Effect of Feed Additives as Alternatives to In-feed Antimicrobials on Production Performance and Intestinal Clostridium perfringens Counts in Broiler Chickens

Simple Summary For many years, antibiotics were added to chicken feed to prevent disease and promote growth. This practice has been banned or voluntarily abolished in many countries. However, most countries still allow the use of in-feed ionophorous coccidiostats, which are drugs that possess both antiparasitic and antibacterial properties. Concerns related to antimicrobial resistance have led to increased focus on broiler chickens raised without the use of any antimicrobial agents, and the interest in non-antibiotic feed additives with beneficial effects on gastrointestinal health and productivity is growing. In this study, feed additives with active components belonging to the product classes probiotics, prebiotics, phytogenics and/or organic acids were assessed for their effect on intestinal health and production performance in broiler chickens. Collectively, the group of non-antibiotic feed additives improved gut health and performance, but not to the same extent as the ionophorous coccidiostat narasin. Probiotics and prebiotics had the overall best performances during coccidia challenge, phytogenics improved overall feed conversion and reduced counts of the intestinal bacterium Clostridium perfringens, and organic acids increased weight gain independent of age. This study provides comparable and unbiased results from testing of alternatives to antibiotics in a uniform experimental model highly relevant to commercial conditions. Abstract Numerous non-antibiotic feed additives (alternatives to antibiotics, ATAs) have been marketed, but few have been evaluated under uniform testing conditions modelling commercial flocks. We compared 24 ATA treatments and the ionophorous coccidiostat narasin against a diet without any feed additives. Feed conversion ratio and body weight gain were registered from day 0 to 28 in Ross 308 chickens housed on litter floor. The chickens were challenged with Eimeria spp., and cecal Clostridium perfringens (CP) counts were investigated. Active components from all ATA classes had a positive impact on intestinal health or production performance. Whereas narasin had a strong CP-reducing effect in combination with performance-promoting impact, only two ATA treatments achieved significantly beneficial effects on CP counts as well as feed conversion during the time span following Eimeria challenge. Active components present in these two treatments include a Bacillus subtilis probiotic strain, short- and medium-chain fatty acids and Saccharomyces cerevisiae components. Different ATA classes had beneficial impact during distinct rearing phases and on specific performance targets, suggesting that optimizing combinations and use of active components can make ATAs even more useful tools in broiler rearing without the use of in-feed antimicrobials. Further studies of promising ATAs and ATA combinations are required.

Simple Summary: For many years, antibiotics were added to chicken feed to prevent disease and promote growth. This practice has been banned or voluntarily abolished in many countries. However, most countries still allow the use of in-feed ionophorous coccidiostats, which are drugs that possess both antiparasitic and antibacterial properties. Concerns related to antimicrobial resistance have led to increased focus on broiler chickens raised without the use of any antimicrobial agents, and the interest in non-antibiotic feed additives with beneficial effects on gastrointestinal health and productivity is growing. In this study, feed additives with active components belonging to the product classes probiotics, prebiotics, phytogenics and/or organic acids were assessed for their effect on intestinal health and production performance in broiler chickens. Collectively, the group of non-antibiotic feed additives improved gut health and performance, but not to the same extent as the ionophorous coccidiostat narasin. Probiotics and prebiotics had the overall best performances during coccidia challenge, phytogenics improved overall feed conversion and reduced counts of the intestinal bacterium Clostridium perfringens, and organic acids increased weight gain independent of age. This study provides comparable and unbiased results from testing of alternatives to antibiotics in a uniform experimental model highly relevant to commercial conditions. Abstract: Numerous non-antibiotic feed additives (alternatives to antibiotics, ATAs) have been marketed, but few have been evaluated under uniform testing conditions modelling commercial flocks. We compared 24 ATA treatments and the ionophorous coccidiostat narasin against a diet without any feed additives. Feed conversion ratio and body weight gain were registered from day 0 to 28 in Ross 308 chickens housed on litter floor. The chickens were challenged with Eimeria spp., and cecal Clostridium perfringens (CP) counts were investigated. Active components from all ATA classes had a positive impact on intestinal health or production performance. Whereas narasin had a strong CP-reducing effect in combination with performance-promoting impact, only two ATA treatments achieved significantly beneficial effects on CP counts as well as feed conversion during the time span following Eimeria challenge. Active components present in these two treatments include a Bacillus subtilis probiotic strain, short-and medium-chain fatty acids and Saccharomyces cerevisiae components. Different ATA classes had beneficial impact during distinct rearing phases and on The aims of the study were to (a) evaluate the performance of the collective ATA group, (b) compare effects of classes of ATAs (probiotics/prebiotics/phytogenics/organic acids) and (c) identify active components or component combinations with beneficial effects on production performance and CP counts, with emphasis on the time span following Eimeria challenge.

Animals and Housing
Six trials were carried out at Scandinavian Poultry Research in Våler, Hedmark, Norway, using one-day-old Ross 308 broiler chickens obtained from a commercial hatchery (Nortura Samvirkekylling, Våler, Norway). The chickens were housed in floor pens of 5.6 m 2 on new wood shavings in a climate-controlled poultry research facility, with a 50/50 female-to-male ratio per pen. Water and pelleted feed were given ad libitum. The chickens were exposed to light for 23 h a day on the first two days. For the rest of the experimental period, the chickens were exposed to light during 2 × 8 h a day, interrupted by 4 h periods of darkness. Apart from a 10-fold dose of Paracox-5 vet. on day 17 or 18, no vaccines were administered throughout the study. The study period lasted from day of hatch until day 28. Animal experiments were approved by the national animal research authority (Norwegian Food Safety Authority, approval ID 8179), and performed in accordance with national and international guidelines for the care and use of experimental animals.

Experimental Design
In each of the six trials, a total of 5280 one-day-old Ross 308 broiler chickens were randomly allocated into six experimental groups, each group comprising 11 replicate pens with 80 chickens per pen. All trials had similar design, and included four treatment groups receiving feed with a specific ATA product or a combination of two ATA products, a positive control group (NAR) receiving feed with the polyether ionophore and coccidiostat narasin (Monteban, Elanco Animal Health, Greenfield, IN, USA), and a negative control group (NEG) receiving feed with neither antimicrobial feed additives (AGPs or coccidiostats) nor ATA products. Feed additives were added to the feeds at an inclusion rate recommended by the manufacturers. No AGP products were included in this study, and narasin was used as a sole coccidiostat in the NAR group. The chickens were fed wheat-based starter and grower diets based on Ross Broiler Nutrition Specifications adapted to Norwegian broiler production from 0 to 14 and 14 to 28 days of age, respectively (Table 1).
In the five initial trials, 20 commercially available ATA products were evaluated individually for their effect on production performance and cecal CP counts. In the sixth trial, combinations of two ATA products per treatment group were evaluated using the same outcome variables. Products included in the sixth experiment were selected for testing due to promising impact on either production performance or CP counts in the five initial experiments. Products with positive effects on production performance were combined with products with CP reducing effect in order to study potential synergy effects. Descriptions of active ingredients and dose levels of the feed additives and feed additive combinations tested are listed in Table 2. Composition of the products and dosage levels are based on information given by the feed additive manufacturers on their web sites or as a response to our request.
On day 17 (one trial) or 18 (five trials) post hatch, all treatment groups in all six trials were challenged with a 10-fold dose of Paracox-5 vet. (  1 Treatment ID number. 2 NEG = negative control, NAR = positive control, PRO = probiotics, PRE = prebiotics, PFA = phytogenics, and OA = organic acids. 3 Based on available information from the product manufacturers. 4 Amount added product given as grams/ton feed in starter and grower diets.

Clostridium Perfringens Quantification
On days 4, 5 and 6 after Eimeria challenge, 11 chickens per treatment group (1 chicken from each replicate pen) were randomly selected and humanely euthanized by cranial stunning immediately followed by cervical dislocation before necropsy. Samples of cecal contents were collected in sterile stomacher bags and directly subjected to cultivation in order to quantify CP. In brief, the samples were diluted 1:100 in peptone saline water (0.1% peptone, Difco Laboratories Inc., Detroit, US and 0.85% NaCl) and homogenized for 30 s in a stomacher (Bagmixer 400 CC, Interscience, Saint Nom, France). Serial dilutions were made with non-buffered peptone water until a dilution of 10 -6 was reached. Aliquots of 100 µL from the dilutions 10 -2 , 10 -4 and 10 -6 were plated onto sheep blood agar plates (Oxoid Blood Agar Base No.2 and 5% sheep blood, manufactured by the Norwegian Veterinary Institute, Oslo, Norway). The plates were incubated anaerobically at 37 • C for 24 h (Genbox anaer, Biomérieux, Marcy-l'Étoile, France). Single colonies with double hemolysis were counted, and colony-forming units per gram (cfu/g) cecal contents were calculated based on the given dilution. Typical colonies were selected for pure cultivation and later confirmed as CP by a matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometer (Bruker Daltonics, Bruker Corp., Billerica, MA, USA).

Post Mortem Examination
The small intestine of all chickens that were sampled for CP quantification was opened longitudinally and examined for pathological changes indicating NE, and scored as follows (modified from [34]): necrotic enteritis negative with no macroscopic mucosal ulcers or pseudomembranes, or necrotic enteritis positive with minimum one mucosal ulcer or pseudomembrane.

Production Performance Measurements
The amount of feed per pen was weighed when allocated and remaining feed was weighed before being discarded at feed change and at the end of the experiment. Accumulated feed intake per pen from days 0 to 14, 14 to 28 and 0 to 28 was calculated. Total live chicken weights per pen were recorded on days 0, 14 and 28, and mean body weight gain (BWG, g/chicken) and mean feed conversion ratio (FCR, g feed intake/g weight gain) per pen were calculated.

Statistical Analysis
Data on production performance and CP counts were examined on three different levels; (a) the impact of ATAs as one collective group (group level), (b) the impact of classes of ATAs (class level), and (c) the impact of individual ATA treatments (treatment level). On all levels, ATAs and the positive control with narasin-supplemented feed (NAR) were compared against the negative control with no feed additive (NEG). Frequencies of broilers with NE lesions were analyzed only on group level using Pearson's chi-squared test in Stata version 14.2 (StataCorp LLC, College Station, TX, USA). Production performance and CP count data were analyzed using regression analyses in R version 3.5.3 (R Foundation for Statistical Computing, Vienna, Austria).
Production performance data were analyzed with pen as the unit of concern. Body weight gain and feed conversion ratio was obtained in the periods 0-14 days, 14-28 days and 0-28 days for groups, classes and treatments tested in six trials. The outcome from the six different trials could not be compared directly due to intertrial variability. In order to validly compare results from six different trials, it was necessary to control for the effect of trial in the statistical analysis. The principle approach to achieve such control was to use the results from NEG in each of the six trials as indicators of trial effect. A mixed-effects model (1) with only intercept (a) was used to obtain a trial-specific random effect (ε Trial ) for each outcome variable (y Neg ) per trial based on results from NEG using the package lme4 in R [35].
For each of the outcome variables (y), results achieved in the different trials were adjusted with a value equal to the random effect obtained for the respective trial. Results across trials were compared using regression analysis (2) with ATA group/class/treatment (x) as fixed-effect variable and trial-specific random-effects from NEG as offset variable (ε Trial ). b represents the estimated parameters in the model.
The necessity of adjustment for trial effect was calculated by the intraclass correlation coefficient (ICC), which is variance explained by the random effect divided by total variance of the residuals for the model based on all observations from NEG. Extreme outlier pens that were highly influential on the estimated regression results were identified using the function outlierTest from the package car in R [36]. Residuals from the regression models were visually inspected using the functions qqnorm and qqline in R and found to follow a normal distribution. The production performance results were reported in tables as means with standard deviation. Differences from NEG with p < 0.05 were accepted as statistically significant differences.
CP counts in cecal samples were analyzed with individual chicken samples as unit of concern. Since the residuals from the regression model did not follow a normal distribution, the CP count numbers were log transformed in order to fulfil this requirement. The effect of trial was controlled by adjusting for obtained random effect as described above, and subsequently regression analysis with ATA group/class/treatment as fixed-effect variable and trial-specific random-effects from NEG as offset variable was conducted. The results were reported in tables as mean log 10 colony forming units per gram cecal content. Estimated mean log 10 CP counts with 95% confidence interval for each treatment were presented in a graph where feed additive classes are indicated with different colors.

Impact of the Collective ATA Group on Necrotic Enteritis, Intestinal CP Counts and Production Performance
Broilers with necrotic enteritis lesions during days 4-6 after Eimeria challenge constituted 8.1% among chickens from the NEG group (no feed additive, n = 198 chickens), 4.4% in the collective ATA group (24 ATA treatments, n = 792 chickens) and 0.5% in the NAR group (in-feed narasin, n = 198 chickens). Statistical analyses indicated significant difference in NE occurrence between the NEG group and the ATA group (p < 0.05), and between the ATA group and the NAR group (p < 0.01).
The ATA group reduced CP counts in intestinal contents from log 10 6.09 to log 10 5.63 cfu/g (p = 0.005), corresponding to a 65% reduction in non-transformed counts (Table 3). This substantial reduction was, however, moderate as compared to the very strong effect of narasin (from log 10 6.09 to log 10 2.92 cfu/g (p < 0.001), corresponding to a 99.9% reduction in non-transformed counts).  1 Results are reported as means ± standard deviation. Body weight gain (BWG) in grams/chicken, feed conversion ratio (FCR) in grams feed intake/grams weight gain and Clostridium perfringens (CP) counts as log 10 colony forming units/gram cecal content. 2 Negative control (no feed additive); production performance data based on n = 66 pens, and CP data based on n = 198 individual chicken samples. 3 Narasin; production performance data based on n = 66 pens, and CP data based on n = 198 individual chicken samples. 4 Alternatives to antibiotics treatments; production performance data based on n = 264 pens, and CP data based on n = 792 individual chicken samples. 5 Intraclass correlation coefficient.

Impact of ATA Classes on Intestinal CP Counts and Production Performance
Four ATA classes (probiotics, PRO; prebiotics, PRE; phytogenics, PFA; organic acids, OA), a set of treatments each based on more than one ATA class (mixed products, MIX) and NAR (i.e., narasin) were compared with NEG (i.e., no feed additive) ( Table 4). Although all ATA classes demonstrated a reducing effect on numbers of CP per gram intestinal contents, only two classes (PFA and PRO) showed statistically significant reduction (p < 0.05). The estimated reducing impacts of PFA and PRO were 87% and 75% in non-transformed CP counts, respectively, when compared to NEG.  1 Results are reported as means ± standard deviation. Body weight gain (BWG) in grams/chicken, feed conversion ratio (FCR) in grams feed intake/grams weight gain and Clostridium perfringens (CP) counts as log 10 colony forming units/gram cecal content. 2 Negative control (no feed additive); production performance data based on n = 66 pens, and CP data based on n = 198 individual chicken samples. 3 Narasin; production performance data based on n = 66 pens, and CP data based on n = 198 individual chicken samples. 4 Probiotics (PRO); production performance data based on n = 33 pens, and CP data based on n = 99 individual chicken samples. 5 Prebiotics (PRE), phytogenics (PFA), organic acids (OA); production performance data based on n = 44 pens, and CP data based on n = 132 individual chicken samples. 6 Mixed products (MIX), i.e., treatments based on more than one ATA class; production performance data based on n = 99 pens, and CP data based on n = 297 individual chicken samples. 7 Intraclass correlation coefficient.

ID-Class
In the sixth trial, two-product combinations of treatments with predominantly CP-reducing impact (ID 5, 16 and 21) and treatments with predominantly production performance-promoting impact (ID 7, 11 and 13) were evaluated (comprising treatment ID 22-25 in Table 2). Treatment 16 did not appear to reduce the FCR-improving effect of treatments 11 and 13 (Table 5) but tended to diminish the growth promoting impact of these treatments. Treatment 5 seemed to diminish the FCR-improving effect and remove the growth-promoting effect of treatment 7. Treatment 21 appeared to reduce or remove the improvement in FCR and to remove the growth-promoting effect of treatment 7. On the other hand, treatment 7 seemed to remove the CP-reducing impact of treatments 5 and 21, and treatment 13 appeared to remove the CP-reducing impact of treatment 16. In contrast to these results, treatment 11 did not appear to impair the CP-reducing impact of treatment 16. As a result of these interactions between predominantly CP-reducing and production performance-promoting single treatments, treatment 24 was the only one of four tested product combinations (ID 22-25, Table 5) with beneficial effects on production performance variables as well as CP counts.

Discussion
The collective group of 24 ATA treatments tested in this study reduced the occurrence of NE and reduced intestinal CP counts after Eimeria challenge. Production performance measured as BWG and FCR was improved by the collective ATA group, but not significantly during the phase prior to Eimeria challenge. These results indicate a beneficial effect of several ATAs when chickens are exposed to mild to moderate intestinal stress. The favorable effects of the polyether ionophore narasin on CP counts and production performance were numerically and in part significantly stronger than the effect of the collective group of ATA treatments.
In this study, the chickens were challenged orally with five precocious lines of Eimeria spp. as a predisposing factor for NE. NE is expected to appear during oocyst excretion, which begins between three and four days following inoculation with precocious Eimeria acervulina [37] and Eimeria mitis lines, and presumably later with precocious Eimeria maxima and Eimeria tenella lines [38]. In a previous study with a similar type of Eimeria challenge, most gut damage was detected four to six days after inoculation [39]. Postmortem examinations in this study confirmed the presence of NE during this time span after Eimeria challenge. Intestinal CP counts are strongly associated with NE [19,34,40]. Furthermore, increased occurrence of NE in commercial broiler flocks has been associated with impaired accumulated FCR at slaughter [20]. Weakened production performance is likely to be most pronounced during the part of the rearing phase that is affected by NE. Based on the considerations mentioned above we chose to emphasize CP counts on days four to six after Eimeria challenge and FCR [14][15][16][17][18][19][20][21][22][23][24][25][26][27][28] in our analyses of effect of ATAs on intestinal health.
The evaluation of ATAs can rest on different criteria, depending on point of view, practical circumstances and current health problems. Disease conditions associated with increased metabolism and rapid growth in broilers (in particular cardiovascular and musculoskeletal disorders) are generally important, and it has been claimed that such conditions cause greater economic loss than infectious agents [41]. In a recent study, it was found that chickens with higher body weight and BWG were predisposed to develop more severe NE lesions when challenged with CP [42]. Although attractive in the short run, increased weight gain may therefore come at a cost not only to chicken health and welfare but also to the farmers' economy and a sustainable use of feed resources. In light of this consideration, we have emphasized the effect on FCR as production performance parameter, because it is an indicator of intestinal health as well as resource efficiency.
The only two ATA treatments (ID 3 and 24) with a combined beneficial effect on CP counts (84 to 89% reduction, p < 0.05) and FCR 14-28 (2.3% improvement, p < 0.05) were based on different types of active components. One of the treatments (ID 3) was a mono-strain (Bacillus subtilis strain PB6) spore-forming bacterial probiotic. This probiotic strain has been reported to inhibit CP in vitro [43] and improve FCR [44], which is in agreement with our findings.
The other treatment (ID 24) was based on a heterogeneous collection of active components including short-and medium-chain fatty acids, a phenolic compound and dehydrated whole cells and metabolites of the yeast Saccharomyces cerevisiae (SC). This treatment comprised two commercial products that were also tested individually (ID 11 and 16). Whereas the yeast product (ID 16) alone demonstrated a 95% reduction (p < 0.001) in non-transformed CP counts, the product containing a blend of organic acids and a phenolic compound (ID 11) had no reducing impact on CP. Viewed against this background it seems probable that the CP reducing effect of treatment 24 was mainly associated with one or several yeast components found in treatment 16. In addition to treatment 24, three other ATA treatments based on the yeast SC were tested. These treatments, which were based on SC cell wall extracts (ID 6, 7 and 19), did not reduce CP counts to an extent that was significant with the sample size and/or feed additive dosage used in our study, whilst the treatments based on SC whole cells and metabolites (ID 16 and 24) did. No previous reports on the effect of SC metabolites and SC whole cells on CP counts in broilers have been found. Regarding yeast cell wall extracts, previously published literature has indicated both significant [45] and non-significant [46] CP-reducing impact. Our results indicate that whole cells and/or metabolites of SC inhibited intestinal CP growth more efficiently than SC cell wall extracts with the product inclusion levels used in this study.
One of the active components in treatment 11 was lauric acid (C12), a MCFA that has been demonstrated to inhibit CP in vitro [47]. Treatment 11 did, as mentioned above, not reduce CP counts when used as sole feed additive in this study. Possible explanations include too low concentration of lauric acid and/or interfering effects by other treatment components.
The combination of SCFAs (including butyric acid-C4), MCFAs (including lauric acid-C12) and a phenolic compound in treatment 11 generated the numerically highest weight gain (BWG 0-14 and BWG [14][15][16][17][18][19][20][21][22][23][24][25][26][27][28] ) of all ATA treatments in this study but had no apparent impact on CP counts. This result suggests that rapid growth is possible in the presence of relatively high cecal CP counts. A possible explanation could be that this treatment reduced the counts of virulent CP strains (e.g., strains harbouring the netB gene) or the expression of virulence factors (e.g., the NetB toxin), but not the total CP counts. Treatment 11 might also have influenced the intestinal microbiota in a way that neutralized the negative impact of high CP counts.
Treatments 22-25 were tested in a final trial intended to evaluate two-product combinations of treatments improving production performance and treatments with CP-reducing impact. Treatment 24 was the only combination with beneficial impact on both CP counts and production results. These results suggest that the interaction between predominantly CP-reducing and production-promoting components vary substantially. Among three tested CP-reducing treatments in the final trial (ID 5, 16 and 21), a treatment based on dehydrated SC culture with whole cells and metabolites (ID 16) was the least impairing with regard to the production-promoting effects of its combination treatment. Among the three tested production performance-improving treatments (ID 7, 11 and 13), treatment 11 based on short-and medium-chain fatty acids and other components was the only one that did not impair the CP-reducing impact of its combination treatment. More work is needed to identify the role of the different components in treatments 11 and 16, and whether the beneficial interaction of these components also can be extended to include other CP-reducing and production-promoting components.
Our findings indicate that a reduction of CP counts induced by ATAs was not always associated with improved production performance. Lack of a positive impact on feed efficiency and growth rate has also been documented with regard to ionophores under certain conditions [54], in spite of these compounds' suppressing effect on CP counts. However, when used at recommended concentrations in broiler flocks exposed to coccidia, the net effect of ionophores is usually improved performance. In our study, considerably improved performance combined with a strong CP-reducing effect of the ionophore narasin was present. These results confirm that our challenge model worked as expected, and that the in-feed concentration of narasin was within the optimal range. The reason why some of the ATAs with CP-reducing effect in our study did not induce a significantly positive net impact on production performance under the same test conditions remains unclear. Possible explanations may be that the inhibiting effect on CP was accompanied by reduced ability to utilize feed efficiently and/or establishment of another performance-impairing intestinal microbiota.
Two other treatments (ID 8 and 9) with favorable effect on FCR 14-28 without significant reduction of CP counts contained essential plant oils. Essential oil components in treatments 8 and 9 included thymol (in both treatments), eugenol and piperine (in treatment 8) and carvacrol, anethol and limonene (in treatment 9). In treatment 8, essential oils were combined with benzoic acid. Published results on effects and mode of action of essential oils suggest that several of these compounds inhibit the growth of CP [66][67][68], although the findings are not always clear cut [69], or they show no effect on CP counts [70]. Reports on mitigation of gut lesions in chickens challenged with CP [67,71] underpin the view that at least thymol and carvacrol suppress the pathogenic action of CP. Studies on the effect of essential oil components on production performance reveal variable results. One study reports a negative effect on broiler performance using a blend of thymol, eugenol, curcumin and piperine [70], another describe a non-significant tendency of improved FCR 14-28 using a blend of carvacrol and thymol [71], and a third study presents significant improvement of FCR 0-28 by carvacrol but not by thymol [72]. The lack of standardization of studies, including variable feed additive dosage and different combinations of active compounds, makes comparison of results from different studies difficult. The interpretation of results is further complicated by the multiple suggested effects of different essential oils, including antibacterial and antioxidant properties, enhancement of the immune system, and stimulation of digestive secretions and blood circulation [73,74]. Regardless of mechanism, these two predominantly phytogenic feed additives (ID 8 and 9) had a pronounced beneficial effect on FCR 14-28 on par with narasin in the current study.
The apparent lack of a CP-reducing effect of yeast cell wall extracts, essential oils and other active components associated with improved FCR [14][15][16][17][18][19][20][21][22][23][24][25][26][27][28] in this study may in part be related to experimental design. As observed from our results (Table 5), estimated CP count reductions of 76% or less (e.g., treatment ID 2 and 14) returned non-significant (>0.05) p-values. The main reason for this low statistical power was high variance of CP counts in individual observations within each treatment, leading to imprecise estimates. Our experiments were designed with 33 replicates of individual CP counts per ATA treatment, and this sample size returned relatively wide confidence intervals (as shown in Figure 1). The statistical analysis involving the whole ATA group (Table 3) indicated that when 792 individual samples with a log 10 5.63 CP estimate were compared with the NEG group with 198 individual samples and a log 10 6.09 estimate (corresponding to a 65% difference), this difference was significant (p = 0.005).
The ATAs did not suppress CP counts to the same extent as the ionophorous coccidiostat narasin. The superior results of narasin in this respect were most likely due to the strong antibacterial effect of this compound. Narasin has been reported with inhibitory effect on CP growth similar to or better than antibiotics used as drinking water medication for poultry [75].
Different ATAs can add value to the broiler chicken industry in several ways. Some improve BWG and/or FCR, others inhibit growth of CP or have a beneficial effect on both production performance parameters and intestinal CP counts. The use of specific ATAs could possibly be targeted to specific age intervals or current health status in the flock. Future studies of the impact of ATAs on intestinal CP counts would most likely benefit from modified sampling protocols and quantification methods. Study designs that were useful for investigating the effect of AGPs and ionophorous coccidiostats should not be copied without reservation when studying non-antibiotic alternative feed additives. Finally, a less pronounced effect than narasin of selected ATAs on production performance and/or CP counts in this study does not necessarily mean that the impact is of no importance to broiler health and production economy.
In this study, ATA classes displayed distinct performance profiles. The probiotic class reduced CP counts and improved production performance during the time period with intestinal stress (days 14-28), but impaired weight gain during days 0-14. The prebiotic class improved production performance during days 14-28 and had a non-significantly reducing impact on CP counts. The phytogenic class had a markedly reducing impact on CP counts and improved FCR 0-28 . The organic acid class increased weight gain throughout the study period and improved FCR 0-14 but did not reduce CP counts significantly. These findings suggest that employing ATA classes for specific purposes may be useful. As an example, combining probiotic and organic acid treatments might boost production performance throughout the grow-out period and at the same time reduce CP counts during intestinal stress. In this study, we tested other ATA class combinations with variable results, indicating the need for testing of specific combinations of active components within the ATA classes.