Extending shelf life of desalted cod by high pressure processing

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.


Introduction
Salt-curing and drying of fish have been used as a traditional way of preserving fish since ancient times. Fully salted cod (saltfish) has a high salt content (~20%) and is low in water (~50%). The water content can be further reduced by drying (< 50%), and then dried saltcured cod (clipfish) is obtained (Bjorkevoll, Olsen, & Skjerdal, 2003). Saltfish and clipfish can be stored for a long time even at abuse temperatures. Salted cod is split or fileted and then pickle salted or brine cured (Andres, Rodriguez-Barona, & Barat, 2005). To obtain clipfish, the salted fish is dried. Due to the very high salt content, a desalting or rehydration process, where the fish is soaked in water, must be performed before human consumption. Following the rehydration process substantial changes of the muscle are observed, and the product get the water content increased to 70-85% (w/w) and reduced its salt content to 2-3 % (w/w) (Lorentzen, Ytterstad, Olsen, & Skjerdal, 2010;Thorarinsdottir, Arason, Geirsdottir, Bogason, & Kristbergsson, 2002). Salt curing implies prevention of bacterial growth, and after a desalting process these products are favorable conditions for bacterial growth. Several studies (Barat et al., 2006;Bjorkevoll et al., 2003;Magnusson, Sveinsdottir, Lauzon, Thorkelsdottir, & Martinsdottir, 2006) have reported high bacterial counts (> 6 log 10 cfu/g) in desalted products after 6 days of chilled storage (1-4°C), and with sensory rejection after 7-J o u r n a l P r e -p r o o f 8 concentration in the water (%) was measured in each desalting unit, after stirring, and before any water change using a manual salt analyzer, an analog refractometer (Kern Optics, Balingen, Germany). Duplicate samples of water were analysed. The drip loss during storage (%), the weight of exudates in the trays after storage in relation to the original weight of the portion, was calculated from triplicate samples. pH was measured in room tempered, homogenised samples using a pH-meter instrument PHM210 (Meterlab, Copenhagen, Denmark).

Colour
The surface colour (L*, a* and b* values, CIELAB) was assessed by a calibrated digital photo imaging colour-measuring system (DigiEye full system, VeriVide Ltd., Leicester, UK). Samples were placed in a standardized light-box with daylight (6400 K) and photographed with a calibrated digital camera (Nikon D80, 35 mm lens, Nikon Corp., Japan). Pictures were analysed with DigiPix software (VeriVide Ltd., Leicester, UK) and the various color parameters were quantified. L* describes brightness (L* = 100 = white, L* = 0 = black), a* describes the intensity of the color in red-green axis (a* <0 = green, a*> 0 = red), while b* describes the intensity of the yellow-blue axis (b* <0 = blue, b*> 0 = yellow).

Texture
The texture analysis was performed with a TA.XTplus equipped with a 50 kg weigh cell and the software Exponent ver: 6.1.16.0 (Stable Micro Systems, Surrey, UK). A flat cylinder probe J o u r n a l P r e -p r o o f 9 temperature for 1 h prior to analysis. Mean values of the two assessments were used for statistical analysis of the data.

Microbiological Analyses
Samples of muscle (25 g) were diluted 1 : 10 in peptone water (BactoPeptone, Merck, Darmstadt, Germany) added 0.85 % NaCl, and homogenised for 2 min in a Stomacher 400 Laboratory Blender (Seward Medical, London, UK). Aerobic quantification of psychrotrophic and heat labile microorganisms (colony forming units [cfu] ml -1 ) was performed by surface plating on Long & Hammer agar (L&H) and Tryptic Soy Agar added 2 % NaCl (TSA-NaCl) plates. The plates were incubated for 5-7 days at 15 C. A mechanical spiral plater (EddyJet, IUL Instruments, Barcelona, Spain) was mainly used for this purpose. However, some manual plating was performed for low dilutions. Aerobic count was also performed on Iron Agar to determine hydrogen sulphide producing bacteria. A pour plate technique was used, adding Iron Agar with 0.8 % L-cysteine to 1 ml of sample. Plates were incubated for 3-4 days at 20 C. Both the total number (aerobic count) and the number of black colonies were counted (hydrogen sulphide producing bacteria). The detection level on L&H/TSA and Iron Agar was 10 2 and 10 cfu g -1 , respectively. For plates with no colonies detected, the level was set to half of the detection limit.

Statistical analyses
Statistical analysis included analysis of variance (one way ANOVA), general linear modelling (GLM) and Tukey's HSD test (p<0.05). All data processing was carried out on Minitab v19 The following analysis has been conducted with a minimum of three technical replicates for each production if nothing else is stated.

Weight change after rehydration and salt content
The rehydration time was different for the clipfish and saltfish. This must be seen in context with the size and the thickness of the fish, and the differences in processing (salting, drying).
For the dried and salted clipfish, the average weight gain for 16 random fish pieces (2 experiments), followed during the rehydration process (48 h  pH in the clipfish varied from 6.12-6.30 at day 0,

Colour and texture after processing and during storage of clipfish
The colour and texture analysis were only performed for the clipfish samples, since they were relatively even pieces of the loin with no bones. At day 0 there was no significant difference between the non-pressurised control and the HPP samples. In a previous study on fresh cod, statistically significant differences (p<0.05) in whiteness was seen for the nonpressurised samples, the lightness (L*) value was 75.9, compared with the samples treated at 500 MPa, where the whiteness was increased to 91.6 (Christensen et al., 2017). In this study the non-pressurised desalted cod showed L* values of 74.9, while HPP at 500 MPa increased the whiteness to 78.1 (not shown). This indicates small differences in whiteness of untreated (control) fresh cod compared to desalted cod. When comparing these studies, exposure to 500 MPa showed that differences is whiteness was very prominent for the raw cod while non-significant changes was detected for the desalted cod compared with the non-pressurised samples. A suggestion can be that some of the pressure sensitive proteins have already been denatured due to the salting and drying process of the cod, and therefore a non-significant change in lightness was detected when pressurizing at 500 MPa compared with the control (0.1 MPa). The colour of the pressurised samples was not significantly For all compression rates, HPP made the samples significantly (p<0.047) harder than the untreated samples (data not shown). When comparing the data for the HPP samples, both storage time and pressure came out as significant main effects. All texture data are shown in Figure 5. Significant changes (p<0.018) at 20 % shows that the surface got a softer texture during the storage period of 49 days. However, only 600 MPa were analysed at day 49. Storage of fish over time generally gives softer texture, either due to enzyme activity or bacterial growth, or both. Due to the low bacterial growth at day 49, the softer tissue that was observed was most likely due to increased enzymatic activity. The latter is confirmed by others reporting that the texture of cod-like fish species is largely influences by enzymes, and can also be changes by HPP (Matser, Stegeman, Kals, & Bartels, 2000). No significant difference was found between day 0 and 14 on any of the samples. The texture analysis showed that the samples pressurised at 500 MPa gave the hardest texture among the other pressure levels tested.

Microbiological quality
There is often a relation between increased rehydration time and increased number of bacteria in the rehydrated product. The clipfish samples showed low bacterial counts after the rehydration process. And the processing at 500 and 600 MPa gave bacterial counts around 2 log cfu/g, Figure 3. bacterial levels throughout the whole storage period of 49 days was observed. All these clipfish samples showed growth below 4 log cfu/g, with an average of 2.5 ± 1.5 log cfu/g. We have from previous cod studies also seen that HPP can be used to extend the shelf life of fresh cod processed at 500 MPa.
In the saltfish experiment, where different packaging regimes were used, the bacterial load after the rehydration process was 4.0 ± 1.3 log cfu/g. The relatively high standard deviation was due to significant differences (p<0.001) in the two rehydration processes, with an average of 2.7 and 5.3 log cfu/g. This shows that it is possible to obtain very low bacterial numbers in rehydrated saltfish and clipfish (2-3 log cfu/g), and this will be a very good starting point and will greatly affect the shelf life of a rehydrated product.
Despite the relatively high average bacterial start levels, the HPP combined with different packaging lowered the bacterial numbers to below 2 log cfu/g for some of the treatments.
We did not observe any significant difference in packaging among the non-processed rehydrated saltfish samples. After 15 days the average bacterial numbers for saltfish packed with CO 2 -emitter and MAP had reached almost 6 log cfu/g, but with relatively high standard deviation indicating that some of the parallel samples were not yet spoiled, Figure 4.
For the HPP-samples, there was no significant difference between the two rehydration batch processes during storage. This indicates that there was a high inactivation J o u r n a l P r e -p r o o f

Conclusion
Rehydrated clipfish and saltfish can increase shelf life from days to several weeks, depending on the packing and processing conditions. High-pressure processing of rehydrated clipfish and saltfish has proven to provide very long shelf life, at least 49 days can be achieved if processed at 600 MPa for 5 minutes. The drip loss in the HPP samples were somewhat high during storage, but this can possibly be solved with the use of an absorbent in the packages.
J o u r n a l P r e -p r o o f