3.1. Chemical Profile of SEO
Initially, GC-MS was performed for chemical characterization of SEO and terpenoids content with their respective retention times given in
Table 1. Results suggested that
α-thujone (35.50%) was the most dominant terpenoid in the SEO sample. It was in agreement with the literature [
38]. Several other compounds detected in particularly high content were camphor (20.73%), eucalyptol (12.13%),
β-thujone (6.57%), borneol (4.26%) and bornyl acetate (2.42%), suggesting that oxygenated monoterpenes were the major terpenoid subgroup present in SEO. Similar chemical profile of these compounds in SEO, as well as substantially high content of oxygenated monoterpenes, was observed by Radivojac et al. [
39].
On the other hand, monoterpene hydrocarbons such as
β-pinene (1.66%),
p-cymene (1.79%) and d,l-limonene (1.19%), together with sesquiterpenes such as
α-humulene (1.69%), caryophyllene oxide (1.18%), viridiflorol (1.61%) and humulene oxide (1.63%), were present in lower content, while all other compounds were detected in content lower than 1%. Even though climate effects, plant properties and geographical origin could affect chemical profile of SEO, hydrodistillation is still the recognized method of choice to produce pure volatile oil with high content of oxygenated terpenoids with strong antioxidative potential [
18]. Even though polyphenols and diterpene polyphenols were often designated as the most powerful antioxidants from sage, SEO rich in oxygenated monoterpenes has generally strong antioxidant activity. This action could be due to its predominant compounds (α-thujone, camphor, eucalyptol and β-thujone) separately [
40,
41]. However, antioxidant capacity of EOs does not necessarily depend on the bioactive potential of their main components. Several studies suggested that bioactivity could be modulated, often towards synergism and additive effect, by the presence of minor EO compounds [
40]. On the other hand, oxygenated monoterpenes from SEO were often identified as the more potent antimicrobials compared to the sage polyphenols. According to Delamare et al. [
42], camphor, thujone and eucalyptol expressed strong antimicrobial activity towards
Aeromonas hydrophila,
Aeromonas sobria,
B. megatherium,
B. subtilis,
B. cereus and
Klebsiella oxytoca strains.
3.3. pH, Instrumental Parameters of Color and TBARS Values of DFS
pH values of DFS are shown in
Table 3. The fat content had a significant (
p < 0.05) effect on the pH values. It was in accordance with the study of Mora-Gallego et al. [
44] and Fonseca et al. [
45], who determined that increase of fat content decreases pH in fermented sausages. This phenomenon could be the result of the more intensive lipolysis in fermented sausages produced with higher fat contents [
44]. The sodium nitrite addition (0, 75 and 150 mg/kg) did not significantly (
p > 0.05) affect the pH values. Regarding SEO, the pH values of the samples produced with SEO addition were significantly (
p < 0.05) higher compared to their counterpart. It could be the result of the chemical profile of SEO. Similarly, the results of our previous study [
14] suggested that SEO increases pH in fresh pork sausages. Moreover, Zhang et al. [
46] observed that bioactive compounds (e.g., polyphenolics) contribute to the growth of ammonia-producing bacteria and consequently increase the pH values of DFS. Moreover, pH values inconsistently increased (
p < 0.05) during 225 days of storage. The increase of pH values during storage is in relation to the formation of amino acids, peptides and amines, and simultaneous reduction of lactic acid compounds during process of proteolysis in DFS [
47,
48]. The two-way (F × SEO, SN × SEO, S × SEO), three-way (F × SN × S, F × S × SEO) and four-way (F × SN × S × SEO) interactions significantly (
p < 0.05–0.001) affected the pH values (
Table 4). pH fluctuated in intervals from 5.18 (F = 25%, SN = 75 mg/kg, S = 0 days, SEO = 0.10 µL/g) to 5.66 (F = 15%, SN = 0 mg/kg, S = 75 days, SEO = 0.01 µL/g). The relatively high pH values could be the consequence of relatively lower lactic acid bacteria (LAB) counts (<7 log CFU/g) (
Table 5). Similar findings were reported by Nikolić et al. [
2] and Ozaki et al. [
7].
The instrumental parameters of color are depicted in
Table 3. Fat content had a significant (
p < 0.05) effect on the color of DFS. In the case of lightness and redness, the samples manufactured with lower fat content were darker and redder. It was in accordance with the findings of Lorenzo and Franco [
49] and Fonseca et al. [
45] for similar meat products. Also, Mora-Gallego et al. [
44] reported that back fat particles in the DFS had a noticeable white color which conferred them a lighter appearance. The yellowness decreased while increasing the fat content, which agrees with the results of Rubio et al. [
50] for similar meat products. The sodium nitrite addition significantly (
p < 0.05) affected the yellowness of DFS. It was in accordance with the finding of Van Ba et al. [
9] for fermented sausages. The inclusion of SEO significantly (
p < 0.05) decreased the lightness and increased the yellowness. These phenomena could be the consequences of myoglobin interaction with the bioactive compounds of SEO, including terpenoids, phenolics, etc. [
20]. Also, Zhang et al. [
51] observed that the sage phenolic compounds could be oxidized into equivalent quinines, which resulted in formation of darker color of meat products. It was in strong agreement with results of our previous study [
4,
22]. The storage time significantly (
p < 0.05) decreased in lightness and increased in redness and yellowness. Similar behavior of vacuum-packed DFS was reported by Pateiro et al. [
52]. The increase of redness could be connected with the ability of
Staphylococcus species (e.g.,
S. carnosus or
S. xylosus) to produce the enzymes of metmyoglobin reductase and nitrate reductase, which possess a significant potential to modify brown metmyoglobin to a red myoglobin in cured meat products [
53]. Moreover, the increase of yellowness could be related with the rancidity during long storage periods [
54]. Also, the F × SN × S × SEO interaction had a significant (
p < 0.01) effect on the lightness (
Table 4). The lightness ranged in intervals from 44.2 (F = 15%, SN = 0 mg/kg, S = 225 days, SEO = 0.01 µL/g) to 56.6 (F = 25%, SN = 0 mg/kg, S = 0 days, SEO = 0 µL/g). The obtained results were in accordance with literature data for similar meat products [
2,
7,
9]. The two-way (N × S) and four-way (F × SN × S × SEO) interactions had a significant (
p < 0.01–0.001) effect on the redness (
Table 4). The lowest level of redness was determined in the sample: F = 25%, SN = 0 mg/kg, S = 0 days, SEO = 0 µL/g (9.38), while the highest level of redness were determined in the sample: F = 25%, SN = 0 mg/kg, S = 225 days, SEO = 0.05 µL/g (16.1). The obtained results suggested that SEO enhanced the redness of DFS produced without sodium nitrite. Accordingly, in our previous studies [
15,
16], we determined that essential oils obtained from coriander and organic peppermint improved the redness of cooked sausages manufactured with reduced levels (0 and 50 mg/kg) of sodium nitrite. The following interactions: F × SN, F × S, SN × S, SN × SEO, SN × S × SEO and F × SN × S × SEO, had a significant (
p < 0.05–0.001) effect on the yellowness. The yellowness ranged in intervals from 5.89 (F = 25%, SN = 0 mg/kg, S = 0 days, SEO = 0.10 µL/g) to 10.02 (F = 15%, SN = 75 mg/kg, S = 225 days, SEO = 0.05 µL/g). The obtained results were in strong agreement with the literature data [
2,
9] for DFS.
TBARS values of DFS are presented in
Table 3. Fat content did not significantly (
p > 0.05) affect the TBARS values. The addition of sodium nitrite had a significant (
p < 0.05) effect to decrease TBARS values. This was in accordance with strong antioxidative potential of this additive [
6]. Also, SEO significantly (
p < 0.05) decreased the TBARS values. It was in accordance with our previous study [
14,
22], where we found that SEO, in similar concentrations, had a strong antioxidative potential in processing of fresh [
14] and cooked [
22] sausages. Moreover, no significant differences (
p > 0.05) were found among SEO-treated samples, hence the lowest concentration of SEO (0.01 µL/g) would be appropriate to enhance oxidative stability of DFS. Due to complex chemical profile with numerous bioactives diverse in volatility, polarity and bioactivity, several extraction procedures could be applied in order to provide selectivity towards target compounds. According to Radivojac et al. [
39], the hydrodistillation procedure (conventional or microwave-assisted) could be an excellent tool to produce volatile SEO with high selectivity towards mono- and sesqui-terpenes. A recent study suggested that sage extracts rich in polyphenols and SEO with high content of terpenoids exhibit strong antioxidant activity in in vitro model systems and cell culture, as well as in in vivo studies on animals [
55]. It has been suggested that SEO components seem to show a main role expression of antioxidant potential. Poulios et al. [
55] suggested that
α-thujone,
β-thujone, camphor, linalool and eucalyptol are oxygenated monoterpenes mostly responsible for this action. Furthermore, presence of these compounds in mixture could provide synergistic effects, contributing to improved antioxidant action compared to effects caused by each compound separately [
56]. As expected, TBARS values significantly (
p < 0.05) increased throughout storage, probably as the consequence of lipid oxidation. The following interactions: F × S, SN × S, S × SEO, F × SN × S, SN × S × SEO and F × SN × S × SEO, significantly (
p < 0.05–0.001) affected the TBARS values (
Table 4). TBARS values ranged in intervals from 0.01 mg MDA/kg (F = 15%, SN = 0 mg/kg, S = 0 days, SEO = 0.05 µL/g) to 0.40 mg MDA/kg (F = 25%, SN = 0 mg/kg, S = 225 days, SEO = 0.05 µL/g). Melton [
57] marked the TBARS values under 0.3 mg MDA/kg as a threshold for meat rancidity. In order to suppress the lipid oxidation, the synthetic (sodium nitrite) and natural additives (SEO) were used. The following combinations of sodium nitrite and SEO added in the DFS batters formulated with two levels of back fat enable TBARS values < 0.3 mg MDA/kg: F = 15–25%, SN = 150 mg/kg, S = 0–225 days, SEO = 0–0.10 µL/g; F = 15–25%, SN = 75 mg/kg, S = 0–225 days, SEO ≥ 0.05 µL/g and F = 15%, SN = 0 mg/kg, S = 0–225 days, SEO ≥ 0.05 µL/g. The obtained results suggested that the SEO (≥0.05 µL/g) in interaction with reduced level of sodium nitrite (75 mg/kg) decreased the lipid oxidation in DFS produced with both levels of back fat, 15% and 25%. Moreover, it should be highlighted that total replacement of sodium nitrite (0 mg/kg) by SEO (≥0.05 µL/g) provided satisfying oxidative stability (TBARS < 0.3 mg MDA/kg) in DFS produced with 15% of fat. SEO could express its antioxidant action through different mechanisms; however, prevention of lipid peroxidation is highlighted as the most important aspect for exhibiting protective effects in meat products.
3.4. Microbiological Profile of DFS
The microbiological profile of DFS is presented in
Table 5.
The fat content and sodium nitrite addition did not have a significant (
p > 0.05) effect on total plate counts (TPC) and lactic acid bacteria (LAB) in DFS. In the case of natural antioxidants, the SEO addition had a tendency to decrease TPC and LAB growth, but the values among the samples were not significantly (
p > 0.05) different. The obtained results suggested that further optimization is essential. Also, our previous study [
14] suggested that the usage of novel extraction techniques (e.g., supercritical fluid extraction—SFE) could be a good tool to improve the antimicrobial potential of SEO. The higher antimicrobial effect of extract obtained by SFE could be related to a synergistic effect of oxygenated terpenes and diterpene polyphenols which could be co-extracted by SFE [
14]. Regarding storage time, a significant (
p < 0.05) effect on the growth of TPC and LAB was observed. Population of LAB decreased during the storage, and this was in accordance with the findings reported by Tomović et al. [
4] and Sucu and Turp [
58] for similar meat products. TPC decreased during the first 75 days of storage (4.56 log CFU/g), then increased (
p < 0.05) until 150 days of storage (5.97 log CFU/g), remaining constant until the end of storage (5.89 log CFU/g). It can be highlighted that throughout the storage, average TPC was lower than 6 log CFU/g. Ozaki et al. [
7] suggested that TPC under 6 log CFU/g could be considered as a threshold for quality and shelf-life of meat products. Also, all two-way, three-way and four-way interactions had no significant (
p > 0.05) effect on the TPC and LAB counts. Moreover, the following foodborne pathogenic bacteria:
E. coli,
L. monocytogenes,
Salmonella spp. and sulfite-reducing clostridia, were not detected in any DFS sample throughout the storage. Therefore, all DFS samples were safe for consumption according to requirements of European Commission Regulation: No 2073/2005 [
59].