8,160
edits
Changes
From Fellrnr.com, Running tips
no edit summary
Lactate Threshold is a key component of running performance, and is a better predictor of race performance than [[VO2max|V̇O<sub>2</sub>max]]. You can think of Lactate Threshold can be thought of a as reflecting a change from largely aerobic exercise to largely anaerobic exercise. Lactate Threshold is often used to determine the correct pace for [[Tempo Runs]], though the science indicates shows such training is ineffective at best. However, Lactate Threshold provides an excellent way of monitoring the effectiveness of your training, and provides an objective estimate of your race pace. Unfortunately, measuring an athlete's Lactate Threshold is time -consuming and expensive, with the gold standard MLSS test requiring three to five 30 minute tests on separate days. All other approaches to measuring an athlete's Lactate Threshold seem deeply flawed. A better approach to your anaerobic threshold is [[Critical Power]].
=What is Lactate?=
''Main article: [[Lactate]]''
At one time , athletes viewed [[Lactate]] was viewed as a harmful waste product due of anaerobic exercise, but research since the early 2000s has shown that Lactate is an intermediate fuel in the metabolism of carbohydrates. Muscles will burn Lactate in preference to Glucose, and will convert Lactate back to Glucose at rest. The level of Lactate in the blood is primarily dependent depends on exercise intensity, rather like [[Heart Rate]]. Lactate is used as a fuel source by for working muscles, and injecting extra lactate into the blood results in increased lactate metabolism and carbohydrate sparing<ref name="MillerFattor2002"/> without impairing performance<ref name="EllisSimmons2009"/>. Note that Lactate forms Lactic Acid in the blood, and literature uses the terms are used interchangeably.
=What is the Lactate Threshold?=
The Lactate Threshold (LT) is the point at which the lactate level in the your blood will rise even if you keep the work intensity is kept constant. This is sometimes can be referred to as the Anaerobic Threshold (AT), or the Onset of Blood Lactate Accumulation (OBLA), though the most accurate term is Maximal Lactate Steady State (MLSS). Even within the scientific community terminology is confusing<ref name="Binder-2008"/>. It is sometimes believed claimed that the MLSS represents the maximum clearance of Lactate, but this may not be the case<ref name="EllisSimmons2009"/>. Note that you normally measure Lactate is normally measured in the blood streambloodstream, so the Lactate level reflects the net of the muscles releasing and absorbing Lactate. Lactate Threshold is important as it is an excellent good predictor of race performance<ref name="Billat-1996"/><ref name="Palmer-1999-perform"/><ref name="Baldari-2000"/><ref name="Jones-2000"/><ref name="Lehmann-1983"/><ref name="Tanaka-1984"/>, and may be a better predictor than [[VO2max|V̇O<sub>2</sub>max]]<ref name="Allen-1985"/>. You can think of Lactate Threshold can be thought of as the percentage of VO2max that you can be maintained maintain for a protracted time<ref name="Costill-1973"/>, though it. (It's not clear what the limiting factor is for exercise above the Lactate Threshold<ref name="Baron-2008"/>.)
=Lactate Threshold Training =
''Main article: [[Tempo Runs]]''
There is good evidence that endurance training changes Lactate Threshold. However, the idea that training at threshold intensity, such as [[Tempo Runs]], is particularly effective has no evidence<ref name="Beneke-2011"/>"/><ref name="Guellich-2010"/>, and polarized training is a better approach<ref name="StögglSperlich2014"/><ref name="Muñoz-2014"/>. For trained athletes, Tempo runs are ineffective<ref name="Londeree-1997"/> and may actually be counterproductive<ref name="Evertsen-2001"/><ref name="Guellich-2010"/>. See [[Tempo Runs]] for more details. Detraining shifts will reduce the your Lactate Threshold<ref name="Coyle-1985"/>, and your Lactate levels can be higher at a given intensity after just a few days without training<ref name="Mujika-2001"/>, suggesting rapid detraining effects. The improvements in Lactate Threshold pace are largely due to because of a greater rate of Lactate removal rather than a reduced rate of production<ref name="Phillips-1995"/><ref name="MacRae-1992"/><ref name="Donovan-1989"/><ref name="Donovan-1983"/><ref name="Bergman-1999"/>.
=The Usefulness of Lactate Threshold=
One of the primary goals of Lactate Threshold testing is has been to determine the correct pace for [[Tempo Runs]]. However, even if Tempo training is ineffective , there are two good reasons for knowing your Lactate Threshold. FirstlyFirst, monitoring your Lactate Threshold is a great way of for evaluating the effectiveness of a training regime. SecondlySecond, Lactate Threshold can be used to validate race pace goals. If your Lactate Threshold suggests indicates a much faster race pace than you've been able to achieve, it suggests either a lack of resistance to muscular damage or a lack of mental fortitude. If the Lactate Threshold suggests a race pace that is slower than your target goal, it suggests you objective your goal is wrong and you should aim for a slower finish time. This is especially important in the marathon, where "hitting the wall" is a common issue.
=Lactate Curve=
It is common to plot exercise intensity against the lactate level to produce a blood lactate curve similar to the one below, showing an exponential rise in lactate level with intensity. It's generally accepted that a shift of the curve to the right indicates shows an improved athletic performance<ref name="Bosquet-2002"/><ref name="Yoshida-1990"/>, and training can improve performance because of this shift without a change in aerobic capacity ([[VO2max|V̇O<sub>2</sub>max]])<ref name="Acevedo-1989"/>. There is some limited evidence from radio-isotope studies in animals that a benefit of endurance training may be in Lactate clearance<ref name="Donovan-1983"/>. Note that above Above the Lactate Threshold , the Lactate level is not at a steady -state, but rises even though the intensity remains constant, so the typical curve that is shown below is rather misleading. Some Lactate Curves are plotting Blood Lactate against time during an Incremental Power Test (see below), which is more reasonable, but can is still be rather misleading.
[[File:LactateThreshold.png|none|thumb|500px|A blood lactate curve with lactate level plotted against intensity. Note that above the Lactate Threshold there is no fixed relationship between Lactate and work intensity, so the curve is misleading at best. ]]
=Determining Your Lactate Threshold=There are various ways of determining the Lactate Threshold, each with their own problems. * '''Critical Power'''. A better alternative is to ignore Lactate and focus on power output. With [[Stryd]] this can be applied to running, though Critical Power tests for runners may have a higher risk for injury. Before Stryd, the approach was to use "Critical Speed" on level ground. See [[Critical Power]] for details. * '''MLSS'''. The gold standard test for Lactate Threshold (and the only one that appears to be valid) is to measure Maximal Lactate Steady State (MLSS). The test requires 3 to 5 constant intensity trials of at least 30 minutes' duration, each performed on separate days. Each test is at a different exercise intensity, and the highest intensity that does not have a rise in blood lactate in the last 20 minutes is the MLSS. While this is the best way of determining the Lactate threshold, it's obviously time -consuming and interferes with the athlete's regular training. Because of Even though MLSS is the best approach to determining the effort of Lactate Threshold, there are issues with MLSS testing, various shortcuts have been triedand Critical Power is probably a better approach. * '''Fixed Blood Lactate Accumulation'''. A simple approach is to assume that the Lactate Threshold always occurs at the same Lactate level. Sadly, this assumption is wrong, as Lactate Threshold can occur at vastly different levels. * '''Lactate Patterns'''. There are various Some approaches that look at the pattern of change in Lactate in an attempt to create a simple test. So far, I've seen little evidence to support any of these approaches. * '''Heart Rate Deflection'''. An indirect way of finding the Lactate Threshold is to look for the [[Heart Rate Deflection]], sometimes called the "Conconi test". This test only requires a heart rate monitor to perform rather than blood draws, so it is much easier than the above approaches. However, the validity of the Conconi test has many issues and seems of dubious value<ref name="Cook2011"/>. See [[Heart Rate Deflection]] for details.
* '''Respiratory gasses'''. Another method for estimating Lactate Threshold is to measure the respiratory gasses<ref name="Binder-2008"/><ref name="Laplaud-2006"/>, but given this is impractical for most athletes, it's not covered here.
==Lactate Threshold & Maximal Lactate Steady State ==
The best approach to determine an athlete's Lactate Threshold is to measure the Maximal Lactate Steady State (MLSS)<ref name="Faude-2009"/>. The test is actually several constant load trials of at least 30 minutes' duration on different days at various exercise intensities (in the range of between 50–90% [[VO2max|V̇O<sub>2</sub>max]]. The highest workload that results in an increase of less than 1 mmol/L of lactate between the 10 and 30 minute mark defines the MLSS<ref name="Urhausen-1993"/><ref name="De SouzaGrossl2012"/><ref name="Beneke-2003"/>. Lactate is typically measured using a blood sample, either using a pinprick or a catheter. Note that MLSS for a given an individual will vary by sport<ref name="Figueira-2008"/>, probably based on the mass of muscle engaged<ref name="Beneke-1996"/><ref name="Beneke-2003-2"/>. The difficulty of performing this test makes it impractical in most situations. Researchers have raised concerns that even when performed correctly, there are issues with MLSS and suggest that [[Critical Power]] is a better approach<ref name="JonesBurnley2019"/>. The primary issue is that MLSS doesn't truly represent the upper limit of aerobic capacity. [[File:LactateMLSS.jpg|none|thumb|500px|Blood lactate levels over plotted against time at various different workloads. The lower lines are at lower intensities, with the line marked MLSS being the highest intensity that produces stable lactate levels.]]
==Lactate Threshold & Incremental Power Test ==
A common approach to determine the Lactate Threshold is the Incremental Power Test. The subject exercises in stages of increasing intensity, with lactate measured at the end of each stage, with stages typically lasting 3 to 10 minutes. However, blood lactate takes 20-30 minutes to stabilize for a given an intensity<ref name="Billat-1996"/>. This means that the incremental power test tends to be is of limited value<ref name="Foxdal-1996"/>, with 3 minute stages giving low reproducibility<ref name="Morton-2012"/>, the stage length changing the lactate values<ref name="Faude-2009"/>, and even longer stages lengths of 8 minutes having low reproducibility<ref name="Gavin-2014"/>. The lactate level can drop between the 4<sup>th</sup> and 12<sup>th</sup> minute of exercise at a constant intensity<ref name="RieuMiladi1989"/>. Some have suggested using the lactate value measured as an indication a sign of the prior stage's intensity, as it takes more than over 3 minutes for lactate to stabilize<ref name="Orok-1989"/>, but this rather arbitrary approach can might be used as a guideline (at best)<ref name="Baldari-2000"/>. For running, it is common to pause the exercise for 30 seconds to take a blood sample. These breaks only make a non-significant difference to the testing, though the slight difference tends to be greater at higher intensities<ref name="GullstrandSjüdin2007"/>.
==Lactate Threshold & Fixed Blood Lactate Accumulation==
Because MLSS is time consuming and expensive, a shortcut is often used to estimate MLSS by assuming that it occurs at a fixed Lactate Level (Fixed Blood Lactate Accumulation, or FBLA)<ref name="De SouzaGrossl2012"/>, unusually 4.0 mmol/l<ref name="Heck-1985"/> though sometimes 3.5 mmol/l<ref name="DenadaiFigueira2004"/><ref name="Denadai-2005"/>. However, while the MLSS may average around 4.0 mmol/l<ref name="Denadai-2005"/>, there are significant differences for individuals<ref name="Mamen-2009"/>, with variations between 3.0 and 5.5 in small sample sizes<ref name="Heck-1985"/> and has been shown to have a range as wide as 2.0 to 10.0 mmol/l<ref name="Faude-2009"/><ref name="Stegmann-1981"/>. This approach also typically uses a blood test, but in some sports (like running), the athlete has to pause to have a pinprick blood sample taken, further confusing the test<ref name="Heck-1985"/>. The term "Individual Anaerobic Threshold" (IAT) has been used to emphasize that the Lactate Threshold is specific to each individual rather than using a FBLAFixed Blood Lactate Accumulation, though this can refer to a specific protocol for estimating MLSS<ref name="Stegmann-1981"/>. The Fixed Blood Lactate Accumulation is sometimes called "Onset of Blood Lactate Accumulation" (OBLA)<ref name="Mamen-2009"/>, a particularly misleading term in this context.
==Lactate Threshold Estimation From Lactate Patterns==
There have been several approaches to determining MLSS without the difficulty of the full protocol<ref name="Harnish-2001"/><ref name="Palmer-1999-1day"/><ref name="Tegtbur-1993"/><ref name="Dickhuth-1999"/><ref name="Urhausen-1993"/>, but their validity is limited<ref name="Kilding-2005"/><ref name="De SouzaGrossl2012"/><ref name="Beneke-1995"/><ref name="Jones-1998"/>. These approaches generally look for some pattern in the change of Lactate level. * For example, one approach called the "Lactate Minimum Speed Test" (LMST) uses an initial sprint to elevate blood lactate followed by an incremental power test<ref name="SoteroPardono2009"/><ref name="Sotero-2009"/><ref name="MiyagiLeite2013"/>. However, the effectiveness of the LMST is profoundly impacted by the starting speed of the incremental portion of the test<ref name="Carter-1999"/>and so the results may be coincidence<ref name="Carter-2000"/>. This is not entirely surprising given the initial sprint phase disrupts the metabolism<ref name="Carter-2000"/>. * Some trivial approaches have been tried, such as looking for a 1 mmol/l increase followed by another 1 mmol/l increase<ref name="CTS"/>. So for an athlete performing an incremental load test with [[Lactate]] readings of 1.7, 2.3, 2.6, 3.7, & 5.6 the conclusion would be their Lactate Threshold is 3.7 (3.7 is more than over 1.0 more than over 2.6 and followed by another increment of more than over 1.0.) However, given that most portable Lactate meters have a Typical Error of 0.4-1.0 mmol/l<ref name="Tanner-2010"/>, a fractional error in the reading gives a different result. In the previous example, if the 2.6 reading was 2.8, then the Lactate Threshold would jump from 3.7 to 5.6. * Part of the problem with these approaches may be that MLSS may not represent the point of maximum lactate clearance<ref name="Beneke-2003-2"/>, as injecting additional lactate into the blood of athletes exercising above MLSS did not significantly increase lactate levels<ref name="EllisSimmons2009"/>.* Some of the tests could be "p-hacking", where the study looks at a sufficiently large number of variables that some correlation occurs randomly<ref name="HeadHolman2015"/>.
[[File:LactateComp.jpg|none|thumb|500px|The correlation (or lack thereof) between MLSS and the lactate levels at the MLSS intensity seen during 3 or 5 minute incremental power tests<ref name="Heck-1985"/>.]]
* One approach that looks promising uses three tests to estimate MLSS<ref name="Billat-1994"/>. First, a standard incremental test is used to give a rough estimate of MLSS. Then two 30 minute tests are performed, one above and one below the rough estimate of MLSS. The relative difference in the rise between the two tests is then used to estimate the crossover point. For instance , assume running at 7:00 min/mile produced a blood lactate level that fell from 4 mmol/l at 5 min to 3 mmol/l at 20 min, a 1 mmol/l drop. Then a run at In the run at 6:20 min/mile the blood lactate rose from 4.0 mmol/l at 5 min to 6.5 mmol/l at 20 min, a 1.5 mmol/l rise. The interception point would then be about a an MLSS pace of 6:26 min/mile. This is not much less effort than the full MLSS test, but it is an improvement. ==Lactate Threshold and Near -Infrared Spectroscopy==A promising technology for measuring Lactate Threshold is Near-infrared spectroscopy (NIRS) which shines infrared light into the skin above an active muscle and measures the reflected light. NIRS measures the oxygen saturation in the capillaries of the muscle and has the potential to test for Lactate Threshold without any with no blood sampling. Because NIRS can monitor continually, it is possible that it may be able to determine the Lactate Threshold during an incremental test rather than requiring the multiple tests of MLSS.
===Introduction to NIRS and SmO2===
Near-infrared spectroscopy (NIRS) has been shown to measure the oxygen saturation of blood in muscle (SmO<sub>2</sub>) or other body tissues (StO<sub>2</sub>)<ref name="Kek-2008"/><ref name="Torricelli-2004"/><ref name="Mancini-1994"/>. (This works on similar principles to a [[Pulse Oximeter]].) Medical NIRS systems for monitoring StO<sub>2</sub> use Infrared LED or Lasers at 2, 3, or 4 frequencies<ref name="Hyttel-Sorensen-2011"/>. SmO<sub>2</sub> reflects the balance of oxygen delivery and consumption during exercise<ref name="Chance-1992"/>. There are some initial indications that relative SmO<sub>2</sub> may reflect changes in performance capacity<ref name="Neary-2005"/>. There is generally a four -phase response of smo2 during incremental exercise from rest to maximum intensity and the following recovery<ref name="Belardinelli-1995a"/>:
# An initial increase in SmO<sub>2</sub> above resting levels to supply the now active muscles. (This may be due to increased blood flow<ref name="Bhambhani-1997"/>, but computer models do not support this<ref name="Fuglevand-1997"/>.)
# SmO<sub>2</sub> decreases linearly or exponentially with increasing intensity, followed by a leveling off as the subject approaches maximum intensity. There is some evidence of a breakpoint where the rate of decline increases (see below).
# During the first 1-2 minutes of recovery , there is a rapid increase in SmO<sub>2</sub> which usually exceeds resting levels.
# SmO<sub>2</sub> then declines to resting levels over a further few minutes.
Skin should not impact SmO<sub>2</sub> readings more than 5%<ref name="Hampson-1988"/>, but surface fat can interfere with smo2<ref name="van Beekvelt-2001"/><ref name="Homma1996"/><ref name="BenaronMatsushita1998"/><ref name="BenaronYamamoto1998"/>. Because the penetration depth of NIRS is about 50-60% of the distance between the emitter and receiver, the site must be selected so that the fat layer is much thinner than this depth<ref name="Bhambhani-2004"/>. (It's been suggested that SmO2 is probably only viable in lean individuals.)
As the intensity increases during incremental exercise SmO<sub>2</sub> will remain constant or decline, with the rate of decline being greater near the Lactate Threshold<ref name="Belardinelli-1995a"/><ref name="Bhambhani-2004"/><ref name="Belardinelli-1995b"/>. This has led to several studies using the concept of an SmO<sub>2</sub> "breakpoint"<ref name="Grassi-1999"/>. This breakpoint is a change in the slope of the line plotting SmO<sub>2</sub> against work intensity in an incremental intensity test. This increase in the rate of desaturation can either be visually determined or based on bilinear regression. (The bilinear regression iterates over different combinations of two regression lines to find the lowest sum of squares of the residuals. I could not find the details of the constraints placed on this approach.)
[[File:SmO2 Breakpoint1.jpg|none|thumb|500px|A graph showing the SmO<sub>2</sub> breakpoint.]]
Another approach used by<ref name="Snyder-2009"/> was defined as the workload immediately prior to before a drop of 15% that lead to a continuous decline in SmO<sub>2</sub>. This is shown in the image below, showing a recording with and without a defined breakpoint.
[[File:SmO2 Breakpoints With-Without.jpg|center|thumb|300px|]]
===SmO<sub>2</sub> and Lactate Threshold===
* A study of 40 sedentary undergraduates showed a correlation between SmO<sub>2</sub> returning to resting levels and Ventilatory Threshold (VT) in 65% of subjects during an incremental cycling test<ref name="Bhambhani-1997"/>. While the text refers to Lactate Threshold as the point at which Lactate rises above resting levels (aerobic threshold) the method used to determine VT appears to be the anaerobic threshold. This study did not use the breakpoint mentioned above, but the point where the SmO<sub>2</sub> drops below the level detected at rest.
* A study of 11 subjects of varying fitness levels showed a correlation between SmO<sub>2</sub> breakpoint determined visually and Ventilatory Threshold (VT) during an incremental cycling test<ref name="Belardinelli-1995a"/>.
* A comparison between 12 healthy subjects and 7 suffering from chronic heart failure (CHF) showed a correlation between the SmO<sub>2</sub> breakpoint and Ventilatory Threshold (VT) during an incremental cycling test<ref name="Belardinelli-1995c"/>.
* A 2012 study compared the results from NIRS on the calf (Gastrocnemius Lateralis) and quads (Vastus Lateralis) in 31 active but not highly trained college students during an incremental cycling test<ref name="Wang-2012"/>. The study found a correlation between Lactate Threshold (determined from the log-log method<ref name="Davis-2007"/>) and the SmO<sub>2</sub> breakpoint (determined from bi-linear regression) in both locations, but the quads corresponded better. (I suspect the results from running could be quite different.)
* A 2009 study used MLSS (the gold standard for Lactate Threshold) with running to evaluate test NIRS<ref name="Snyder-2009"/>. The 16 athletes performed between 2 and 5 tests of 30 minutes each to determine MLSS, separated by at least 48 hours each. The subjects then performed an incremental treadmill test using 6 -minute stages with the 4th stage at the pace they estimated they could maintain for an hour (around Lactate Threshold). The first 3 stages where then 0.66, 0.44, & 0.22 meter/second slower, and the subsequent stages were 0.22 meters/second faster each time. SmO<sub>2</sub> breakpoint was defined as the workload immediately prior to before a drop of 15% that lead to a continuous decline in SmO<sub>2</sub>. A Lactate breakpoint was also determined based on the incremental test using the workload prior to before an increase of 1 mmol/l as the criteria. Both the SmO<sub>2</sub> and the Lactate breakpoint were determined visually. Of the 16 subjects, 1 did not reach MLSS, 2 did not have both a SmO<sub>2</sub> breakpoint or a Lactate breakpoint (based on the criteria used) and 1 did not have eitherhad neither. The study found that SmO<sub>2</sub> is as effective as Lactate breakpoint tests for determining true Lactate Threshold (MLSS). The table below shows the values for each of 12 subjects, with the paces shown as KPH, then min/mile, then the error as a percentage. This shows that while smo2 is as good as the lactate breakpoint, the individual differences from MLSS are not insignificant. For instance, subject 5 had an MLSS of 6:21, but an SmO<sub>2</sub> breakpoint of 6:57 and a lactate breakpoint of 6:59, which is a big difference.
{| class="wikitable" style="margin-left: auto; margin-right: auto; border: none;"
! Subject
| 12.85
| 12.75
| | | | |
|-
| SD
| 1.51
| 1.55
| | | | |
|}
===SmO2 and MLSS===
[[File:SmO2 MLSS Test.jpg|center|thumb|300px| A graph of lactate and SmO<sub>2</sub> above and below the MLSS threshold.]]
===Thoughts on SmO2 and Lactate Threshold===
I think that the currently available research indicates shows that NIRS and SmO<sub>2</sub> might hold promise for simplifying the measurement of Lactate Threshold. However, the research is at a fairly early stage, with only one study using the gold standard of MLSS, and the results are mixed at best. In some ways, I feel the MLSS test above indicates shows that SmO<sub>2</sub> is a poor option for evaluating an athlete's Lactate Threshold, but it could equally be argued that perhaps most existing approaches other than a full MLSS test are equally flawed. Currently , there are two consumer products available; [[BSX]] and [[Moxy]]. BSX is a fully automated approach to analyzing the data and estimating Lactate Threshold, whereas the Moxy is intended to provide the end-user with the underlying data to evaluate. (BSX is being discontinued.)=Factors That May Influence An Athlete's Lactate Threshold=
There are a few factors that may change the Lactate Threshold (other than training)
* Because lactate is produced from the metabolism of carbohydrate, a reduction in carbohydrate intake (or [[Glycogen]] depletion) will shift the lactate curve to the right<ref name="Reilly-1999"/><ref name="Yoshida-1984"/><ref name="Maassen-1989"/><ref name="McLellan-1989"/>.
<ref name="Kilding-2005">AE. Kilding, AM. Jones, Validity of a single-visit protocol to estimate the maximum lactate steady state., Med Sci Sports Exerc, volume 37, issue 10, pages 1734-40, Oct 2005, PMID [http://www.ncbi.nlm.nih.gov/pubmed/16260974 16260974]</ref>
<ref name="Harnish-2001">CR. Harnish, TC. Swensen, RR. Pate, Methods for estimating the maximal lactate steady state in trained cyclists., Med Sci Sports Exerc, volume 33, issue 6, pages 1052-5, Jun 2001, PMID [http://www.ncbi.nlm.nih.gov/pubmed/11404673 11404673]</ref>
<ref name="Palmer-1999-1day">AS. Palmer, JA. Potteiger, KL. Nau, RJ. Tong, A 1-day maximal lactate steady-state assessment protocol for trained runners., Med Sci Sports Exerc, volume 31, issue 9, pages 1336-41, Sep 1999, PMID [http://www.ncbi.nlm.nih.gov/pubmed/10487377 10487377]</ref>
<ref name="Tegtbur-1993">U. Tegtbur, MW. Busse, KM. Braumann, Estimation of an individual equilibrium between lactate production and catabolism during exercise., Med Sci Sports Exerc, volume 25, issue 5, pages 620-7, May 1993, PMID [http://www.ncbi.nlm.nih.gov/pubmed/8492691 8492691]</ref>
<ref name="Beneke-2003">R. Beneke, Methodological aspects of maximal lactate steady state-implications for performance testing., Eur J Appl Physiol, volume 89, issue 1, pages 95-9, Mar 2003, doi [http://dx.doi.org/10.1007/s00421-002-0783-1 10.1007/s00421-002-0783-1], PMID [http://www.ncbi.nlm.nih.gov/pubmed/12627312 12627312]</ref>
<ref name="Foxdal-1996">P. Foxdal, A. Sjödin, B. Sjödin, Comparison of blood lactate concentrations obtained during incremental and constant intensity exercise., Int J Sports Med, volume 17, issue 5, pages 360-5, Jul 1996, doi [http://dx.doi.org/10.1055/s-2007-972861 10.1055/s-2007-972861], PMID [http://www.ncbi.nlm.nih.gov/pubmed/8858408 8858408]</ref>
<ref name="HeadHolman2015">Megan L. Head, Luke Holman, Rob Lanfear, Andrew T. Kahn, Michael D. Jennions, The Extent and Consequences of P-Hacking in Science, PLOS Biology, volume 13, issue 3, 2015, pages e1002106, ISSN [http://www.worldcat.org/issn/1545-7885 1545-7885], doi [http://dx.doi.org/10.1371/journal.pbio.1002106 10.1371/journal.pbio.1002106]</ref>
<ref name="Palmer-1999-perform">GS. Palmer, LB. Borghouts, TD. Noakes, JA. Hawley, Metabolic and performance responses to constant-load vs. variable-intensity exercise in trained cyclists., J Appl Physiol (1985), volume 87, issue 3, pages 1186-96, Sep 1999, PMID [http://www.ncbi.nlm.nih.gov/pubmed/10484594 10484594]</ref>
<ref name="Jones-2000">AM. Jones, H. Carter, The effect of endurance training on parameters of aerobic fitness., Sports Med, volume 29, issue 6, pages 373-86, Jun 2000, PMID [http://www.ncbi.nlm.nih.gov/pubmed/10870864 10870864]</ref>
<ref name="Donovan-1983">CM. Donovan, GA. Brooks, Endurance training affects lactate clearance, not lactate production., Am J Physiol, volume 244, issue 1, pages E83-92, Jan 1983, PMID [http://www.ncbi.nlm.nih.gov/pubmed/6401405 6401405]</ref>
<ref name="Donovan-1983">CM. Donovan, GA. Brooks, Endurance training affects lactate clearance, not lactate production., Am J Physiol, volume 244, issue 1, pages E83-92, Jan 1983, PMID [http://www.ncbi.nlm.nih.gov/pubmed/6401405 6401405]</ref>
<ref name="Bergman-1999">BC. Bergman, EE. Wolfel, GE. Butterfield, GD. Lopaschuk, GA. Casazza, MA. Horning, GA. Brooks, Active muscle and whole body lactate kinetics after endurance training in men., J Appl Physiol (1985), volume 87, issue 5, pages 1684-96, Nov 1999, PMID [http://www.ncbi.nlm.nih.gov/pubmed/10562610 10562610]</ref>
<ref name="JonesBurnley2019">Andrew M. Jones, Mark Burnley, Matthew I. Black, David C. Poole, Anni Vanhatalo, The maximal metabolic steady state: redefining the 'gold standard', Physiological Reports, volume 7, issue 10, 2019, pages e14098, ISSN [http://www.worldcat.org/issn/2051-817X 2051-817X], doi [http://dx.doi.org/10.14814/phy2.14098 10.14814/phy2.14098]</ref>
</references>