4. Discussion
4.1. General information
The 4,671,062 animals serviced by the nutritionists interviewed in this survey represented close to 75% of the cattle finished in a feedlot in Brazil in 2019 based on estimation of ABIEC (ABIEC, 2020). The interviewed consulting nutritionists reported visits to their clients every 34.2 days, on average, which is close to the interval of 38.2 days reported by Pinto and Millen (2019) , but longer than the one of 26.3 days reported by Oliveira and Millen (2014) . When compared with the first survey ( Millen et al., 2009 ) conducted 10 years ago, the percentage of nutritionists serving feedlots that fed <5000 animals annually decreased (71.0% vs. 63.9%). In addition, it was the first time that nutritionists reported servicing feedlots with more than 20,000 heads capacity, on average, which is an evidence that feedlot operations still have been growing in Brazil since the first survey was conducted ( Millen et al., 2009) .
Out of the 36 surveyed nutritionists, 47.2% were representatives of a corporate feed manufacturing company, which is in agreement with previous surveys ( Pinto and Millen, 2019 ; Oliveira and Millen, 2014 ), in which most of the participants were from corporate feed manufacturing companies (45.5 and 45.4% respectively). Furthermore, the percentage of nutritionists who work with feedlot cattle with for more than 10 years of practice rose from 42.4% ( Oliveira and Millen, 2014 ) to 83.3%. It is noteworthy that 48% of the nutritionists interviewed in this survey participated at least in one of the three previous surveys ( Millen et al., 2009 ; Oliveira and Millen, 2014 ; Pinto and Millen, 2019 ). However, unlike previous surveys ( Millen et al., 2009 ; Oliveira and Millen, 2014 ; Pinto and Millen, 2019 ), there was no interviewee with less than five years of practice.
In agreement with the first ( Millen et al., 2009 ) and second Brazilian surveys ( Oliveira and Millen, 2014 ), a Bachelor degree with graduate certificate courses in cattle nutrition was the most common terminal degree among the nutritionists surveyed, which is differently from what was reported in the previous survey by Pinto and Millen (2019) , in which nutritionists with a Bachelor degree with graduate certificate courses in cattle nutrition represented only 15.1%. Most of the participating nutritionists (42.3%) in the current survey reported that their degree was obtained in universities of the state of São Paulo.
4.2. Information on grains and energy levels
Corn was the most utilized grain (97.2%), as reported previously ( Millen et al., 2009 ; Oliveira and Millen, 2014 ; Pinto and Millen, 2019 ). In agreement with previous surveys by Millen et al. (2009) and Pinto and Millen (2019) , fine grinding remained the primary grain processing method adopted by feedlot nutritionists in Brazil (44.4%). On the other hand, the percentage of nutritionists who adopted only grain cracking has drastically reduced over the years, from 57.6% ( Oliveira and Millen, 2014 ) to only 2.8% of nutritionists in this survey.
The average granulometry of 3.2 mm for corn in Brazilian feedlots, reported by the interviewed nutritionists, showed a slightly reduction when compared with the one of 3.6 mm reported by Oliveira and Millen (2014) . It seems that grain-processing methods improved over the years in feedlot operations in Brazil, showing that nutritionists and feedlot owners were determined to increase starch availability and also the energy content of finishing diets in Brazil. In this survey, high-moisture harvesting and storage was the secondary preferred grain processing method by more than 50.0% of the participating nutritionists. However, Brazilian feedlots still have a long road to improve efficiency of grain processing, and adopt a more extensive method, as first their option, such as steam-flaking or high-moisture, the primary choices of grain processing methods for 87.5% of the nutritionists (70.6 and 16.7%, respectively) in the US ( Samuelson et al., 2016 ). As described before, flint type was the most fed corn in Brazil, and the greater adoption of steam-flaking and high-moisture processing methods would increase the proportion of starch that becomes available for ruminal fermentation at similar levels of dent corn ( Zinn et al., 2002 ).
The average inclusion level of grains from 51 to 65% was close to the ones of 51.5, 51.6, and 51.5% reported by Millen et al. (2009) , Oliveira and Millen (2014) , and Pinto and Millen (2019) , respectively. The percentage of nutritionists reporting inclusion levels of grains from 36 to 50% decreased over the last 10 years from 22.6% ( Millen et al., 2009 ) to 5.6% in this survey; however, the percentage of consulting nutritionists reporting that their finishing diets contained more 66% of grains increased from 6.5 to 38.9% ( Figure 1 ), showing that inclusion levels of grains in finishing diets in Brazil increased over the last decade.
Regarding the inclusion level of concentrate in the diets, 97.2% of the participants recommended from 71 to 90% per kg of diet DM, which was 9.3% higher than the value reported in the previous survey by Pinto and Millen (2019) , and 39.1% higher than those reported in the very first Brazilian survey ( Millen et al., 2009 ) 10 years ago ( Figure 2 ). Unlike the data for feedlots covered by the American survey, in which 78.2% of nutritionists recommended inclusions higher than 60% of grains ( Samuelson et al., 2016 ), mostly steam-flaked grains (70.8%), in the finishing diets, the present survey indicated that there is still a window for improvement, not only in the extent of grain processing, but also in the inclusion of grains in the finishing diets in Brazil. The increasing levels of grains in finishing diets, as well as the slight improvement on grain processing methods, were indicators that the energy content of finishing diets has increased over the last 10 years in Brazil.
4.3. Sources of information on feed energy values
As in previous surveys, TDN was the most recommended energy unit (52.8%); however, the use of TDN by feedlot cattle nutritionists decreased over the years (83.8%, Millen et al., 2009 ; 78.8%, Oliveira and Millen, 2014 ; 69.7%, Pinto and Millen, 2019 ). Furthermore, the use of NEg became more popular among feedlot nutritionists, as its utilization increased about 10% over the last 10 years (16.7 vs. 6.5%; Millen et al., 2009 ). Nevertheless, despite the increasing inclusion level of grains in the finishing diets, as well as the improvement of the grain processing methods, the energy content of finishing diets in Brazil is still 12.5% lower than the recommended by the US consulting nutritionists (1.33 vs. 1.52 Mcal of NEg/kg of diet DM; Samuelson et al., 2016 ) when considering the 16.7% of the nutritionists who used NEg as preferred energy unit to formulate their diets.
Regarding the main sources of information on feed energy values, RLM (Ração de Lucro Máximo, 2014) remained as the most cited source by 36.1% of the surveyed nutritionists. The RLM was also reported by Oliveira and Millen (2014) and Pinto and Millen (2019) as the most popular source of information for feed energy values among nutritionists in Brazil. The eighth revised version of the NRC ( NASEM, 2016 ) was cited by 11.1% of the surveyed nutritionists, despite its relatively recent publication.
4.4. Use of coproducts
Whole cottonseed remained as the primary coproduct included in finishing diets in Brazil, as reported by Millen et al. (2009) , Oliveira and Millen (2014) , and Pinto and Millen (2019) . The inclusion level of whole cottonseed decreased 0.4% (14.6 vs. 15.0%) when compared with the survey conducted 10 years ago ( Millen et al., 2009 ). In addition, the inclusion level of citrus pulp pellets was 27.8% ( Table 3 ), which was lower when compared with inclusions reported in previous surveys (33.8%, Millen et al., 2009 ; 40.0%, Oliveira and Millen 2014 ; 29.2%, Pinto and Millen, 2019 ). In this survey, it was reported for the first time the inclusion of DDG into feedlot diets (at 15.0% of diet DM, on average) due to the expansion of the grain milling industry for ethanol. Although the percentage of customers using some type of coproduct increased when compared with the last survey (82.3 vs. 70.6%; Pinto and Millen, 2019 ), the level of inclusion of these ingredients decreased slightly.
4.5. Roughage sources and levels and fiber analysis methods
The typical level of roughage inclusion in finishing diets recommended by the nutritionists surveyed was 16.8% of diet DM ( Table 4 ), which shows a 3.8% decrease when compared with the previous survey conducted by Pinto and Millen (2019) , confirming that the level of roughage inclusion decreased, since the first survey was conducted 10 years ago ( Millen et al., 2009 ; Figure 3 ). The current level of roughage inclusion recommended by feedlot nutritionists in Brazil was still above the levels used by US feedlot-consulting nutritionists ( Samuelson et al., 2016 ), in which 50 and 41.7% of nutritionists reported the use of inclusion levels between 8 to 10% in summer and winter, respectively. As reported earlier, finishing diets changed consistently in Brazil over the last 10 years, where roughage and coproducts were partially replaced by cereal grains. Corn silage remained as the primary source of roughage in finishing diets ( Pinto and Millen, 2019 ). Samuelson et al. (2016) also reported that corn silage was the roughage of choice of 37.5% of the nutritionists surveyed, followed by Alfalfa hay (20.8%), for finishing diets in US. Since large feedlot operations are still expanding in Brazil, fresh feeds, such as chopped sugarcane and sugarcane bagasse, were gradually replaced by conserved feeds, e.g., corn silage, as reported in this survey. Furthermore, the greater use of corn silage contributed for increasing the energy content of finishing diets in Brazil since it presents the highest energy value among the roughages cited in this survey ( Table 4 ; Valadares Filho et al., 2016).
The methods of fiber analysis preferred by nutritionists in this survey was peNDF, which was in agreement with Pinto and Millen (2019) , who reported peNDF as the preferred method as well. However, Millen et al. (2009) and Oliveira and Millen (2014) reported in the past that NDF was the method of choice of most nutritionists in Brazil, which was a significant change. Since energy content in finishing diets have increased in the last 10 years, the use of peNDF is more accurate than NDF to monitor the particle size of feedlot diets, which is mandatory to assure a minimum level of fiber to stimulate rumination and rumen buffering ( NASEM, 2016 ). In addition, the use of peNDF allows the nutritionist to identify and prevent particle sorting, which may negatively impact health and performance of feedlot cattle ( Rivera et al., 2005 ).
The average dietary peNDF inclusion level of 14.3% was close to the 14.4% reported by Pinto and Millen (2019) four years ago. This result agrees with Goulart and Nussio (2011) , who recommended finishing diets containing from 10 to 18% dietary peNDF for Nellore cattle in Brazil. For nutritionists who reported the use of dietary NDF, the current recommendation of dietary NDF content in finishing diets was 21,8%, on average, which was 1.9% lower when compared with the previous survey conducted by Pinto and Millen (2019) .
4.6. Cattle adaptation methods
Receiving programs became more popular among feedlot nutritionists, based on the fact that 10 years ago, 35.5% of the nutritionists interviewed did not adopt any type of receiving program ( Millen et al., 2009 ), which was really different compared with this survey, in which only one nutritionist (2.8%, Table 5 ) did not recommend receiving programs. The use of receiving programs is important so that cattle can recover from psychological agents and physical stressors associated with management procedures and transporting ( Marques et al., 2012 ), such as feed and water deprivation ( Swanson and Morrow-Tesch, 2001 ) that often lead to impaired health and productivity during feedlot receiving ( Araujo et al., 2010 ).
In agreement with previous surveys ( Millen et al., 2009 ; Oliveira and Millen, 2014 ; Pinto and Millen, 2019 ), including US surveys ( Vasconcelos and Galyean, 2007 ; Samuelson et al., 2016 ), multiple step-up diets were the preferred adaptation method reported by 22 (61.1%) participants. Millen et al. (2009) , Oliveira and Millen (2014) , and Pinto and Millen (2019) reported that when the multiple step-up diets program was the choice of the nutritionists, cattle spent, on average, 17.1, 18.6, and 16.2 days, respectively, to adapt to finishing diets. However, in the current survey, probably due to increasing energy contents of finishing diets reported earlier, nutritionists were recommending an extended stay on adaptation to high-concentrate diets, which were, on average, 19.2 days. Since the level of roughage inclusion in finishing diets has been decreasing over the years, cattle may take a little longer to reach the expected dry matter intake to receive the finishing diet ( NASEM, 2016 ). However, this little increase in adaptation length when multiple step-up diets method was used, still falls into the recommendation that feedlot cattle, either in Brazil or US, should not be adapted in less than 14 days to high-energy diets ( Brown et al., 2006 ; Parra et al., 2019 ), especially newly received cattle coming from either nutritional restriction or previously exposed to supplementation of concentrate feedstuffs ( Pereira et al., 2020 ).
The average initial roughage concentration recommended by the nutritionists surveyed, when the multiple step-up diets method was adopted, was 36.9%, which was lower than previous recommendations made in the past: 54.7% ( Millen et al., 2009 ), 50.5% ( Oliveira and Millen, 2014 ), and 45.1% ( Pinto and Millen, 2019 ). The greater level of grains inclusion, a longer adaptation to high-concentrate diets, as well as the consistent adoption of peNDF among feedlot cattle nutritionists in Brazil, may explain this reduction in the average initial level of roughage for the adaptation period.
4.7. General feeding and bunk management
The percentage of nutritionists’ clients who used truck-mounted mixers and horizontal mixers increased 38.5% (79.0 vs. 40.5%; Millen et al., 2009 ) and 47.7% (81.6 vs. 33.9%; Millen et al., 2009 ), respectively, over the last 10 years. The percentage of nutritionists’ clients not using any mixer was 10%, which was lower than the 33.4 and 16.2% reported by Millen et al. (2009) and Pinto and Millen (2019) , respectively. Results obtained in the present survey showed that feedlot operations in Brazil evolved in terms of use of mixers, but there is still window for growing, since some clients of the nutritionists still do not mix properly the feed delivered to cattle. Certainly, the quality of ration mixing is variable among mechanic mixers available in Brazil; however, it is reasonable to assume that the use of any mechanic mixer promotes better mixing when compared with manual mixing or simply to unmixed rations, which is the case of nutritionists’ clients who did not use any mixer. Pinto and Millen (2019) reported a significant negative correlation between feedlot size and percentage of clients who did not use any type mixer; however, in this survey, this correlation was no longer significant (r = −0.24, P = 0.15), which may indicate that even small feedlot operations (<5,000 animals) are improving their feeding management by acquiring mixers that, in many cases, are equipped with scales and automated systems.
The increasing use of truck-mounted mixers allowed an increase in the percentage of nutritionists’ clients using programmed delivery per pen, which increased 9.2% in the last four years ( Pinto and Millen, 2019 ) and 23.9% in the last 10 years ( Millen et al., 2009 ; Figure 4 ). In this survey, the correlation between the percentage of clients who did not use any mixer and percentage of clients who adopted continuous delivery was significant (r = 0.32, P = 0.05), as well as the correlation between feedlot size and percentage of clients who adopted continuous delivery (r = −0.37, P<0.05), showing that it is difficult to adopt a ration delivery system using a programmed delivery per pen at small operations and when no mixer is available. Therefore, nutritionists may be cautious to recommend diets containing higher energy content in feedlot operations that adopted continuous feed delivery, especially those that fed less than 1,000 animals. On the other hand, the greater use of truck-mounted mixers, peNDF, and programmed delivery per pen allowed nutritionists to increase the energy content of finishing diets in Brazil. As a result, consulting nutritionists are able to collect more accurate data related to feeding management and delivery from their client’s operations, which helps consultants to make decisions and nutritional recommendations to improve performance and avoid digestive disturbances, such as acidosis ( Owens et al., 1998 ).
The water added at 12% to finishing diets during the process of mixing represented a slight increase when compared with previous surveys (9.3%, Millen et al., 2009 ; 11.0%, Oliveira and Millen, 2014 ; 10.2%, Pinto and Millen, 2019 ), but the percentage of clients who added water to feed increased about fourfold (45.1 vs. 10.2%, Millen et al., 2009 ) over the last decade. The increasing percentage of clients that added water to finishing diet was associated with TDN content (r = 0.66, P<0.01), but not with inclusion level of concentrate (r = 0.19, P = 0.26), showing that the addition of water to finishing diets may be alternative to decrease diet DM content when needed. Despite this increase in the percentage of clients who added water to feed, DM in finishing diets increased from 59.9 ( Millen et al., 2009 ) to 65.4% ( Table 7 ), which was caused by a higher inclusion of grains. The average mixing time recommended by the nutritionists was 6.1 min, which is 2.4 min shorter when compared with the first survey conducted in Brazil 10 years ago ( Millen et al., 2009) and may be related to the decreased level of roughage inclusion in the finishing diets reported in previous surveys ( Millen et al., 2009 ; Oliveira and Millen, 2014 ; Pinto and Millen, 2019) .
About 67% of the nutritionists surveyed reported that most of their clients fed cattle four times a day, which agrees with the survey conducted by Pinto and Millen (2019) four years ago. Apparently, the greater use of truck-mounted mixers, as well as programmed deliveries per pen and the lower roughage inclusion, makes the planning of feeding frequency more organized without requiring extra deliveries eventually. Ten years ago, Millen et al. (2009) reported that 36.8% of the nutritionists interviewed declared that their clients used to feed cattle five times a day or more. It has been reported in the literature that feedlot performance improved when Nellore cattle, the predominant breed at Brazilian feedlot, were fed from three to four times a day compared with one to two times daily ( Silva et al., 2018 ). Furthermore, the correlation between feeding frequency and feedlot size was not significant (r = 0.07, P = 0.68), showing that feedlot size is not the main factor affecting feeding frequency planning. On the other hand, a correlation was found to be significant (r = 0.42, P<0.01) between feeding frequency and inclusion level of concentrate, which is excellent to avoid overconsumption of feed in some of the meals, which could possibly increase DMI fluctuation and lead to ruminal acidification (Schwartzkopf-Geinswein et al., 2004).
As energy content of the finishing diets has increased and level of roughage inclusion has decreased, bunk space per animal reported by nutritionists interviewed decreased 5 cm ( Table 7 ) when compared with previous survey (36 vs. 41 cm; Pinto and Millen, 2019 ). Moreover, a significant correlation (r = −0.50, P<0.01) was observed between feedlot size and bunk space, demonstrating that larger feedlot operations are adopting diets containing higher inclusions of grains, which allows them to provide a shorter bunk space per animal. Likewise, there was a correlation between bunk space and percentage of clients who adopted continuous delivery (r = 0.50, P = 0.02). However, the correlation between bunk space and feeding frequency, and bunk space and inclusion level of roughage, was not significant in this study (r = 0.02, P = 0.87; and r = 0.27, P = 0.12; respectively). The average area per animal in a pen reported by the nutritionists interviewed was 15.2 m2.
With respect to bunk management, about 45% of nutritionists reported that their clients use clean-bunk management, which was an evolution when compared with Pinto and Millen (2019) , in which nutritionists reported that most of their clients left 1-3% orts. Furthermore, a correlation was found between bunk management and feedlot size (r = −0.30, P = 0.05), confirming that shifting to a clean-bunk management was a required evolution when feedlot operations became larger. However, bunk management was not directly associated to the energy content of finishing diet, as the correlation between TDN content and bunk management was not significant (r = −0.24, P = 0.32).
The use of sprinklers in feedlot pens was reported by 44.4% (n = 16) of the nutritionists, a 15.6% decrease when compared with the previous survey ( Pinto and Millen, 2019 ). However, if this decreased use of sprinklers is associated with increasing respiratory diseases issues, it remains unknown; since nutritionists were not questioned about it.
4.8. Management of newly received cattle
The use of only BW to sort cattle at feedlot arrival was the primary method adopted by 73.9% of the participants’ clients in this survey, and in past surveys as well ( Millen et al., 2009 ; Oliveira and Millen, 2014 ; Pinto and Millen, 2019 ), including the US survey (56.3%; Samuelson et al., 2016 ).
4.9. Cattle performance information
As reported in past surveys ( Millen et al., 2009 ; Oliveira and Millen, 2014 ; Pinto and Millen, 2019 ), bulls remained as the predominant type of cattle fed in Brazilian feedlots (by 87.5% of the clients; Table 8 ). In addition, the final weight of bulls increased about 30 kg when compared with the previous survey ( Pinto and Millen, 2019 ), and 55 kg over the last 10 years ( Millen et al., 2009 ). Heavier final weights were also observed for heifers, calves, cull cows, and steers. However, the initial weights remained practically unaltered over the last 10 years ( Millen et al., 2009 ). Considering that the mature weight of Nellore cattle, which is the predominant breed in Brazil and also in Brazilian feedlots (fed by 75.3% of nutritionists’ clients), ranges between 560 and 580 kg ( Fox et al., 1992 ), nutritionists have been taking advantage of feed efficiency of cattle slaughtered at lighter weights when compared with mature weight of Nellore animals ( Owens et al., 1995 ) to explore carcass deposition and increase dressing percentage. Certainly, the increasing energy content in finishing diets in Brazil was one of the factors that led to heavier weights at slaughter. In addition, bulls, and heifers, calves, cull cows, and steers, have been fed for longer periods, which also contributes to increase slaughter weights. The number of days on feed increased about 23 days for bulls and 21 for heifers when compared with the first Brazilian survey conducted 10 years ago ( Millen et al., 2009 ). In the last four years, the Brazilian markets and the packing plants increased the requirement for minimum fat cover (5 mm on backfat of carcasses) and started paying bonuses for it. However, there is no national grading system established by the Ministry of Agriculture and Livestock, and each packing plant has developed its own system for carcass grading according to their markets.
4.10. Recommended nutrient composition for finishing diets
The average dietary fat concentration in finishing diets (DM basis) was 5.22%, which is greater than the 5.0% reported by Pinto and Millen (2019) , and even greater than the 4.7% observed by Millen et al. (2009) ten years ago. For maximum inclusion of dietary fat, nutritionists interviewed in this survey recommended 6.64%, which represents an 8.8% increase in the last decade ( Millen et al., 2009 ). Levels of dietary fat higher than 7.0% (DM basis) may negatively impact fiber digestion in the rumen ( Oldick and Firkins, 2000 ); however, it seems that Brazilian nutritionists were aware of this fact by recommending both average and maximum dietary fat concentrations below 7% of diet DM. Along with increasing levels of both grains and concentrate, the greater dietary fat concentration recommended in the current survey is also contributing to increase energy content of finishing diets in Brazil.
Protein recommendations did not change much on the course of the last ten years, and values reported in this survey were similar to previous ones ( Millen et al., 2009 ; Oliveira and Millen, 2014 ; Pinto and Millen, 2019 ). As discussed earlier, Brazilian cattle has typically been slaughtered before reaching the mature weight, which supports the unaltered nutritionists’ recommendations of protein over the last decade, since this type of cattle requires more metabolizable protein for muscle protein synthesis ( NASEM, 2016 ). Moreover, in US feedlot system, the average CP concentration recommended by consulting nutritionists was 13.4% (DM basis; Samuelson et al., 2016 ), which is quite similar to the 13.7% reported in this survey. In the last and eighth revised version of the NRC ( NASEM, 2016 ), the requirements of RDP have been revisited, and from now on, nutritionists in US and Brazil may want to revise their protein recommendations to meet a lower ammonia and RDP requirements, based on the fact that the current model assumes that all RDP will be used for microbial protein synthesis. Feedlot performance is usually limited by energy, and adjustments in finishing diets in that direction (less protein and more energy) may increase slaughter weights beyond maturity, and positively impact fat cover of carcasses of feedlot cattle in Brazil ( NASEM, 2016 ). In this case, formulation of diets for RDP would be less important, and this may explain why only one-third of feedlot nutritionists in US formulated for RDP ( Samuelson et al., 2016 ). It is noteworthy that soybean meal remained as the primary source of plant-based protein used in finishing diets in Brazil; however, it was the first time that distillers grains were cited among the main sources of plant-based protein used by the nutritionists ( Table 9 ).
The average Ca concentration recommended by the surveyed nutritionists for finishing diets was 0.63% of diet DM, value that is greater than 0.58% (DM basis) reported by Millen et al. (2009) ; however, both values still met the NASEM (2016) recommendations. Furthermore, the 0.30% of diet DM recommended for P was similar to the value reported by Millen et al. (2009) 10 years ago and agreed with recommendations of NASEM (2016) and Samuelson et al. (2016) . The K concentration decreased slightly after 10 years (0.75 vs. 0.83% of diet DM; Millen et al., 2009 ); however, S concentrations increased about 12% (0.19 vs. 0.17%; Millen et al., 2009 ), which was consistent with increasing levels of energy in finishing diets, since increased synthesis of microbial protein requires more S, especially for amino acids such as methionine, cysteine, and cystine ( NASEM, 2016 ). Recommendations of nutritionists for Mg in finishing diets increased when compared with the reported by Millen et al. (2009) ten years ago (0.18 vs. 0.16%).
In this survey, the recommended concentrations of trace minerals values increased slightly ( Table 10 ), but it still met NASEM (2016) recommendations, and were less than those reported by feedlot cattle nutritionists in US ( Samuelson et al., 2016 ), which may be a result of the use of different feedstuffs and energy levels in finishing diets between Brazil and the US.
The values for vitamin concentrations in this survey ( Table 10 ) were greater than those reported by Millen et al. (2009) and were in accordance with the recommendations of NASEM (2016) , differently from what was reported in 2009, in which these values were below the recommendations ( Millen et al., 2009 ). In terms of vitamin A, the American nutritionists recommended 4,715 IU/kg ( Samuelson et al., 2016 ), which was 82% greater than the recommended by Brazilian nutritionists. The higher inclusion of roughages in finishing diets in Brazil compared with USA may explain these differences in vitamin A recommendations, since tropical roughages contain significant amount of this vitamin in their composition ( Reynoso et al., 2004 ).
Sodium monensin, the most popular feed additive among feedlot consulting nutritionists ( Table 10 ), is an ionophore, which has been cited in all past Brazilian ( Millen et al., 2009 ; Oliveira and Millen, 2014 ; Pinto and Millen, 2019 ) and American surveys ( Samuelson et al., 2016 ) as the primary type of feed additive used in finishing diets. Virginiamycin has been cited as the secondary feed additive used by their clients in finishing diets, since the combination of monensin and virginiamycin in finishing diets has become popular in the past four years based on promising results related to increasing carcass weight ( Rigueiro et al., 2020 ). However, we cannot guarantee, based on the results of this survey, that all nutritionists’ clients that use virginiamycin as secondary feed additive also use sodium monensin as their primary feed additive.
4.11. Sources of information
When asked about nutritional models they used, the RLM was the most common nutritional model used by the nutritionists (30.6%), as also reported previously by Pinto and Millen (2019) four years ago (36.7% of the responses); however, considering new and old versions of the NRC publications, 33.4% of the nutritionists reported that they used some of its versions as the nutritional model of choice.
Regarding source of scientific or general information, the Journal of Animal Science remained as the most popular source among Brazilian nutritionists over the past 10 years ( Millen et al., 2009 ; Oliveira and Millen, 2014 ; Pinto and Millen, 2019 ).
4.12. Problems reported by the nutritionists
Respiratory diseases in general, as well as acidosis, were already cited in previous surveys ( Millen et al., 2009 ; Oliveira and Millen, 2014 ; Pinto and Millen, 2019 ) as the primary and secondary issues related to animal health in Brazilian feedlots. We point out that the percentage of the nutritionists reporting respiratory diseases as the major health problem increased 19.7% when compared with the last survey ( Pinto and Millen, 2019 ). However, mortality rate reported by the nutritionists in the feedlots they serviced was a little lower (0.57 vs. 0.76%) than the reported by Pinto and Millen (2019) .
Past surveys indicated the lack of trained employees (63.0%; Millen et al., 2009 ) and availability and accuracy of equipment (58.1%; Oliveira and Millen, 2014 ) as the major challenges; however, in this survey, nutritionists (61.1% of the responses) indicated administration and management as the most challenging issues to put their nutritional recommendations into practice. Due to the fact that Brazilian feedlots increased their investments on feeding technologies, we assumed that employees have been subjected to more rigorous training to operate machinery and computer-based devices; however, the major challenge now is to manage and organize all the processes present in the feedlot and coordinate these with nutritional recommendations, which include time of feeding, mixing time, feeding frequency, and so on.