Essential Oils as a Dietary Additive for Small Ruminants: A Meta-Analysis on Performance, Rumen Parameters, Serum Metabolites, and Product Quality

Simple Summary Essential oils can be used to improve animal performance as well as the health and quality of livestock products. The aim of this study was to evaluate the effects of essential oil supplementation on animal performance, ruminal fermentation, blood metabolites, and meat and milk quality of small ruminants through a meta-analysis. Supplementation with essential oils improved weight gain, milk production and composition, oxidative stability of meat, and blood serum antioxidant enzyme activity. Additionally, essential oils decreased methane emissions. This suggests that the inclusion of essential oils in the diets of small ruminants could be used to improve animal performance and the quality of meat and milk, in addition to reducing the environmental impact and oxidative stress of the animals. Abstract There is an increasing pressure to identify natural feed additives that improve the productivity and health of livestock, without affecting the quality of derived products. The objective of this study was to evaluate the effects of dietary supplementation with essential oils (EOs) on productive performance, rumen parameters, serum metabolites, and quality of products (meat and milk) derived from small ruminants by means of a meta-analysis. Seventy-four peer-reviewed publications were included in the data set. Weighted mean differences (WMD) between the EOs treatments and the control treatment were used to assess the magnitude of effect. Dietary inclusion of EOs increased (p < 0.05) dry matter intake (WMD = 0.021 kg/d), dry matter digestibility (WMD = 14.11 g/kg of DM), daily weight gain (WMD = 0.008 kg/d), and feed conversion ratio (WMD = −0.111). The inclusion of EOs in small ruminants’ diets decreased (p < 0.05) ruminal ammonia nitrogen concentration (WMD = −0.310 mg/dL), total protozoa (WMD = −1.426 × 105/mL), methanogens (WMD = −0.60 × 107/mL), and enteric methane emissions (WMD = −3.93 L/d) and increased ruminal propionate concentration (WMD = 0.726 mol/100 mol, p < 0.001). The serum urea concentration was lower (WMD = −0.688 mg/dL; p = 0.009), but serum catalase (WMD = 0.204 ng/mL), superoxide dismutase (WMD = 0.037 ng/mL), and total antioxidant capacity (WMD = 0.749 U/mL) were higher (p < 0.05) in response to EOs supplementation. In meat, EOs supplementation decreased (p < 0.05) the cooking loss (WMD = −0.617 g/100 g), malondialdehyde content (WMD = −0.029 mg/kg of meat), yellowness (WMD = −0.316), and total viable bacterial count (WMD = −0.780 CFU/g of meat). There was higher (p < 0.05) milk production (WMD = 0.113 kg/d), feed efficiency (WMD = 0.039 kg/kg), protein (WMD = 0.059 g/100 g), and lactose content in the milk (WMD = 0.100 g/100 g), as well as lower somatic cell counts in milk (WMD = −0.910 × 103 cells/mL) in response to EOs supplementation. In conclusion, dietary supplementation with EOs improves productive performance as well as meat and milk quality of small ruminants. In addition, EOs improve antioxidant status in blood serum and rumen fermentation and decrease environmental impact.


Introduction
In ruminants, antibiotics have been used for several years to prevent and cure diseases, as well as to improve growth and the efficiency of conversion of ingested feed into products for human consumption, such as meat and milk [1]. However, due to the inappropriate use of antibiotics, the emergence of bacteria with resistance to their effects is currently among the main threats to global health [2]. In addition, the extensive use of antibiotics in ruminants can generate antibiotic residues in meat and milk, which when consumed by humans can affect their health [3]. Therefore, in recent years, the interest in the use of natural products to improve the health and productivity of livestock has increased [4]. Among these, EOs are plant-derived products that have gained greater economic relevance [5]. EOs extracted from plants are obtained by distillation and are composed of mixtures of low-weight molecules, such as terpenes (monoterpenes and sesquiterpenes), terpenoids, ketones, aldehydes, and alcohols [6].
It has been reported that EOs and their bioactive metabolites have diverse biological effects, such as antimicrobial, anti-inflammatory, antioxidant, and antiparasitic, among others [6,7]. In previous studies, the effects of EOs as dietary additives have been evaluated mainly in non-ruminants [8][9][10]. Therefore, information on the effects of dietary inclusion of EOs in ruminants is still limited. However, there is evidence that EOs can improve the efficiency of energy utilization and nutrient intake by ruminants [11]. Moreover, moderate doses of some EOs in a diet can improve volatile fatty acid production and protein metabolism in ruminants [12], while, at high doses, some EOs can decrease methane production [13]. In contrast, EOs have been reported to have no positive effects on productive performance and nutrient utilization efficiency in ruminants [1]. Furthermore, a large part of the positive effects of EOs on ruminal fermentation have been obtained from in vitro studies and using high doses [11], which, when applied to in vivo studies in ruminants, are likely to negatively affect feed intake [14] and ruminal fermentation [11].
Particularly in small ruminants, some studies have evaluated the effect of dietary inclusion of EOs on animal performance [15,16], nutrient digestibility, ruminal fermentation [17,18], blood biochemistry [19,20], and meat quality [21,22], as well as milk production and composition [23,24]. However, no conclusive results have been obtained until now, perhaps due to the variability that exists among these studies regarding the experimental periods, primary bioactive compounds, and the doses of EOs used [25]. Thus, identifying and controlling this variability could help to obtain EOs that can be used in small ruminants to improve animal performance, rumen fermentation, health, and product quality.
Several review articles have been published to date [1,5,[12][13][14] concluding that dietary supplementation with EOs can be used to improve animal productive performance, rumen parameters, animal health, and product quality (meat and milk) in ruminants. These positive effects of EOs have been confirmed in beef cattle and dairy cows using metaanalytical methods [25,26]. In a previous meta-analysis (MA), Khiaosa-ard and Zebeli [27] evaluated the effects of dietary inclusion of EOs on rumen fermentation in small ruminants. However, that study [27] only included six references applied to small ruminants in their database and did not evaluate productive performance, blood metabolites, or meat and milk quality. Although the meta-analytical approach has been used mainly in research related to human health, its application in research on natural food additives in domestic animals is still limited [28,29]. MA allows to combine and quantitatively synthesize previously published results from multiple independent studies [30]. In addition, with the use of MA, it is possible to identify sources of heterogeneity among studies performed on the same subject [31].
Considering the mentioned antecedents, the hypothesis of the present study proposes that the addition of EOs in diets for small ruminants will benefit productive performance,

Data Extraction
Based on the inclusion criteria, only 74 articles were included in the database for the final analysis (Table A1). Only response variables that were reported in at least three studies were included in the database [25,30,31]. Therefore, among the variables included in the present meta-analysis were the following: dry matter intake, dry matter and nutrient digestibility (protein and ethereal extract, among others), daily weight gain, feed conversion ratio, carcass characteristics (hot carcass weight and yield, among others), ruminal parameters (volatile fatty acids, protozoa and bacteria, and pH and ammonia nitrogen), serum metabolites (urea, glucose and cholesterol, among others), antioxidant enzymes in blood serum (superoxide dismutase and catalase, among others), characteristics related to meat quality (color, chemical composition, pH, and malondialdehyde content, among others), as well as milk production and composition (protein and fat and lactose content).
Finally, when available, from the 74 selected publications, the publication reference (author and year), the country where the study was conducted, the amount of forage and concentrate in the diet (g/kg DM), the nutritional composition of the diet (g/kg DM), the period of supplementation with EOs (days), the dose of EOs in the diet (mg/kg DM), and the primary bioactive metabolite of the Eos were obtained. From these publications, the number of replicates means and standard deviations (SD) for each of the control and Vet. Sci. 2022, 9,475 4 of 31 experimental treatments were extracted. In articles where the SD was not reported, it was calculated using the standard errors of the means (SEM), using the equation [33]: SD = SEM × √ n, where n = number of replicates.

Calculations and Statistical Analysis
To perform the meta-analysis, as well as for heterogeneity, publication bias, and metaregression analyses, the metaphor package [34] in the statistical software R (version 4.1.2) was used. The effects of EOs as an additive in small ruminant diets were evaluated by weighted mean differences (WMD) between treatments with EOs (diets with EOs) and control treatments (diets without EOs). Treatment means were weighted with the inverse of the variance, following methods previously proposed by DerSimonian and Laird [35] for random effects models. WMD was used because it allows interpretation of the results in the original units of measurement [28].
On the other hand, the MEANS procedure of the statistical software SAS [36] was used to obtain descriptive statistical values of the nutritional composition of the diets. In addition, the MIXED procedure of SAS was used to evaluate the differences in the nutritional composition of the diets of the EOs treatments and the control treatments. For this, random effect studies were used, as well as the Tukey test to detect differences between treatments, as previously reported by Orzuna-Orzuna et al [25,30,31].

Heterogeneity and Publication Bias
In the meta-analysis, the heterogeneity of treatment effect was assessed with the chisquare (Q) test, in which, due to its relatively low power, a significance level of p ≤ 0.10 was used [37]. In addition, the I 2 statistic was used to measure the percentage of variation due to heterogeneity [29]. Negative values of I 2 (percent variation) were assigned as zero, while values less than 25, 25 to 50, and greater than 50% indicated low, moderate, and high heterogeneity, respectively [28,29].
On the other hand, publication bias was assessed using Egger's regression asymmetry test [38], which was considered significant (publication bias) when p ≤ 0.05 was observed. In addition, when Egger's test was statistically significant (p ≤ 0.05), the "trim and fill" method of Duval and Tweedie [39] was used with the aim of estimating the possible number of missing observations.

Meta-Regression and Subgroup Analysis
A meta-regression analysis was performed to identify sources of heterogeneity in the variables evaluated. The meta-regression criteria were (1) p ≤ 0.10 for the Q test [37] or I 2 greater than 50% [28]; (2) p ≥ 0.05 for Egger's test [38]; and (3) response variables reported in at least 10 studies [40]. In all cases, meta-regression was performed using the method of moments of DerSimonian and Laird [35], which is well-established to estimate the variance between studies. When covariates were significant, with p ≤ 0.05, WMD was evaluated by subgroup analysis. The primary bioactive compound (carvacrol, eugenol, thymol, limonene, and linalool, among others) was used as a categorical covariate, whereas the duration of the experimental phase (days) and the dose of EOs (mg/kg DM) were used as continuous covariates. When covariates were significant, with p ≤ 0.10; these were evaluated by subgroup analysis [25,30,31]. Each of the primary bioactive metabolites was considered a single category. Moreover, when meta-regression was significant (p ≤ 0.05) for each Eos supplementation period (days) and dietary dose of Eos (mg/kg DM), these two covariates were evaluated by the subgroups: supplementation period (≤70 and >70 days) and dietary dose of Eos (≤500, 501-1000 and >1000 mg/kg DM). Table 1 shows that there were no differences (p < 0.05) between the control treatment and the different treatments with Eos for forage, concentrate, nutrients, and metabolizable Vet. Sci. 2022, 9, 475 5 of 31 energy content of the diets. This suggests that, for the data set analyzed, it is possible to exclude the effects of these components on the response of small ruminants to dietary inclusion of Eos. In the present meta-analysis, the studies included were performed in 17 different countries, mainly in Egypt (16.2%), China (13.5%), Spain (12.2%), Iran (9.5%), Turkey (8.1%), and Tunisia (8.1%). Regarding the animal species, sheep were used in 75.7% of the studies, and goats were used in the rest (24.3%). Table A1 shows that the experimental periods ranged from 14 to 288 days, while the experimental doses of Eos ranged from 10 to 40,000 mg/kg DM. The Eos were grouped based on the primary bioactive metabolite, and in total, 20 different types of primary bioactive metabolites were observed. EOs with mixtures of primary bioactive metabolites in similar proportions were the most commonly used in the treatments (44.8%). Moreover, a significant proportion of the treatments used EOs with carnosic acid (11.6%), carvacrol (6.1%), thymol (4.9%), and limonene (4.9%) as a primary bioactive metabolite, while, in the remaining treatments (27.7%), EOs with 15 other different primary bioactive metabolites were used (Table A1). Table 2 shows that dry matter intake increased (p < 0.001) in response to dietary supplementation of EOs. Similarly, dietary inclusion of EOs increased (p < 0.05) dry matter digestibility (DMD), organic matter digestibility (OMD), crude protein digestibility (CPD), neutral detergent fiber digestibility (NDFD), and acid detergent fiber digestibility (ADFD). However, ether extract (EE) digestibility was similar among treatments (p > 0.05). Table 3 shows that daily weight gain (DWG), hot carcass yield (HCY), and Longissimus dorsi muscle area (LMA) increased in response to dietary supplementation with EOs (p < 0.05). On the other hand, dietary inclusion of EOs decreased the feed conversion ratio (FCR; p = 0.045). However, hot carcass weight (HCW), cold carcass weight (CCW), and backfat thickness (BFT) were similar among treatments (p > 0.05).   Table 4 shows that ruminal pH was similar between treatments (p > 0.05). Moreover, dietary supplementation with EOs did not affect (p > 0.05) the rumen concentration of acetate, butyrate, and the count of Entodinium, Diplodinium, Isotrichae, total bacteria, Ruminococcus albus (R. albus) and Fibrobacter succinogenes (F. succinogenes). However, dietary inclusion of EOs reduced (p < 0.05) ruminal ammonia nitrogen (NH 3 -N) concentration, total protozoan populations, Epidinium, methanogens, and daily enteric methane (CH 4 ) emissions. On the other hand, ruminal propionate concentration and relative amount of Ruminococcus flavefaciens (R. flavefaciens) increased in response to EOs supplementation. Table 5 shows that dietary supplementation with EOs decreased (p < 0.05) the serum concentration of urea, cholesterol, triglycerides, non-esterified fatty acids (NEFA), and beta-hydroxybutyrate (BHB). On the other hand, dietary supplementation with EOs did not affect (p > 0.05) the serum concentration of glucose, albumin, globulin, total protein, malondialdehyde (MDA), and glutathione peroxidase (GPx). However, higher serum concentrations of thyroxine, catalase (CAT), superoxide dismutase (SOD), and total antioxidant capacity (TAC) were observed in response to the dietary inclusion of EOs (p < 0.05).

Milk Yield and Quality
Dietary supplementation with EOs increased (p < 0.05) milk yield, feed efficiency (FE), and protein and lactose content in milk (Table 7). On the other hand, dietary supplementation with EOs decreased somatic cell count (SCC) and milk urea content (p < 0.001). However, there was no significant impact (p > 0.05) of dietary inclusion of EOs on milk fat content and milk pH (Table 7).

Publication Bias and Meta-Regression
Tables 2-7 show that Egger's asymmetry regression test was non-significant (p > 0.05) for all variables evaluated, which indicates that there was no publication bias.
On the other hand, Tables 2-7 show significant (p ≤ 0.10) heterogeneity (Q) for DMD, OMD, CPD, EED, NDFD, ADFD, ADG, LMA, ruminal pH, NH 3 -N, acetate, propionate, butyrate, total protozoa, Entodinium protozoa, total bacteria, R. flavefaciens, R. albus, F. succinogenes, methanogens, glucose, cholesterol, triglycerides, NEFA, GPx, TAC, meat pH, MDA content in meat on days 3, 6 and 14 of storage, TVC, milk yield, and protein, fat, lactose, SCC, and urea content in milk. However, to obtain reliable results, it is recommended to use meta-regression only when the variable of interest is reported in at least 10 studies [40]. Consequently, meta-regression was only performed for the variables: DMD, OMD, CPD, NDFD, ADFD, ADG, LMA, rumen pH, NH 3 -N, acetate, propionate, butyrate, total protozoa, glucose, cholesterol, triglycerides, meat pH, and milk yield, as well as protein, fat, and lactose content in milk. Table 8 shows that EOs dose explained (p = 0.004) 16.39% of the observed heterogeneity for ruminal NH 3 -N concentration. Similarly, EOs dose explained (p < 0.001) 2.65 and 28.20% of the observed heterogeneity for serum triglyceride concentration and milk yield, respectively (Table A2). On the other hand, the experimental period only explained (p = 0.003) 3.52 and 7.30% of the heterogeneity observed for rumen pH and total rumen protozoa (Table 8). Moreover, Table A2 shows that the experimental period explained (p = 0.011) 10.41% of the heterogeneity observed for milk protein content. Table 8 shows that the primary bioactive metabolite explained (p < 0.05) between 7.63 and 62.55% of the observed heterogeneity for DMD, NDFD, ADFD, ruminal pH, NH 3 -N, acetate, propionate, butyrate, and total protozoa. Likewise, the primary bioactive metabolite of the EOs explained (p < 0.05) 30.28, 80.30, 47.17, and 32.38% of the observed heterogeneity for serum glucose concentration, triglycerides, milk yield, and milk protein content, respectively (Table A2). Moreover, Table 8 shows that there was no significant relationship (p > 0.05) between the covariates used and the response variables ADG, OMD, and CPD. In addition, serum cholesterol concentration and milk lactose content had no significant relationship (p > 0.05) with any of the covariates used (Table A2).  Figure 1a shows that ruminal NH 3 -N concentration decreased (WMD = −2.065 mg/dL; p = 0.005) when moderate doses of EOs (501-1000 mg/kg DM) were used. However, low (≤500 mg/kg DM) and high (>1000 mg/kg DM) doses of EOs did not affect ruminal NH 3 -N concentration. Dietary inclusion of EOs at low (501-1000 mg/kg DM) and moderate (501-1000 mg/kg DM) doses did not affect serum triglyceride concentration (Figure 1b; p > 0.05). However, serum triglyceride concentration decreased (WMD = −4.793 mg/dL; p < 0.001) when doses of EOs greater than 1000 mg/kg DM were used (Figure 1b). Figure 1c shows that milk yield increased (p < 0.001) regardless of the dose of EOs used. However, the effect was greater (WMD = 0.226 kg/d) when low doses of EOs (≤ 500 mg/kg DM) were used compared with doses between 501 and 1000 mg/kg DM (WMD = 0.080 kg/d) and doses greater than 1000 mg/kg DM (WMD = 0.083 kg/d).

Dry Matter Intake and Digestibility
Supplementation with EOs increased DMI. Similar responses were previously reported by Orzuna-Orzuna et al. [25] in a meta-analysis with beef cattle supplemented with EOs. It has been reported that EOs can improve the taste and palatability of livestock foods [14], which could result in increased DMI. Furthermore, in small ruminants, various EOs have been shown to increase the relative abundance of fungi and ruminal bacteria (R. flavefaciens, R. albus, and F. succinogenes) that are related to fiber degradation in the rumen [41,42]. This could result in a higher rate of feed particle passage in the rumen and higher DMI. Consequently, similar effects of EO consumption in the present meta-analysis would partly explain the observed increase in DMI.
According to Clouard and Val-Laillet [43], the first stimulus perceived by animals when exposed to feed is its aroma. Therefore, EOs should be carefully dosed because some of them have primary bioactive compounds with strong aroma [14], which could limit DMI in ruminants. In addition, a meta-analysis conducted by Orzuna-Orzuna et al. [25] showed that, in beef cattle, the effects of dietary inclusion of EOs on DMI depend on the dose and experimental period used. Nevertheless, in the present meta-analysis, the heterogeneity test for DMI was not significant. This suggests that EOs could be used to stimulate DMI in small ruminants independently of the primary bioactive metabolite, dose, and supplementation period used.
It has been reported that there is a close relationship between the relative abundance of some ruminal microorganisms and the digestibility of dietary nutrients [44]. Thus, one of the objectives of dietary inclusion of EOs is to increase the relative abundance of ruminal microbial populations that utilize efficient fermentation pathways [45], which could improve the efficiency of nutrient utilization. In the present study, dietary supplementation of EOs increased the relative abundance of R. flavefaciens, which would explain the increase in DMD, NDFD, and ADFD. Zhou et al. [41] and Kim et al. [46] observed that, in sheep and cattle, dietary supplementation of EOs increased the presence of rumen fungi. These produce high levels of hemicellulases and cellulases [46] and have the ability to penetrate the cell wall to enhance cellulose degradation [47]. Likewise, Zhang et al. [48] reported a higher ruminal concentration of cellulase, lipase, and β-glucosidase in the ruminal fluid of beef cattle supplemented with EOs. In vitro studies [49,50] have shown that EOs increase the relative abundance of Succinivibrio bacteria, which have a positive correlation with DMD, NDFD, and ADFD in dairy cows [51]. Furthermore, Cobellis et al [52] reported that, in lambs, dietary supplementation of plants high in EOs decreases the ruminal abundance of Prevotella bacteria, which are negatively correlated with CPD in cattle [51]. In the present meta-analysis, similar effects of EO consumption would partially explain the increases observed for DMD, OMD, NDFD, ADFD, and CPD.

Growth Performance and Carcass Characteristics
In the present meta-analysis, dietary inclusion of EOs increased DMI, DMD, OMD, CPD, NDFD, and ADFD, which would partially explain the increase and decrease in ADG and FCR, respectively. In vitro studies [49,50] have reported that EOs increase the relative abundance of bacterial families (Lachnospiraceae, Rikenellaceae, and Christensenellaceae) that are positively and negatively correlated with ADG and FCR, respectively [44,53]. Likewise, some EOs reduce the relative abundance of Veillonellaceae bacteria [49], which are negatively correlated with ruminal production of TVFA and ADG in sheep [54,55]. Other additives containing EOs can increase up to 17 and 23% the efficiency of dietary energy utilization for maintenance and weight gain in lambs, respectively [4,16]. Ann et al. [56] and Wu et al. [57] reported that dietary supplementation with EOs in sheep increases the serum concentration of immunoglobulins IgA, IgG, and IgM. This could improve the health status of the animals and consequently increase their productive performance. In addition, previous studies [56,58] have shown that in lambs the dietary inclusion of low doses (50, 80, and 250 mg/kg DM) of EOs increases serum levels of IGF-1 (insulin-like growth factor 1), which is positively correlated with ADG in sheep [59]. For this reason, similar effects of EO consumption in the present study would partially explain the observed improvements in ADG and FCR [59].
Dietary supplementation of EOs increases the relative abundance of Lachnospiraceae bacteria in bovine rumen fluid [49], which correlates positively with the length of ruminal papillae in sheep [53,55]. Similar effects of EO consumption used in the present metaanalysis could increase ruminal absorption of TVFA and result in higher ADG and lower FCR. Moreover, dietary inclusion of EOs (150 and 300 mg/kg DM) has been reported to increase villus length in the duodenum, jejunum, and ileum of lambs by 37-75% [60]. This could increase the intestinal absorption of amino acids and other nutrients by the animal, which would explain the observed improvements in ADG and FCR.
Dietary inclusion of EOs did not affect HCW, CCW, and BFT but increased HCY and LMA. Similar to our results, a meta-analysis conducted by Orzuna-Orzuna et al. [25] reported that dietary supplementation with EOs increased HCW and LMA in beef cattle, without negatively affecting other carcass characteristics. The increase in HCY observed in the present meta-analysis could be associated with the increase in LMA because there is a positive correlation between these carcass characteristics [61]. Furthermore, according to Laliotis et al. [62], ruminal acetate is the main lipogenic precursor in ruminant adipose tissue. In the present study, the dietary inclusion of EOs did not affect the ruminal acetate concentration, which would explain the absence of significant changes in BFT.
Many of the EOs used in livestock feed are high in terpenoids [5]. It has been reported that terpenoids can promote muscle stem cell differentiation as well as muscle tissue synthesis in mammals [63]. Although there is little information on the mechanisms of action of EOs and their bioactive metabolites on muscle development in ruminants, there is evidence that terpenoids reduce proteolytic degradation in muscle tissue in rats [64] and increase hypertrophy in skeletal muscle cells [65]. Likewise, in beef cattle supplemented with EO mixtures, Monteschio et al. [66] observed increases of 7 and 16% in the diameter and fiber area of Longissimus dorsi muscle, respectively. Consequently, similar effects of EO consumption in the present study would partially explain the higher LMA observed.

Ruminal Fermentation and Ruminal Microorganisms
In the present study, dietary supplementation with EOs did not affect rumen pH but reduced rumen NH 3 -N concentration. Like our results, a meta-analysis conducted by Orzuna-Orzuna et al. [25] reported that, in beef cattle, dietary supplementation with EOs reduced rumen NH 3 -N concentration without affecting rumen pH. In the present meta-analysis, the results observed for ruminal pH suggest that ruminal functions of small ruminants were performed under stable conditions because rumen pH serves as an indicator of the internal homeostasis of the ruminal environment [25]. On the other hand, the lower rumen NH 3 -N concentration observed suggests that EOs reduced rumen protein degradation. It has been reported that EOs can reduce the rate of amino acid deamination in the rumen and inhibit the ruminal growth of some hyper ammonia-producing bacteria (Clostridium sticklandii and Peptostreptococcus anaerobius) [67]. This could reduce rumen ammonia production, which would explain the observed reduction in the ruminal NH 3 -N.
Dietary supplementation with EOs increased the rumen concentration of propionate and reduced the total rumen protozoa population but did not affect the concentration of acetate and butyrate. Like our results, a meta-analysis conducted by Orzuna-Orzuna et al. [25] reported that, in beef cattle, dietary inclusion of EOs reduced the number of protozoa in the rumen and increased the ruminal concentration of propionate, without negatively affecting the other ruminal parameters. The observed increase in ruminal propionate concentration suggests that EOs may increase the availability of energy for growth and production because propionate serves as an energy source for some anabolic functions in ruminants [68]. It has been reported that, under in vitro conditions, dietary inclusion of EOs decreases the relative abundance of Succiniclasticum bacteria [49], which are negatively correlated with propionate concentration in rumen fluid [69]. Zhang et al. [48] reported that, in beef cattle, dietary supplementation with EOs increased the relative abundance of Parabacteroides distasonis and Bacteroides thetaiotaomicron bacteria, which increased ruminal propionate concentration. Similar effects of EO consumption in the present meta-analysis would partially explain the higher rumen propionate concentration observed.
The reduction in the number of total protozoa in the rumen could be favorable because an increase in the population of ruminal protozoa increases ruminal protein degradation [70] and CH 4 emissions [71]. This results in lower utilization efficiency of protein and energy consumed and, consequently, limits the productivity of small ruminants. According to Franzolin and Dehority [72], ruminal pH plays an important role in the survival of rumen protozoa. In the present study, ruminal pH was not affected by the dietary inclusion of EOs. This suggests that the reduction in total protozoa and rumen Epidinium might be associated with antimicrobial effects of EOs rather than rumen pH. However, Benchaar et al. [13] mentioned that it is possible that ruminal microorganisms adapt to the effects of EOs when they are used for long periods, which could diminish their positive effects. In the present study, subgroup analysis revealed that total rumen protozoa decreased regardless of the supplementation period used.
In a review article, Cobellis et al. [73] mentioned that, although EOs appear to be effective in reducing the rumen abundance of methanogens, they could also negatively affect the relative abundance of R. flavefaciens, R. albus, and F. succinogenes. However, in the present meta-analysis, EOs increased the abundance of R. flavefaciens without affecting the abundance of R. albus and F. succinogenes. This could be associated with the observed reduction in the total rumen protozoan population, which has been reported to be negatively correlated with the rumen abundance of R. flavefaciens in dairy goats [74].
Supplementation of EOs in beef cattle has been reported to increase the relative rumen abundance of the bacterial family Succinivibrionaceae [49,50], which has a strong negative correlation with the relative abundance of Methanobacteriaceae microorganisms [75]. Similar effects of EO consumption in the present meta-analysis would partially explain the reduction observed for rumen abundance of methanogens. On the other hand, dietary supplementation of EOs reduced enteric CH 4 emissions in small ruminants. Like our results, a meta-analysis conducted by Belanche et al. [26] reports lower CH 4 production in dairy cows supplemented with EOs. Wallace et al. [76] demonstrated that, in beef cattle, there is a strong positive correlation between enteric CH 4 emissions and the relative abundance of rumen methanogens and protozoa. In the present study, EOs reduced the rumen abundance of protozoa and methanogens, which would explain the observed reduction in CH 4 .

Blood Metabolites
In the present meta-analysis, the lower serum urea concentration was observed in response to EOs supplementation. In a previous meta-analysis, Orzuna-Orzuna et al. [25] also reported lower serum urea concentration in beef cattle supplemented with EOs. Additionally, it has been demonstrated that under in vitro conditions, EOs increase the relative abundance of the bacterial family Lachnospiraceae [49], which has a negative correlation with serum urea concentration in beef cattle [69]. Similar effects of EO consumption in the present meta-analysis would partially explain the reduction in serum urea concentration. In addition, the lower serum concentration of urea in the present study could be related to the reduction observed in the ruminal concentration of NH 3 -N, because in ruminants these two parameters are positively correlated [77].
According to Ran et al. [78], serum concentrations of glucose, NEFA, and BHB can be used as reliable indicators of energy status in ruminants. In the present meta-analysis, EOs did not affect serum glucose concentration but reduced serum NEFA and BHB concentration. This suggests that dietary supplementation with EOs improves energy balance in small ruminants. Similar responses were previously reported by Orzuna-Orzuna et al. [25] in a meta-analysis with beef cattle supplemented with EOs. The absence of significant changes in serum glucose concentration was not expected because EOs increased the ruminal concentration of propionate, which is the main glucose precursor in ruminants [79]. Additionally, it has been reported that EOs increase the relative abundance of Lachnospiraceae and Bifidobacterium bacteria [49], which are negatively correlated with serum levels of NEFA [69] and BHB [53] in beef cattle and sheep, respectively. Therefore, similar effects of EO consumption in the present meta-analysis would partially explain the reduction in serum NEFA and BHB concentration.
EOs supplemented in the diet did not affect the serum concentration of albumin, globulin, and total protein, suggesting that supplementation with EOs has no negative effects on protein catabolism and nutritional status of small ruminants [25]. The serum concentration of cholesterol and triglycerides decreased in response to EOs supplementation. There is limited information on the mechanisms of action of EOs on lipid metabolism in small ruminants. However, it has been reported in mice that terpenoids from EOs inhibit hepatic cholesterol biosynthesis and decrease the expression of genes (Fas, Scd1, and Acc1) that are involved in fatty acid synthesis [80]. Similar effects of EO consumption in small ruminants would explain the lower serum cholesterol and triglyceride concentrations in the present study.
In ruminants, excessive accumulation of prooxidant substances such as reactive oxygen species (ROS) can cause oxidative stress [81]. According to Vasta and Luciano [82], it is possible to use EOs as natural antioxidants in the diet of small ruminants because they contain bioactive metabolites (terpenes and terpenoids) with antioxidant properties. In the present study, dietary supplementation with EOs increased TAC. This indicates that EOs reduce ROS in blood serum due to the negative correlation between TAC and ROS in blood [83]. On the other hand, Gessner et al. [84] reported that CAT, SOD, and GPx are antioxidant enzymes that can reduce oxidative stress because they convert ROS into other compounds less harmful to biological macromolecules in the organism. Therefore, in the present meta-analysis, the observed increase in CAT and SOD indicates that EOs reduce oxidative stress in small ruminants.

Meat Quality
Dietary supplementation with EOs did not affect meat pH but reduced CL and ShF. In a previous meta-analysis, Orzuna-Orzuna et al. [25] also found that supplementation with EOs in beef cattle reduced CL and ShF without affecting meat pH. It has been stated that CL can be used as an indicator of water holding capacity (WHC) in ruminant meat [61]. Ablikim et al. [85] reported that CL was negatively correlated (r = −0.894) with WHC in sheep meat. Consequently, the lower CL observed in the present study suggests that supplementation with EOs improves WHC in small ruminant meat. The lower ShF observed in the present meta-analysis suggests that EOs improve tenderness in small ruminant meat [86]. Likewise, the lower ShF observed in response to EOs supplementation could be associated with reduced CL, because in small ruminant meat there is a positive correlation (r = 0.42) between ShF and CL [87].
In the present meta-analysis, EOs reduced b* and MDA in small ruminant meat but did not affect other color parameters (L*, a*) or the chemical composition of meat. Similar responses were previously reported by Orzuna-Orzuna et al. [25] in a meta-analysis with beef cattle supplemented with EOs. Meat color can be used to evaluate the quality of ruminant meat because the color is the first characteristic considered by consumers when choosing fresh meat [88,89]. It has been reported that in lamb meat L* values are associated with the fat content of the meat [90]. Likewise, in beef, Węglarz [91] reported that color parameters (L* and a*) are negatively correlated with meat pH. In the present study, EOs did not affect pH or fat content in small ruminant meat, which would partially explain the absence of changes observed for L* and a*. Furthermore, the observed reduction in b* suggests that EOs improve the quality of fresh small ruminant meat, because consumers generally do not expect to find b* too high in fresh meat [89].
The reduction in MDA in stored meat (on days 1, 3, 6, 9, and 14) suggests that EOs reduce lipid peroxidation of small ruminant meat [92]. According to Pateiro et al. [93], oxidation reactions that occur in meat during storage can cause physicochemical changes and unpleasant odors, which negatively affect meat quality and shelf life. Therefore, the reduction observed for MDA in the present study suggests that EOs can be used as a nutritional strategy to improve the quality and shelf life of small ruminant meat. In addition, previous studies in non-ruminants [94,95] have reported that dietary supplementation with EOs increases antioxidant activity in smooth and skeletal muscle due to increased mRNA for SOD, CAT, and GPx. Likewise, the antioxidant capacity of EOs has been attributed mainly to the terpenoids they contain [5], which after consumption by sheep can be absorbed and deposited in muscle tissues [96]. Similar effects of EO consumption in the present study partially explain the reduction in MDA.
The values observed for the chemical composition of meat suggest that supplementation with EOs does not affect the nutritional value of meat in small ruminants because the nutritional value of meat is related to the content of minerals, fats, and proteins [86]. Thus, the composition of small ruminant meat can be modified by changes in dietary components [61]. In the present study, the dietary inclusion of EOs did not significantly affect the chemical composition of the diets, which would partially explain the observed similarity in the chemical composition of the meat.
According to Dave and Ghaly [97], microbial spoilage of meat can affect its quality and shelf life by negatively affecting pH and appearance, as well as causing off-odors and degradation of structural components. In the present study, supplementation with EOs reduced TVC, ENT, PSY, and MY in meat, suggesting that EOs improve the quality and shelf life of small ruminant meat. Some natural additives have bacteriostatic effects on lamb meat because they reduce its pH [88]. However, in the present meta-analysis, supplementation with EOs did not affect the pH of the meat. It has been shown that terpenoids can be absorbed and deposited in muscle tissues when administered through feed [96]. Terpenoids are known to cause cell lysis and cell death in pathogenic bacteria, as well as inhibit the growth of yeasts and molds [98]. Similar effects of EO consumption in the present study would partially explain the reduction observed for TVC, ENT, PSY, and MY.

Milk Production and Quality
According to Kholif et al. [99], to increase milk yield and FE in small ruminants it is necessary to reduce protein and energy losses during ruminal fermentation, in addition to improving the efficiency of utilization of consumed nutrients. In the present metaanalysis, EOs reduced ruminal NH 3 -N concentration and CH 4 emissions but increased OMD, CPD, and NDFD. This suggests lower energy and protein loss during ruminal fermentation and higher utilization efficiency of ingested nutrients, which explains the observed increase in milk yield and FE. In addition, under in vitro conditions, EOs increase the relative abundance of microorganisms of the genus Ruminococcus [49], which are positively correlated with milk yield in dairy goats [74,100]. It has also been reported that EOs reduce the ruminal abundance of Clostridium bacteria in goats [101], which has been negatively correlated with FE in dairy cows [102]. Hence, similar effects of EO consumption in the present study partially explain the observed increases in milk yield and FE [102].
Milk fat content was not affected by supplementation with EOs. In the mammary gland of ruminants, the main precursor for de novo fatty acid synthesis is rumen acetate [103]. Seymour et al. [104] showed that there is a positive correlation (r = 0.31) between milk fat content and ruminal acetate concentration. In the present study, EOs did not affect rumen acetate concentration, which explains the absence of significant effects on milk fat content. On the other hand, higher milk lactose content was observed in response to dietary supplementation with EOs. This could be associated with the observed increase in ruminal propionate concentration because propionate is the main rumen volatile fatty acid required for lactose biosynthesis [99]. Furthermore, in vivo studies with sheep [52] and goats [101] have reported that dietary supplementation with EOs reduces the ruminal concentration of Prevotella, which has been negatively correlated with the percentage of lactose in the milk of dairy cows [102]. Zhou et al. [49] observed that in bovine rumen fluid, EOs decrease the relative abundance of Eubacterium and methanogens, which is also negatively correlated with the protein content in milk of dairy goats [100]. Consequently, similar effects of EO consumption in the present meta-analysis partially explain the increases observed in lactose and protein content in milk.
SCC can be used as an indicator of udder health and milk quality in ruminants [105]. According to Malik et al. [106], in the milk of healthy ruminants, SCC includes 75-85% immune cells and 15-25% epithelial cells. It has been reported that an increase in SCC is associated with poorer udder health [107] and lower milk quality in ruminants [108]. In the present study, supplementation with EOs reduced SCC, suggesting that EOs could be used to improve udder health and milk quality in small ruminants. In addition, it has been reported that there is a negative correlation between the concentration of antioxidant enzymes and SCC in ruminant milk [106]. In the present meta-analysis, a higher serum concentration of CAT and SOD was observed in small ruminants supplemented with EOs, which would partially explain the observed reduction in SCC.

Conclusions
The results of the present study indicate that EOs can be used as natural growth promoters in small ruminants and, at the same time, improve feed intake and feed efficiency. Furthermore, dietary supplementation with EOs improves nutrient digestibility, meat quality, and shelf life, as well as milk production and quality. The best result for milk production is obtained with EOs doses lower than 500 mg/kg DM and when the primary bioactive metabolite of the EOs is linalool, limonene, or mixtures of metabolites. Likewise, the best protein content in milk is obtained with supplementation periods longer than 70 days and with the use of EOs that have mixtures of bioactive metabolites.
Dietary supplementation with EOs improves fermentation and reduces environmental impact by increasing ruminal propionate concentration and by reducing methane emissions, ruminal ammonia nitrogen concentration, and the number of total protozoa and methanogens. The best ruminal propionate concentration is obtained when using EOs containing limonene, linalool, or eucalyptol as primary bioactive metabolites. The best result for rumen ammonia nitrogen is obtained with moderate doses (501-1000 mg/kg DM) of EOs and when the primary bioactive metabolite of the EOs is limonene, diallyl disulfide, or mixtures of metabolites. The best results for total protozoa were obtained with supplementation periods longer than 70 days and with the use of EOs having linalool, alpha-pinene, diallyl disulfide, or mixtures of bioactive metabolites. Finally, the results of serum metabolites indicate that EOs improve the antioxidant status in the blood of small ruminants.     Carvacrol (n = 2), thymol (n = 2), blend (n = 2) 100 (n = 3), 300 (n = 3)