Estimating the total Ca requirement and optimizing dietary Ca for meat-type, broiler chickens is not a new concept. In 1961, Simco and Stephenson (1961) estimated the total Ca requirement to optimize gain, feed to gain, and toe ash could be as low as 0.50%. Whereas, in 1963, Edwards Jr. and colleagues estimated the total Ca requirement to optimize gain and feed to gain was between 0.99% and 1.07% and much greater at 1.39% to 1.47% total Ca to optimize bone ash. The differences in the estimated total Ca requirements between the previous authors was presumed to be related to differences in growth rates and feed efficiencies of the broilers, differences in protein and energy concentrations in the diets, and fat type and content in the experimental rations (Edwards et al., 1963).

Today, approximately 60-years later, discrepancies still exist when estimating the total Ca requirements for modern broilers; and especially when comparing the newly developed standardized ileal digestible (SID) Ca requirements for broiler chickens. Driver et al. (2005b) and Bai et al. (2022) estimated the requirement to maximize body weight gain (BWG) during the starter phase was 0.625% or 0.590% total Ca, respectively. Whereas, during the grower phase (d 19 to 42), Driver et al. (2005b) reported no significant effect of increasing total Ca level on BWG or bone ash. However, Rama Rao et al. (2003) suggested 0.756% total Ca was required to maximize BWG from d 1 to 35 post-hatch. In 3 of the 4 studies, approximately 1.0% total Ca was required to support Ca metabolism and optimal skeletal development (Bai et al., 2022; Driver et al., 2005b; Rama Rao et al., 2003), indicating the Ca requirement to optimize tibia ash is greater than that needed to optimize growth and feed efficiency, but the exact total Ca requirement remains elusive.

More recently, David et al. (2021) and Walk et al. (2021b) set about estimating the SID Ca requirements of fast-growing broilers from hatch to d 10. In P-adequate diets, David et al. (2021) estimated the SID Ca requirement for maximum weight gain and bone mineralization was 0.332% and 0.436% to 0.478%, respectively. Whereas Walk et al. (2021b) reported no effect of graded levels of SID Ca on BWG and tibia ash was maximized at 0.530% SID Ca. In both experiments, apparent metabolizable energy (AME), crude protein, and digestible lysine in the experimental diets were formulated to meet fast-growing broiler requirements, soybean oil was used as the supplemental fat source, and growth rates, feed intake (FI), and feed conversion ratio (FCR) were comparable between the studies. Therefore, the discrepancies in the estimated SID Ca requirements were not related to differences in growth rates or dietary composition of protein, energy, or fat sources as previously hypothesized by Edwards et al. (1960, 1963).

Other factors that might be influencing the results are the concentration of phytate and phytase in the experimental diets (Selle et al., 2009), the SID Ca coefficients used to formulate the experimental diets, which can be quite variable (Walk et al., 2021a), differences in the source, quality, and particle size of the limestone (Kim et al., 2019) used to create the graded concentrations of SID Ca, different graded levels of total Ca, and differences in broiler genetics, with Ross 308 broilers used in David et al. (2021) and Arbor Acres Plus broilers used in Walk et al. (2021b). Therefore, the current study aims to re-evaluate the previously determined SID Ca requirement for fast-growing broilers, using Ross 308 genetics, 3 different non-linear regression models including the quadratic (QP), straight-broken line (SBL) and quadratic-broken line (QBL), and a reference diet formulated using total Ca to meet or exceed nutrient recommendations, from hatch to d 14 post-hatch.

Calcium and P are the most abundant minerals in the body and are predominantly found in the skeleton. Dietary imbalances or deficiencies of Ca or P and excess dietary Ca can lead to reduced growth performance and nutrient digestion (particularly P and amino acids), as well as poor feed efficiency and bone mineralization and may even lead to death in severe cases. Understanding the amount of dietary Ca and P that is digested, absorbed, and used by the animal is vitally important for bird health and welfare and promotes efficient use of inorganic P. Formulating broiler diets to meet digestible, rather than total nutrient requirements has resulted in improvements in BWG, feed efficiency, and reduced nutrient excretion. For example, Rostango et al. (1995) found that birds fed diets based on digestible amino acids were more efficient and gained more compared with birds fed diets based on total amino acids. Interestingly, at that time, the digestible amino acid diet was more expensive; but because birds were more efficient and performed better, feed costs per kilogram of BW were lower in birds fed the digestible amino acid diet (Rostango et al., 1995). Similarly, formulating diets closer to the bird's requirement for P can reduce the amount of P excreted by the bird by 30% to 40% (Applegate and Angel, 2008).

The long-term objective of this work is to optimize the inclusion of Ca in broiler diets, thereby improving growth performance, feed efficiency, and P and amino acid utilization through the implementation of a SID Ca formulation system. The specific objective of this study was to estimate the SID Ca requirement of fast-growing broilers from hatch to d 14 post-hatch using QP, SBL, and QBL models, to provide additional and comparative information on the SID Ca requirement of broilers. The models were selected based on previously published methods to estimate nutrient requirements of broilers. Each non-linear model has benefits and limitations and these should be considered when setting nutrient requirements. For example, the QP model estimates the requirement at the concentration of nutrient that results in the maximum response. There is a single maximum, not a plateau, and anything below or above the maximum is considered lower predicted performance or nutrient utilization levels (Pesti et al., 2009). In this regard, the QP model may be useful to estimate deficiencies or toxicities, but generally results in a greater estimated requirement compared with the SBL or QBL (Pesti et al., 2009; Walk et al., 2022). When using the QP, it may be possible to estimate the requirement at a value other than the maximum, such as 90% to 99% of the maximum (Pesti et al., 2009), to account for the higher estimated requirements compared with other non-linear models. The SBL model presumes that the response to a nutrient dose is linear until a plateau and no further changes in the requirement are noted (Robbins et al., 2006). In this case, the linear assumption may simplify the impact of dietary Ca level on performance or nutrient utilization by assuming the rate of change is the same with each change in nutrient concentration in the diet. In the current study, the significant quadratic effects of dietary SID Ca on growth performance, mFCR, tibia ash, ileal digestibility, and retention of Ca and P indicate the response of the bird differs depending on the SID Ca concentration in the diet and a SBL might simplify this effect. In addition, the SBL assumes additional supplementation of the nutrient at concentrations above the break point or plateau has no negative impacts on growth performance or nutrient utilization. This may be the case for some nutrients, but high concentrations of total dietary Ca have been shown to reduce growth performance (Amerah et al., 2014; Bai et al., 2022; Simco and Stephenson, 1961) and P and amino acid digestibility (Amerah et al., 2014) of broilers, especially when dietary P is limiting. Therefore, use of the SBL may over-simplify the effect of dietary Ca on broiler performance, efficiency, and nutrient utilization. Finally, the QBL model was considered as a method to estimate both the curvilinear (Robbins et al., 2006) and potential toxic (Pesti et al., 2009) effects of dietary SID Ca concentrations on growth performance, tibia ash, and nutrient utilization. In the current study, the QBL model resulted in the lowest estimated SID Ca requirements, especially compared with the QP model. This was also true for the SID Ca requirements estimated by David et al. (2021) for tibia ash concentration (0.451%). Walk et al. (2022) also found the QBL model generally estimated the SID Ca requirements at lower dietary concentrations when compared with the SBL and QP. Whereas Robbins et al. (2006) reported the QBL provided a greater estimated isoleucine requirement for growing pigs when compared with the SBL. Unfortunately, and as found with other nutrient requirement studies, choosing the best model to estimate the requirement is not simple. It may be possible to select the best regression model using the coefficient of determination statistics, such as R², adjusted R², root mean square error, and models with lower Akaike information criterion (AIC) or Bayesian information criterion (BIC) coefficients, assuming all factors are significant (Pesti et al., 2009). Another option to select the optimum requirement could be to use practical considerations, such as feed cost savings or cost per bird. In this case, the best feeding level is the nutrient concentration that maximizes profits (Pesti et al., 2009). The economic model has been considered for digestible amino acids (Kidd and Tillman, 2016; Rostango et al., 1995) and may be something to consider when utilizing the SID Ca requirement recommendations in practice.

When using the mean of the parameters and non-linear models in the current study, the estimated formulated SID Ca or measured AID Ca requirement was 0.471% or 0.454%, respectively, for broilers from hatch to d 14. This is equivalent to a SID Ca to nPP ratio of 1.024 or 0.987, respectively, or a measured apparent ileal digested P ratio of 1.198 or 1.185 for formulated SID Ca or measured AID Ca, respectively. The SID or AID Ca to digested P ratio is similar to that reported by Angel (2017; 1.15) but higher than that reported by David et al. (2021; 0.66 to 0.96) in broilers from hatch to d 10. The estimated SID Ca requirements in the current study were slightly lower than those obtained by Walk et al. (2021b) and similar to or higher than those obtained by David et al. (2021) for optimal tibia ash (0.478%) and BWG (0.332%), respectively, when comparing data from the QP model and male broilers from hatch to d 10.

Differences in the findings between the studies may be related to differences in the SID Ca coefficients of the ingredients, the use of high doses of phytase in the current study, differences in the concentrations of total Ca in the diets, or differences in the P formulation systems, particularly the digestible P (dgP) coefficients. Upon comparison of the previous work by David et al. (2021) and the current study, we found the SID Ca coefficients were similar between the studies and the analyzed total Ca concentrations were wider in the current study but within the range of that tested by David et al. (2021; 1.04% to 0.39% and 1.16% to 0.69%, respectively). In fact, when the ingredient composition from David et al. (2021) was input into the feed formulation software used to formulate the diets for the current experiment, the SID Ca values only differed by 0.01 to 0.03 percentage units. The total P and nPP concentrations were also very similar between the 2 studies, whereas there were large differences in the dgP coefficients used between the different experimental diets. Therefore, the differences in the estimated requirements between the studies were most likely not related to differences in the SID Ca coefficients, total P, or nPP, but rather differences between the formulated dgP concentrations in the experimental diets. For example, when comparing the formulation systems and using the diet and ingredient composition of the treatments formulated to contain 0.40% SID P from David et al. (2021), total P was listed as 0.53% in the paper and 0.56% in the current study formulation software. However, dgP was only calculated at 0.29% in the current formulation software and this is nearly 30% below the expected 0.40% SID P. The dgP coefficients used in the David et al. (2021) study for corn, soybean meal, and inorganic P were based on data from the same lab and the coefficients used were 0.70, 0.75, and 0.79, respectively. Whereas, the dgP coefficients in the current study are from CVB (2021) values, which estimate the P digestibility of corn, soybean meal, and inorganic phosphates at much lower coefficients of 0.30, 0.42, and 0.78, respectively. In addition, based on CVB (2021), dgP coefficients are determined from fecal samples and not adjusted for endogenous P losses, which may also be a big factor influencing the differences in the coefficients. However, when using standardized experimental methods and the same soybean meal, previous authors reported the dgP concentration of soybean meal ranged from 19% to 51% (Rodehutscord et al., 2017). This difference in dgP concentration in the diet formulations, plus the inclusion of high doses of phytase in the diets of the current study may have resulted in the big differences in the estimated SID Ca requirements between the different labs.

The impact of dietary P concentration on the Ca requirement is well known. For example, in 1961, Simco and Stephenson found when a lower level of total P was fed (0.42%), feeding excessive total dietary Ca (1.0%) had a greater detrimental effect on growth performance compared with feeding lower concentrations of total Ca. Similarly, widening the Ca to total P ratio in broiler diets containing phytase had detrimental effects on growth performance (Driver et al., 2005a). David et al. (2021) reported the SID Ca requirement to optimize BWG or tibia ash was different depending on the dgP concentration in the diet. To avoid differences between the digestibility systems in the different studies and its impact on the interpretation of the requirements, the authors of the current study chose to present the estimated SID Ca requirements as a mean of 3 regression models, using various response variables, and as a ratio to the nPP concentration in the experimental diets. Non-phytate P is useful because it can be analyzed in the diets using standardized assays of total P and phytate P and then subtracting phytate P from the total P. Furthermore, because the apparent ileal digested P was also measured in the current study, it was possible to determine the apparent ileal digested P at the estimated SID Ca (0.471%) or measured AID Ca (0.454%) requirements and therefore present the mean digestible Ca to dgP ratio from the results. We hope using nPP and/or measured apparent ileal digested P at the estimated SID Ca requirement will aid in establishing aligned digestible Ca requirements and recommendations in the future.

In conclusion, the formulated SID Ca and measured AID Ca requirement of fast-growing broilers from hatch to d 14 was estimated at 0.534%, 0.480%, or 0.398%, and 0.495%, 0.430%, or 0.436% when QP, SBL, or QBL regression models were employed, respectively. Taking the average of the estimates from all response parameters and the 3 non-linear models, the formulated SID Ca requirement of fast-growing broilers from hatch to d 14 post-hatch was estimated at 0.471% and this is equivalent to a SID Ca to nPP ratio of 1.024 or a SID Ca to measured apparent ileal digested P ratio of 1.198. When considering the measured apparent ileal digested Ca, the Ca requirement was estimated at 0.454% and equivalent to a nPP or measured apparent ileal digested P ratio of 0.987 or 1.185, respectively. These results support the use of digestible Ca in feed formulation for broilers from hatch to d 14. They are validated by previously published data as well as the use of the nutrient adequate reference diet. The estimated SID Ca requirements in the current study were similar to previously published data from the same lab, but higher than those previously reported by David et al. (2021) and this may be due to differences in the dgP concentration between the studies. Therefore, the use of nPP or measured apparent ileal digested P may be useful to include in studies in the future to compare SID Ca requirements at measurable and comparable levels of dietary P.