Richard Paratte, Ph.D. — Global Ruminant Technical & Innovation Manager, Vetagro S.p.A.
Bioavailability is not a fixed property of a product; it is the measured outcome of a release-and-absorption process, captured within a defined analytical window, for a product with a specific intestinal kinetic profile. Comparisons between rumen-protected amino acid products are routinely reduced to a single figure — a bioavailability percentage obtained with one method, on one reference, in one sampling window — as if that number were an intrinsic property of the product, like its methionine content. It is not: the number depends as much on the method as on the product, affecting how technologies are ranked and how decisions are justified on farm. The discussion centers on methionine.
What Each Method Actually Measures
The methods fall into two groups. The first group measures a signal that depends on the timing of amino acid appearance in blood, not only on the total amount absorbed. In vitro and in situ methods estimate ruminal bypass and intestinal digestibility within a fixed incubation window (6 vs. 8 h changes the estimate) and do not measure absorption. The plasma free amino acid (FAA) dose-response method (Whitehouse et al., 2017, 2024) compares the plasma response slope of the test product to a reference (either free methionine infused into the abomasum, and therefore fully absorbed, or a commercial product with an assumed bioavailability), sampling at 2–8 h post-feeding; it was built and validated on fast-release polymeric products. The AUC method (Graulet et al., 2005) integrates the plasma concentration-time curve within the sampling window, so lag and window length set the result. The second group integrates over days rather than hours. The selenomethionine (Se-Met) method (Weiss and St-Pierre, 2009) uses Se-yeast selenium as a methionine tracer: more absorbed methionine dilutes the tracer and lowers milk Se per unit of milk N. Sampling is a multi-day steady-state composite, so the method integrates over time rather than reading a peak; its estimates are relative to an assumed bioavailability for the reference product. Stable isotope dilution (Estes et al., 2018) infuses a labeled amino acid and estimates the fraction reaching plasma, corrected for splanchnic first-pass extraction; it is the most direct approach. Only the first group is sensitive to release kinetics by design. Each method answers a different biological question, so divergent results are not necessarily errors. But when a peak-based method ranks products, part of that ranking reflects how fast each product releases methionine, not how much methionine it delivers in total.
Figure 1. Main methods for evaluating bioavailability of rumen-protected amino acids: what each measures and its main limitation.
Polymeric Coating vs. Lipid Matrix
Products are not built the same way; structure determines release kinetics. Polymeric or pH-sensitive coatings remain stable at ruminal pH and release the amino acid in the acidic abomasum, producing a sharp, concentrated plasma peak (Graulet et al., 2005; Kihal et al., 2021) — the profile the plasma FAA method targets. Microencapsulated lipid-matrix products, such as Timet® (Vetagro S.p.A.), are designed to release amino acid progressively as the lipid matrix is digested in the small intestine; this would produce a lower, longer-duration plasma signal even when total absorption is comparable or greater. A third paradigm — HMBi methionine analogs — is absorbed largely through the rumen wall, by passive diffusion or monocarboxylate transport, peaking at 1.4 h (Kihal et al., 2021).
Figure 2. Schematic plasma methionine kinetics: fast-release, polymeric-coated products (rapid, sharp peak) vs. lipid-matrix products such as Timet® (slow, distributed release that short-window methods are not built to capture).
Why the Sampling Window Matters
A slow-release product cannot generate a plasma peak comparable to a fast-release one: amino acids are cleared from plasma in proportion to their concentration (Estes et al., 2018), so the concentration at any moment tracks the rate of absorption at that moment. Distributed absorption never reaches a high enough instantaneous concentration to register within a short window. Whitehouse et al. (2024) showed that, for a fast-release product, estimates were consistent across 2–8 h, 10–16 h, 18–24 h and 2–24 h windows; lipid-matrix products have been evaluated only within the standard 2–8 h window (Whitehouse et al., 2020), never with an extended one. Graulet et al. (2005) documented a 10-hour lag even for a pH-sensitive product; a short window cannot capture a slower, distributed release.
When Methods Diverge, the Final Word Is Biological
Two Latin-square studies show how loosely analytical ranking maps onto production response. Whitehouse et al. (2020) fed four rumen-protected methionine products at 24 g/d methionine; relative to the polymeric reference product, the other three were estimated at 9%, 24% and 35% effectiveness in supplying metabolizable methionine; Timet® (Vetagro) was the product estimated at 24%. Milk yield and dry matter intake did not differ, although 7-day periods are too short to test production responses. Parales-Giron et al. (2022) compared four products at 15 g/d metabolizable methionine using the Se-Met method; estimates again diverged sharply. DMI, milk yield, and fat did not differ (P > 0.10). Timet® (Vetagro), a lipid-matrix product, showed lower milk protein content and milk N (P < 0.05) than two of the three polymeric products, with the third statistically intermediate, plus a trend toward lower protein yield (P = 0.08); that lower N mechanically inflates the Se/N ratio the bioavailability estimate is built on, even though milk Se itself did not differ among treatments (P = 0.40). Timet® was assigned 37.4% bioavailability there — a relative figure that rests on an assumed 80% for the reference product and on a pre-trial rather than concurrent reference Se/N. Evaluated instead using the same stable isotope dilution technique (Estes et al., 2018), internal trial data on Timet® show 71.78% plasma methionine appearance and 76.80% bioavailability after correcting for 7% splanchnic extraction — consistent with production data showing gains in milk, protein yield (~50 g/d) and ECM (King et al., 2021), and parallel gains in milk protein and fat, with a numerical but non-significant reduction in milk urea nitrogen (P = 0.43; Sáinz de la Maza et al., 2019). Isotopic estimate and production data point the same way.
Figure 3. Timet® (Vetagro), a commercially available lipid-matrix product: plasma FAA dose-response (Whitehouse et al., 2020: 24% relative to the polymeric reference product) vs. stable isotope dilution (internal trial data: 76.80%).
Practical Implications
Three practical conclusions follow. First, a bioavailability figure is meaningless without its method, sampling window, and reference product. Second, peak-based methods (plasma FAA dose-response, AUC) are valid within a release profile (fast, polymeric) but not across structurally different technologies, where they rank release kinetics rather than total amino acid supply; steady-state and isotope methods are less sensitive to release kinetics, but their relative estimates depend on the bioavailability assumed for the reference product. Third, when technologies diverge, the best comparison integrates digestive potential (in vitro/in situ), systemic appearance (plasma/isotopic tracers), and functional outcomes (protein yield, nitrogen efficiency). Economically, cost per gram of metabolizable methionine is only as valid as the bioavailability estimate behind it; a ranking from a mismatched method should not underpin it. When methods disagree, functional outcomes remain the most reliable arbiter.
References
Estes, K. A., et al. 2018. J. Dairy Sci. 101:7040–7060.
Graulet, B., et al. 2005. J. Dairy Sci. 88:3640–3649.
Kihal, A., et al. 2021. JDS Communications 2:182–185.
King, L., et al. 2021. Animals 11(12):3545.
Parales-Giron, J., et al. 2022. J. Dairy Sci. 105 (Suppl. 1):Abstr. 2146M.
Sáinz de la Maza, V., et al. 2019. J. Dairy Sci. 102 (Suppl. 1):350 (Abstr. W85).
Weiss, W. P., and N. R. St-Pierre. 2009. J. Dairy Sci. 92:2835–2842.
Whitehouse, N. L., et al. 2017. J. Dairy Sci. 100:9585–9601.
Whitehouse, N. L., et al. 2020. J. Dairy Sci. 103 (Suppl. 1):122 (Abstr. 315).
Whitehouse, N. L., et al. 2024. JDS Communications 5:539–542.
Timet® is a registered trademark of Vetagro S.p.A. All other trademarks are the property of their respective owners.
