Advertisement

Comparison of different test protocols to determine maximal lactate steady state intensity in swimming

      Abstract

      Objectives

      This study compared step test, lactate minimum (LM) test and reverse lactate threshold (RLT) test protocols with maximal lactate steady state (MLSS) in free-swimming. All test protocols used fixed duration increments and high work-rate resolution (≤ 0.03 m·s−1) to ensure high sensitivity.

      Design

      Validation study.

      Methods

      23 swimmers or triathletes (12 male and 11 female) of different ages (19.0 ± 5.9 yrs) and performance levels (400 m personal best 1.38 ± 0.13 m·s−1, FINA points 490 ± 118) completed an incremental step test (+0.03 m·s−1 every 3 min) to determine speed at 4 mmol·L−1 and at modified maximal distance method, a LM test, a RLT test and two to five 30 min tests (±0.015 m·s−1) to determine MLSS. Following a 200 m all-out and a 5 min rest, LM was determined during an incremental segment (+0.03 m·s−1 every 2 min) as the nadir of the speed-lactate curve. After a priming segment with four increments (+0.06 m·s−1), RLT was determined as the lactate apex during a reverse segment (−0.03 m·s−1) every 3 min.

      Results

      The mean differences (± limits of agreement) to speed at MLSS were +1.0 ± 4.1% (speed at 4 mmol·L−1), +1.5 ± 3.5% (modified maximum distance method), −0.2 ± 4.7% (LM) and 2.0 ± 3.1% (RLT). All threshold concepts showed good agreement with MLSS pace (intraclass correlation coefficient ≥ 0.886).

      Conclusions

      Test protocols with a fixed step duration and fine increments allowed high accuracy in estimating MLSS pace. With similar criterion agreement to the LM and RLT tests, incremental step tests appear more practicable due to less prior knowledge required and derivation of individual training zones.

      Abbreviations:

      LMT (Lactate minimum test), RLT (Reverse lactate threshold)

      Keywords

      Practical Implications

      • Narrow speed and time resolution enable sensitive MLSS determination in swimming.
      • Test protocols with fixed step duration and fine increments allow high accuracy in estimating MLSS pace.
      • ST is more practicable than LMT and RLT test because little prior knowledge is required and individual training zones can be derived.

      1. Introduction

      In swimming, the spectrum of submaximal speeds is narrow due to the nonlinear relationship between active drag and swimming speed. Thus, slight increases in external workload (i.e., speed) may lead to marked alterations of internal workload, e.g., physiological alterations including blood lactate concentration (bLa), heart rate (HR) and oxygen uptake (V̇O2).
      • Pelarigo J.G.
      • Greco C.C.
      • Denadai B.S.
      • et al.
      Do 5% changes around maximal lactate steady state lead to swimming biophysical modifications?.
      ,
      • Espada M.C.
      • Reis J.F.
      • Almeida T.F.
      • et al.
      Ventilatory and physiological responses in swimmers below and above their maximal lactate steady state.
      For example, a speed increase from the first rise in bLa above baseline (LT1) of only 0.06 m·s−1 may already mark the maximal metabolic steady state (MMSS),
      • Greco C.C.
      • De Oliveira M.F.M.
      • Caputo F.
      • et al.
      How narrow is the spectrum of submaximal speeds in swimming?.
      which can be sustained almost entirely by oxidative phosphorylation and therefore still yields stable V̇O2 and bLa.
      • Iannetta D.
      • Ingram C.P.
      • Keir D.A.
      • et al.
      Methodological reconciliation of CP and MLSS and their agreement with the maximal metabolic steady state.
      Therefore, precise and sensitive testing measures are required to accurately determine training zones and performance development for each athlete.
      • Greco C.C.
      • De Oliveira M.F.M.
      • Caputo F.
      • et al.
      How narrow is the spectrum of submaximal speeds in swimming?.
      ,
      • Espada M.C.
      • Alves F.B.
      • Curto D.
      • et al.
      Can an incremental step test be used for maximal lactate steady state determination in swimming? Clues for practice.
      MMSS has often been estimated as the highest speed that can be maintained at a constant bLa, i.e., maximal lactate steady state (MLSS). When swimming slightly above the MLSS pace, marked physiological alterations, like increases in bLa,
      • Pelarigo J.G.
      • Greco C.C.
      • Denadai B.S.
      • et al.
      Do 5% changes around maximal lactate steady state lead to swimming biophysical modifications?.
      ,
      • Espada M.C.
      • Reis J.F.
      • Almeida T.F.
      • et al.
      Ventilatory and physiological responses in swimmers below and above their maximal lactate steady state.
      ,
      • Dekerle J.
      • Nesi X.
      • Lefevre T.
      • et al.
      Stroking parameters in front crawl swimming and maximal lactate steady state speed.
      minute ventilation,
      • Pelarigo J.G.
      • Greco C.C.
      • Denadai B.S.
      • et al.
      Do 5% changes around maximal lactate steady state lead to swimming biophysical modifications?.
      and V̇O2 approaching maximal oxygen uptake (V̇O2peak)
      • Espada M.C.
      • Reis J.F.
      • Almeida T.F.
      • et al.
      Ventilatory and physiological responses in swimmers below and above their maximal lactate steady state.
      have been reported. However, the MLSS concept has been criticized due to methodological aspects (e.g., arbitrary temporal criteria together with cutoff values for the permitted increase in bLa) and limited test sensitivity due to the inevitable selection of discrete workloads, which may lead to an underestimation of the ‘true’ MMSS.
      • Iannetta D.
      • Ingram C.P.
      • Keir D.A.
      • et al.
      Methodological reconciliation of CP and MLSS and their agreement with the maximal metabolic steady state.
      ,
      • Jones A.M.
      • Burnley M.
      • Black M.I.
      • et al.
      The maximal metabolic steady state: redefining the ‘gold standard.’.
      ,
      • Dotan R.
      Reverse lactate threshold: a novel single-session approach to reliable high-resolution estimation of the anaerobic threshold.
      Since direct MLSS determination is already time-consuming (i.e., at least two 30 min trials on separate days are necessary) and a higher resolution to refine MLSS determination would require even more sessions, various single-session tests have been introduced in swimming research to estimate MLSS more feasibly.
      • Espada M.C.
      • Alves F.B.
      • Curto D.
      • et al.
      Can an incremental step test be used for maximal lactate steady state determination in swimming? Clues for practice.
      ,
      • Nikitakis I.S.
      • Toubekis A.G.
      Lactate threshold evaluation in swimmers: the importance of age and method.
      Incremental step tests (ST) are particularly popular in swimming for determining breakpoints in the speed-lactate curve as practicable MLSS derivatives.
      • Espada M.C.
      • Alves F.B.
      • Curto D.
      • et al.
      Can an incremental step test be used for maximal lactate steady state determination in swimming? Clues for practice.
      ,
      • Nikitakis I.S.
      • Toubekis A.G.
      Lactate threshold evaluation in swimmers: the importance of age and method.
      ,
      • Fernandes R.J.
      • Sousa M.
      • Machado L.
      • et al.
      Step length and individual anaerobic threshold assessment in swimming.
      Yet some of the ST protocols have not been validated against physiological criteria
      • Toubekis A.G.
      • Tsami A.P.
      • Tokmakidis S.P.
      Critical velocity and lactate threshold in young swimmers.
      ,
      • Faude O.
      • Meyer T.
      • Scharhag J.
      • et al.
      Volume vs. intensity in the training of competitive swimmers.
      and are limited by their resolution and a fixed step distance. In detail, most of the ST protocols use speed increments of ≥ 0.05 m·s−1,
      • Espada M.C.
      • Alves F.B.
      • Curto D.
      • et al.
      Can an incremental step test be used for maximal lactate steady state determination in swimming? Clues for practice.
      ,
      • Nikitakis I.S.
      • Toubekis A.G.
      Lactate threshold evaluation in swimmers: the importance of age and method.
      ,
      • Fernandes R.J.
      • Sousa M.
      • Machado L.
      • et al.
      Step length and individual anaerobic threshold assessment in swimming.
      ,
      • Faude O.
      • Meyer T.
      • Scharhag J.
      • et al.
      Volume vs. intensity in the training of competitive swimmers.
      ,
      • Pelarigo J.G.
      • Fernandes R.J.
      • Ribeiro J.
      • et al.
      Comparison of different methods for the swimming aerobic capacity evaluation.
      which can hardly meet the requirements of a high resolution of intensity zones as depicted above.
      • Greco C.C.
      • De Oliveira M.F.M.
      • Caputo F.
      • et al.
      How narrow is the spectrum of submaximal speeds in swimming?.
      In addition, unlike tethered swimming, in which athletes are attached to a load cell with a cord and are required to maintain their position against gradually increasing resistance,
      • Papoti M.
      • Da Silva A.S.R.
      • Araujo G.G.
      • et al.
      Aerobic and anaerobic performances in tethered swimming.
      most free-swimming protocols use increments with fixed distances, particularly often 200 m.
      • Espada M.C.
      • Alves F.B.
      • Curto D.
      • et al.
      Can an incremental step test be used for maximal lactate steady state determination in swimming? Clues for practice.
      ,
      • Nikitakis I.S.
      • Toubekis A.G.
      Lactate threshold evaluation in swimmers: the importance of age and method.
      ,
      • Fernandes R.J.
      • Sousa M.
      • Machado L.
      • et al.
      Step length and individual anaerobic threshold assessment in swimming.
      ,
      • Faude O.
      • Meyer T.
      • Scharhag J.
      • et al.
      Volume vs. intensity in the training of competitive swimmers.
      ,
      • Pelarigo J.G.
      • Fernandes R.J.
      • Ribeiro J.
      • et al.
      Comparison of different methods for the swimming aerobic capacity evaluation.
      Despite high practicability (i.e., blood sampling always at the same place), this fixed distance approach bears the risk of overestimating breakpoints and derived training zones due to decreasing step durations with increasing speeds.
      • Papoti M.
      • Da Silva A.S.R.
      • Araujo G.G.
      • et al.
      Aerobic and anaerobic performances in tethered swimming.
      • Zinner C.
      • Krueger M.
      • Wahl P.
      • et al.
      Comparison of three different step test protocols in elite swimming.
      • Ribeiro L.
      • Balikian P.
      • Malachias P.
      • et al.
      Stage length, spline function and lactate minimum swimming speed.
      Besides these methodological issues, lactate thresholds derived from STs have been criticized in various sports for their empirical/mathematical rather than physiological association with MLSS.
      • Dotan R.
      Reverse lactate threshold: a novel single-session approach to reliable high-resolution estimation of the anaerobic threshold.
      ,
      • Wahl P.
      • Manunzio C.
      • Vogt F.
      • et al.
      Accuracy of a modified lactate minimum test and reverse lactate threshold test to determine maximal lactate steady state.
      In consequence, two single-session tests based on the blood lactate accumulation-elimination-equilibrium, namely the lactate minimum (LM) and reverse lactate threshold (RLT) tests, have been compared to MLSS in cycling and running, showing a high agreement, e.g.,
      • Wahl P.
      • Manunzio C.
      • Vogt F.
      • et al.
      Accuracy of a modified lactate minimum test and reverse lactate threshold test to determine maximal lactate steady state.
      • Wahl P.
      • Zwingmann L.
      • Manunzio C.
      • et al.
      Higher accuracy of the lactate minimum test compared to established threshold concepts to determine maximal lactate steady state in running.
      • Wahl P.
      • Manunzio C.
      • Zwingmann L.
      • et al.
      Reverse lactate threshold test accurately predicts maximal lactate steady state and 5 km performance in running.
      . Interestingly, the lactate minimum test (LMT) has already been used in swimming in some investigations,
      • Ribeiro L.
      • Balikian P.
      • Malachias P.
      • et al.
      Stage length, spline function and lactate minimum swimming speed.
      ,
      • Kalva-Filho C.A.
      • Zagatto A.M.
      • Araújo M.I.C.
      • et al.
      Relationship between aerobic and anaerobic parameters from 3-minute all-out tethered swimming and 400-m maximal front crawl effort.
      ,
      • De Barros Sousa F.A.
      • Rodrigues N.A.
      • Messias L.H.D.
      • et al.
      Aerobic and anaerobic swimming force evaluation in one single test session for young swimmers.
      but unlike tethered swimming, free-swimming protocols have so far lacked aspects similar to those already described for ST protocols (i.e., fixed step distance and coarse intensity resolution). In contrast, the RLT has not yet been implemented in swimming at all.
      Therefore, this study aimed to test the level of agreement of different protocols and derived threshold concepts with MLSS established with a narrow speed and temporal resolution in free-swimming. To overcome some of the mentioned limitations, all tests used a higher resolution of the speed gradation and increments with a fixed duration.

      2. Methods

      Participants underwent five to eight separate test sessions: a ST, a LMT, a RLT test and two to five 30 min constant speed trials for MLSS determination. After completing their individual routine warm-up procedure on land and in the water, athletes performed all tests using the front crawl technique, in-water starts and flip turns. All sessions took place simultaneously for a maximum of two participants on a side lane (separated by a lane rope) in the same 50 m pool for each participant. The velocity (calculated by total lap time) was set by a visual pacing device on the bottom of the pool equipped with three flashing LED lights every 10 cm (Virtual Swim Trainer, Indico Technologies, Torino, Italy; precision: 0.02 s). Due to higher speeds after the turns, athletes were instructed to swim as close to the lights as possible by performing smooth turns at submaximal speeds. All tests, including athlete instruction, speed monitoring and data collection, were always performed by at least two experienced diagnosticians, at least one of whom was one of the authors. All athletes were instructed to avoid intense exercise for 24 h preceding each test and instead to perform low to moderate intensity work only when needed. In addition, the single test sessions were separated by at least 24 h, limited to a maximum of three tests per week and the testing time for each participant was kept constant within a time frame of ±2 h (only 3 of 23 participants completed one of the tests outside this time frame due to training time constraints).
      All testing protocols consisted of stages with a fixed duration and short resting periods (30 s) in-between for blood sampling. Thus, as speed increased, larger distances were covered to keep the time for physiological responses constant. This was achieved by programming the pacing lights to indicate to the athletes the start and end of each stage at the corresponding point wherever in the pool with a 5 s start-stop countdown. Before the first test, the 400 m personal best (v400m) was requested from the athletes or coaches to define individual initial speeds.
      • Greco C.C.
      • De Oliveira M.F.M.
      • Caputo F.
      • et al.
      How narrow is the spectrum of submaximal speeds in swimming?.
      ,
      • Dekerle J.
      • Nesi X.
      • Lefevre T.
      • et al.
      Stroking parameters in front crawl swimming and maximal lactate steady state speed.
      However, since pilot testing revealed difficulties in selecting appropriate initial intensities for the LMT and RLT test based on v400m alone, the order of the tests was kept constant (i.e., ST, LMT and RLT) to incorporate prior knowledge from the ST. During all sessions, HR was recorded continuously (HRM-Swim™, Garmin Deutschland GmbH, Garching) and 20 μl of capillary blood for lactate analysis (Biosen C-line; EKF Diagnostic Sales, Magdeburg, Germany) was taken from the dried earlobe.
      23 healthy front crawl swimmers or triathletes of different ages and performance levels (recreational to regional level) volunteered to participate in the study (Table 1). Inclusion criteria were cardiovascular health, systematic training and competition experience for at least 3 years and a target v400m of >1.33 m·s−1. These rather broad criteria were chosen to validate the test protocols for a wide range of athletes. The participants and their parents (when necessary) were informed about the benefits and risks of the investigation and provided written informed consent. The study was conducted according to the Declaration of Helsinki and was approved by the university's ethical committee (124/2020).
      Table 1Descriptive characteristics of the participants (n = 23) presented as mean ± standard deviation along with the range for the whole group and separated by sex.
      SexnAge [yrs]Height [cm]Mass [kg]V̇O2peak [ml·min−1·kg−1]v400m [m·s−1]FINA points [AU]Training volume [h·wk.−1]
      Male1220.2 ± 5.7 (13.2–33.3)179 ± 7 (166–194)67.9 ± 10.0 (51.9–84.0)60.9 ± 7.2 (47.2–74.4)1.41 ± 0.15 (1.14–1.59)484 ± 142 (247–669)14 ± 6 (7–21)
      Female1117.7 ± 6.2 (13.1–30.0)169 ± 7 (154–176)57.4 ± 5.7 (48.6–68.0)52.6 ± 6.3 (38.7–59.2)1.34 ± 0.08 (1.21–1.46)499 ± 89 (368–636)14 ± 5 (5–20)
      All2319.0 ± 5.9 (13.1–33.3)174 ± 8 (154–194)62.9 ± 9.7 (48.6–84.0)56.9 ± 7.9 (38.7–74.4)1.38 ± 0.13 (1.14–1.59)490 ± 118 (247–669)14 ± 5 (5–21)
      Abbreviations: V̇O2peak: maximal oxygen uptake (assessed after a 200 m all-out); v400m: 400 m personal best.
      The 3 min fixed duration incremental ST started at 88% of v400m minus four increments (0.12 m·s−1),
      • Greco C.C.
      • De Oliveira M.F.M.
      • Caputo F.
      • et al.
      How narrow is the spectrum of submaximal speeds in swimming?.
      ,
      • Dekerle J.
      • Nesi X.
      • Lefevre T.
      • et al.
      Stroking parameters in front crawl swimming and maximal lactate steady state speed.
      and increased by 0.03 m·s−1 until exhaustion (average number of steps 8.6 ± 1.0). The swimming speed in the last stage (vpeak) was recorded and corrected for the effective duration if the 3 min were not completed. For threshold determination, bLa was plotted against swimming speed and fitted by a third-order polynomial (R2 ≥ 0.98).
      • Zwingmann L.
      • Strütt S.
      • Martin A.
      • et al.
      Modifications of the Dmax method in comparison to the maximal lactate steady state in young male athletes.
      The speed corresponding to a bLa of 4 mmol·L−1 (OBLA), a widely used threshold concept in swimming,
      • Espada M.C.
      • Alves F.B.
      • Curto D.
      • et al.
      Can an incremental step test be used for maximal lactate steady state determination in swimming? Clues for practice.
      ,
      • Arsoniadis G.G.
      • Nikitakis I.S.
      • Botonis P.G.
      • et al.
      Validating physiological and biomechanical parameters during intermittent swimming at speed corresponding to lactate concentration of 4 mmol.L-1.
      and the modified maximum distance method (mDmax), recently shown to be valid in swimming,
      • Nikitakis I.S.
      • Toubekis A.G.
      Lactate threshold evaluation in swimmers: the importance of age and method.
      were determined as estimates of MLSS pace. mDmax was determined as the point on the third-order polynomial with the maximal perpendicular distance to the straight line formed by the LT1 and the final data point. Based on the method proposed by Bishop et al.,
      • Bishop D.
      • Jenkins D.G.
      • Mackinnon L.T.
      The relationship between plasma lactate parameters, W(peak) and 1-h cycling performance in women.
      LT1 was defined as the point on the fitted curve with a slope equal to 13.3, which corresponds to an increase in bLa of 0.4 mmol·L−1 but allows for higher resolution.
      • Zwingmann L.
      • Strütt S.
      • Martin A.
      • et al.
      Modifications of the Dmax method in comparison to the maximal lactate steady state in young male athletes.
      The LMT comprised a priming segment to assess V̇O2peak and to induce hyperlactatemia and an incremental segment to determine the LM. The 300 m priming segment consisted of 100 m at progressively increased speed directly followed by a 200 m all-out.
      • Kalva-Filho C.A.
      • Zagatto A.M.
      • Araújo M.I.C.
      • et al.
      Relationship between aerobic and anaerobic parameters from 3-minute all-out tethered swimming and 400-m maximal front crawl effort.
      Immediately after striking the wall, the spirometry mask was firmly applied to the participant (< 60 s) who rested in the water in an upright position immersed to the mid-sternum.
      • Chaverri D.
      • Schuller T.
      • Iglesias X.
      • et al.
      A new model for estimating peak oxygen uptake based on postexercise measurements in swimming.
      The measured V̇O2 was corrected by the post-exercise HR decline and the highest 5 s average value was considered the V̇O2peak.
      • Chaverri D.
      • Schuller T.
      • Iglesias X.
      • et al.
      A new model for estimating peak oxygen uptake based on postexercise measurements in swimming.
      Following a resting period of 5 min (including V̇O2 assessment), the second segment started at the average of the three values 88% of v400m,
      • Greco C.C.
      • De Oliveira M.F.M.
      • Caputo F.
      • et al.
      How narrow is the spectrum of submaximal speeds in swimming?.
      ,
      • Dekerle J.
      • Nesi X.
      • Lefevre T.
      • et al.
      Stroking parameters in front crawl swimming and maximal lactate steady state speed.
      OBLA and mDmax minus four increments (0.12 m·s−1) and speed was increased by 0.03 m·s−1 every 2 min until exhaustion (average number of steps 8.8 ± 1.2). The shorter stage duration compared to the ST was chosen after pilot tests revealed underestimation of speed at MLSS with longer durations (3 min), a known weakness of the LMT design, attributable to the greater cumulative time below the MMSS that does not necessarily lead to the highest possible lactate equilibrium.
      • Dotan R.
      Reverse lactate threshold: a novel single-session approach to reliable high-resolution estimation of the anaerobic threshold.
      Following previous work,
      • Ribeiro L.
      • Balikian P.
      • Malachias P.
      • et al.
      Stage length, spline function and lactate minimum swimming speed.
      ,
      • Wahl P.
      • Manunzio C.
      • Vogt F.
      • et al.
      Accuracy of a modified lactate minimum test and reverse lactate threshold test to determine maximal lactate steady state.
      swimming speed at LM was calculated by the first derivative of a third-order polynomial fitting of the speed-lactate curve (R2 ≥ 0.98).
      The RLT test comprised an incremental priming segment with four coarse speed increments (0.06 m·s−1) starting at the average of the three values 88% of v400m,
      • Greco C.C.
      • De Oliveira M.F.M.
      • Caputo F.
      • et al.
      How narrow is the spectrum of submaximal speeds in swimming?.
      ,
      • Dekerle J.
      • Nesi X.
      • Lefevre T.
      • et al.
      Stroking parameters in front crawl swimming and maximal lactate steady state speed.
      OBLA and mDmax minus 0.12 m·s−1 and increasing up to the average of the three values plus 0.06 m·s−1. The reverse segment immediately followed with a speed reduced by 0.03 m·s−1, which decreased by another 0.03 m·s−1 every 3 min until the initial speed was reached again (total number of steps ten). Analogous to the LM, swimming speed corresponding to the lactate apex reached during the reverse segment (RLT) was determined using the first derivative of a third-order polynomial fitting of the speed-lactate curve (R2 ≥ 0.99).
      • Dotan R.
      Reverse lactate threshold: a novel single-session approach to reliable high-resolution estimation of the anaerobic threshold.
      MLSS was determined by two to five 6 × 5 min constant speed tests, the first one set at the average of OBLA, mDmax, LM and RLT. The 5 min intervals were chosen to allow determination of additional temporal criteria (i.e., between the 15th and 30th and 20th and 30th min) besides the traditional interval (i.e., between the 10th and 30th min) for the permitted ≤ 1 mmol·L−1 increase in bLa by blood sampling every 5 min during the 30 s resting phases as suggested recently.
      • Iannetta D.
      • Ingram C.P.
      • Keir D.A.
      • et al.
      Methodological reconciliation of CP and MLSS and their agreement with the maximal metabolic steady state.
      A narrow speed gradation, i.e., 0.015 m·s−1 steps, allowed high sensitivity in the detection of the ‘true’ MMSS.
      For the ST, LMT and RLT test, bLa and HR corresponding to the respective thresholds were calculated by third-order polynomial and linear interpolation, respectively. For the MLSS tests, bLa and HR were averaged over the respective time intervals.
      Homoscedasticity and normality were checked by visual inspection of residual histograms, residual plots and Q-Q-plots. Linear mixed-effects models were constructed to examine differences between speeds corresponding to each threshold concept using the lme4 package in R.
      • R Core Team. R
      A language and environment for statistical computing.
      As fixed effects, threshold concepts (five levels), sex (two levels) and age (covariate) were entered into the model (with and without interaction plus random intercepts for participants) and retained if maximum likelihood ratio test indicated a significant alteration (p < 0.05). Consequently, multiple post-hoc pairwise comparisons were conducted with the “Bonferroni” adjustment method using the emmeans package. Cohen's effect sizes (ES) along with 95% confidence intervals (CI) to indicate the accuracy of ES estimates were determined using the effsize package to assess the magnitude of differences between threshold concepts. ES estimates were interpreted as follows: ES < 0.2 = trivial, 0.2 ≤ ES < 0.6 = small, 0.6 ≤ ES < 1.2 = moderate and ES ≥ 1.2 = large.
      • Hopkins W.G.
      A scale of magnitudes for effect statistics.
      A Bland-Altman analysis was performed using the BlandAltmanLeh package to check the agreement (bias along with limits of agreement [LoA], i.e., 1.96-fold standard deviation [SD]) between the measures presented in [m·s−1] and in [%] of MLSS pace. To account for correlation and agreement between measurements, intraclass correlation coefficients (ICC) along with 95% CI were calculated using the icc package with single measure two-way mixed-effects models and the “absolute agreement” as the type of analysis.
      • Wahl P.
      • Manunzio C.
      • Zwingmann L.
      • et al.
      Reverse lactate threshold test accurately predicts maximal lactate steady state and 5 km performance in running.
      ,
      • Koo T.K.
      • Li M.Y.
      A guideline of selecting and reporting intraclass correlation coefficients for reliability research.
      The agreement was interpreted as follows: ICC ≤ 0.50 = poor, 0.50 ≤ ICC < 0.75 = moderate, 0.75 ≤ ICC < 0.90 = good and ≥ 0.90 = excellent.
      • Koo T.K.
      • Li M.Y.
      A guideline of selecting and reporting intraclass correlation coefficients for reliability research.
      Further, associations between LT1 and MLSS pace were examined using the Pearson product-moment correlation coefficient r along with 95% CI calculated with the stats package. For all tests, statistical significance was accepted at p < 0.05. All data are presented as mean ± SD.

      3. Results

      Determination of RLT was only possible in 17 of the 23 participants due to the lack of the characteristic blood lactate curve. Only marginal differences were observed between the MLSS determination methods. Thus, mean MLSS paces were 1.225 ± 0.112, 1.230 ± 0.110 and 1.231 ± 0.110 m·s−1 when determined as the highest speed with an increase in bLa ≤ 1 mmol·L−1 between the 10th and 30th, 15th and 30th, and 20th and 30th min, respectively. Given the marginal variation between the MLSS determination methods (< 0.5%) and the fact that 14 of the 23 participants could not finish the next 30 min trial, the highest value, determined between the 20th and 30th min, was accepted as MLSS pace according to Iannetta et al.
      • Iannetta D.
      • Ingram C.P.
      • Keir D.A.
      • et al.
      Methodological reconciliation of CP and MLSS and their agreement with the maximal metabolic steady state.
      The average speed, HR and bLa corresponding to each threshold concept are depicted in Table 2. For HR and bLa, statistically significant main effects for threshold concepts were found (p < 0.001). HR at OBLA (p < 0.05) and LM (p < 0.001) and bLa at OBLA (p < 0.05), LM (p < 0.05) and RLT (p < 0.05) differed from those at MLSS as detailed in Table 2. Adding sex as another fixed effect improved the model only for bLa (p < 0.05), with male athletes showing higher bLa than females. For speed, a statistically significant main effect for threshold concept was observed (p < 0.001), however adding either sex or age as another fixed effect (with and without interaction) did not further improve the model. Across all participants, pairwise comparisons revealed significant differences between speed at MLSS and mDmax (p < 0.05) and RLT (p < 0.05), albeit with trivial to small effects (ES ≤ 0.23) as detailed in Table 2. Excellent measures of agreement (ICC ≥ 0.926) demonstrate high conformity of all threshold concepts with MLSS in male and female athletes (Table 2).
      Table 2Speed (absolute value and difference [Δ] to MLSS pace), blood lactate concentration (bLa) and heart rate (HR) corresponding to each threshold concept presented as mean ± standard deviation along with the range for the whole group and separated by sex. In addition, effect size (ES) and intraclass correlation coefficient (ICC) indicate the magnitude of difference and agreement, respectively, between each speed estimate and MLSS pace and are presented along with 95% confidence intervals (CI).
      ParameterSexnSpeed [m·s−1]Δ Threshold-MLSS [m·s−1]ES (95% CI)ICC (95% CI)bLa [mmol·L−1]HR [b·min−1]
      MLSSMale121.271 ± 0.123 (1.040–1.445)5.28 ± 1.91 (2.83–9.23)184 ± 10 (162–197)
      Female111.187 ± 0.078 (1.050–1.335)4.89 ± 1.60 (2.43–7.08)188 ± 11 (168–205)
      All231.231 ± 0.110 (1.040–1.445)5.10 ± 1.74 (2.43–9.23)186 ± 10 (162–205)
      OBLAMale121.277 ± 0.130 (1.066–1.457)0.006 ± 0.0240.05 (−0.06–0.15)0.982⁎⁎⁎ (0.942–0.995)4.00 ± 0.00179 ± 10 (161–193)
      Female111.208 ± 0.089 (1.080–1.380)0.020 ± 0.0270.22 (0.04–0.41)0.926⁎⁎⁎ (0.644–0.981)4.00 ± 0.00186 ± 16 (150–209)
      All231.244 ± 0.115 (1.066–1.457)0.013 ± 0.0260.11 (0.02–0.21)0.968⁎⁎⁎ (0.915–0.987)4.00 ± 0.00182 ± 13 (150–209)
      mDmaxMale121.294 ± 0.113 (1.101–1.452)0.023 ± 0.0240.18 (0.06–0.29)0.963⁎⁎ (0.716–0.991)4.48 ± 0.95 (3.13–5.97)181 ± 10 (163–194)
      Female111.201 ± 0.069 (1.085–1.326)0.013 ± 0.0200.16 (0.11–0.31)0.951⁎⁎⁎ (0.780–0.987)3.86 ± 0.86 (2.48–5.00)185 ± 14 (155–209)
      All231.249 ± 0.104 (1.085–1.452)0.018 ± 0.0220.16 (0.08–0.25)0.966⁎⁎⁎ (0.822–0.989)4.18 ± 0.94 (2.48–5.97)183 ± 12 (155–209)
      LMMale121.273 ± 0.124 (0.994–1.427)0.001 ± 0.0360.01 (−0.16–0.19)0.960⁎⁎⁎ (0.868–0.988)4.71 ± 2.12 (1.79–7.75)178 ± 8 (168–189)
      Female111.180 ± 0.073 (1.072–1.304)−0.007 ± 0.021−0.10 (−0.27–0.08)0.960⁎⁎⁎ (0.864–0.989)3.05 ± 0.88 (1.79–4.68)179 ± 15 (152–202)
      All231.228 ± 0.111 (0.994–1.427)−0.003 ± 0.030−0.03 (−0.14–0.09)0.965⁎⁎⁎ (0.920–0.985)3.92 ± 1.82 (1.79–7.75)178 ± 11⁎⁎⁎ (152–202)
      RLTMale91.264 ± 0.127 (1.057–1.424)0.024 ± 0.0190.17 (0.08–0.27)0.971⁎⁎ (0.505–0.995)8.36 ± 2.36 (5.38–11.49)185 ± 11 (168–196)
      Female81.196 ± 0.085 (1.035–1.307)0.023 ± 0.0210.23 (0.07–0.40)0.928⁎⁎ (0.376–0.987)5.67 ± 1.70 (3.29–7.86)186 ± 14 (164–199)
      All171.232 ± 0.112 (1.035–1.424)0.024 ± 0.0190.19 (0.11–0.27)0.961⁎⁎ (0.522–0.991)7.09 ± 2.44 (3.29–11.49)186 ± 12 (164–199)
      Abbreviations: MLSS: speed at maximal lactate steady state; OBLA: speed at a blood lactate concentration of 4 mmol·L−1; mDmax: speed at modified maximal distance method; LM: speed at lactate minimum; RLT: speed at reverse lactate threshold. Statistically significant differences between values (speed, bLa and HR) corresponding to estimates and MLSS: p < 0.05, ⁎⁎⁎ p < 0.001 and statistically significant ICC values: ⁎⁎ p < 0.01, ⁎⁎⁎ p < 0.001.
      Bland-Altman plots confirm the absolute agreement between all threshold concepts and MLSS (Fig. 1). LM showed the smallest bias (−0.2%), but the largest LoA (±4.7%). OBLA (+1.0%), mDmax (+1.5%) and RLT (+2.0%) showed larger biases (overestimation of MLSS pace) but with smaller LoA (±4.1%, ±3.5% and ±3.1%). The coefficients (intercept and slope) for each proportional bias (linear regression between the mean and the difference between the speeds at each threshold concept and MLSS) were −0.045 and 0.046 for OBLA, 0.092 and −0.059 for mDmax, −0.013 and 0.009 for LM and −0.085 and 0.089 for RLT. Apart from the slope for RLT (p < 0.05), all other coefficients were not statistically significant.
      Fig. 1
      Fig. 1Bland-Altman plots: differences between the swimming speeds at a blood lactate concentration of 4 mmol·L−1 (OBLA) (A), modified maximal distance method (mDmax) (B), lactate minimum (LM) (C) and reverse lactate threshold (RLT) (D) as well as maximal lactate steady state (MLSS) pace in male (open) and female (filled circles) athletes. The solid line indicates the mean difference (fixed bias), the dotted lines indicate the limits of agreement (fixed bias ±1.96-fold standard deviation) and the dashed line indicates the linear regression line (proportional bias) of the data.
      In addition to MLSS estimates, it was found that speed at LT1 (1.173 ± 0.105 m·s−1) was highly associated with speed at MLSS (r = 0.97, CI: 0.92–0.99, p < 0.001) across all participants.

      4. Discussion

      This study investigated the level of agreement of various threshold concepts (OBLA, mDmax, LM, RLT) with MLSS as a physiological criterion representing MMSS in free-swimming. While previous investigations used speed increments of ≥ 0.05 m·s−1 along with fixed step distances (frequently 200 m),
      • Espada M.C.
      • Alves F.B.
      • Curto D.
      • et al.
      Can an incremental step test be used for maximal lactate steady state determination in swimming? Clues for practice.
      ,
      • Nikitakis I.S.
      • Toubekis A.G.
      Lactate threshold evaluation in swimmers: the importance of age and method.
      ,
      • Fernandes R.J.
      • Sousa M.
      • Machado L.
      • et al.
      Step length and individual anaerobic threshold assessment in swimming.
      ,
      • Faude O.
      • Meyer T.
      • Scharhag J.
      • et al.
      Volume vs. intensity in the training of competitive swimmers.
      ,
      • Pelarigo J.G.
      • Fernandes R.J.
      • Ribeiro J.
      • et al.
      Comparison of different methods for the swimming aerobic capacity evaluation.
      finer increments of 0.03 m·s−1 and fixed step durations (2 or 3 min) were applied in the present study. Likewise, MLSS was determined using additional narrower temporal criteria (i.e., intervals between the 15th and 30th and 20th and 30th min) and a higher resolution (0.015 m·s−1) than before (≥ 0.03 m·s−1 and interval between the 10th and 30th min)
      • Pelarigo J.G.
      • Greco C.C.
      • Denadai B.S.
      • et al.
      Do 5% changes around maximal lactate steady state lead to swimming biophysical modifications?.
      ,
      • Espada M.C.
      • Alves F.B.
      • Curto D.
      • et al.
      Can an incremental step test be used for maximal lactate steady state determination in swimming? Clues for practice.
      ,
      • Dekerle J.
      • Nesi X.
      • Lefevre T.
      • et al.
      Stroking parameters in front crawl swimming and maximal lactate steady state speed.
      to alleviate granularity concerns related to an underestimation of MMSS.
      • Jones A.M.
      • Burnley M.
      • Black M.I.
      • et al.
      The maximal metabolic steady state: redefining the ‘gold standard.’.
      Compared to previous studies,
      • Espada M.C.
      • Alves F.B.
      • Curto D.
      • et al.
      Can an incremental step test be used for maximal lactate steady state determination in swimming? Clues for practice.
      ,
      • Nikitakis I.S.
      • Toubekis A.G.
      Lactate threshold evaluation in swimmers: the importance of age and method.
      ,
      • Pelarigo J.G.
      • Fernandes R.J.
      • Ribeiro J.
      • et al.
      Comparison of different methods for the swimming aerobic capacity evaluation.
      ,
      • Ribeiro L.
      • Balikian P.
      • Malachias P.
      • et al.
      Stage length, spline function and lactate minimum swimming speed.
      a higher agreement was observed between all threshold concepts and MLSS as indicated by small biases and LoAs (≤ 2% and < 5%, respectively).
      This observation applies to a broad sample including male and female swimmers and triathletes aged from 13 to 33 years. As modeling speed differences between threshold concepts and MLSS was not improved by including sex or age in the model, the agreement between estimates and MLSS pace appears to be valid regardless of both covariates. Apart from sex and age, participants' performance levels were heterogeneous along with wide variation in training routines from 5 to 21 h per week, as indicated by the ranges of V̇O2peak and v400m, i.e., 38.7 to 74.4 ml·min−1·kg−1 and 1.14 to 1.59 m·s−1, respectively. However, as evident from the small proportional biases between the speeds corresponding to the estimates and MLSS shown in Fig. 1 (ranging from −0.059 to 0.089), the accuracy of the threshold concepts was independent of performance level represented by largely different MLSS paces (i.e., 1.040 to 1.445 m·s−1). Together, linear mixed models and Bland-Altman analyses confirmed the agreement of threshold estimates and MLSS pace in participants with widely varying characteristics including sex, age and performance level.
      The average speed corresponding to MLSS observed in the present study was in the range of earlier findings in trained swimmers (e.g., 1.22 ± 0.09 m·s−1,
      • Dekerle J.
      • Nesi X.
      • Lefevre T.
      • et al.
      Stroking parameters in front crawl swimming and maximal lactate steady state speed.
      1.25 ± 0.06 m·s−1).
      • Ribeiro L.
      • Balikian P.
      • Malachias P.
      • et al.
      Stage length, spline function and lactate minimum swimming speed.
      In accordance with Iannetta et al.,
      • Iannetta D.
      • Ingram C.P.
      • Keir D.A.
      • et al.
      Methodological reconciliation of CP and MLSS and their agreement with the maximal metabolic steady state.
      high sensitivity in MLSS testing was achieved by using narrower temporal criteria and small increments between 30 min trials, which together allowed precise estimation of the physiological criterion. Considering that 14 of the 23 participants could not complete the next trial above MLSS pace, MLSS, defined as the highest speed with an increase in bLa ≤ 1 mmol·L−1 between the 20th and 30th min of constant exercise, appears to represent well the ‘true’ MMSS, above which contrasting physiological responses occur.
      • Iannetta D.
      • Ingram C.P.
      • Keir D.A.
      • et al.
      Methodological reconciliation of CP and MLSS and their agreement with the maximal metabolic steady state.
      ,
      • Jones A.M.
      • Burnley M.
      • Black M.I.
      • et al.
      The maximal metabolic steady state: redefining the ‘gold standard.’.
      These findings highlight the importance of a high work-rate and temporal resolution for MLSS testing due to the narrow intensity range in swimming, which allows one to remain in a physiological steady state for at least 30 min at a given speed, whereas a speed of only 0.015 m·s−1 higher can quickly lead to exhaustion. At the expense of higher temporal resolution, the intermittent MLSS test design, which is required for regular blood sampling, may have yielded a slightly higher speed than the traditional continuous protocol. However, given the low ratio of exercise to resting time of 1:10,
      • Stockhausen W.
      • Grathwohl D.
      • Bürklin C.
      • et al.
      Stage duration and increase of work load in incremental testing on a cycle ergometer.
      the influence of the intermittent protocol appears to be rather small, as also suggested by previous findings in running.
      • Wahl P.
      • Zwingmann L.
      • Manunzio C.
      • et al.
      Higher accuracy of the lactate minimum test compared to established threshold concepts to determine maximal lactate steady state in running.
      ,
      • Wahl P.
      • Manunzio C.
      • Zwingmann L.
      • et al.
      Reverse lactate threshold test accurately predicts maximal lactate steady state and 5 km performance in running.
      Despite the frequent usage of incremental tests in swimming, few studies have validated ST protocols and associated threshold concepts against MLSS.
      • Espada M.C.
      • Alves F.B.
      • Curto D.
      • et al.
      Can an incremental step test be used for maximal lactate steady state determination in swimming? Clues for practice.
      ,
      • Nikitakis I.S.
      • Toubekis A.G.
      Lactate threshold evaluation in swimmers: the importance of age and method.
      ,
      • Fernandes R.J.
      • Sousa M.
      • Machado L.
      • et al.
      Step length and individual anaerobic threshold assessment in swimming.
      ,
      • Pelarigo J.G.
      • Fernandes R.J.
      • Ribeiro J.
      • et al.
      Comparison of different methods for the swimming aerobic capacity evaluation.
      Even fewer studies have reported exact agreement with MLSS pace.
      • Espada M.C.
      • Alves F.B.
      • Curto D.
      • et al.
      Can an incremental step test be used for maximal lactate steady state determination in swimming? Clues for practice.
      ,
      • Nikitakis I.S.
      • Toubekis A.G.
      Lactate threshold evaluation in swimmers: the importance of age and method.
      ,
      • Pelarigo J.G.
      • Fernandes R.J.
      • Ribeiro J.
      • et al.
      Comparison of different methods for the swimming aerobic capacity evaluation.
      These studies showed a deviation from speed at MLSS (bias ± LoA) of −0.3 ± 5.1% (OBLA)
      • Espada M.C.
      • Alves F.B.
      • Curto D.
      • et al.
      Can an incremental step test be used for maximal lactate steady state determination in swimming? Clues for practice.
      and −1.3 ± 6.8% (mDmax),
      • Nikitakis I.S.
      • Toubekis A.G.
      Lactate threshold evaluation in swimmers: the importance of age and method.
      respectively. However, in the study by Espada et al.,
      • Espada M.C.
      • Alves F.B.
      • Curto D.
      • et al.
      Can an incremental step test be used for maximal lactate steady state determination in swimming? Clues for practice.
      MLSS was only determined with a coarse resolution of ≥ 0.06 m·s−1. In contrast, our findings indicate narrower LoA (OBLA: +1.0 ± 4.1% and mDmax: +1.5 ± 3.5%) in the context of a finely resolved MLSS. Given the excellent agreement between the speeds at OBLA and mDmax with MLSS pace, our findings suggest that constant step duration and high work-rate resolution seem to alleviate MLSS overestimation previously reported by OBLA,
      • Fernandes R.J.
      • Sousa M.
      • Machado L.
      • et al.
      Step length and individual anaerobic threshold assessment in swimming.
      ,
      • Zinner C.
      • Krueger M.
      • Wahl P.
      • et al.
      Comparison of three different step test protocols in elite swimming.
      and hence generally provide a more accurate estimation of MLSS. Despite the high agreement between the speeds at OBLA and MLSS in our sample, the high variability found for bLa at MLSS (range for the whole sample: 2.43–9.23 mmol·L−1) generally suggests not to focus on fixed bLa but rather to interpret bLa kinetics such as at mDmax. Since bLa at MLSS already shows a high day-to-day variation (16.6%),
      • Hauser T.
      • Bartsch D.
      • Baumgärtel L.
      • et al.
      Reliability of maximal lactate-steady-state.
      it does not seem necessary or possible to approach bLa at MLSS with other threshold concepts.
      • Wahl P.
      • Zwingmann L.
      • Manunzio C.
      • et al.
      Higher accuracy of the lactate minimum test compared to established threshold concepts to determine maximal lactate steady state in running.
      This is further supported by the bLa at LM and RLT in the present study, which deviate even more from that at MLSS compared to OBLA and mDmax, whereas the respective speeds are highly consistent with that at MLSS.
      In addition to the MLSS estimates, the ST additionally provides a measure of LT1. Due to the high correlation with MLSS pace (r = 0.97), LT1 might be monitored regularly as a submaximal indicator of aerobic performance without the need to perform tests until exhaustion. However, the suitability of LT1 as a submaximal indicator of MLSS in swimming remains to be verified in longitudinal studies. Together, both LT1 and mDmax/OBLA may serve to define training intensity zones, namely high volume < LT1, threshold ≤ mDmax/OBLA and high intensity training > mDmax/OBLA, on an individual basis.
      • Greco C.C.
      • De Oliveira M.F.M.
      • Caputo F.
      • et al.
      How narrow is the spectrum of submaximal speeds in swimming?.
      While high agreement between RLT and MLSS has been shown in cycling and running,
      • Wahl P.
      • Manunzio C.
      • Vogt F.
      • et al.
      Accuracy of a modified lactate minimum test and reverse lactate threshold test to determine maximal lactate steady state.
      ,
      • Wahl P.
      • Manunzio C.
      • Zwingmann L.
      • et al.
      Reverse lactate threshold test accurately predicts maximal lactate steady state and 5 km performance in running.
      this is the first investigation to apply the RLT test in swimming. RLT showed a similar overestimation of MLSS pace as OBLA and mDmax, which may be explained by the short resting periods inevitable for blood sampling, leading to an earlier decline in bLa and thus a shift in RLT towards higher speeds.
      • Dotan R.
      Reverse lactate threshold: a novel single-session approach to reliable high-resolution estimation of the anaerobic threshold.
      ,
      • Wahl P.
      • Manunzio C.
      • Zwingmann L.
      • et al.
      Reverse lactate threshold test accurately predicts maximal lactate steady state and 5 km performance in running.
      However, the major challenge in RLT testing was the prior selection of individual intensities that exceed MMSS while avoiding premature test termination due to fatigue. In contrast to running and cycling, the narrow speed range due to the aquatic environment made it difficult to accurately determine swimming speeds. Thus, in accordance with Greco et al.
      • Greco C.C.
      • De Oliveira M.F.M.
      • Caputo F.
      • et al.
      How narrow is the spectrum of submaximal speeds in swimming?.
      reporting a narrow range of submaximal speeds (i.e., between LT1 and MLSS), we observed a similarly narrow range between the speed associated with MLSS and vpeak obtained during the ST (0.090 ± 0.028 m·s−1). Given the increments of 0.03 m·s−1, a maximum of two stages remains to obtain the characteristic blood lactate curve of the RLT test. Despite detailed prior knowledge of athlete performance based on the ST results, no ‘valid’ RLT (i.e., no further increase in bLa during the reverse segment) could be recorded in six out of 23 participants. Therefore, since the RLT test did not show higher accuracy compared to the ST, it may be less practicable because of the need for accurate prior knowledge of athlete performance.
      Based on the experience from previous investigations with the LMT in cycling and running,
      • Wahl P.
      • Manunzio C.
      • Vogt F.
      • et al.
      Accuracy of a modified lactate minimum test and reverse lactate threshold test to determine maximal lactate steady state.
      ,
      • Wahl P.
      • Zwingmann L.
      • Manunzio C.
      • et al.
      Higher accuracy of the lactate minimum test compared to established threshold concepts to determine maximal lactate steady state in running.
      we chose a short rest interval between both segments (5 min, because pilot testing revealed that peak bLa was often reached after this time and athletes should restart with the highest possible bLa) and targeted a similar duration below MMSS (~8 min) by using 2 min stages to avoid underestimation of MLSS pace. In addition, we applied fixed step durations in combination with finer increments (0.03 instead of 0.05 m·s−1) compared to previous studies in swimming.
      • Ribeiro L.
      • Balikian P.
      • Malachias P.
      • et al.
      Stage length, spline function and lactate minimum swimming speed.
      ,
      • Kalva-Filho C.A.
      • Zagatto A.M.
      • Araújo M.I.C.
      • et al.
      Relationship between aerobic and anaerobic parameters from 3-minute all-out tethered swimming and 400-m maximal front crawl effort.
      Indeed, LM was highly consistent with MLSS pace, albeit unlike previous investigations in running and cycling,
      • Wahl P.
      • Manunzio C.
      • Vogt F.
      • et al.
      Accuracy of a modified lactate minimum test and reverse lactate threshold test to determine maximal lactate steady state.
      ,
      • Wahl P.
      • Zwingmann L.
      • Manunzio C.
      • et al.
      Higher accuracy of the lactate minimum test compared to established threshold concepts to determine maximal lactate steady state in running.
      it did not show higher agreement with MLSS pace compared to the other threshold concepts. Nevertheless, compared to the only study presenting the absolute agreement between the speeds at LM and MLSS in free-swimming, our protocol design allowed for a halving of the observed variation (i.e., 2.7 ± 10.0%).
      • Ribeiro L.
      • Balikian P.
      • Malachias P.
      • et al.
      Stage length, spline function and lactate minimum swimming speed.
      When interpreting the results, some limitations need to be considered and addressed in future research. It should be considered that the tests were not conducted in randomized order since preliminary testing showed that detailed prior knowledge (from the ST) was necessary to select appropriate initial intensities to obtain valid results in the LMT and RLT test. Therefore, learning effects cannot be completely excluded, although we estimate them to be small, if any, since no novice swimmers or triathletes were studied. Still, no valid RLT could be determined for six participants, which limits the comparability with the other tests and highlights the difficulty of this test design in swimming. Furthermore, unlike the ST and RLT test, we chose a step duration of 2 instead of 3 min for the LMT as detailed above underlining the impact of step duration on threshold determination. However, since stage duration should always be considered together with the increment per step,
      • Wahl P.
      • Zwingmann L.
      • Manunzio C.
      • et al.
      Higher accuracy of the lactate minimum test compared to established threshold concepts to determine maximal lactate steady state in running.
      ,
      • Stockhausen W.
      • Grathwohl D.
      • Bürklin C.
      • et al.
      Stage duration and increase of work load in incremental testing on a cycle ergometer.
      2 min steps might still be appropriate for an increment of 0.03 m·s−1 during the LMT, especially compared with previous studies that used increments of ≥ 0.05 m·s−1 for a step duration of < 3 min (i.e., 200 m steps).
      • Ribeiro L.
      • Balikian P.
      • Malachias P.
      • et al.
      Stage length, spline function and lactate minimum swimming speed.
      ,
      • Kalva-Filho C.A.
      • Zagatto A.M.
      • Araújo M.I.C.
      • et al.
      Relationship between aerobic and anaerobic parameters from 3-minute all-out tethered swimming and 400-m maximal front crawl effort.
      Since our study was the first to use tests with such a high resolution of 0.015 and 0.03 m·s−1, the reliability of the presented test protocols and associated MLSS estimates needs to be verified in future studies. In particular, athletes' compliance with the required speeds needs to be quantified (e.g., by video analysis), to ensure that, given the narrow intensity range in swimming, even small changes in speed are detected. Although we cannot yet quantify whether athletes met the required speeds, the fact that 14 of 23 participants were able to complete a 30 min test at a given pace but not the test that was only 0.015 m·s−1 faster clearly indicates that athletes can follow such small adjustments.

      5. Conclusion

      Test protocols with a fixed step duration and fine increments allowed high accuracy in estimating speed at MLSS, which was likewise determined with high temporal and speed resolution, in swimmers or triathletes of different levels, ages and sexes. As LMT and RLT test protocols did not provide higher accuracy than STs, the latter appear more practicable because little prior knowledge is required and individual training intensity zones can be derived from LT1 and mDmax.

      Funding Information

      This work was supported by a grant from the Federal Institute of Sports Science ( Bundesinstitut für Sportwissenschaft ), Bonn, Germany ( PR020181200125 , 01/2019–12/2020 ).

      Declaration of Interest Statement

      The authors declare no conflict of interest.

      Confirmation of Ethical Compliance

      The study was conducted according to the Declaration of Helsinki and was approved by the ethical committee of the German Sport University Cologne (approval number: 124/2020).

      References

        • Pelarigo J.G.
        • Greco C.C.
        • Denadai B.S.
        • et al.
        Do 5% changes around maximal lactate steady state lead to swimming biophysical modifications?.
        Hum Mov Sci. 2016; 49: 258-266https://doi.org/10.1016/j.humov.2016.07.009
        • Espada M.C.
        • Reis J.F.
        • Almeida T.F.
        • et al.
        Ventilatory and physiological responses in swimmers below and above their maximal lactate steady state.
        J Strength Cond Res. 2015; 29: 2836-2843https://doi.org/10.1519/JSC.0000000000000504
        • Greco C.C.
        • De Oliveira M.F.M.
        • Caputo F.
        • et al.
        How narrow is the spectrum of submaximal speeds in swimming?.
        J Strength Cond Res. 2013; 27: 1450-1454https://doi.org/10.1519/JSC.0b013e318265a6ce
        • Iannetta D.
        • Ingram C.P.
        • Keir D.A.
        • et al.
        Methodological reconciliation of CP and MLSS and their agreement with the maximal metabolic steady state.
        Med Sci Sports Exerc. 2022; 54: 622-632https://doi.org/10.1249/MSS.0000000000002831
        • Espada M.C.
        • Alves F.B.
        • Curto D.
        • et al.
        Can an incremental step test be used for maximal lactate steady state determination in swimming? Clues for practice.
        Int J Environ Res Public Health. 2021; 18: 477https://doi.org/10.3390/ijerph18020477
        • Dekerle J.
        • Nesi X.
        • Lefevre T.
        • et al.
        Stroking parameters in front crawl swimming and maximal lactate steady state speed.
        Int J Sports Med. 2005; 26: 53-58https://doi.org/10.1055/s-2004-817854
        • Jones A.M.
        • Burnley M.
        • Black M.I.
        • et al.
        The maximal metabolic steady state: redefining the ‘gold standard.’.
        Physiol Rep. 2019; 7: 1-16https://doi.org/10.14814/phy2.14098
        • Dotan R.
        Reverse lactate threshold: a novel single-session approach to reliable high-resolution estimation of the anaerobic threshold.
        Int J Sports Physiol Perform. 2012; 7: 141-151https://doi.org/10.1123/ijspp.7.2.141
        • Nikitakis I.S.
        • Toubekis A.G.
        Lactate threshold evaluation in swimmers: the importance of age and method.
        Int J Sports Med. 2021; 42: 818-824https://doi.org/10.1055/a-1342-7446
        • Fernandes R.J.
        • Sousa M.
        • Machado L.
        • et al.
        Step length and individual anaerobic threshold assessment in swimming.
        Int J Sports Med. 2011; 32: 940-946https://doi.org/10.1055/s-0031-1283189
        • Toubekis A.G.
        • Tsami A.P.
        • Tokmakidis S.P.
        Critical velocity and lactate threshold in young swimmers.
        Int J Sports Med. 2006; 27: 117-123https://doi.org/10.1055/s-2005-837487
        • Faude O.
        • Meyer T.
        • Scharhag J.
        • et al.
        Volume vs. intensity in the training of competitive swimmers.
        Int J Sports Med. 2008; 29: 906-912https://doi.org/10.1055/s-2008-1038377
        • Pelarigo J.G.
        • Fernandes R.J.
        • Ribeiro J.
        • et al.
        Comparison of different methods for the swimming aerobic capacity evaluation.
        J Strength Cond Res. 2018; 32: 3542-3551https://doi.org/10.1519/jsc.0000000000001873
        • Papoti M.
        • Da Silva A.S.R.
        • Araujo G.G.
        • et al.
        Aerobic and anaerobic performances in tethered swimming.
        Int J Sports Med. 2013; 34: 712-719https://doi.org/10.1055/s-0031-1291250
        • Zinner C.
        • Krueger M.
        • Wahl P.
        • et al.
        Comparison of three different step test protocols in elite swimming.
        J Exerc Physiol Online. 2011; 14: 43-48
        • Ribeiro L.
        • Balikian P.
        • Malachias P.
        • et al.
        Stage length, spline function and lactate minimum swimming speed.
        J Sports Med Phys Fitness. 2003; 43: 312-318
        • Wahl P.
        • Manunzio C.
        • Vogt F.
        • et al.
        Accuracy of a modified lactate minimum test and reverse lactate threshold test to determine maximal lactate steady state.
        J Strength Cond Res. 2017; 31: 3489-3496https://doi.org/10.1519/JSC.0000000000001770
        • Wahl P.
        • Zwingmann L.
        • Manunzio C.
        • et al.
        Higher accuracy of the lactate minimum test compared to established threshold concepts to determine maximal lactate steady state in running.
        Int J Sports Med. 2018; 39: 541-548https://doi.org/10.1055/s-0044-102131
        • Wahl P.
        • Manunzio C.
        • Zwingmann L.
        • et al.
        Reverse lactate threshold test accurately predicts maximal lactate steady state and 5 km performance in running.
        Biol Sport. 2021; 38: 285-290https://doi.org/10.5114/biolsport.2021.99326
        • Kalva-Filho C.A.
        • Zagatto A.M.
        • Araújo M.I.C.
        • et al.
        Relationship between aerobic and anaerobic parameters from 3-minute all-out tethered swimming and 400-m maximal front crawl effort.
        J Strength Cond Res. 2015; 29: 238-245https://doi.org/10.1519/JSC.0000000000000592
        • De Barros Sousa F.A.
        • Rodrigues N.A.
        • Messias L.H.D.
        • et al.
        Aerobic and anaerobic swimming force evaluation in one single test session for young swimmers.
        Int J Sports Med. 2017; 38: 378-383https://doi.org/10.1055/s-0043-100101
        • Zwingmann L.
        • Strütt S.
        • Martin A.
        • et al.
        Modifications of the Dmax method in comparison to the maximal lactate steady state in young male athletes.
        Phys Sportsmed. 2019; 47: 174-181https://doi.org/10.1080/00913847.2018.1546103
        • Arsoniadis G.G.
        • Nikitakis I.S.
        • Botonis P.G.
        • et al.
        Validating physiological and biomechanical parameters during intermittent swimming at speed corresponding to lactate concentration of 4 mmol.L-1.
        Sports. 2020; 8: 23https://doi.org/10.3390/sports8020023
        • Bishop D.
        • Jenkins D.G.
        • Mackinnon L.T.
        The relationship between plasma lactate parameters, W(peak) and 1-h cycling performance in women.
        Med Sci Sports Exerc. 1998; 30: 1270-1275https://doi.org/10.1097/00005768-199808000-00014
        • Chaverri D.
        • Schuller T.
        • Iglesias X.
        • et al.
        A new model for estimating peak oxygen uptake based on postexercise measurements in swimming.
        Int J Sports Physiol Perform. 2016; 11: 419-424https://doi.org/10.1123/ijspp.2015-0227
        • R Core Team. R
        A language and environment for statistical computing.
        R Foundation for Statistical computing, Vienna, Austria2020
        • Hopkins W.G.
        A scale of magnitudes for effect statistics.
        Sportscience. 2006;
        • Koo T.K.
        • Li M.Y.
        A guideline of selecting and reporting intraclass correlation coefficients for reliability research.
        J Chiropr Med. 2016; 15: 155-163https://doi.org/10.1016/j.jcm.2016.02.012
        • Stockhausen W.
        • Grathwohl D.
        • Bürklin C.
        • et al.
        Stage duration and increase of work load in incremental testing on a cycle ergometer.
        Eur J Appl Physiol Occup Physiol. 1997; 76: 295-301https://doi.org/10.1007/s004210050251
        • Hauser T.
        • Bartsch D.
        • Baumgärtel L.
        • et al.
        Reliability of maximal lactate-steady-state.
        Int J Sports Med. 2013; 34: 196-199https://doi.org/10.1055/S-0032-1321719/ID/R2469-0016