**4. Discussion**

Firstly, the obtained results of total phenolic content agrees with previous findings by other authors using the Folin-Ciocalteau method or by HPLC [30–32]. Results obtained in the present spectrophotometric determination were not strictly correlated with the data obtained by HPLC analysis, which is due to the different response factors of each of the polyphenol structures present in the extracts (punicalagins, rosmarinic acid, carnosic acid, carnosol and hydroxytyrosol) regarding the pattern used, as the gallic acid in this case. It is difficult to make a structural interpretation of the results obtained for the antioxidant capacity measurements using the studied methods, although, clearly, some factors are related with the molecular structure of the active substances: the presence of phenols, their conjugation and polymerisation, cathecol and/or gallate groups presence, etc. In both methods, P shows the best results, probably, due to the presence of some conjugated polyphenol structures and a significant amount of gallic acid groups (tri-hydroxy phenol structures). Regarding the olive extracts, HYT-L, with its lower level of hydroxytyrosol than HYT-F, as principal active compound, showed a higher antioxidant capacity in both models, making it one of the most powerful extracts. This fact could originate from the presence of flavonoid compounds in combination with the hydroxytyrosol, providing a synergistic effect in terms of antioxidant activity. RA is the most structurally similar extract to olive extracts, due to the presence of rosmarinic acid as an active compound. This substance could be termed "double-hydroxytyrosol," only for their structural similarity, perhaps for this reason both extracts showed proximate chelating activity values. The difference among rosemary extracts was significant. The water-soluble extract (RA) was more active in the ABTS method, while liposoluble extracts (NOS, and NOVS) showed a higher activity in the DPPH model. This behaviour could be explained by the different chemical structure of the radical used in each technique and by the different properties of the molecular structures of phenylpropanoids (rosmarinic acid) and diterpens (carnosic acid and carnosol). However, both structures have a cathecol group and a carborxylic acid group. These results can be compared with previous research. For example, Hmid et al. [33] and Elfalleh et al. [34] obtained similar values for pomegranate extracts using the same methods, as well as hydroxytyrosol [35] and rosemary extracts [36,37].

Hydrophilic ORAC is one of the most widely used methods for evaluating antioxidant capacity, but it is clear that the results may be conditioned not only by the antioxidant capacity of each compound, but also by the physical and chemical properties, particularly its water solubility. Pomegranate and olive extracts obtained similar values for their antioxidant activity unrelated to their origin (leaves, fruit or vegetation water). In this case, the different of cathecol and gallate groups did not seem to be significant. Despite this, HYT-L again showed a higher activity. The antioxidant capacity of RA was lower than that of the above (–15%), although it followed the same order. It seems obvious that the structural similarity goes on establishing a parallellism in the antioxidant activity, also in this model. If not, the lower ORAC activity of the fat-soluble rosemary extracts (diterpens) compared with the hydrosoluble extracts that were already described. Previous researchers, such as Azaizeh et al. [38], obtained similar results analysing hydroxytyrosol in olive (*Olea europaea*) vegetation waters, while Sueishi et al. [39] obtained results that were 50% higher when measuring the seasonal variations of oxygen radical scavenging ability in rosemary leaf extract using the same method. In research carried out by Durante et al. [40], the authors measured the antioxidant activity of diferent extracts from tomato, grape and pomegranate seeds, obtaining similar results as the last. In the same way, previous research obtained similar results to that obtained results by the FRAP method regarding to rosemary [36], pomegranate [33] and hydroxytyrosol [35,41].

Regarding antimicrobial activity, the terpenoid structure did not provide good results, although this does not mean that this compound has a lower antioxidant capacity. While not significant, it is interesting to point out that NOVS, which contains lecitin as an emulsionant, shows higher antioxidant activity than NOS, which does not contain this excipient. This method has been used in much research to test the antimicrobial capacity of many drugs and natural extracts. For example, Laincer et al. [39] measured the antimicrobial activity of several olive phenolics, including HYT, against *E. Coli* and *S. Aureus,* obtaining similar results. Weckesser et al. [42] analysed the antimicrobial activity of plant extracts, such as *Rosmarinus officinalis* L. against bacteria of dermatological relevance, among them *E. Coli* and *S. Aureus,* obtaining similar results using the diffusion disk method. Regarding the P extract, Kharchoufi et al. [7] obtained similar results for pomegranate peel extracts against *Pseudomonas putida,* *Penicillium digitatum* and *Saccharomyces cerevisae,* but not against any strains used in the present study. Finally, applying rosemary extracts, Santomauro et al. [43] obtained similar results (more than 10 mm of inhibition) in different strains.

Regarding the oxidative and antimicrobial damage of fish products under refrigerated storage for 11 days, all the natural extracts showed an antioxidative effect against formation of volatile compounds related to lipid oxidation.

Table 4 shows that all the volatile compounds analysed are the main components that contribute the most to the emergence of unpleasant notes of flavour, due to the low flavour threshold [44]. In general, the presence of natural extract (especially NOVS) in the fish patties delayed the formation of all the volatile lipid-derived compounds. In the same line, Nieto et al. [45,46] reported lower hexanal values, rancid odour and rancid flavour scores in lamb meat from ewes fed thyme leaves or rosemary by-products, respectively.

As it can be appreciated in Table 5, natural extracts also acted as antimicrobial agents against TVC and TCC proliferation. This behaviour has been previously observed by other researchers using other natural extracts. For example, Del Nobile et al. [47] studied the combined effect of different gas mix compositions (MAP 30:40:30 O2:CO2:N2, 50:50 O2:CO2, and 5:95 O2:CO2) and three essential oils (thymol, lemon and grapefruit seed extracts) on fresh blue fish burgers. Results obtained showed as the combination of 110 ppm of thymol, 100 ppm of grapefruit seed extract, or 120 ppm of lemon extract with MAP 5:95 O2:CO2 was able to maintain the microbial quality of fish burgers for 28 days under refrigerated storage. In the same way, the combined effect of antimicrobial mixtures of chitosan, nisin and sodium lactate with MAP 55:45 CO2:N2 was able to guarantee the microbial acceptance of hake burgers for 30 days of refrigerated storage [48]. On the other hand, Smaldone et al. [49] have observed that only MAP 5:60:35 O2:CO2:N2 application can extend the microbiological shelf-life of hake burgers for 15 days after elaboration. However, in the present study, modified atmosphere package treatment was not assessed, neither in previous research on fish products using natural extracts from pomegranate, rosemary or olive tree (*Olea europaea*). Likewise, with obtained results it can be concluded that bioactive compounds from studied extracts (P, RA, NOS, NOVS, HYT-L and HYT-F) act as antimicrobial agents, which has also demonstrated in vitro and it is due to their high concentration of phenolic compounds (punicalagin, carnosic acid, carnosol, rosmarinic acid and hydroxytyrosol) with known antimicrobial activity, as it has been exposed in the introduction of this work. For this reason, it is not surprising that their application avoided *L. monocytogenes* or *E. Coli* growth. Nevertheless, is important to know that samples that incorporated rosemary extracts presented TVC growth higher than the control sample at day 11, similar to hydroxytyrosol extracts that showed higher TCC growth than the control. This fact can be explained by the grea<sup>t</sup> amount of antioxidant compounds in combination with spices and spice extracts that the commercial mix contained, and which can produce a synergism between them, increasing the shelf-life of fish products.
