Effect of fermentation of herbal products on growth performance, breast meat quality, and intestinal morphology of broiler chickens: a meta-analysis

Abstract

Although it has widely been applied in human applications for therapeutic purposes, the use of fermented herbal products to improve growth performance in broiler chickens is still disputable. This study aimed to compare the effectiveness of fermented versus unfermented herbal products in improving the growth performance, breast meat quality, and small intestinal morphology of broiler chickens and to determine the optimal conditions according to meta-analysis and response surface methodology. The database was developed based on 23 articles comprising 123 data points collected in 2023. The metadata was analysed using OpenMEE and R, with the inclusion of fermented herbal products as fixed factors and individual studies as random factors. The optimal dosage was determined using response surface methodology (RSM). A significant difference (p < 0.05) was detected between unfermented and fermented herbal products in terms of broiler body weight gain, feed intake, and feed conversion ratio (FCR) during the starter phase. The principal notable impact is associated with the utilisation of Zingiber officinale and fermenting agents, such as Lactobacillus sp. and Saccharomyces cerevisiae. Both treatments resulted in a substantial reduction (p < 0.01) in the FCR. According to the meta-regression, growth performance exhibited a consistent pattern with the results of the meta-analysis, including overall production parameters during the starter phase (p < 0.05), such as body weight, average daily gain, and daily feed intake. Significant differences were detected in breast meat quality, particularly in terms of polyunsaturated fatty acids (p < 0.05), as well as in the morphology of the digestive tract, such as duodenum crypt depth, villus height, and their ratio (p < 0.05). These findings indicate that fermenting dietary herbs can enhance the growth, breast meat quality, and intestinal morphology of broiler chickens, particularly during their initial growth phase. It is envisaged that broiler producers ferment herbs before adding them to chicken feed at ∼26.3 g/kg.

HIGHLIGHTS

• Fermented herbal products increase the body weight of broiler chickens.

• Fermented herbal products tend to suppress FCR in broiler chickens.

• There are trends in the increase in PUFA content after supplementation with fermented herbal products.

• Fermented herbal products affect villus height and crypt depth in broiler chickens.

• The addition of fermented herbal products reaches the optimum dosage at 26.3 g/kg.

Keywords:

• Breast meat quality
• broilers
• feed conversion ratio
• fermented herbal products
• Zingiber officinale

Introduction

The ban on the utilisation of antibiotic growth promoters (AGP) has prompted broiler chicken producers to explore substitutes for AGP, aiming to ensure the sustainability of broiler production. Alternatives to AGP, such as clay minerals, exogenous enzymes, herbal products, hyperimmune compounds, phytobiotics, prebiotics, probiotics, and symbionts, have all been proven to be effective in enhancing the digestive health status of broiler chickens, thereby boosting production (Adli et al. 2023; Sholikin et al. 2023). Herbal products have emerged as commonly adopted AGP alternatives in broiler production. Active compounds in herbal products have been shown to improve the ecological conditions, morphology and functions of the digestive tract, physiological conditions, antioxidant status, carcase characteristics, and immune competences of broiler chickens (Teymouri et al. 2021; Sugiharto and Ayasan 2023). Owing to these properties, herbal products are commonly used as antibacterial agents and sources of antioxidants for broiler chickens. However, the use of herbal products as alternatives to AGPs for broiler chickens is not consistent. In addition to showing positive results from the use of herbal products, Sugiharto (2021) revealed that many studies have reported no effect of herbal products on the growth performance of broiler chickens. Even the use of herbal products may impair feed palatability and consumption due to their bitter taste and pungent smell. The active ingredients found in herbal products, such as polyphenols, flavonoids, bioactive peptides, and essential oils, are highly susceptible to damage from light, heat, and oxygen during storage. This sensitivity can diminish the effectiveness of herbal products in enhancing the growth performance of broiler chickens (Jeong and Kim 2015; Bahadori et al. 2022; Sugiharto and Ayasan 2023). The effectiveness of herbal products as alternatives to AGPs is also often hindered by their low bioavailability for broiler chickens. Indeed, low absorption, biotransformation, and fast excretion may be attributed to the low bioavailability of herbal products for the host (Kikusato 2021).

Fermentation is a simple method involving microorganisms and has been widely used in the livestock industry (Sugiharto and Ranjitkar 2019). With respect to herbal products in particular, fermentation has been confirmed to enhance the active components present in herbal products, increase antioxidant activity, and improve the effectiveness of herbal products for therapeutic applications (Hussain et al. 2016). In line with a previous study, Zhang et al. (2013) reported that fermentation had positive effects on the bioavailability and bioactivity of the active compounds in herbal products. Therefore, fermentation can be applied to improve the effectiveness of herbal products as an alternative to AGP to improve broiler chicken growth performance.

Although it has widely been applied in human applications for therapeutic purposes, the use of fermented herbal products to improve growth performance in broilers is still disputable. A study conducted by Wang et al. (2021) utilising an herbal blend powder containing Astragalus, Panax notoginseng, and liquorice demonstrated enhanced growth performance in broiler chickens. Concurrently, Qiao et al. (2018) successfully enhanced growth performance by utilising fermented Astragalus, resulting in a reduction in the feed conversion ratio and an increase in the average daily feed intake and average daily gain. In contrast, studies by Cao et al. (2012) and Park et al. (2016) reported that the effects of fermented herbal products were not significant. Cao et al. (2012) reported that fermented Ginkgo biloba did not significantly affect the growth performance (body weight gain, feed intake, or feed conversion ratio) of broiler chickens. Similarly, Park et al. (2016) found no significant difference in the growth performance of broiler chickens when fermented ginger was used. Despite the inconsistent results from various studies, no specific recommended dosage has been reported for use with broiler chickens.

This indicates a need to understand the various perceptions of the use of fermented herbal products that exist among parameters that can be determined using meta-analysis. Using this information as a foundation, this meta-analysis aimed to evaluate the effectiveness of fermented and unfermented herbal products in a comparative manner. We also sought to identify the ideal dosage for improving the growth performance, quality of breast meat, and intestinal morphology of broiler chickens. Additionally, this study aimed to establish the optimal dosage of fermented herbal products.

Materials and methods

Eligibility criteria, search strategy, and data extraction

Science Direct (www.sciencedirect.com) and Scopus (www.scopus.com) were used to conduct the literature search. During the search, the keywords ‘broiler’, ‘fermented’, ‘herb/s’, ‘growth performance’, ‘breast meat quality’, and ‘intestinal morphology’ were used. In addition, the following criteria were utilised to identify literature: (1) full-text publications published in English; (2) peer-reviewed published journals; (3) indexed conference proceedings; (4) direct comparison of unfermented and fermented herbal products; (5) broiler chicken studies; and (6) measurements of body weight gain, feed intake, and FCR. The preliminary searches yielded 456 potential references. Following the screening, 41 references were removed since the study did not directly compare unfermented and fermented herbal products for broiler (Table 1). As shown in Figure 1, a total of 23 papers were utilised for data encoding and statistical examination, with adherence to the PRISMA-P guidelines throughout the entire process (Moher et al. 2015; Page et al. 2021).

Figure 1. The procedure for retrieving and choosing references for compiling metadata on fermented herbal products in broiler production.

Data coding

The study coding included characteristics, such as rearing period, broiler strains, chicken sex, herbs employed, and microbes used to ferment the herbal products. The mean values of the main selected parameters (growth performance, breast meat quality, and intestinal morphology) were recorded and tabulated, and the units of measurement were homogenised for further data analysis. The details of the growth performance parameters included body weight (BW; in grams), daily weight gain (DWG; in grams/heads/day), daily feed intake (DFI; in grams/heads/day), the feed conversion ratio (FCR), the mortality rate (%), the liveability rate (%), and the productive efficiency index (PEI).

The quality of breast meat includes the water holding capacity (% of the total live weight), protein content (%), fat content (%), pH (measured after 45 to 60 min), brightness (L*), redness (a*), and yellowness (b*). Additionally, the composition of fatty acids in the breast was considered: these included myristic acid (C14:0), palmitic acid (C16:0), stearic acid (C18:0), palmitoleic acid (C16:1), oleic acid (C18:1), eicosanoid acid (C20:0), linoleic acid (C18:2), α-linolenic acid (C18:3), arachidonic acid (C20:4), saturated fatty acids (SFAs), monounsaturated fatty acids (MUFAs), polyunsaturated fatty acids (PUFAs), the ratio of unsaturated fatty acids to saturated fatty acids (UFAs vs SFAs), and the ratio of polyunsaturated fatty acids to saturated fatty acids (PUFAs vs SFAs), was considered. The unit for breast fatty acids is the percentage of total fatty acids, except for the comparison parameters. Furthermore, the morphological characteristics of the gastrointestinal tract include crypt depth (measured in µm), villus height (measured in µm), and the villus height-to-crypt depth ratio in the duodenum, jejunum, and ileum.

Table 1. List of comparison studies used in the meta-analysis.

Herbs	Fermenter	Broilers	Sex	Rearing period	N	References
Alisma canaliculatum	Bacillus subtilis, Enterococcus faecium, Lactobacillus acidophilus, and Saccharomyces cerevisiae	Ross	Unsexed	1–35	5	Hossain et al. 2012
Allium sativum	Weissella koreensis	Arbour acres	Male	1–35	5	Ao, Yoo, et al. 2011
Allium sativum	Leuconostoc citreum	Ross 308	Male	1–35	5	Lee et al. 2016
Astragalus membranaceus	Lactobacillus plantarum	Avian	Unsexed	1–42	6	Qiao et al. 2018
Chenopodium album	Bacillus subtilis, Lactobacillus plantarum, and Saccharomyces cerevisiae	Arbour acres	Unsexed	1–42	5	Xie et al. 2021
Curcuma longa and Piper nigrum	Lactobacillus sp.	Lohmann	Unsexed	7–35	7	Sugiharto et al. 2021
Ginkgo biloba	Aspergillus niger	Arbour acres	Unsexed	1–42	6	Cao et al. 2012
Ginkgo biloba	Lactobacillus plantarum	Ross	Unsexed	1–35	8	Kim, Bostami, et al. 2016
Ginkgo biloba	Aspergillus niger and Candida utilis	Arbour acres	Unsexed	1–42	6	Niu et al. 2017
Ginkgo biloba	Aspergillus niger and Candida utilis	Arbour acres	Unsexed	1–42	6	Niu, Zhang, et al. 2019
Ginkgo biloba	Aspergillus niger and Candida utilis	Arbour acres	Unsexed	1–42	6	Niu, Wan, et al. 2019
Ginkgo biloba	Aspergillus niger	Arbour acres	Unsexed	1–42	6	Zhang et al. 2012
Ginkgo biloba	Aspergillus niger	Arbour acres	Male	1–21	5	Zhang et al. 2013
Ginkgo biloba	Aspergillus niger and Candida utilis	Arbour acres	Mixed	1–42	6	Zhang et al. 2015
Kaempferia galanga	Lactobacillus casei	Lohmann	Unsexed	8–35	5	Astuti et al. 2022
Panax ginseng	Bifidobacterium sp.	Arbour acres	Male	1–35	6	Ao, Zhou, et al. 2011
Panax ginseng	–	Arbour acres	Mixed	1–28	4	Kim, Lee, et al. 2016
Sauropus androgynus	Lactobacillus sp., Neurospora crassa, and Saccharomyces cerevisiae	–	Unsexed	1–35	4	Santoso et al. 2015b
Sauropus androgynus	Lactobacillus sp., Neurospora crassa, and Saccharomyces cerevisiae	–	Unsexed	1–35	4	Santoso et al. 2015a
Tamarindus indica	Lactobacillus sp. and Saccharomyces cerevisiae	Ross 308	Unsexed	1–28	3	Ferdouse et al. 2020
Zingiber officinale	Actinomycetes sp., Aspergillus sp., and Bacillus subtilis	Marshall chunky	Male	7–42	4	Incharoen et al. 2010
Zingiber officinale	Actinomycetes sp., Aspergillus sp., and Bacillus subtilis	Arbour acres	Male	7–49	5	Khonyoung et al. 2012
Zingiber officinale	Lactobacillus paracasei	Ross 308	Unsexed	1–21	10	Park et al. 2016

N: number of sample replications.

Testing metadata and determining the optimal dosage

The broad summary of the conducted tests revolves around investigating the connections between production factors (such as daily weight gain, daily feed intake, and FCR) in unfermented and fermented herbal products through general and subgroup meta-analyses. Meta-regression was utilised to determine how effective different dosages of fermented herbal products are in improving growth performance, breast meat quality, and intestinal morphology parameters. At the same time, we pinpointed the ideal point for using fermented herbal products using response surface methodology (RSM). A summary of the data tabulation utilised for analysis is provided in Table 2.

Table 2. Summary statistics for the metadata.

Response variable	Control					Fermented Herb
	NDP	Mean	SD	Max	Min	NDP	Mean	SD	Max	Min
Fermented herbal products, g/kg as fed	29	0	0	0	0	39	16.4	37.6	200	0.02
Growth performance
Starter
Body weight, g	17	676	148	956	313	41	717	110	981	332
Average daily gain, g/h/d	17	32.70	11.50	65.40	12.50	38	35	9.24	67.10	13.40
Daily feed intake, g/h/d	17	47.70	11.80	75.60	30.20	38	47.80	10.20	75.50	31.70
Feed conversion ratio	17	1.58	0.26	2.43	1.16	38	1.51	0.24	2.87	1.13
Finisher
Body weight, g	13	2169	748	4255	969	37	2197	518	4491	1058
Average daily gain, g/h/d	13	69.10	15.50	89.50	31.30	37	69.80	12.60	91.90	34.60
Daily feed intake, g/h/d	13	122	26.50	157	67.80	37	126	24.80	165	64.70
Feed conversion ratio	13	1.85	0.32	2.56	1.27	37	1.88	0.34	2.63	1.19
The entirety of the raising phase
Body weight, g	17	2177	690	4255	969	53	2224	497	4491	1058
Average daily gain, g/h/d	17	51.70	11.90	76.10	21.90	53	53.70	9.60	83.40	24
Daily feed intake, g/h/d	17	96.30	25.10	161	49	53	101	23.70	156	51.50
Feed conversion ratio	17	1.84	0.25	2.24	1.22	53	1.86	0.25	2.63	1.16
Mortality, %	4	10.80	9.18	20	0	14	6.61	6.86	18	0
Liveability, %	4	89.20	9.18	100	80	14	93.38	6.86	100	82
Productive efficiency index	16	365	186	792	97.40	48	390	182	807	112
Breast meat
Breast meat characteristics
Breast weight, % of total live weight	6	24.50	6.72	35.70	16.10	21	25.60	4.59	35.50	17.30
Water holding capacity, %	3	59.10	37.70	87.30	16.20	5	39.60	36	85	12.90
Protein, %	3	21.90	3.02	24.70	18.70	8	20.40	2.83	25.90	18.50
Fat, %	3	9.06	14.10	25.30	0.27	8	18.90	11.10	25.40	0.24
pH (45 to 60 min)	8	6.11	0.34	6.79	5.72	20	6.18	0.36	6.9	5.76
Drip loss (24 h), %	4	2.63	1.93	5.46	1.23	14	2.55	1.64	4.99	1.02
Cooking loss, %	5	21.70	4.81	27.50	14.30	17	21.30	4.68	26.70	11.90
Shear force, N	3	37.60	19.70	59	20.30	11	33.80	12.60	52.90	18.30
Breast colours
Lightness	8	49.80	1.90	52.40	47.40	20	49.50	2.55	54.60	45.40
Redness	8	6.85	3.65	13.90	3.24	20	6.92	4.32	14.90	2.98
Yellowness	8	11.10	3.52	16.80	6.61	20	12.40	2.54	17.60	7.51
Breast fatty acids, % of total fatty acid
Myristic acid, C14:0	3	0.76	0.07	0.84	0.70	10	0.63	0.16	0.88	0.31
Palmitic acid, C16:0	3	24.20	1.38	25.50	22.70	10	24.50	1.84	27.10	22.60
Stearic acid, C18:0	3	6.07	0.58	6.65	5.49	10	5.75	0.64	6.49	4.17
Palmitoleic acid, C16:1	2	4.02	1.27	4.91	3.12	4	4.87	1.55	5.69	2.55
Oleic acid, C18:1	3	45.10	6.92	52.70	39.20	10	42.40	4.11	52.50	38.60
Eicosanoic acid, C20:0	2	0.81	0.11	0.89	0.73	4	0.79	0.08	0.92	0.73
Linoleic acid, C18:2	3	13.90	3.96	17	9.44	10	14.30	3.57	19.70	9.32
α-Linolenic acid, C18:3	3	1.40	0.79	2.04	0.51	10	1.17	0.84	2.88	0.46
Arachidonic acid, C20:4	2	2.43	1.12	3.22	1.63	4	3.25	0.89	3.78	1.91
SFA	3	31.10	1.99	33.10	29.10	10	31	1.68	33.30	28.80
MUFA	2	50.90	8.22	56.70	45.10	4	48.10	5.20	55.90	45.20
PUFA	2	18	5.42	21.90	14.20	4	22.10	4.93	25.30	14.90
UFA	3	65.50	6.24	70.90	58.70	10	61.50	7.82	70.80	50.60
UFA vs SFA	3	2.12	0.28	2.44	1.89	10	2	0.32	2.43	1.67
PUFA vs SFA	2	0.57	0.12	0.66	0.48	4	0.74	0.16	0.86	0.51
Intestinal morphology (at finisher phase)
Crypt Depth, µm
Duodenum	4	173	57.40	229	120	8	193	49.90	229	112
Ileum	5	104	55.40	190	53.1	11	107	36.50	174	48.8
Jejunum	5	152	49.30	193	96.6	11	146	38.90	185	89.20
Villus height, µm
Duodenum	6	1348	440	1688	756	17	1627	311	1952	840
Ileum	7	609	180	865	409	20	650	158	879	413
Jejunum	7	894	255	1141	518	20	1088	217	1295	629
Villus height per crypt depth
Duodenum	4	6.82	0.67	7.37	5.88	8	7.88	0.26	8.14	7.30
Ileum	5	6.76	1.89	7.86	3.4	11	7.09	2.2	8.78	3.58
Jejunum	5	5.58	0.43	5.92	5.02	11	6.97	0.35	7.53	6.19
Villus area, mm^2
Duodenum	2	0.39	0.34	0.63	0.146	9	0.32	0.25	0.68	0.13
Ileum	2	0.07	0.05	0.11	0.039	9	0.07	0.04	0.15	0.04
Jejunum	2	0.17	0.12	0.26	0.089	9	0.19	0.12	0.38	0.09
Cell area, µm^2
Duodenum	2	218	64.90	264	172	9	221	70.60	322	158
Ileum	2	113	12.70	122	104	9	118	29.20	162	92.70
Jejunum	2	175	23.50	191	158	9	176	40.20	246	138
Cell mitosis, numbers
Duodenum	2	852	426	1153	550	9	1099	316	1564	660
Ileum	2	346	215	499	194	9	485	159	670	266
Jejunum	2	612	399	894	330	9	792	264	1081	440

MAX: upper limit of the data; MIN: lower limit from dataset; NDP: number of data points; SD: standard deviation; MUFAs: monounsaturated fatty acids; PUFAs: polyunsaturated fatty acids; SFAs: saturated fatty acids; UFAs: unsaturated fatty acids.

Meta-analysis of both overall and subgroup data

In a general context, a standard meta-analysis was used to examine the disparity in distance between unfermented and fermented herbal products, which was established by computing Hedges’ d effect size. According to Marín-Martínez and Sánchez-Meca (2010), this approach was chosen because it can assess the effect of paired treatments and determine the effect size despite variations in sample size, measurement unit, and statistical test findings. The fermented herbal product group was pooled to form the experimental group (E), whereas the unfermented herbal product group was pooled to form the control group (C). The effect size (d) was calculated using the following formula: (1) $$d=\frac{{\mathrm{{\rm X}}}^E-{\mathrm{{\rm X}}}^C}{S_J}$$(1) where X^C represents the control group mean value and X^E represents the experimental group mean value. When the effect size was positive, the measured parameter was greater in the unfermented herbal product group, and vice versa. The small sample size correction factor, denoted by S_J, is as follows: (2) $$S_J=1-\frac{3}{4\left(N^C+N^E-2\right)-1}$$(2)

S denotes the pooled standard deviation, which is formulated as: (3) $$S=\sqrt{\frac{\left(N^E-1\right){\left(S^E\right)}^2+\left(N^C-1\right){\left(S^C\right)}^2}{N^E+N^C-2}\times S_j}$$(3) where S^E denotes the standard deviation of the experimental group, S^C denotes the standard deviation of the control group, N^E denotes the sample size of the experimental group, and N^C denotes the sample size of the control group. The following is a description of Hedges’ d (V_d) variation: (4) $$V_d=\frac{N^C+N^E}{N^C{N}^E}+\frac{d^2}{2\left(N^C+N^E\right)}$$(4)

The cumulative effect size (d₊₊) was computed as follows: (5) $$d_{++}=\frac{{\sum }_{i=1}^n{W}_i{d}_1}{{\sum }_{i=1}^n{W}_i}$$(5) where W_i is the sampling variance’s inverse: $W_i=\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$V_d$}\right..$ The precision of the effect size was reported as a 95% confidence interval (CI), denoted by the term $d_{++}$ ± (1.96 × SE), where SE represents the standard error of the sum effect size. The equations provided above were derived from Sánchez-Meca and Marín-Martínez (2010). If the computed effect size did not match the null effect size, the result was considered to be statistically significant. To uncover publication bias caused by nonsignificant papers that were excluded from the study, a fail-safe number (Nfs) was generated. A robust meta-analysis model was supposed to include Nfs (fail-safe number), which is greater than five times the study effect size used to compute the initial effect size (N) plus 10. Nfs was calculated using the Rosenthal (1979) approach. The N. Cohen benchmarks and the smallest sample size available from the relevant studies were utilised as reference points for effect size estimation. The minor, medium, and high effect sizes were 0.2, 0.5, and 0.8, respectively. All of the aforementioned effect size computations were carried out with OpenMEE 2.0 (Wallace et al. 2017).

The practical application of subgroup meta-analysis resembled that of general meta-analysis in the past, but it entailed utilising categorical factors to categorise and contrast each of these groupings. The categorical factors employed in subgroup meta-analysis included the rearing period (starter, finisher, and total phase), herbals (Alisma canaliculatum, Allium sativum, Astragalus membranaceus, Chenopodium album, Curcuma longa, Piper nigrum, G. biloba, Kaempferia galanga, Panax ginseng, Sauropus androgynus, Tamarindus indica, and Zingiber officinale), and fermenters (Actinomycetes sp., Aspergillus niger, Aspergillus sp., Bacillus subtilis, Bifidobacterium sp., Candida utilis, Enterococcus faecium, Lactobacillus acidophilus, Lactobacillus casei, Lactobacillus paracasei, Lactobacillus plantarum, Lactobacillus sp., Leuconostoc citreum, Neurospora crassa, Saccharomyces cerevisiae, and Weissella koreensis).

Meta-regression

In meta-regression, some principles are similar to those of regular meta-analysis except for the fixed factor being the dosage of fermented herb. The meta-regression equation follows the following formula (St-Pierre 2001; Sauvant et al. 2008): (6) $$Y_{\mathit{\text{ijk}}}=\mu +S_i+{\tau }_j+{S\tau }_{\mathit{\text{ij}}}+{\beta }_1{X}_{\mathit{\text{ij}}}+b_i{X}_{\mathit{\text{ij}}}+{\beta }_2{X^2}_{\mathit{\text{ij}}}+b_i{X^2}_{\mathit{\text{ij}}}+e_{\mathit{\text{ijk}}}$$(6)

$Y_{\mathit{\text{ijk}}}$ represents the dependent variable, $\mu$ represents the overall mean value (intercept value), $S_i$ denotes the random effect of the ith study, assumed to be $\sim N_{\mathit{\text{iid}}} (0,{\sigma }_S^2),$ ${\tau }_j$ represents the fixed effect of the jth of $\tau$ factors, and ${S\tau }_{\mathit{\text{ij}}}$ represents the random interaction effect between the ith and jth dosage of the $\tau$ factor, also assumed to follow a normal distribution with mean 0 and variance ${\sigma }_{S\tau }^2.$ ${\beta }_1$ represents the overall value of the linear regression coefficient for Y in relation to X, serving as a fixed effect or slope; ${\beta }_2$ denotes the general coefficient of the quadratic regression for Y concerning X, functioning as a fixed effect or slope; and $X_{\mathit{\text{ij}}}$ and ${X^2}_{\mathit{\text{ij}}}$ represent the continuous values of the predictor variable in both linear and quadratic forms, respectively. $b_i$ represents the random effect specific to each study on the regression coefficient of Y with respect to X, assumed to be $\sim N_{\mathit{\text{iid}}} (0,{\sigma }_b^2).$ Finally, $e_{\mathit{\text{ijk}}}$ represents the residual value arising from unpredictable error.

The validation test was carried out utilising the root mean square error (RMSE) and coefficient of determination (R²) as metrics. The following equation represents the RMSE and R². (7) $$\mathit{\text{RMSE}}=\sqrt{\frac{\sum {(O-P)}^2}{\mathit{\text{NDP}}}}$$(7) (8) $$R^2=\frac{({{\sigma }^2}_f+\sum \left({{\sigma }^2}_l\right))}{({{\sigma }^2}_f+\sum ({{\sigma }^2}_l)+{{\sigma }^2}_e+{{\sigma }^2}_d)}$$(8)

In this scenario, $O$ represents the actual value, $P$ represents the estimated value, $\mathit{\text{NDP}}$ denotes the number of data points, ${{\sigma }^2}_f$ represents the variant of a fixed factor, $\sum \left({{\sigma }^2}_l\right)$ is the sum of component variances, ${{\sigma }^2}_e$ signifies the variance attributed to predictor dispersion, and ${{\sigma }^2}_d$ characterises the specific distribution of the variance. All meta-regression analyses were carried out using R version 4.

Determining the optimal point

Response surface methodology was employed to determine the ideal concentration of fermented herbal products. This approach involved four factors: a denoting the dosage of fermented herbal products (in g/kg as fed), b representing metabolisable energy (ME, in kcal/kg), c standing for crude protein (CP, in %), and d indicating the treatment type (coded as 1 for control and 2 for fermented herbal products). Four responses were chosen for assessment: response 1 for final body weight (g), response 2 for average daily gain (g/h/d), response 3 for daily feed intake (g/h/d), and response 4 for feed conversion ratio. These responses reflect the performance parameters throughout the entire treatment period. It is important to note that not all responses were suitable for analysis using RSM; only the mentioned responses were considered. The optimisation process aimed to determine the optimal values for each factor and response. The entire RSM and optimisation procedures were executed using Design Expert version 13.

Results and discussion

Growth performance of broiler chickens

Tables 3–5 display the cumulative effect size (d++) along with Figure 2, which illustrates the FCR for all phases. The meta-analysis revealed notable differences in the performance of broiler chickens under treatments employing unfermented and fermented herbal products, specifically in terms of daily weight gain, daily feed intake, and FCR. Notably, the DWG and FCR parameters exhibited substantial effect sizes, while the DFI demonstrated a moderate effect. In the context of the broiler DWG (Table 3), Marshall Chunky exhibited the greatest response to treatment, followed by Ross, Ross 308, and Avian strains. Throughout all maintenance periods (starter and finisher) and for all genders of broilers, there was a moderate to large effect resulting from the use of fermented herbal products. Tamarindus  indica had the greatest effect among the herbs, followed by Alisma canaliculatum and Chenopodium album. Moreover, the combination of Lactobacillus sp. and Saccharomyces cerevisiae had the greatest effect, followed by Lactobacillus paracasei compared with the other fermenters. The daily feed consumption of Arbor acres and Lohmann was responsive to treatment (effect size > 0.8; Table 4). Both starter and overall treatments exhibit a substantial effect size, as does the unsexed category. The effects of the herbal products from Kaempferia galanga and the fermentation process involving Aspergillus niger and candida utilis were greater than those of their nonfermented counterparts. Examining the FCR parameters (Table 5), Marshall Chunky and Ross were the two strains with the greatest effect sizes. Furthermore, all maintenance phases and genders of the broilers displayed a high effect size. Additionally, the herbal product Zingiber. officinale and the combination of Actinomycetes sp., Aspergillus sp., and Bacillus subtilis also had strong effects.

Figure 2. Broiler chickens feed conversion ratio (FCR) under the influence of fermented herbal products during the starter, finisher, and overall phases. A significant difference (p < 0.05) is observed only in the FCR for the total phase, with an R-squared value of ∼0.97, while other phases show no significant differences.

Table 3. Impact of incorporating fermented herbal products and its subsequent subgroup meta-analysis on broiler chickens daily weight gain.

	Study estimator (95% C.I.)					Heterogeneity
	d++	Lower	Upper	SE	p-Value	I²	p-Value
Daily weight gain, g/h/d
Overall	1.010	0.792	1.230	0.113	<0.001	71	<0.001
Broilers
Arbour acres	0.978	0.721	1.240	0.131	<0.001	72	<0.001
Avian	1.240	0.518	1.950	0.366	0.001	0	0.612
Lohmann	0.898	−0.409	2.210	0.667	0.178	63.5	0.065
Marshall chunky	5.740	−0.616	12.100	3.250	0.077	90.5	<0.001
Ross	1.110	0.732	1.490	0.194	<0.001	0	0.524
Ross 308	0.913	0.092	1.730	0.418	0.029	77.1	<0.001
Period of rearing
Starter	1.100	0.702	1.490	0.202	<0.001	68.2	<0.001
Finisher	0.591	0.263	0.918	0.167	<0.001	58.8	<0.001
Total phase	1.280	0.891	1.670	0.199	<0.001	76.7	<0.001
Sex
Male	0.514	0.203	0.825	0.159	0.001	60.5	<0.001
Mixed	0.958	0.531	1.390	0.218	<0.001	9.2	0.358
Unsexed	1.350	1.020	1.670	0.164	<0.001	74.9	<0.001
Herbs
Alisma canaliculatum	1.770	0.771	2.770	0.510	0.001	25.1	0.263
Allium sativum	0.432	0.124	0.741	0.157	0.006	3.3	0.415
Astragalus membranaceus	1.240	0.518	1.950	0.366	0.001	0	0.612
Chenopodium album	1.710	0.685	2.730	0.523	0.001	74.9	<0.001
Ginkgo biloba	1.180	0.854	1.500	0.165	<0.001	73.1	<0.001
Kaempferia galanga	0.898	−0.409	2.210	0.667	0.178	63.5	0.065
Panax ginseng	−0.441	−0.823	−0.058	0.195	0.024	0	0.999
Tamarindus indica	8.990	0.972	17	4.090	0.028	84.4	0.002
Zingiber officinale	1.270	0.619	1.920	0.330	<0.001	74.9	<0.001
Fermenter
Actinomycetes sp., Aspergillus sp., and Bacillus subtilis	1.125	0.449	1.801	0.345	0.001	72.7	<0.001
Aspergillus niger	0.999	0.564	1.434	0.222	<0.001	62.5	<0.001
Aspergillus niger and Candida utilis	1.501	0.870	2.133	0.322	<0.001	82.2	<0.001
Bacillus subtilis, Enterococcus faecium, Lactobacillus acidophilus, and Saccharomyces cerevisiae	1.771	0.771	2.771	0.510	0.001	25.1	0.263
Bacillus subtilis, Lactobacillus plantarum, and Saccharomyces cerevisiae	1.709	0.685	2.733	0.523	0.001	75.9	<0.001
Bifidobacterium sp.	−0.441	−0.823	−0.058	0.195	0.024	0	0.998
Candida utilis	0.406	−0.258	1.070	0.339	0.231	0	0.641
Lactobacillus casei	0.898	−0.409	2.210	0.667	0.178	63.5	0.065
Lactobacillus paracasei	2.280	0.284	4.300	1.020	0.025	82.6	0.017
Lactobacillus plantarum	1.030	0.667	1.400	0.187	<0.001	0	0.908
Lactobacillus sp. and Saccharomyces cerevisiae	8.990	0.972	17	4.090	0.028	84.4	0.002
Leuconostoc citreum	0.117	−0.354	0.589	0.241	0.626	18.8	0.278
Weissella koreensis	0.761	0.329	1.190	0.220	0.001	0	0.911

d++: sum of effect size; I²: heterogeneity level between studies; Lower/Upper: the 95% CI for d++ provides a range capturing the likely true population parameter, enhancing result reliability and precision; SE: standard error.

Table 4. Cumulative effect size resulting from the integration of fermented herbal products and subsequent subgroup meta-analysis on broiler chickens feed consumption.

	Study estimator (95% C.I.)					Heterogeneity
	d++	Lower	Upper	SE	p-Value	I²	p-Value
Daily feed intake, g/h/d
Overall	0.483	0.200	0.766	0.144	0.001	81.8	<0.001
Broilers
Arbour acres	0.767	0.428	1.110	0.173	<0.001	82.5	<0.001
Avian	1.170	−1.030	3.360	1.120	0.297	87	<0.001
Lohmann	1.440	0.634	2.250	0.412	0.001	0	0.759
Marshall chunky	−2.280	−3.830	−0.741	0.787	0.004	50.9	0.130
Ross	−0.836	−1.820	0.149	0.503	0.096	83	<0.001
Ross 308	−0.221	−0.731	0.288	0.260	0.394	54.4	0.008
Period of rearing
Starter	0.660	0.256	1.060	0.206	0.001	71.2	<0.001
Finisher	−0.206	−0.633	0.220	0.218	0.343	74.8	<0.001
Total	0.929	0.385	1.470	0.277	0.001	86.8	<0.001
Sex
Male	0.045	−0.323	0.414	0.188	0.809	70.7	<0.001
Mixed	−0.283	−0.668	0.103	0.197	0.151	0	0.548
Unsexed	0.877	0.437	1.320	0.224	<0.001	85.9	<0.001
Herbs
Alisma canaliculatum	0.733	−0.473	1.940	0.615	0.234	59	0.087
Allium sativum	−0.051	−0.454	0.352	0.206	0.804	42.6	0.029
Astragalus membranaceus	1.170	−1.030	3.360	1.120	0.297	87	<0.001
Chenopodium album	0.375	−0.630	1.380	0.513	0.464	78.3	<0.001
Ginkgo biloba	0.744	0.251	1.240	0.252	0.003	87.1	<0.001
Kaempferia galanga	1.440	0.634	2.250	0.412	<0.001	0	0.759
Panax ginseng	−0.276	−0.740	0.188	0.237	0.244	29.5	0.183
Tamarindus indica	0.071	−0.854	0.996	0.472	0.88	0	0.956
Zingiber officinale	0.604	−0.233	1.440	0.427	0.157	84.1	<0.001
Fermenter
Actinomycetes sp., Aspergillus sp., and Bacillus subtilis	0.561	−0.397	1.520	0.488	0.251	84.6	<0.001
Aspergillus niger	−0.624	−1.180	−0.067	0.284	0.028	6.61	<0.001
Aspergillus niger and Candida utilis	2.590	1.940	3.240	0.331	<0.001	78.8	<0.001
Bacillus subtilis, Enterococcus faecium, Lactobacillus acidophilus, and Saccharomyces cerevisiae	0.733	−0.473	1.940	0.615	0.234	59	0.087
Bacillus subtilis, Lactobacillus plantarum, and Saccharomyces cerevisiae	0.375	−0.630	1.380	0.513	0.464	78.3	<0.001
Bifidobacterium sp.	−0.276	−0.740	0.188	0.237	0.244	29.5	0.183
Candida utilis	−0.051	−0.704	0.603	0.333	0.879	0	0.999
Lactobacillus casei	1.440	0.634	2.250	0.412	<0.001	0	0.759
Lactobacillus paracasei	1.250	0.575	1.930	0.347	<0.001	0	0.555
Lactobacillus plantarum	−0.748	−1.820	0.321	0.545	0.17	85.6	<0.001
Lactobacillus sp. and Saccharomyces cerevisiae	0.071	−0.854	0.996	0.472	0.88	0	0.956
Leuconostoc citreum	−0.696	−1.130	−0.264	0.220	0.002	0	0.721
Weissella koreensis	0.584	0.153	1.020	0.220	0.008	0	0.496

d++: sum of effect size; I²: heterogeneity level between studies; Lower/Upper: the 95% CI for d++ provides a range capturing the likely true population parameter, enhancing result reliability and precision; SE: standard error.

Table 5. The sum of the effect sizes resulting from the integration of fermented herbal products and subsequent subcategory meta-analyses on broiler chickens feed conversion.

	Study estimator (95% C.I.)					Heterogeneity
	d++	Lower	Upper	SE	p-Value	I^2	p-Value
FCR
Overall	−1.170	−1.520	−0.834	0.173	<0.001	84.5	<0.001
Broilers chickens
Arbour acres	−1.020	−1.400	−0.647	0.192	<0.001	83.9	<0.001
Avian	0.516	−1.250	2.280	0.899	0.566	82.9	0.003
Lohmann	7.780	−31.700	47.300	20.200	0.700	95.9	<0.001
Marshall chunky	−14.300	−21.700	−6.940	3.750	<0.001	65.7	0.054
Ross	−3.140	−4.470	−1.820	0.676	<0.001	83.9	<0.001
Ross 308	−0.779	−1.250	−0.313	0.238	0.001	41.7	0.051
Period of rearing
Starter	−1.157	−1.602	−0.713	0.227	<0.001	73.4	<0.001
Finisher	−1.351	−2.023	−0.680	0.343	<0.001	86	<0.001
Total	−1.032	−1.645	−0.418	0.313	0.001	87.3	<0.001
Sex
Male	−0.886	−1.500	−0.270	0.314	0.005	85	<0.001
Mixed	−1.020	−1.460	−0.573	0.226	<0.001	14.6	0.312
Unsexed	−1.380	−1.840	−0.921	0.233	<0.001	85.2	<0.001
Herbs
Alisma canaliculatum	−2.144	−3.942	−0.347	0.917	0.019	72.3	0.027
Allium sativum	−0.329	−0.629	−0.030	0.153	0.031	0	0.709
Astragalus membranaceus	0.516	−1.246	2.278	0.899	0.566	81.9	0.003
Chenopodium album	−1.406	−2.155	−0.657	0.382	<0.001	58.2	0.014
Ginkgo biloba	−1.393	−1.845	−0.940	0.231	<0.001	83.8	<0.001
Kaempferia galanga	7.775	−31.711	47.260	20.146	0.700	95.9	<0.001
Panax ginseng	0.138	−0.656	0.933	0.405	0.733	73.6	<0.001
Tamarindus indica	−2.386	−3.730	−1.043	0.686	0.001	13.1	0.316
Zingiber officinale	−2.793	−4.732	−0.853	0.989	0.005	92.3	<0.001
Fermenter
Actinomycetes sp., Aspergillus sp., and Bacillus subtilis	−2.920	−5.530	−0.305	1.330	0.029	92.9	<0.001
Aspergillus niger	−1.710	−2.150	−1.280	0.222	<0.001	55.9	<0.001
Aspergillus niger and Candida utilis	−0.617	−1.390	0.157	0.395	0.118	87.4	<0.001
Bacillus subtilis, Enterococcus faecium, Lactobacillus acidophilus, and Saccharomyces cerevisiae	−2.140	−3.940	−0.347	0.917	0.019	72.3	0.027
Bacillus subtilis, Lactobacillus plantarum, and Saccharomyces cerevisiae	−1.410	−2.160	−0.657	0.382	<0.001	58.2	0.014
Bifidobacterium sp.	0.138	−0.656	0.933	0.405	0.733	72.6	<0.001
Candida utilis	−0.470	−1.140	0.197	0.340	0.168	0	0.588
Lactobacillus casei	7.780	−31.70	47.300	20.200	0.700	95.9	<0.001
Lactobacillus paracasei	−0.968	−2.70	0.766	0.885	0.274	84.4	0.011
Lactobacillus plantarum	−2.200	−3.750	−0.645	0.793	0.006	90.7	<0.001
Lactobacillus sp. and Saccharomyces cerevisiae	−2.390	−3.730	−1.043	0.686	0.001	13.1	0.316
Leuconostoc citreum	−0.460	−0.885	−0.035	0.217	0.034	0	0.708
Weissella koreensis	−0.201	−0.623	0.222	0.216	0.353	0	0.512

d++: sum of effect size; I²: heterogeneity level between studies; Lower/Upper: the 95% CI for d++ provides a range capturing the likely true population parameter, enhancing result reliability and precision; SE: standard error.

The meta-regression findings indicated statistically significant increases (p < 0.05) in the production parameters, including BW and DWG, during the starter phase, with changes in slope values of ∼4.59 g and 0.0568 g/h/d, respectively (see Table 6). In the finisher phase, only BW significantly changed (p = 0.017) to ∼8.65 g for a single dose (g/kg as fed) of fermented herbal products. Moreover, over the entire treatment period, production parameters, such as BW, BWG, and FCR were significantly improved (p < 0.05) due to the use of fermented herbal products. Additional parameters, including mortality, viability, and PEI, also showed noteworthy increases (p < 0.05). Improvements in chicken meat production are credited to favourable reactions to fermented herbal products. Fermentation is widely recognised to be linked to the increased presence of beneficial compounds in herbs, which serve as antibacterial, antioxidant, and immune system regulators for animals (Hussain et al. 2016). Indeed, fermentation may also address a major problem in the use of herbal products related to the limited bioavailability and bioactivity of the active components in the herbal products. In this respect, fermentation could enhance the bioavailability and bioactivity of the active compounds in the herbs (Zhang et al. 2023). On this basis, fermentation may improve the effectiveness of herbal products as alternative growth promoters for broilers.

Table 6. Meta-regression analysis of the influence of including fermented herbal products on broiler chickens performance, breast meat quality, and gastrointestinal morphology.

Response variable	M	Intercept ($\mu$)		Slope (${\beta }_{1\mathit{\text{ or }}2}$)		p-Value	RMSE	R^2
		Value	SE	Value	SE
Growth performance
Starter
Body weight, g	L	684	34.30	4.5900	1.1700	<0.001	15.30	0.98
Daily weight gain, g/h/d	L	32.90	2.80	0.2730	0.0568	<0.001	0.70	0.97
Daily feed intake, g/h/d	L	47.90	2.72	0.2240	0.0912	0.019	1.12	0.95
Feed conversion ratio	L	1.57	0.07	−0.0054	0.0043	0.223	0.05	0.96
Finisher
Body weight, g	L	2194	198	8.6500	3.4400	0.017	50.40	0.98
Daily weight gain, g/h/d	L	69.10	4.19	0.0608	0.1450	0.677	2.13	0.97
Daily feed intake, g/h/d	L	122	7.11	0.0012	0.1950	0.995	2.87	0.98
Feed conversion ratio	L	1.86	0.10	−0.0026	0.0045	0.565	0.07	0.95
The entirety of the raising phase
Body weight, g	Q	2191	163	19.8000	4.6000	<0.001	1.55	0.97
				−1.1000	0.3060	0.001
Daily weight gain, g/h/d	Q	52.10	2.86	0.4260	0.1410	0.004	1.45	0.98
				−0.0241	0.0093	0.013
Daily feed intake, g/h/d	L	98.90	6.14	−0.1210	0.1190	0.315	2.45	0.97
Feed conversion ratio	L	1.86	0.07	−0.0043	0.0021	0.046	0.04	0.97
Mortality, %	L	9.52	2.92	−0.0355	0.0700	0.630	3	0.80
Liveability, %	L	90.50	2.90	0.0360	0.0714	0.626	2.95	0.75
Productive efficiency index	L	363	46.10	1.7300	0.7360	0.023	14.80	0.97
Breast meat (at finisher phase)
Breast meat characteristics
Breast weight, % of total live weight	L	24.40	2.14	0.1370	0.0607	0.036	0.66	0.98
Water holding capacity, %	L	60.10	4.70	0.2540	0.2670	0.385	0.67	0.98
Protein, %	L	22.40	1.92	0.0031	0.0136	0.823	0.51	0.96
Fat, %	L	9.12	8.15	−0.0138	0.0080	0.128	0.30	0.98
pH (45–60 min)	L	6.10	0.12	0.0058	0.0055	0.304	0.07	0.94
Drip loss (24 h), %	L	2.56	0.84	−0.0792	0.0248	0.007	0.23	0.97
Cooking loss, %	L	21.60	2.28	−0.1290	0.0413	0.006	0.39	0.98
Shear force, N	L	36	9.77	−0.2860	0.2180	0.214	1.91	0.98
Breast colours
Lightness	L	49.70	0.83	0.0435	0.0883	0.627	1.18	0.68
Redness	L	6.92	1.38	−0.0162	0.0304	0.602	0.39	0.99
Yellowness	L	11.10	1.10	0.0779	0.0528	0.156	0.68	0.93
Breast fatty acids, % of total fatty acid
Myristic acid, C14:0	L	0.75	0.05	−0.0045	0.0019	0.039	0.12	–
Palmitic acid, C16:0	L	23.80	0.80	0.0191	0.0265	0.491	1.30	0.30
Stearic acid, C18:0	L	5.85	0.26	−0.0011	0.0094	0.906	0.60	–
Palmitoleic acid, C16:1	L	3.98	1.26	0.0447	0.0473	0.414	0.28	0.95
Oleic acid, C18:1	L	45.30	3.96	−0.0425	0.0236	0.106	0.93	0.97
Eicosanoic acid, C20:0	L	0.81	0.08	0.00495	0.0013	0.037	0.01	0.95
Linoleic acid, C18:2	L	14.40	2.71	−0.0525	0.0237	0.053	0.94	0.95
α-Linolenic acid, C18:3	L	1.59	0.56	0.0000	0.0046	0.998	0.18	0.95
Arachidonic acid, C20:4	L	2.45	0.82	0.0578	0.0086	0.007	0.05	0.97
SFA	L	30.40	0.64	0.0296	0.0238	0.239	1.50	–
MUFA	L	51	5.40	−0.0333	0.0687	0.662	0.40	0.96
PUFA	L	17.90	4.21	0.3360	0.0783	0.023	0.45	0.98
UFA	L	66.70	3.79	−0.1020	0.0430	0.041	1.71	0.91
UFA vs. SFA	L	2.19	0.17	−0.0032	0.0024	0.216	0.10	0.86
PUFA vs. SFA	L	0.57	0.12	0.0188	0.0047	0.029	0.03	0.95
Intestinal morphology (at finisher phase)
Villus height, µm
Duodenum	L	173	29.60	−1.0900	0.3260	0.013	2.22	0.96
Ileum	L	97.80	20.60	−0.8280	1.4000	0.567	10.90	0.91
Jejunum	L	143	20.20	−2.0600	1.8300	0.286	14.40	0.85
Crypt depth, µm
Duodenum	L	1411	162	10.1000	3.6900	0.016	58.50	0.97
Ileum	L	593	60.80	4.5400	2.5300	0.089	41.80	0.92
Jejunum	L	942	92.70	12	3.7800	0.005	62.20	0.92
Villus height per crypt depth
Duodenum	L	7.06	0.34	0.1290	0.0393	0.012	0.28	0.81
Ileum	L	6.95	0.90	0.0914	0.0329	0.019	0.25	0.98
Jejunum	L	6.13	0.20	0.1570	0.0504	0.008	0.57	–
Villus area, mm^2
Duodenum	L	0.31	0.16	0.0019	0.0011	0.135	0.02	0.98
Ileum	L	0.06	0.03	0.0018	0.0003	0.002	0.01	0.98
Jejunum	L	0.15	0.07	0.0047	0.0012	0.006	0.02	0.97
Cell area, µm^2
Duodenum	L	203	40.90	2.0400	0.8510	0.048	11.80	0.96
Ileum	L	108	14.90	1.2900	0.6240	0.079	8.75	0.84
Jejunum	L	169	22.60	0.747	0.8930	0.430	12.50	0.85
Cell mitosis, numbers
Duodenum	L	1016	217	8.2500	7.5400	0.310	106	0.88
Ileum	L	455	115	3.1000	2.2200	0.205	30.90	0.96
Jejunum	L	735	186	5.6400	3.5100	0.152	48.80	0.96

L: linear; M: model; MUFAs: monounsaturated fatty acids; PUFAs: polyunsaturated fatty acids; Q: quadratic; R²: the proportion of variation in the dependent variable that can be predicted by the independent variable; RMSE: root mean square error; SE: standard error; SFA: saturated fatty acids; UFAs: unsaturated fatty acids.

The improvement in the growth performance of broilers administered with fermented herbal products can be attributed to the improved digestibility of feed in the respective chickens (Xie et al. 2021). Indeed, the improved feed digestibility of broilers administered with fermented herbal products could be due to improved gastrointestinal functions and increased digestive enzyme activity. Compared to unfermented herbal products, fermented herbal products contain more polyphenols that can stimulate bile production and digestive enzyme activity, resulting in improved feed digestibility (Xie et al. 2021). Moreover, fermentation may eliminate the content of antinutritional factors (Emenalom et al. 2021; Xie et al. 2021) while increased the contents of polysaccharides and organic acids in herbs (Zhang et al. 2013; Qiao et al. 2018). The increase in polysaccharides (functions as prebiotics) and organic acids may contribute to improved intestinal functions in feed digestion and absorption (Qiao et al. 2018; Lv et al. 2023).

One of the problems related to the use of herbal products as additives for broiler chickens is their bitter taste and pungent smell, which can result in a decrease in feed palatability (Sugiharto 2021). This meta-analysis indicated that fermented herbal products, particularly those containing K. galanga and L. casei fermenter, can notably enhance the dietary consumption of broiler chickens (Table 4). Several factors can explain the increase in feed intake following the administration of fermented herbal products. Xiang et al. (2019) reported that compared to unfermented products, fermented products have better palatability, smell, and taste and contain fewer antinutrient compounds. This condition results in increased feed consumption in broilers. Furthermore, improvements in gastrointestinal morphology, function, and digestive activity as a result of fermented herbal product treatment can positively influence feed intake in broiler chickens (Zhang et al. 2013). In the context of FCR, the use of fermented herbal products led to improved FCR in broilers, as observed in this meta-analysis. Examples of such herbal products include Z. officinale and the fermenting agent Actinomycetes sp., Aspergillus sp., and B. subtilis, as indicated in Table 5. It seems that the better digestibility of feed due to fermented herbal products compared to that of unfermented herbal products may be associated with the improved FCR of broilers (Park et al. 2016; Ding et al. 2021).

Breast meat of broiler chickens

The outcomes of the meta-regression strongly support the assertion that the use of fermented herbal products significantly influences the breast meat quality of broiler chickens, as indicated by a study estimator within 95% confidence intervals. As illustrated in Table 6, there was a significant increase (p < 0.05) in breast weight, as well as in characteristics, such as cooking loss and drip loss. Conversely, the colour parameter remained unchanged due to the fermented herbal product treatments. The utilisation of fermented Chinese chive and fermented garlic powder, as observed in both Ao, Zhou, et al. (2011) and Lee et al. (2022), does not affect the colour of breast meat in broiler chickens. On the contrary, consumer preference for visual parameters, such as colour, plays a crucial role in determining the quality of chicken meat. In addition to visual aspects, characteristics, such as cooking loss and drip loss also contribute to the quality of chicken breast meat. Experimental findings suggest that cooking loss and drip loss (at 24 and 48 h) tend to decrease due to treatment with G. biloba fermented with A. niger and C. utilis (Niu et al. 2017).

There was a notable increase (p < 0.05) in the content of significant fatty acids, including myristic acid, eicosanoids acid, and arachidonic acid. Similarly, the levels of polyunsaturated and unsaturated fatty acids significantly increased (p < 0.05), followed by a significant increase (p < 0.05) in the PUFA vs SFA ratio due to the treatment (Figure 3). With respect to fatty acids, unsaturated fat parameters, such as arachidonic acid increased. In contrast, the fermentation process involving A. canaliculatum was observed to result in a reduction in arachidonic acid concentrations, differing from the findings of a previous meta-analysis (Hossain and Yang 2014). This is presumed to be a result of differences in the fatty acid composition of the fermented herbal products used as a feed supplement.

Figure 3. The analysis pertains to the composition of saturated fatty acids (SFAs) and polyunsaturated fatty acids (PUFAs) and the ratio between them in broiler chickens breast meat. The p values for both the PUFA parameters and the ratio of PUFA to SFA were <0.05, with corresponding R² values of ∼0.91 and 0.95, respectively.

The high PUFA content in chicken meat is attributed to energy conversion from feed facilitated by probiotics, such as Saccharomyces cerevisiae, which increases the PUFA dosage in breast meat. Subsequently, the probiotic demonstrates tolerance to bile salt hydrolase activity, which beneficially impacts nutrient absorption in the process (Hossain et al. 2012). As previously stated, the use of fermented herbal products has been shown to increase the PUFA content in breast meat. This finding is consistent with previous research conducted by Mazur-Kuśnirek et al. (2019), who also reported higher dosages of PUFAs in individuals treated with polyphenols derived from herbals. Similarly, Sosnówka-Czajka et al. (2017) observed elevated PUFA content in the muscle tissue of birds. The active compounds found in fermented herbal products, including polyphenols, such as genistein and hesperidin, thymol, and other essential components, have been shown to improve the omega-3 to omega-6 ratio and increase the PUFA dosage in eggs, breast meat, and back meat in pigs, as supported by recent studies (Bozkurt et al. 2013; Starčević et al. 2015; Mahfuz et al. 2021). Furthermore, as noted in the review by Jachimowicz et al. (2022), these active substances from herbs function as antioxidants, reducing the oxidation of unsaturated fatty acids and thereby increasing the PUFA content in breast muscles. Additionally, it is well established, as mentioned by Li and Liu (2012), that incorporating exogenous antioxidants into animal diets can enhance the stability of meat lipids. These antioxidants play a crucial role in mitigating the negative effects of oxidative stress on muscle tissue, particularly when exposed to factors, such as heat, as observed in several studies (Amaral et al. 2018; Domínguez et al. 2019; Jachimowicz et al. 2022). In a broader context, strategies aimed at minimising lipid oxidation in meat generally involve modifying the lipid composition of animal feeds and supplementing them with antioxidants, such as tocopherol from plants, as discussed in studies by Estévez (2015) and Niu et al. (2022). Once again, this finding serves as a positive reference regarding herbal fermentation products, wherein the composition of compounds, notably antioxidants, increases following the ensiling process (Zhang et al. 2013, 2020).

Fermentation is a useful tool for promoting health in broiler chickens. Microbial fermentation, according to this meta-analysis, can enrich the quality of breast meat from chickens. Various fermented herbal products, including Chinese chives, garlic powder, ginger, and G. biloba, enhance the quality of breast meat (Ao, Yoo, et al. 2011; Ao, Zhou, et al. 2011; Cao et al. 2012; Lee et al. 2022). Reportedly, G. biloba consists of flavonoid aglycones, which are easily and rapidly absorbed in the intestine after fermentation (Cao et al. 2012; Niu et al. 2017; Niu, Wan, et al. 2019).

Intestinal morphology of broiler chickens

There is a potential to decrease the depth of crypts in the duodenum of both chicks and laying hens by incorporating specific activating compounds extracted from herbal plants. It is crucial to acknowledge that conflicting results have been reported in certain studies, as evident in the introduction of wine pomace polyphenols to piglets (Sehm et al. 2007; Viveros et al. 2011). Moreover, the incorporation of polyphenols or other flavonoids from herbs, particularly through fermentation, not only increases villi height but also improves villi height (Cao et al. 2012; Zhang et al. 2012, 2013; Iqbal et al. 2020). Consequently, there tends to be an increase in the ratio between villus height and crypt depth, a consistent observation highlighted in this meta-regression. Notably, significant increases (p < 0.05) in crypt depth, villus height, and their ratio in the duodenum were detected. Additionally, the villus height in the ileum and jejunum also increased significantly (p < 0.05). Similarly, other parameters, such as the villus area and cell area, showed substantial increases (p < 0.05) in the duodenum, jejunum, and ileum. Similar to prior treatments with fermented herbal products that enhance villus height, crypt depth, villus area, and cell area, these findings underscore the positive impact of such treatments (Incharoen et al. 2010; Zhang et al. 2013, 2015; Lv et al. 2023).

As noted, the height of the villi in the intestines serves as a valuable indicator of the absorptive surface area within the digestive system (Güleş and Yildiz 2021). When these villi are tall and well-developed, they significantly expand the surface available for absorbing nutrients (Ningsih et al. 2019). This enhancement in nutrient absorption capacity results in the more efficient utilisation of nutrients obtained from their diet, ultimately facilitating healthier growth and increased production in broiler chickens. Moreover, robust villi are closely linked to improved nutrient absorption, a critical factor in achieving optimal weight gain and enhancing feed conversion efficiency (Sugiharto and Ayasan 2023). In summary, villus height and crypt depth play pivotal roles in evaluating the gut health of broiler chickens. Striking the right balance between these two factors is indispensable for maximising nutrient absorption, stimulating growth, and ensuring the overall well-being of birds. In essence, when the digestive system is in good health, broiler chickens are poised for better performance and overall thriving.

Optimising the addition of fermented herbal products

Optimising the administration of fermented herbal products is a crucial consideration because it involves careful dosage calculations. The results obtained from RSM indicate that to achieve the lowest FCR, feed supplemented with fermented herbal products should contain a minimum of 3274 kcal/kg of metabolisable energy and 21.9% protein. The optimal dosage of fermented herbal products was determined to be 26.3 g/kg as fed (at the total phase of broiler production). RSM also estimated the BWG and FCR at 67.5 g/h/d and 1.87, respectively (Figure 4). It is important to note that the specificity of the fermented herbal products used can significantly affect the accuracy of the optimal dosage, as the scope of this study is limited in terms of the types of herbs or specific compounds investigated by individual researchers. Therefore, optimisation strategies should be broadly applied to ‘fermented herbal products’ to enhance the productivity and quality of breast meat in broiler chickens (Figure 5).

Figure 4. The optimisation value is determined through the application of response surface methodology (RSM). Factor a represents the quantity of fermented herbal products (g/kg as fed), factor b denotes metabolisable energy (ME; kcal/kg), factor c indicates crude protein (CP; %), and factor d signifies the treatment of control versus fermented herbal products. Response 1 corresponds to the final body weight, response 2 to the average daily gain, response 3 to daily feed intake, and response 4 to the feed conversion ratio (FCR). Within this framework, (a) identifies the optimal state resulting from the interactions between factors and responses, while (b) specifies the achieved optimal value. RSM optimisation yielded optimal values of 26.3 g/kg for fermented herbal products, 3274 kcal/kg for ME, and 21.9% for CP. These values correspond to predicted output values of 2804 g for body weight, 67.5 g/h/d for average daily gain, 136 g/h/d for daily feed intake, and an FCR of ∼1.87 in the total phase of broiler chickens production.

Figure 5. Fermented herbal products were incorporated into broiler chicken diets. By utilising a blend of various herbs, a fermented herbal product enriched with a combination of beneficial microbes is employed to improve broiler performance, health, and meat quality.

Conclusions

Feeding broiler chickens with fermented herbal products has positive effects on growth performance, breast meat quality, and intestinal morphology. This improvement is particularly relevant in terms of enhancing FCR, PUFA, and the ratio of villus height to crypt depth. Hence, it is recommended that broiler farmers ferment herbs before adding them to poultry feed, using a recommended quantity of ∼26.3 g/kg as fed.

Ethical approval

This research does not require ethical approval because of the methodology of the meta-analysis, which relies on secondary data. However, all references mention the use of appropriate ethical clearance.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The information provided in this article is accessible without restrictions, as it has been sourced from previously published articles that are duly referenced.