1. Introduction
Ruminant diet formulation is essential when it comes to animal productivity. In order to fulfil the energy requirements for a productive dairy cow, highly fermentable diets are commonly fed [
1]. These highly fermentable diets are often high in starch and limited in the amount of effective fibre they contain, which may alter the volatile fatty acid (VFA) levels in the rumen [
2]. VFAs, including lactic acid, are produced during the fermentation of feedstuff in the rumen; however, accumulation of these acids may lead to a drop in rumen pH [
3]. Increased rumen acidity can result in metabolic disorders, with sub-acute rumen acidosis (SARA) being one of the most common [
1,
4]. SARA is of concern amongst dairy farmers, as it reduces cow productivity, increases the risk of adverse health conditions [
2,
5,
6] and can lead to the animal’s death [
1]. The definition of SARA is a depression in rumen pH for more than 3 consecutive hours per day below a pH of 5.6 [
7]. To maintain a healthy rumen pH and reduce the risk of rumen acidosis, dietary alterations must be put in place [
8,
9]. The rumen pH depends on the Henderson–Hasselbalch equilibrium (which describes buffering and acid–base balance) and is given by the equation pH = pK
a + log ([A
−]/[HA]); ([HA] and [A
−] refer to the equilibrium concentrations of the conjugate acid–base pair), where pK
a is the negative log of the acid constant, K
a. Lactic acid has a lower pK
a (3.86) than acetic acid (pK
a = 4.76), propionic acid (pK
a = 4.87) or butyric acid (pK
a = 4.82) [
10], meaning it is a stronger acid; therefore, rumen pH is more affected by the production of lactic acid. Thirty percent of VFAs are neutralized by salivary sodium bicarbonate (SB) and phosphate, whose production is stimulated by chewing and the amount of effective dietary fibre [
5,
11]. If the capacity of the endogenous buffer system is exceeded due to a highly fermentable diet with low fibre content, the production of VFA can exceed removal via the rumen wall, resulting in a drop in rumen pH [
12]. The lower pH favours the growth of lactic acid-producing bacteria in the rumen, and the increased accumulation of low pK
a lactic acid results in a shift in VFA profiles and a decrease in pH occurs, resulting in acute acidosis [
12].
The inclusion of rumen buffers to the ruminant diet is one intervention that can be implemented to regulate rumen pH [
13,
14] and suppress SARA in dairy cows [
4,
15] while simultaneously combating milk fat depression [
16]. Dietary buffers reduce rumen acidity and create a more desirable environment for microbial activity [
17,
18]. This can lead to enhanced rumen microbial growth, increased enzymatic activity and higher microbial diversity and thus contribute to improved fermentation efficiency and nutrient availability [
19]. Rumen buffers are typically composed of mineral salts and calcareous marine algae (CMA) products [
14,
16,
20].
Sodium Bicarbonate (SB) is widely used as a rumen buffer [
13,
21,
22,
23,
24]. SB buffering action is short lived as a fully soluble buffer in the rumen and cannot buffer ongoing acid production [
15,
23]. Furthermore, Cruywagen et al. reported CMA (
Lithothamnion sp.) had a higher buffering capacity and higher minimum rumen pH when compared to (SB) when fed to dairy cows over the course of sixty-six days [
20]. An in vivo study conducted by Neville et al. explored the effect of different rumen buffers on rumen pH and milk production in mid-lactation dairy cows fed a high-starch TMR based on ryegrass silage and corn silage [
15]. The study reported differences in in vivo rumen pH, milk production and milk quality depending on the type of buffer fed. Similar data are available highlighting the milk production benefits of CMA when fed to pasture grazing dairy cows [
25]. The influence of rumen buffers on in vivo dairy cow productivity has therefore been well reported.
However, laboratory-based rumen buffer testing protocols can produce results that are contradictory to the reported in vivo performance. Historic laboratory testing protocols were based on total acid neutralisation capacity of feed and/or rumen buffers [
10,
26,
27]. However, these methods do not take into account the rate of acid production in the rumen or the actual rate of change of pH for different rumen buffer materials over a fixed period of time. This study aims to provide an in-depth evaluation of analytical protocols to calculate the buffering capacity of commercially available rumen buffer materials (e.g., CMA, calcium carbonate and SB), along with assessing how their composition and particle size might affect buffering performance, in order to determine the analytical factors that affect the prediction of buffering capacity and their role in developing an in vitro protocol that correlates better with in vivo performance.
4. Discussion
The focus of this study was to evaluate and optimise laboratory protocols to consistently and robustly evaluate carbonate-based rumen buffering materials in order to provide benchmarking and prediction of their in vivo efficacy and ultimately animal performance, through more efficient rumen pH management. Compositional and physical characterisation of the different calcium carbonate materials was performed in order to assess the variation between the materials and to assess the contribution of this variation to efficacy. It is evident from the data presented in
Table 1 and
Table 2 that significant variation exists between individual CMA materials and calcium carbonate for the tested parameters. This variation is particularly apparent in the organic matter, calcium and magnesium content. All four CMA materials have statistically significant differences in organic matter, with CMA Cal 2 and CMA Glac having the highest content. The presence of organic matter in CMA has been previously reported. Components of this organic matter have been described to have a key role in the biomineralisation process of red seaweed during its vegetative growth phase [
30]. A number of different species of CMA exist, with the distribution of species being specific to geographical location, water temperature, depth of water and ocean currents [
31]. Differences between the organic, calcium and magnesium content have been reported for different CMA species [
30,
32]. The variation in compositional data reported in this study likely reflects the use of different CMA species for the manufacture of rumen buffers. The composition of calcium carbonate is in agreement with previously reported values. The particle size distribution of the materials is similar, but there are statistically significant differences in the distribution. The differences in the particle size is likely associated with the processing of the materials rather than material-specific characteristics. The relationship between acid neutralisation capacity and calcium carbonate-based particle size has been reported [
33], with a range of particle size from 37.7 to 3067.9 µm for limestone [
33].
Three different methods were evaluated for determining the buffering potential of rumen buffer materials: (a) 2 and 8 h static pH, (b) 8 h fixed HCl acid load addition and (c) 3 h acidotic diet simulation using acetic acid. All methods are based on acid–base titration, with the 2 h static pH based on previously published methods for buffer capacity (BC) and buffer value index (BVI) methods, with modification to focus on evaluation of the buffers only over a duration rather than total diets and/or rumen fluid [
27,
34]. The uncertainty of measurement (UoM) of the static pH titration methodology was determined at pH 5.5 and 6.0 and was found to range from 0.51–4.39 and 0.93–2.55 mL of HCl, respectively (
Table 3), for five different carbonate buffer materials. Data on UoM for feed buffer evaluation were not found in the published literature for comparison. Calcium carbonate had the highest UoM (2.55–4.39), which is likely due to inherent variability in the material, which has been widely reported for limestones of various origins [
33,
35,
36]. All the other materials had similar UoMs, which indicates the variation of the titration methodology was between 0.51 and 2.1 mL of 0.1 M HCl when performed as outlined in this study. Similar coefficients of variation (1–8%) have been reported in other studies for solubilisation and neutralisation of calcium carbonate-based materials [
37,
38].
Other laboratory testing protocols for assessing rumen buffering capacity are based on total acid neutralisation capacity or acid binding capacity (ABC) of feed and/or rumen buffers [
10,
26,
27]. The results presented in
Table 4 show that the length of time of the titration (2 or 8 h) has a significant effect on ABC and BUF values for all materials, except SB, with values increasing in ABC and BUF over time, which is likely to be relevant in an in vivo rumen environment. A prolonged buffering action is desirable for dairy cows fed a highly fermentable diet to avoid periods of sub-optimal rumen pH; therefore, those materials having prolonged buffering action are likely to be beneficial in vivo [
15]. There was no sizeable increase in ABC or BUF with SB over time, which is likely due to its high solubility and indicates that it has a quick acting buffering action, which has been observed in vivo [
15]. In
Table 4, the best performing buffers are SB and CMA Glac. CMA Glac has approximately 100% higher ABC than CMA Cal 1 and calcium carbonate at pH 5.5 over 2 and 8 h. This result highlights the significant differences between CMA buffers, which is likely associated with the species of CMA, their location of origin and their inherent composition. CMA Glac had significantly higher organic matter and magnesium content than CMA Cal 1 and CMA Cal 3, with the divalent basic forming cation, magnesium, of marine origin likely affecting the acid neutralisation capacity of CMA Glac. Processing of the CMA buffers and their particle size distribution are also likely play a role in the observed ABC values, as increased surface area is likely to enhance ABC. However, the significantly smaller particle size distribution in CMA Cal 3 did not provide values that were higher than CMA Glac, so other factors may be contributors, such as surface area from the inherent pore space in the material, which should be evaluated in the future. Similar trends are also evident from the calculated BUF values. Evaluation of the extended period of 8 h at pH 5.5 confirmed the trend observed at the same pH over 2 h, where the CMA Glac and SB materials provided the highest ABC and BUF values. It was clear in each method that 65–72% of SB was consumed to reach the first end point of pH 7, thus demonstrating the influence of the pK
a value of SB [
34]. Additionally, the high solubility and pK
a value of SB limits the buffering capability of the material below pH 6, as the alkaline species of SB disassociates at pH 6.3 to form the intermediate carbonic acid, with up to 20% of the initial concentration converting to the intermediate and subsequently carbon dioxide at pH 6.0 [
34]. The ABC of all the materials at pH 6 is lower than that reported in the literature for calcareous materials (calcium carbonate = 18,000–20,000 meq/Kg), but this is likely due to the target pH being pH 3–4, in line with monogastric nutrition requirements [
28]. Previous evaluation of the performance of an SB buffer combined with a defined diet in rumen fluid was reported as BC with values of 70–85 meq/L of rumen fluid reported for the control with no buffer and 110–125 meq/L reported for SB [
34]. However comparison of the BC and BVI values with ABC and BUF values as determined in this study is not possible due to different experimental setups and testing objectives, with BC and BVI focused on total feed or rumen contents.
ABC and BUF testing protocols can produce results that are contradictory to the reported in vivo performance [
18,
20]. These methods do not take into account the rate of acid production in the rumen or the actual rate of change of pH for different rumen buffer materials over a fixed period. The fixed HCl acid load titration was developed as an in vitro method to evaluate the performance of buffers using a methodology that has a closer relationship with the rate and dynamics of acid production in the rumen environment over time. The fixed HCl acid load methodology was evaluated to determine the method variables that have a significant effect on predicting a buffer potential to prevent suboptimal rumen pH and SARA (pH 5.8 and 5.5, respectively), reported as area under the curve (AUC, mmol H
+.s), rate of change of pH over time and length of time pH, was greater than the threshold pH for SARA diagnosis. SB and CMA Glac were the two best performing buffer materials in this method, having the highest values for the three calculated variables of AUC, dpH/dt and time above threshold pH, which is in line with ABC and BUF values derived from the static pH methodology. Of the three calculated output measures from the titration curves, AUC was identified to be the most sensitive measure. Buffer material, threshold pH, test duration and interactions between all three variables significantly (
p < 0.001) affected AUC values, and thus this output measure was selected to predict performance of the buffer materials. AUC had been previously shown in vivo to differentiate between pH profiles of low-fibre and high-fibre diets, which further supports it utilisation in an in vitro testing methodology [
39].
HCl is a strong acid whose pKa (pKa = −8.0) is much lower than rumen relevant organic acids and therefore unlikely to accurately model rumen acid–base dynamics. Therefore, an in vitro test system that utilises a rumen relevant organic acid such as acetic acid (pKa = 4.76) is likely to be a better indicator of in vivo buffering performance [
40]. pH and acid load are not static in the rumen; they are constantly changing based on the animal’s diet, eating pattern and production system. Therefore an estimation of the rate of acid addition from in vivo data on pH changes is likely to yield a better estimate of in-practice in vivo performance. The three-hour acidotic diet simulation using acetic acid developed in this work was optimised using data obtained from an in vivo trial using cannulated dairy cows [
15]. The simulation was found to provide a different ranking of materials to the ABC/BUF values and the 8 h fixed HCl acid load methodology. The main difference in the comparison of the materials was the reduction in the predicted buffering capacity of SB in the acidotic diet simulation. The predicted reduced buffering capacity of SB is likely associated with its pKa (SB pKa = 6.3) [
41]. A pKa of SB of 6.3 means that its buffering capacity will become limited at pH levels below 6.3. When SB is not in excess of the rumen acids and the rumen pH is less than the pK
a, it starts to convert to dissolved carbon dioxide (dCO
2; with carbonic acid as an intermediate species) [
37,
41]. Additionally, Kohn and Dunlap describe the influence of the partial pressure of carbon dioxide (pCO
2), where it increases in systems from which CO
2 escape is held up and CO
2 is forced back into solution, which results in lowering of the pH [
10]. Experimental data to validate the build-up of dCO
2 and the subsequent decrease in the rumen pH have been reported, with dCO
2 recognised as being more dominant than VFAs in rumen acidosis [
41]. The increase in dissolved CO
2 has also been shown to be independent of the pCO
2 and dependent on other rumen fluid factors, such as fluid viscosity, surface tension and temperature. Therefore, current research underestimates rumen dCO
2 concentrations, because it assumes a linear relationship with pCO
2 for indicating the concentration and because it employs in vitro conditions. However, liquid CO
2 species are the main source of rumen CO
2, and Henry’s law cannot predict dCO
2 in a non-ideal rumen fluid [
42].
The acidotic diet simulation highlights the limitations of SB as a rumen buffer, which are not demonstrated by the HCl-based methodologies but are correlated with in vivo data [
15]. The AUC values from the in vivo data presented in
Table 7 highlight the correlation of the acidotic diet simulation over an initial 3 h period after the first feeding period of the cows in the day. However there are sizeable differences between the calculated in vivo AUC (14.3 mmol H
+.s) and in vitro AUC (6.63 mmol. H
+.s) for SB, which are likely associated with the starting rumen pH (6.2 and 5.8, respectively).