 2.2. Microbiological analysis 2.2.1. Sample preparation A sample was taken from the flesh of the anterior-dorsal region of each whole fish or fish fillet. About 25 cm 2 of the skin of filleted samples only were rinsed with 70% ethanol and flamed. The skin was aseptically removed and ca. 25 g (25 cm2) of the underlying flesh were sampled to the bone using sterile scalpels and forceps. Tenfold dilutions in 0.1% peptonewater were prepared and each sample was homogenized for 60 s using a Lab Blender 400, stomacher (Seward Medical,

S. Chytiri et al. / Food Microbiology 21 (2004) 157-165

London, UK). 2.2.2. Microbiological media and enumeration For microbial enumeration 0.1 ml samples of serial dilutions (1:10, diluent, 0.1% peptone-water) of fish homogenates were spread on the surface of dry media, except for 1:10 dilution where 1 ml sample was used (see below). Mesophilic counts were determined using Plate Count Agar (PCA, Merck), after incubation for 48 h at 30C; Pseudomonads were enumerated on cetrimide fusidin cephaloridine agar (CFC, Oxoid code CM 559, supplemented with supplement SR 103, Oxoid, Basing-stoke, UK) and incubated at 20C for 2 days (Mead and Adams, 1977); STAA (streptomycin sulphate-thallous acetate-actidione) agar (Oxoid code CM 881, supplemented with selective supplement SR 151) was used for enumeration of Brochothrix thermosphacta, prepared from basic ingredients in the laboratory and incubated at 20C for 3 days (Gardner, 1966). For Enterobacter-iaceae and H2S-producing bacteria (typical of Shewa-nella putrefaciens) enumeration, a 1.0 ml sample was inoculated into 10 ml of molten (45C) violet red bile glucose agar (VRBGA, Oxoid code CM 485, Basing-stoke, UK) and Iron Agar (IA, Oxoid code CM 867, Basingstoke, UK), respectively. After setting, a 10 ml overlay of molten medium was added. For the former (VRBGA), incubation was carried out at 30C for 24h. The large colonies with purple haloes were counted (Mossel et al., 1979). Iron agar plates were incubated at 20 C for 4 days; black colonies formed by the production of H2S were enumerated after 2-3 days (Gennari and Campanini, 1991). Three replicates of at least three appropriate dilutions (10\ 1 0 3 , 1 0 5) depending on the sampling day were enumerated. A detection limit of ca. 1.0 logcfu/cm2 for the types of bacteria analysed in this study was achieved by spreading a total of 1.0 ml inoculum of the 1:10 dilution into 3 Petri plates (approximately 330 ml in each plate) and thus enumerating colonies to give a total count of 1 log cfu/cm2. Microbiological data were transformed into logarithms of the number of colony forming units (cfu)/ cm2. All plates were examined visually for typical colony types and morphology characteristics associated with each growth medium. 2.3. Chemical analysis The anterior-dorsal half of each whole fish or fish fillet (25 g) were homogenized in a Waring blender for 60 s. Appropriate quantities of homogenized fish were used for determination of the following chemical parameters. Trimethylamine (TMA), total volatile basic nitrogen (TVB-N) and thiobarbituric acid (TBA) were determined as follows: TMA analysis was carried out according to the method proposed by Dyer (AOAC,

1990). TVB-N was determined according to the method of Malle and Poumeyrol (1989). TBA was determined according to the method proposed by Pearson (1976). The pH value was recorded using a glass electrode being applied directly to the anterior-dorsal region of fish flesh (Metrohm 691, Herisan, Switzerland). Five readings were averaged for each sample. 2.4. Sensory assessment 2.4.1. Raw fish Sensory analysis was conducted on whole trout according to the European Community (EC) grading scheme by a taste panel consisting of five experienced judges from the laboratory staff, trained in grading fish (Howgate et al., 1992). The appearance of the skin, eyes, gills and internal organs, surface slime, and the odor and texture of each fish (whole) was assessed into four quality grades. In this EC grading scheme, excellent quality (perfect condition), high quality (slight loss of excellent characteristics), good quality (some deterioration, but fit for sale) and unfit for sale were assigned by E, A, B and C grades, respectively. The five panelists were also presented with a freshly thawed fish sample, stored at 30 C throughout the experiment, this serving as the control sample. 2.4.2. Cooked fish The attributes of cooked fish (whole and filleted) were evaluated by a panel of five experienced judges on each day of sampling. Fish samples (200 g of fish fillets) were cooked individually in a microwave oven at full power, for 5 min including defrosting time and immediately presented to the panelists. Panelists were laboratory trained. Sensory evaluation was conducted in individual booths under controlled conditions of light, temperature and humidity. Panelists were asked to score odor, taste and texture of fish using a 0-10 acceptability scale. (Score 10-9 = excellent, 8-7 = very good, 6-5 = acceptable, < 5 = unacceptable). A score of seven (7) was used as the point of first off-odor/off-taste development. 2.5. Statistical analysis Experiments were replicated twice on different occasions with different fish samples. Analyses were run in triplicate for each replicate. Data were subjected to analysis of variance (ANOVA). The least significant difference (LSD) procedure of SPSS (SPSS, 1995) was used to test for differences between means (Steel and Torrie, 1980). Results of microbiological and sensory analyses were reported as mean values 7 standard error. Results of chemical analyses were reported as mean values 7 standard deviation.

S. Chytiri et al. / Food Microbiology 21 (2004) 157-165

3. Results and discussion

3.1. Microbiological analysis

Changes in microbial flora of whole and filleted aquacultured rainbow trout during storage in ice are shown in Figs. 1a-e. Initial mesophilic viable counts (Fig. 1a) of whole ungutted and filleted rainbow trout were ca. 2.5 and 3.8 logcfu/cm2, respectively (day 0). Mesophilic counts reached ca. 7.0 logcfu/cm2 after 18 days of storage for whole ungutted trout and after 10 days for filleted samples. In all cases, counts of filleted
10 8 6
CM

E o o

(a)

0 Days in ice 0 6 2 8 4 10

10 8 10 8 CM CM 4 E6 o6 4 E ) ) O 2 o 2

0 0 2 4 1 0 1 2 1 4 1 6 1 8 2 0 10 8 CM 4 E6 o ) 2 o Days in ice 6 8

12

14

16

18

20 Days in ice

0
T 1

2
1

4
1

6
1

8
1

10 12 14 16 18 20 Days in ice
1 1 1 1 1

10 8 CM E6 4

o 0
0 2 4 6 8 10 12 14 16 18 20

Days in ice

Fig. 1. Changes in mesophilic counts (a), Pseudomonas spp. (b), H2Sproducing bacteria (c), B. thermosphacta (d), Enterobacteriaceae (e) of aquacultured trout stored in ice. Whole ungutted (A) and filleted trout (). Points represent mean values of six determinations 7 standard error (n = 2 x 3).

trout were ca. 1.7-2.3 log cycles higher (P<0.05) than those for the whole ungutted trout samples. Although it is widely accepted that the initial microbial load of freshwater fish varies depending on water conditions and temperature, most available literature on different freshwater fish species (tilapia, striped bass, rainbow trout, silver perch) reports bacterial counts of 102106cfu/g (Acuff et al., 1984; Nedohula and Westhoff, 1997; Gelman et al., 2001; Savvaidis et al., 2002). In this study, initial mesophilic counts of 2.5 logcfu/cm 2 for whole ungutted trout and 3.8 logcfu/cm 2 for filleted samples are indicative of high fish quality and good manufacturing practices, considering the microbiological upper limit for fresh fish, proposed by ICSMF (1986) for aerobic mesophilic plate counts at 30C (7 logcfu/cm 2). The mesophilic count (2.5 logcfu/cm 2) of whole ungutted trout found in the present study is in correlation with results (total viable counts of ca. 102-103 logcfu/cm2) reported for

wild brown and farmed rainbow trout (Gonzalez, 1999), and for vacuum-packaged gravad trout slices reported by Lyhs et al. (2001). The initial mesophilic count (3.8 logcfu/cm2) of filleted trout in the present study is most probably associated with cross-contamination of trout samples during the filleting process. Such possible sources of cross-contamination may include the house aerobic microflora, the various utensils being used (cutting board, knives), the personnel (worker's hands), the ice water being used for the slaughtering of samples. Gonzalez-Rodriguez et al. (2001) reported even higher initial counts (5.3 logcfu/cm2) for rainbow trout fillets prepackaged in trays stored at 3C. Mesophilic counts for whole ungutted and filleted trout exceeded 7 log cfu/ cm2 after 18 and 10 days of storage, respectively. These values are in good agreement with those obtained for European whole sea bass (D. labrax) stored in ice (Kyrana and Lougovois, 2002), and sea bass slices (Lates calcalifer) stored in 60% CO2 atmosphere (Masniyom et al., 2002). Koutsoumanis and Nychas (1999), however, reported counts of 7 log cfu/cm 2 for Mediterranean boque (Boops boops) after 4 days of aerobic storage under refrigeration. Same counts were reported by Ordonez et al. (2000) for hake steaks after 7 days under a CO2 enriched atmosphere and by Savvaidis et al. (2002) for farmed vacuum-packaged rainbow trout after 8 days of refrigerated storage. Recently, Gimenez et al. (2002) reported an aerobic psychrotrophic count of 7 logcfu/cm2 for farmed filleted rainbow trout in overwrap, vacuum packages and under modified atmosphere after 6, 10 and 14-17 days, respectively. Pseudomonads (Fig. 1b) were dominant in both whole ungutted and filleted trout over the entire storage period of 18 days. Initial Pseudomonads counts of whole ungutted and filleted rainbow trout were ca. 1.0 and 3.5 logcfu/cm 2, respectively (day 0). A count of ca. 7.0 logcfu/cm2 was reached for the filleted trout after 10 days of storage whereas the whole ungutted trout reached a count of 6.0 logcfu/cm2 at the end of the experiment. In all cases, Pseudomonads counts were higher (P<0.05) by ca. 2.5-2.8 log cycles for the filleted trout samples as compared to the whole ungutted samples. H2S-producing bacteria (including S. putrefaciens) (Fig. 1c)and B. thermosphacta (Fig. 1d) constituted a large proportion of the microflora of the whole ungutted and filleted rainbow trout samples. Interestingly, initial counts of these bacteria remained at low levels (1 and 2.2-2.7 logcfu/cm2) for both whole ungutted and filleted trout samples respectively over the first 6 days of storage in ice. A H2S-producing bacteria (including S. putrefa-ciens) and B. thermosphacta count of 7 log cfu/cm 2 was reached for the filleted trout after 18 days of storage whereas the whole ungutted trout reached a respective count of 5.5 logcfu/cm 2 at the end of the experiment. In all cases, H 2S-producing bacteria (including S. putrefa-ciens) and B. thermosphacta counts of filleted trout were ca. 1-2 logcfu/cm 2 higher (P<0.05) than those of the whole ungutted trout samples. Pseudomonas spp. and H2S-producing bacteria (including S. putrefaciens) dominated the microflora of whole ungutted and filleted fish during storage in ice. These two bacteria have been reported to be the specific spoilage bacteria in fish from temperate and tropical waters (Gillespie, 1981; Lima dos Santos, 1981; Gram and Huss, 1996) and in fresh Mediterranean fish stored aerobically under refrigeration (Koutsoumanis and Nychas, 1999) or ice storage (Gennari et al., 1999). Our findings agree with those of Drosinos and Nychas (1996), Koutsoumanis and Nychas (1999), Koutsouma-nis et al. (1999),and Gonzalez et al. (1999), who reported that the microbial population of fish stored aerobically consists almost exclusively ofGramnegative oxidase-positive psychrotrophic bacteria, Pseudomonas spp. and H2S-producing bacteria (including S. putrefa-ciens). It is interesting to note that a

S. Chytiri et al. / Food Microbiology 21 (2004) 157-165

Pseudomonas spp. count of ca. 7 logcfu/cm 2 for the filleted trout was reached after 10 days of storage in ice, whereas for H2S-producing bacteria (including S. putrefaciens) and B. thermosphacta after 18 days of storage, probably due to the short generation time of Pseudomonas spp. (Shewan, 1977). On the contrary, lower counts of H2S-producing bacteria (including S. putrefaciens), have been reported by Kyrana and Lougovois (2002) for European sea bass stored in ice (ca. 4.5 x 104) at spoilage time (day 15), based on sensory evaluation. These authors suggested that organisms other than S. putrefaciens had been involved in the spoilage of sea bass, without identifying the main spoilers. Since H2S-producing bacteria, mainly S. putrefaciens, and Pseudomonas spp. are two strongly competitive psychrotrophic micro-organisms (Koutsoumanis et al., 1999), the initial lag phase of H2S-producing bacteria (including S. putrefaciens) may be the result of inhibition by Pseudomonas spp. Indeed, it was reported by Gram and Melchiorsen (1996) that Pseudomonas spp. can inhibit the growth of H2S-producing bacteria (including S. putrefaciens) due to the ability of the former to produce siderophores. With regard to B. thermosphacta, a bacterium more common in meat products, it can be stated that B. thermosphacta contributes to a significant extent to spoilage of whole ungutted and filleted trout, in agreement with the finding of Koutsoumanis and Nychas (1999) who studied the spoilage patterns of bogue (B. boops) stored aerobically between 0C and 10C, and those of Ordonez et al. (2000) who studied the spoilage of hake (Merluccius merluccius) stored under CO2 enriched atmospheres. At this point it should be mentioned than only a few studies have been published on B. thermosphacta and its relation to spoilage of marine and freshwater fish including sea bream (Dro-sinos and Nychas, 1996), hake (Ordonez et al., 2000) and rainbow trout (Savvaidis et al., 2002). Enterobacteriaceae were also found to be part of the spoilage microflora of both whole gutted and filleted trout stored in ice. Enterobacteriaceae counts (Fig. 1e) were higher (P<0.05) for filleted than for whole ungutted trout samples by approximately 1.0-1.2 log cfu/cm2 throughout the entire storage period of 18 days, reaching final levels of ca. 5.5 and 4.2 logcfu/cm 2 respectively. The population of this group was lower than that obtained for other bacteria in this study, which is in agreement with results reported for different fresh Mediterranean fish at the end of the product's shelf-life (Gennari et al., 1999; Koutsoumanis et al., 1999; Ordonez et al., 2000; Tejada and Huidobro, 2002). Although, this group can grow at low temperatures, their abundance decreases during ice storage, possibly because their growth rate is lower than that of other Gramnegative psychrotrophic spoilers. On the basis of the microbiological data presented here, this is the first report on microbiological changes associated with whole ungutted and filleted rainbow trout species (O. mykiss) stored in ice. It is apparent that for all micro-organisms examined, populations of filleted fish were higher than those obtained for whole ungutted trout stored in ice throughout the entire storage period. This may be attributed to post processing cross-contamination of trout samples during the filleting process. 3.2. Chemical analyses The changes in TMA, TVB-N, TBA and pH for whole ungutted and filleted trout during the 18-day storage period in ice are shown in T a b l e s 1 a n d 2 . A n initial 15day lag phase in TMA values was recorded in whole ungutted trout samples during storage in ice
Table 2 Changes in TMA, TVB-N, TBA and pH of filleted trout stored in ice

TBA PH ( m MA/g) g 0 1.11 +0.06a 22.51+ 0.71a 10.43+ 0.38a 6.43+ 0.10a 3 1.28+0.02b 21.17 + 0.82a,b 11.09+ 0.47a 6.56+ 0.20a 6 1.23+0.04b 19.82+ 0.56b 12.64+ 0.56b 6.58+ 0.25a 9 1.23 +0.03b 18.11 + 0.64c 13.23+ 0.62b 6.65+ 0.12a 12 2.05+ 0.07c 18.31 + 0.68c 13.42+ 0.44b 6.42+ 0.20a 15 2.51+0.08d 22.31+ 0.73a 19.85+ 0.75c 6.63+ 0.14a 18 6.38+0.05e 26.06+ 0.62d 19.41+ 0.62c 6.52+ 0.10a Mean values of six samples 7 standard deviation (n = 2 x 3). Mean values within a column with the same letter are not significantly different. TMA: Trimethylamine, TVB-N: total volatile basic nitrogen, TBA: thiobarbituric acid, MA: malonaldehyde.

Days in ice

TMA (mg/100g)

TVB-N (mg N/100 g)

(Table 1). The largest TMA value for whole ungutted samples was recorded on day 18. The respective initial lag phase for filleted fish samples was 9 days, after which TMA values increased by ca. 0.8-5.1 mg/100 g flesh until the end of storage (day 18) (Table 2). These values are considered low as compared to the value of 10 mg/100 g flesh proposed by Teskeredzic and Pfeifer (1987) as the upper acceptable limit for trout stored at 4 C. The small increase in TMA values for ungutted trout over the entire storage period in ice reflects the low level of trimethylamine oxide (TMAO) in the flesh of this fish species. TMA is produced by the decomposition of TMAO due to bacterial spoilage and enzymatic activity, (Hebard et al., 1982). Dalgaard et al. (1993) reported that a population of 108-109cfu/g of S. putrefaciens is considered crucial for TMA production. H2S-producing bacteria (including S. putrefaciens) in this study reached at the end of storage ca. 5.5 and 7 logcfu/cm2 for whole ungutted and filleted trout, respectively. Thus, TMA is not is not a particularly useful indicator of trout freshness. Higher TMA values of filleted vs. whole trout samples, however, may be associated with the post-harvest handling conditions involving bacterial contamination during the invasion of visceral fluid into the fish muscle enhanced by the conditions of fish transportation to the laboratory. Similarly, low TMA values have been reported for whole fresh fish (Kyrana et al., 1997; Koutsoumanis and Nychas, 1999; Rodriguez et al., 1999; Tejada and Huidobro, 2002). The TVB-N content ranged from 14.11 to 20.16mg N/100 g flesh for whole ungutted and from 18.11-26.06 for filleted trout samples, respectively, during the 18-day period of storage in ice (Tables 1 and 2). Given that no limit for acceptability of rainbow trout has been established by Decision 95/149 (EU, 1995), Gimenez et al. (2002) proposed a value of 25 mg N/100 g flesh as the highest acceptable level. In our study, all TVB-N values remained below this limit of acceptability throughout the entire storage period in ice with the exception of filleted trout for which a value of 26.06 mg N/100 g flesh was recorded on day 18 of storage (Table 2). Since, TVB-N is produced mainly by bacterial decomposition of fish flesh, the higher values of total viable counts of filleted vs. whole trout (8.6 vs. 7 logcfu/ cm2) after 18 days of storage in ice could account for the higher TVB-N values of filleted fish samples. TVB-N values showed significant fluctuation for both whole and filleted trout samples as a function of storage period indicating that TVB-N is a poor indicator of fish freshness, as also proposed by Dawood et al. (1986), Kyrana et al. (1997), and Tejada and Huidobro (2002). Similar TVB-N values have been reported for whole fish (Kyrana et al., 1997; Kyrana and Lougovois, 2002), silver perch (Gelman et al., 2001), gutted gilthead sea bream fish (Tejada and Huidobro, 2002), and recently for sea bass (L. calcalifer) slices stored in air (Masniyom et al., 2002), and rainbow trout stored in modified atmospheres (Gimenez et al., 2002). TBA is a widely used indicator for the assessment of degree of lipid oxidation (Nishimoto et al., 1985). TBA values for whole ungutted and filleted trout samples

S. Chytiri et al. / Food Microbiology 21 (2004) 157-165

increased throughout the entire 18-day storage period (Tables 1 and 2). TBA values for the filleted samples were significantly higher (P<0.05) than those for the whole ungutted samples. Present results indicate that oxidative rancidity remained relatively low in whole ungutted trout samples throughout the entire period of storage in ice. Kyrana et al. (1997) reported that icing of whole, ungutted, fish tends to slow down the production of malonaldehyde (MA). Similarly, low TBA values have been reported for other whole fish such as farmed sea bream (Kyrana et al., 1997), European sea bass (Kyrana and Lougovois, 2002), and sea bass (L. calcalifer) slices stored in air (Masniyom et al., 2002). It is important to note that according to Auburg (1993), TBA values may not reveal the actual degree of lipid oxidation since MA can interact with other components of fish body. Such components may be amines, nucleosides and nucleic acid, proteins,

phospholipids, and other aldehydes that are end products of lipid oxidation. Such interactions may vary greatly with fish species. Higher TBA values for filleted fish samples are probably due to the filleting process of samples that affects rancidity levels of fish samples, probably as a result of exposure of the fish lipids to atmospheric oxygen, which accelerates oxidation. Si-meonidou et al. (1998) reported similar TBA values of 7.43-20.98 mg MA/g for Mediterranean fatty fish species, chub mackerel, hake, horse mackerel, Atlantic mackerel, striped mullet, sardine, and bogue stored frozen. Filleting of ice-stored trout, therefore, seemed to have no significant effect on rancidity, in agreement with findings reported for ice-stored sea bream (Tejada and Huidobro, 2002). Changes in pH values for whole ungutted and filleted trout samples during the entire period of storage in ice were not statistically significant (P > 0.05) (Tables 1 and 2). The pH value of live fish muscle is close to 7.0, however post mortem pH can vary from 6.0 to 7.0 depending on season, species and other factors (Simeonidou et al., 1998). The pH values are in agreement with those of Ryder et al. (1993), Simeonidou et al. (1998), Rodriguez et al. (1999),and Kyrana and Lougovois (2002). 3.3. Sensory analyses 3.3.1. Raw fish Changes in sensory attributes of the raw fish (whole ungutted trout) during storage in ice were recorded using the descriptions given by the individual panel members (Table 3). Rigor mortis, metallic sheen and iridescence of the skin as well as glossy, bright red gills possessing fresh odor should be considered as

attributes of extreme freshness, whereas loss of brilliance and iridescence, fading of skin colors and bleaching of the gills in patches would indicate stale fish. It can be seen (Table 3) that excellent and very good grades were scored during the first 9 days of storage for whole ungutted trout. Moderate grades were obtained between days 12-15 of storage. Unfit for sale, raw fish samples were obtained after day 15 of storage.

3.3.2. Cooked fish Acceptability scores for odor, taste and texture of whole ungutted and filleted trout evaluated by the panelists, decreased significantly (P<0.05) with time of storage, as shown in Figs. 2a-d. Odor (Fig. 2a) and taste (Fig. 2b) showed a similar pattern of decreasing scores. Odor and taste scores of whole ungutted trout were similar to those of filleted trout at a given sampling time. The limit of acceptability for odor was reached after 15-17 days of storage for both ungutted and filleted trout samples. The same comments hold for taste scores (Fig. 2b). Texture scores (Fig. 2c) of both whole ungutted and filleted trout decreased at slower rate than odor and taste scores. The limit of acceptability for texture was reached after 18 days for filleted trout samples, while this limit was never reached for whole ungutted trout samples throughout the entire storage period in ice.

4. Conclusion

Based on (a) microbial counts (b) odor and taste scores and (c) evaluation of raw fish according to the EC freshness grading scheme, a shelf-life of approximately 15-16 days for the whole and 10-12

S. Chytiri et al. / Food Microbiology 21 (2004) 157-165

days for the filleted trout

may be expected.

1086 4 o2 0

S. Chytiri et al. / Food Microbiology 21 (2004) 157-165

Days in ice

10 8 6 4 2 0

ra
of trimethylamine oxide and its derivatives in fish and shellfish. In: Martin, R.E., Flick, G.J., Hebard, C.E. (Eds.), Chemistry and Biochemistry of Marine Food Products. AVI, Westport, CT, USA, pp. 149-304. Howgate, P., Johnston, A., Whittle, K.J. (Eds.), 1992. Multilingual Guide to EC Freshness Grades for Fishery Products. Marine Laboratory, Scottish Office of Agriculture, Environment and Fisheries Department, Aberdeen.

i-

Acknowledgements

s 10 8 6 4 2
x

e
I-

We wish to thank GIANNETAS, SA, Ioannina, Greece for providing the fish samples.
0 3 6 9 1 2 1 5 1 8 2 1

0 Days in ice

10 sj-s.886 4 2 0

Oo
ca 6 9 1 2

References
1 Days in ice

Acuff, G., Izat, A.L., Finne, G., 1984. Microbial flora of pond-reared tilapia (Tilapia aurea) held on ice. J. Food Prot. 47, 778-780. Alasalvar, C., Taylor, K.D.A., Oksuz, A., Garthwaite, T., Alexis, M.N, Grigorakis, K., 2001. Freshness assessment of cultured sea bream (Sparus aurata) by chemical, physical, and sensory methods. Food Chem. 72, 33-10. AOAC, 1990. Official Methods of Analysis, 14th Edition. Association of Analytical Chemists, Washington, DC. Auburg, S.P., 1993. Review: interaction of malondialdehyde with biological moleculesnew trends about reactivity and significance. Int. J. Food Sci. Technol. 28, 323-335. Connell, J.J., 1975. Control of

Fish Quality, 1st Edition. Fishing News Books Limited, London. Dalgaard, P., Gram, L., Huss, H., 1993. Spoilage and shelf life of cod fillets packed in vacuum or modified atmospheres. Int. J. Food Microbiol. 19, 283-294. Dawood, A.A., Roy, R.N., Williams, C.S., 1986. Quality of rainbow trout chilled-stored after post-catch holding. J. Sci. Food Agric. 37, 421-427. Drosinos, E.H., Nychas, G.-J.E., 1996. Brochothrix thermosphacta,a dominant organism in Mediterranean fresh fish (Sparus aurata) stored under modified atmosphere. Ital. J. Food Sci. 4, 323-329. Ehira, S., Uchiyama, H., 1986. Determination of fish freshness using the K value and comments on some other biochemical changes in relation to freshness. In: Kramer, D.E., Liston, J. (Eds.), Seafood Quality Determination. Elsevier Science, BV, Amsterdam, pp. 185-207. European Union, 1995. Commission decision 95/149/EC, 8, March 1995. Off. J. Eur. Commun. L97, 84-87. Gardner, G.A., 1966. A selective medium for the enumeration of Microbacterium thermosphactum in meat and meat products. J. Appl. Bacteriol. 29, 455-460. Gelman, A., Glatman, L., Drabkin, V., Harpaz, S., 2001. Effects of storage temperature and preservative treatment on shelf life of the pond-raised freshwater fish, silver perch (Bidyanus bidyanus). J. Food Prot. 64, 1584-1591. Gennari, M., Campanini, R., 1991. Isolamento e caratterizzazione di Shewanella putrefaciens da pesce fresco e alterato, carni fresche e alterate, prodotti lattiero-caseari, acqua e solo. Ind. Aliment. 30, 965-976, 988. Gennari, M., Tomaselli, S., Cotrona, V., 1999. The microflora of fresh and spoiled sardines (Sardina pilchardus) caught in Adriatic (Mediterranean) sea and stored in ice. Food Microbiol. 16, 15-28. Gill, T.A., 1992. Chemical and biochemical indices in seafood quality. In: Huss, H.H., Jacobsen, M., Liston, J. (Eds.), Quality Assurance in the Fish Industry. Elsevier Science Publishers, Amsterdam, pp. 377-387. Gill, T.A., 1997. Advanced analytical tools in seafood science. In: Luten, J.B., Borensen, T., Oehlenschlager, J. (Eds.), Quality Assuranse in the Fish Industry. Elsevier Science Publishers, Amsterdam, pp. 479-490. Gillespie, N.C., 1981. A numerical taxonomic study of Pseudomonas-like bacteria isolated from fish in Southeastern Queensland and their association with spoilage. J. Appl. Bacteriol. 50, 29-44. Gimenez, B., Roncales, P., Beltran, J.A., 2002. Modified atmosphere packaging of filleted rainbow trout. J. Sci. Food Agric. 84, 1154-1159. Gonzalez, C.-J., 1999. Bacterial microflora of wild brown trout (Salmo trutta), wild pike (Esox luciu), and aquacultured rainbow trout (Onchorynchus mykiss). J. Food Prot. 62, 1270-1277. Gonzalez-Rodriguez, M.N., Sanz, J.J., Santos, J.A., Otero, A., GarciaLopez, M.L., 2001. Bacteriological quality of aquacultured freshwater fish portions in prepackaged trays stored at 3C. J. Food Prot. 64, 13991404. Gram, L., Huss, H.H., 1996. Microbiological spoilage offish and fish products. Int. J. Food Microbiol. 33, 121-137. Gram, L., Melchiorsen, J., 1996. Interaction between fish spoilage bacteria Pseudomonas spp. and S. Putrefaciens in fish extracts and on fish tissue. J. Appl. Bacteriol. 80, 589-595. Hebard, C.E., Flick, G.J., Martin, R.E., 1982. Occurrence and significance

S. Chytiri et al. / Food Microbiology 21 (2004) 157-165

Hubbs, J., 1991. Fish: microbiological spoilage and safety. Food Sci. Technol. Today 5, 166-173. International Commission on Microbiological Specifications for Foods (ICMSF), 1986. Sampling plans for fish and shellfish, In: Microorganisms in Foods. Sampling for Microbiological Analysis: Principles and Scientific Applications, Vol. 2, 2nd Edition. University of Toronto Press, Toronto, Canada, pp. 181-196. International Commission on Microbiological Specifications for Foods (ICMSF), 1998. Microorganisms in Foods, 6: Microbial Ecology of Food Commodities. Blackie Academic & Professional, Baltimore. Koutsoumanis, K., Nychas, G.-J.E., 1999. Chemical and sensory changes associated with microbial flora of Mediterranean boque (Boops boops) stored aerobically at 0, 3, 7, and 10C. Appl. Environ. Microbiol. 65, 698-706. Koutsoumanis, K., Lambropoulou, K., Nychas, G.-J.E., 1999. Biogenic amines and sensory changes associated with the microbial flora of Mediterranean gilt-head sea bream (Sparus aurata) stored aerobically at 0, 8, and 15C. J. Food Prot. 62, 398-402. Kyrana, V.R., Lougovois, V.P., 2002. Sensory, chemical, microbiological assessment of farm-raised European sea bass (Dicentrarchus labrax) stored in melting ice. Int. J. Food Sci. Technol. 37, 319-328. Kyrana, V.R., Lougovois, V.P., Valsamis, D.S., 1997. Assessment of shelflife of maricultured gilthead sea bream (Sparus aurata) stored in ice. Int. J. Food Sci. Technol. 32, 339-347. Lima dos Santos, C.A.M., 1981. The storage of tropical fish in icea review. Trop. Sci. 23, 97-127. Liston, J., 1980. Microbiology in fishery science. In: Connell, J.J. (Ed.), Advances in Fish Science and Technology. Fishing News Book Limited, Surrey, Farnham, pp. 138-157. Liston, J., 1992. Bacterial spoilage of seafood. In: Huss, H.H., Jacobsen, M., Liston, J. (Eds.), Quality Assurance in the Fish Industry Developments in Food Science, Vol. 30. Elsevier Science Publishers, Amsterdam, pp. 93-106. Lyhs, U., Lahtinen, J., Fredriksson-Ahomaa, M., Hyytia-Trees, E., Elfing, K., Korkeala, H., 2001. Microbiological quality and shelf life of vacuum-packaged 'gravad' rainbow trout stored at 3 and 8C. Int. J. Food Microbiol. 70, 221-230. Malle, P., Poumeyrol, M., 1989. A new chemical criterion for the quality of fish: trimethylamine/total volatile basic nitrogen (%). J. Food Prot. 50, 419-423. Masniyom, P., Benjakul, S., Visessanguan, W., 2002. Shelf-life extension of refrigerated sea bass slices under modified atmosphere packaging. J. Sci. Food Agric. 82, 873-880. Mead, G.C., Adams, B.W., 1977. A selective medium for the rapid isolation of Pseudomonas associated with poultry meat spoilage. Br. Poultry Sci. 18, 661-670. Mossel, D.A.A., Eelderink, L., Koopmans, M., Rossem, F.V., 1979. Influence of carbon source, bile salts and incubation temperature on recovery of Enterobacteriaceae from foods using macconkey-type agars. J. Food Prot. 42, 470-475.

Nedohula, P.C., Westhoff, D., 1997. Microbiological analysis of striped bass grown in three aquaculture systems. Food Microbiol. 14, 255-264. Nishimoto, J., Suwetja, I.K., Miki, H., 1985. Estimation of keeping freshness period and practical storage life of mackerel muscle during storage at low temperatures. Mem. Fac. Fish. Kagoshima Univ. 34, 8996. Ordonez, J.A., Lopez-Galvez, D.E., Fernandez, M., Hierro, E., Hoz, L., 2000. Microbial and physicochemical modifications of hake (Merluccius merluccius) steaks stored under carbon dioxide enriched atmospheres. J. Sci. Food Agric. 80, 1831-1840. Pearson, D., 1976. The Chemical Analysis of Foods. Chemical Publishing Co Inc, New York. Randell, K., Hattula, T., Ahvenainen, R., 1997. Effect of packaging method on the quality of rainbow trout and Baltic herring fillets. [Link]. Technol. 30, 56-61. Reineccius, G., 1990. Off-flavours in foods. Crit. Rev. Food Sci. Nutr. 29, 381-402. Rodriguez, C.J., Besteiro, I., Pascual, C., 1999. Biochemical changes in freshwater rainbow trout (Oncorhynchus mykiss) during chilled storage. J. Sci. Food Agric. 79, 1473-1480. Ryder, J.M., Fletcher, G.C., Stec, M.G., Seelye, R.J., 1993. Sensory, microbiological and chemical changes in hoki stored in ice. Int. J. Food Sci. Technol. 28, 169-180. Savvaidis, I.N., Skandamis, P.N., Riganakos, K.A., Panagiotakis, N., Kontominas, M.G., 2002. Control of natural microbial flora and Listeria monocytogenes in vacuum-packaged trout at 4 and 10C using irradiation. J. Food Prot. 65, 515-522. Shewan, J.M., 1977. The bacteriology of fresh spoiling fish and biochemical changes induced by bacterial action. In: Proceedings of the Conference on the Handling, Processing and Marketing of Tropical Fish, London, Tropical Products, pp. 511-566. Shewan, J.M., Murray, C.K., 1979. The microbial spoilage offish with special reference to the role of the psychrophiles. In: Russel, A.D., Fuller, R. (Eds.), Cold Tolerant Microbes in Spoilage and the Environment. Academic Press, London, pp. 117-136. Simeonidou, S., Govaris, A., Vareltzis, K., 1998. Quality assessement of seven Mediterranean fish species during storage on ice. Food Res. Int. 30, 479-484. SPSS, 1995. SPSS for Windows: 7.5.2.S. SPSS Inc, Chicago, IL. Steel, R.G.D., Torrie, J.H., 1980. Principles and Procedures of Statistics: A Biometrical Approach, 2nd edition. McGraw-Hill, New York. Urch, M., 1994. Industry grows up in Greece. Seafood International 9, 1921, 23. Tejada, M., Huidobro, A., 2002. Quality of farmed gilthead seabream (Sparus aurata) during ice storage related to the slaughter method and gutting. Eur. Food Res. Technol. 215, 1-7. Teskeredzic, Z., Pfeifer, K., 1987. Determining the degree of freshness of rainbow trout (Salmo gairdneri) cultured in brackish water. J. Food Sci. 52, 1101-1102.