This study was conducted in Xanxerê, Santa Catarina, Brazil (Latitude: −26.8364531149; Longitude: − 52.4079407059). The experimental protocol for evaluation and analysis was submitted to the local animal ethics commission (protocol number 41/2018) and after approval, research activities were initiated.

The additive used in the experiment consisted of a blend of compounds derived from essential oils of cinnamon and oregano. From cinnamon, the bioactive compound was cinnamaldehyde, from oregano, the bioactive molecule was carvacrol. The blend contained 300 g/kg of cinnamaldehyde, 200 g/kg of carvacrol, and 500 g/kg of calcium carbonate as the carrier (Enterosan, Phytobiotec, Canelinha, Brazil). The phytogenics were microencapsulated through coating by a thin lipid layer, through a spray chilling process ( Pereira et al., 2018 ), aiming its dissolution only in the distal portion of broiler jejunum and ileum. Five treatments with varying levels of blend addition on broiler feed were evaluated: control (absence of carvacrol and cinnamaldehyde), 50, 100, 150, and 200 mg/kg of ration with 7% of active ingredients.

Broilers were previously raised (1 to 14 days) in a concrete-floor masonry shed, divided into boxes with wood shaving as litter material, receiving a standard diet without any phytogenic additive, to guarantee non-influence in the digestibility trial evaluation. Animals were raised according to the norms and management of the commercial farms and the strain manual. Diets were formulated according to Rostagno et al. (2017) . Mashed feed ( Table 1 ) and water were provided ad libitum throughout the experimental period. Three hundred and twenty males of the Cobb500 strain were allocated in an entirely randomized, ungraded design consisting of five treatments, and eight replicates (n = 8). At 14-24 days, the animals were placed in metabolic cages for digestibility testing. Birds were maintained in metallic cages, distributed on four floors, and equipped with manual feeders and nipple drinkers. Each cage had a tray for excreta collection, digestibility, and energy metabolism evaluation. After that, in the second phase, birds were allocated to wood shaving pens, and raised until 42 days old for intestinal morphometry evaluation and blood sample collections.

Table 1

	Starter phase (1-24 d)	Finishing phase (25-42 d)
Ingredient (g/kg as-fed)
Corn	545.00	579.66
Soybean meal	361.65	309.00
Soybean oil	27.79	44.89
Dicalcium phosphate	18.30	18.64
Limestone	8.25	8.41
Salt	3.25	3.32
DL-Methionine (99%)	2.60	3.11
L-Lysine HCl	2.25	1.94
Choline chloride (60%)	1.00	1.00
Vitamin supplement1	15.00	15.00
Mineral supplement2	15.00	15.00
Calculated values (g/kg)
Dry matter	884.51	886.00
Metabolizable energy (kcal/kg)	2950.00	3100.00
Crude protein	215.00	194.00
Digestible lysine	12.00	10.50
Digestible methionine	5.44	5.05
Digestible Met + Cys	8.39	7.75
Digestible threonine	7.55	6.84
Digestible tryptophan	2.46	2.13
Digestible arginine	14.14	12.27
Digestible valine	9.25	8.20
Calcium	9.02	8.24
Available phosphorus	4.51	4.10
Sodium	1.70	2.05
Potassium	8.49	7.46
Chloride	3.77	3.56
Ash	88.29	59.13
Neutral detergent fiber	113.37	109.84
Acid detergent fiber	46.77	43.86

¹ Vitamin supplement containing per kg of complete diet: vitamin A, 15,000 IU; vitamin D3, 3,000 IU; vitamin E, 45 IU; vitamin B1, 3.0 mg; vitamin B2, 9.0 mg; vitamin B6, 6.0 mg; vitamin B12, 0.03 mg; pantothenic acid, 18.0 mg; biotin, 0.3 mg; vitamin K3, 4.5 mg; folic acid, 1.5 mg; nicotinic acid, 75.0 mg.

The experimental period of digestibility was 10 days (14-24 age), five for adaptation of the bird to feed and management and five for the traditional total excreta collection method. The quantities of feed consumed and excreta produced per experimental unit were determined, and dry matter (DM; method 934.01; AOAC, 1995 ), nitrogen (method 990.03; AOAC, 1995 ), and crude fat (method 920.39; AOAC, 2012 ) contents were analyzed. Once the results of laboratory analyses of feed and excreta were obtained, retention values of diet protein and lipid content were calculated according to the following formula:

Feed and gross energy were determined in a bomb calorimeter, Leco AC500 (LECO Corporation, St. Joseph, MI, USA). Then, the calculation of apparent metabolizable energy (AME) and apparent metabolizable energy corrected for nitrogen retention (AMEn) was done using equations proposed by Matterson et al. (1965) .

in which GE = gross energy and NB = nitrogen balance (nitrogen ingested − nitrogen excreted).

At the end of the 14-24-day experimental period, birds were relocated to the concrete-floor masonry shed and kept in boxes until 42 days old for intestine and blood sample collection.

After 42 days of the experiment, one animal per replicate was euthanized according to the guidelines and recommendations of CONCEA’s Guide to Euthanasia Practices (Brasil/MCTI, 2013). After jejunum fragment collection for histological analysis, samples were processed and sectioned at 4 to 6 μm and stained with hematoxylin-eosin. Measurements of villus height and crypt depth were performed using the image analyzer Image pro Plus 1.3.2 (1994) (magnification 40X) under an optical microscope. From each slide (one per bird), 30 villi and 30 crypts were selected to obtain the average value of each section presented. To obtain the villus:crypt ratio, the value of the height of the intestinal villi was divided by the depth value of the adjacent crypt.

For the evaluation of biochemical parameters, blood samples were collected from the brachial vein, 1 mL per animal, at 20 and 42 days of age. Serum was separated by centrifugation and stored at −20 °C until further analysis of serum concentrations of total protein, albumin, cholesterol uric acid, and aspartate aminotransferase (AST) activity. The biochemical tests were carried out using commercial kits (Analisa) specific for each of these parameters, measured using a Bioplus 2000 (Bioplus, Barueri, Brazil) semi-automatic analyzer. Globulin levels were calculated as the differences between total protein and albumin levels.

For oxidant evaluation, reactive oxygen species (ROS) production was assessed by determining dichloro-dihydro-fluorescein diacetate (DCFH-DA) oxidation ( LeBel et al., 1992 ), which is hydrolyzed by intracellular esterases to form non-fluorescent DCFH, which is then rapidly oxidized to form highly fluorescent 2′,7′-dichlorofluorescein (DCF) in the presence of ROS. The DCF fluorescence intensity correlates to the amount of ROS formed. Fluorescence was measured using excitation and emission wavelengths of 480 and 535 nm, respectively. The calibration curve was performed with standard DCF (0.1–1 μM), and data were calculated as micromoles DCF formed/mL of serum. Lipid peroxidation was determined as levels of thiobarbituric acid reactive substances (TBARS) in serum according to the method described by Jentzsch et al. (1996) . Results were obtained by spectrophotometry at 535 nm and expressed in nmol of malondialdehyde (MDA)/mL of serum.

To evaluate antioxidant parameters, glutathione peroxidase (GPx) activity was measured using tert-butyl-hydroperoxide as a substrate ( Wendel, 1981 ). The enzyme activity was determined by monitoring the nicotinamide-adenine-dinucleotide-phosphate (NADPH) disappearance at 340 nm in a medium containing 100 mM potassium phosphate buffer/1 mM EDTA, pH 7.7, 2 mM GSH, 0.1 U/mL GR, 0.4 mM azide, 0.5 mM tert-butyl hydroperoxide, 0.1 mM NADPH, and tissue supernatants. The results were calculated and expressed as U/mL of serum. The activity of glutathione S-transferase (GST) was measured according to Mannervik and Guthenberg (1981) with slight modifications; GST activity was measured by the rate of formation of dinitrophenyl-S-glutathione at 340 nm in a medium containing 50 mM potassium phosphate, pH 6.5, 1 mM GSH, 1 mM 1-chloro-2,4-dinitrobenzene (CDNB) as substrate, and tissue supernatants (approximately 0.045 mg of protein); results were calculated and expressed as U/mL of serum. Superoxide dismutase (SOD) activity was determined according to the auto-oxidation principle of pyrogallol, inhibited in the presence of SOD. The optical density change was determined kinetically for two minutes at 420 nm, at ten-second intervals according to the methodology described by Beutler (1984) . The activity was expressed as U/mL of serum.

The results were subjected to the Shapiro-Wilk normality test. When normal, the analysis of variance used was parametric and if significant, subjected to linear and quadratic regression analysis and also to the Dunnett’s test. The different levels of the exogenous enzymes were subjected to regression analysis using linear and quadratic functions. All analyzes were performed at a 5% significance level, using the R statistical program (R Core Team, 2013 ). The statistical model was as follows:

in which Y_ij = dependent variable, μ = variable mean, β_i = effect of broilers, and ε_ij = experimental error associated with observation Y_ij.

There was quadratic effect (P<0.001) for values of AME expressed in DM ( Y = 3492 − 0.646 x + 0.008 x 2 ) and as-fed basis ( Y = 3268 − 0.903 x + 0.008 x 2 ) ( Table 2 ). The feed supplemented with 200 mg/kg showed numerically higher values of AME than the control. There were differences ( Table 2 ) in AMEn, expressed in natural matter basis ( Y = 3065 + 0.219 x + 0.0005 x 2 , P = 0.012) and DM ( Y = 3273 + 0.286 x + 0.0005 x 2 , P = 0.005), showing a quadratic effect (P = 0.005) and significance of samples subjected to the Dunnett’s test (P<0.001) for AMEn in dry matter. Similar behavior was observed for the AMEn values with quadratic effect (P = 0.012) as a fed basis and averages with significance according to the Dunnett’s test (P<0.001).

Table 2

Item	Blend inclusion level1 (mg/kg)					SEM	Linear effect	Quadratic effect	Dunnett
	0	50	100	150	200
Feed intake (15-24 d, g)	325	323	323	340	319	18.23	0.312	0.214	0.323
DM AME (kcal/kg)	3473	3525	3491	3562	3706*	82.26	0.432	<0.001	<0.001
NM AME (kcal/kg)	3252	3281	3251	3313	3450*	76.62	0.142	<0.001	<0.001
DM AMEn (kcal/kg)	3274	3284	3297	3305	3310*	74.71	0.215	0.005	<0.001
NM AMEn (kcal/kg)	3306	3073*	3083*	3088*	3090*	69.58	0.215	0.012	<0.001
PD (g/100 g)	69.99	71.46	67.71	68.48	69.50	5.54	0.294	0.706	0.706
EED (g/100 g)	88.94	91.92	89.77	89.79	91.18	2.11	0.105	0.053	0.053

DM AME - dry matter apparent metabolizable energy; NM AME - natural matter apparent metabolizable energy; DM AMEn - dry matter apparent metabolizable energy corrected by nitrogen retention; NM AMEn - natural matter apparent metabolizable energy corrected by nitrogen retention; PD - protein digestibility; EED - ether extract digestibility; NS - non-significant; SEM - standard error of the mean.

There was no effect (P>0.05) of the addition of carvacrol and cinnamaldehyde on the digestibility of protein and ether extract ( Table 2 ) of feed according to the levels of phytogenics added.

Inclusion of 100 mg/kg of carvacrol and cinnamaldehyde blend increased intestinal villus height according to the Dunnet’s test (P = 0.033), crypt depth (P<0.001), as well as villus:crypt ratio (P<0.001). The inclusion of other levels did not cause changes ( Table 3 ) in villus height, crypt depth, or villus:crypt ratios (P>0.05).

Table 3

Item	Blend inclusion level1 (mg/kg)					SEM	Linear effect	Quadratic effect	Dunnett
	0	50	100	150	200
Feed intake (g)	3299	3315	3283	3304	3290	28.542	0.184	0.216	0.164
Villus height (µm)	637.24	687.59	787.63*	711.48	685.90	34.701	0.354	0.242	0.033
Crypt depth (µm)	59.00	52.92	53.54	76.48*	60.03	12.405	0.635	0.593	<0.001
Villus:crypt ratio	11.45	13.55	15.93*	9.96	12.03	0.436	0.790	0.774	<0.001

SEM - standard error of the mean.

Serum levels of AST did not change in both periods (20 and 42 days) (P>0.05) ( Table 4 ). Levels of AST increased between the two collection periods evaluated in all groups, however, in a proportional way to all evaluated treatments, with no significant differences between them (P>0.05). Albumin levels significantly differed (Dunnett’s test, P = 0.047) at 20 days of age in broilers; nevertheless, the same data did not show a significant quadratic effect ( Table 4 ) (P>0.05). Total protein and globulin levels did not reflect a quadratic effect at the 20th day, but the same data were not significant when subjected to the Dunnett’s test (P = 0.840 and P = 0.266, respectively). At 42 days of age, the means differed for total protein (P<0.001), albumin (P<0.001), and globulins (P<0.001) according to the Dunnett’s test, and quadratic effect for total protein (P = 0.009).

Table 4

Parameter	Blend inclusion level1 (mg/kg)					SEM	Linear effect	Quadratic effect	Dunnett
	0	50	100	150	200
20 days
AST (IU/L)	157.57	165.14	183.28	175.57	188.14	24.151	0.183	0.732	0.183
TP (g/dL)	2.62	2.75	2.95	2.83	2.77	0.562	0.164	0.840	0.840
Albumin (g/dL)	1.38	1.16	1.25	1.11*	1.40	0.213	0.256	0.070	0.047
Globulin (g/dL)	1.24	1.58	1.70	1.72	1.37	0.510	0.351	0.266	0.266
Cholesterol (mg/dL)	76.00	79.37	71.14	71.28	78.37	10.841	0.562	0.400	0.400
Uric acid (mg/dL)	6.21	6.38	5.35	4.02	6.78	2.151	0.260	0.169	0.168
42 days
AST (IU/L)	292.75	350.25	302.85	355.71	335.28	59.61	0.295	0.176	0.176
TP (g/dL)	2.98	3.01	3.64*	2.65	3.20	0.25	0.143	0.009	<0.001
Albumin (g/dL)	1.10	1.27	1.24	1.18*	1.40	0.08	0.264	0.154	<0.001
Globulin (g/dL)	1.88	1.73	2.40*	1.46*	1.80	0.25	0.255	0.175	<0.001
Cholesterol (mg/dL)	80.25	86.62	81.00	70.62	77.12	14.47	0.723	0.603	0.370
Uric acid (mg/dL)	6.68	5.25	5.14	4.46*	3.58*	0.86	0.092	0.219	0.040

AST - aspartate aminotransferase; TP - total protein; SEM - standard error of the mean.

Cholesterol values in broilers fed various levels of carvacrol and cinnamaldehyde microencapsulate showed no changes at 20 and 42 days of age (Dunnet’s test, P = 0.400 and P = 0.370, respectively), with no linear or quadratic effects ( Table 4 ). There were no changes in uric acid levels at 20 days of age (P = 0.168), but at 42 days, there was a tendency for a steady drop in uric acid due to increasing inclusion levels of carvacrol and cinnamaldehyde (P = 0.092), and lower uric acid levels (P = 0.040) in broilers fed 150 and 200 mg/kg of the blend than control ( Table 4 ).

There were no differences in TBARS levels at 20 days (P>0.05), but there was a quadratic effect (P = 0.007) at 42 days ( Y = 3.968 + 0.00005 x + 0.003 x 2 ). Reactive oxygen species at 20 days showed higher levels at 50 mg/kg, and at 42 days, the same dose at 50 mg/kg presented the lowest values for this parameter (Dunnett’s test, P<0.001).

At 20 days of age, there were increases in GPx (Dunnett’s test, P = 0.007) and GST activities (Dunnett’s test, P<0.001) at 50 mg/kg ( Table 5 ). For SOD (P = 0.012), values were lower at 100, 150, and 200 mg/kg of the blend than in the control, with the tendency for quadratic effect (P = 0.060).

Table 5

Parameter	Blend Inclusion levels1 (mg/kg)					SEM	Linear effect	Quadratic effect	Dunnett
	0	50	100	150	200
20 days
TBARS (nmol MDA/mL of serum)	4.045	6.187*	5.500	5.027	6.187*	0.7723	0.544	0.629	0.008
ROS (U DCF/mL of serum)	1.853	3.844*	1.827	1.749	2.299	1.2015	0.933	0.922	<0.001
GPx (U/mL of serum)	1.947	6.197*	3.917	2.818	4.617	0.7602	0.654	0.282	0.007
GST (U/mL of serum)	12.696	18.650*	14.739	12.289	12.372	2.3202	0.563	0.587	<0.001
SOD (U/mL of serum)	1.891	1.430	0.997*	1.181*	1.236*	0.1103	0.065	0.060	0.012
42 days
TBARS (nmol MDA/mL of serum)	3.958	4.067	3.783	3.474	2.775*	0.06	0.192	0.007	0.003
ROS (U DCF/mL of serum)	2.566	1.301*	2.549	1.904	2.515	0.69	0.781	0.598	<0.001
GPx (U/mL of serum)	3.485	2.369	6.414	2.917	3.881	1.41	0.892	0.258	0.051
GST (U/mL of serum)	12.248	11.407	18.139*	11.664	12.307	3.47	0.524	0.071	0.023
SOD (U/mL of serum)	1.361	1.099	3.450	2.033	1.217	0.46	0.401	0.326	0.264

TBARS - thiobarbituric acid reactive substances; MDA - malondialdehyde; ROS - reactive oxygen species; DCF - diclorofluorescin diacetate; GPx - glutathione peroxidase; GST - glutathione-S-transferase; SOD - superoxide dismutase; SEM - standard error of the mean.

At 42 days, there were no differences among the doses for GPx, GST, and SOD activities (P>0.05). For GST, the dose of 100 mg/kg gave higher levels than the other doses studied (Dunnett’s test, P = 0.023).