Where does my first threshold sit against my threshold power?
In 14 well-trained men VT1 was 200 W against 255 W at MLSS (78%, our ratio), at 71-78% HRmax and RPE 11-12 (ramp; CIs of the mean, not one rider's band). One lab, men (B).
Main caveat: Small, single-lab, men only. Fourteen well-trained young men, one ergometer, one ramp protocol.
Pallarés and colleagues asked which capillary-lactate markers from a graded exercise test stand in for the ventilatory thresholds (VT1, VT2) and for the maximal lactate steady state (MLSS), and how well each repeats.
Fourteen well-trained male cyclists (VO2max 62.1 ± 4.6 mL/kg/min) did two identical graded tests (50 W start, +25 W per minute) and then 3-5 constant-load 30-min trials to find MLSS. Power at VT1 was 200 ± 36 W, at MLSS 255 ± 32 W and at VT2 304 ± 39 W; peak power was 388 ± 32 W.
The lactate threshold (LT, the highest workload before lactate rose above baseline, 214 ± 33 W) did not differ from VT1 (p = 0.147; bias −13.6 ± 34.3 W); LT + 0.5 mmol/L matched MLSS (p = 0.449; bias −4.5 ± 23.2 W); LT + 2.0 mmol/L and the 4-mmol/L point matched VT2.
VT1, VT2 and the lactate markers repeated well (CV 2.1-3.7%), DMAX, RER = 1 and V-slope less well (CV 5.0-10.3%). The authors give 95% CIs for heart rate and RPE at each event — VT1 at 71-78% of HRmax, 62-71% of heart-rate reserve and RPE 11-12 — and a zone scheme of their own in which the lowest training zone is 70-90% of VT1 power.
Our reading: for the question "where does easy end below threshold power", this is a clean group-mean anchor in trained men — VT1 sat about 55 W, or about a fifth, below MLSS (our arithmetic) — but the individual spread is wide (VT1 minus MLSS: −55.2 ± 30.3 W), so a fixed ratio misplaces some riders.
Protocol
Participants: n = 14 trained male cyclists; age 26.7 ± 8.2 y, body mass 70.3 ± 4.9 kg, height 173.7 ± 4.2 cm, body fat 12.5 ± 3.0% (DXA), VO2max 62.1 ± 4.6 mL/kg/min, endurance training experience 10.9 ± 4.9 y; medical examination with ECG; anyone with VO2max below 55.0 mL/kg/min at the familiarisation test was to be excluded. Tested January 2014 to July 2015; approved by the Bioethics Commission of the University of Murcia.
Design: a familiarisation test, then two identical graded tests (reliability and threshold detection), then 3-5 MLSS trials; 5-7 visits 2-5 days apart, all between 16:00 and 19:00 at 21-24 °C, 24 h of full recovery before each.
Training held during the study (5-6 weeks): 90-min rides every 48 h at the individual intensity of "nVT1" (undefined in the paper; presumably VT1) with 5-7-min efforts at 90-95% of VT2 every 20 min, to keep fitness constant.
Graded test: a warm-up (the Methods give both "10 min at 50 W" and, for the two experimental tests, three 5-min stages at 45%, 55% and 65% of peak power), 10 min passive rest, then 50 W + 25 W/min to exhaustion at 80-85 rpm (Ergoselect 200 ergometer); breath-by-breath gas exchange (Cosmed Quark B2), 12-lead ECG, heart rate (Polar RS400), capillary lactate every 2 min (Lactate Pro), Borg 6-20 RPE every 2 min.
Definitions: VT1 — rise in VE/VO2 and end-tidal PO2 without a rise in VE/VCO2; VT2 — rise in both VE/VO2 and VE/VCO2 and a fall in end-tidal PCO2; LT — the highest workload not associated with a rise in lactate above baseline; LT + 0.5 to LT + 3.0 mmol/L by interpolation; DMAX (third-order polynomial); OBLA at 4 mmol/L; V-slope; RER = 1. Two independent observers, a third if they disagreed.
MLSS: the highest constant workload at which lactate rose less than 1 mmol/L between minutes 10 and 30 of a 30-min ride; first trial at the rider's LT, then steps of 0.2 W/kg (about 15 W); 3-5 trials per rider.
Statistics: one-way repeated-measures ANOVA with Bonferroni, intraclass correlation, Bland-Altman against VT1, MLSS and VT2; reliability by CV, ICC and Bland-Altman between the two tests.
Funding and interests: "The authors received no specific funding for this work"; the authors declared no competing interests. Equipment makers are named as suppliers only.
Key Results
Power at each marker (mean ± SD), lactate at it, and its difference from VT1 and from MLSS (Table 1); test-retest CV (Table 2):
| Marker | Power (W) | Lactate (mmol/L) | Minus VT1 (W), p | Minus MLSS (W), p | Test-retest CV |
|---|---|---|---|---|---|
| VT1 | 200 ± 36 | 1.1 ± 0.4 | — | −55, p < 0.001 | 3.5% |
| LT (first rise above baseline) | 214 ± 33 | 1.6 ± 0.3 | +14, p = 0.147 | −41, p < 0.001 | 3.7% |
| V-slope | 235 ± 32 | 2.1 ± 0.6 | +35, p < 0.001 | −20, p = 0.006 | 5.0% |
| MLSS | 255 ± 32 | 4.5 ± 0.9 | +55, p < 0.001 | — | not repeated |
| DMAX | 257 ± 40 | 2.8 ± 1.4 | +57 | +2, p = 0.830 | 10.3% |
| RER = 1 | 259 ± 36 | 2.8 ± 1.3 | +59 | +4, p = 0.750 | 6.4% |
| LT + 0.5 mmol/L | 260 ± 36 | 2.1 ± 0.3 | +60 | +5, p = 0.449 | 3.7% |
| LT + 2.0 mmol/L | 300 ± 37 | 3.6 ± 0.3 | +100 | +45 | 3.4% |
| VT2 | 304 ± 39 | 4.2 ± 1.0 | +104, p < 0.001 | +49, p < 0.001 | 2.1% |
| OBLA 4 mmol/L | 304 ± 40 | 4.0 ± 0.0 | +104 | +49 | 3.7% |
| Peak power (VO2max test) | 388 ± 32 | — | — | — | 1.6% |
Agreement (Bland-Altman, bias ± SD): VT1 vs LT −13.6 ± 34.3 W (r = 0.72); VT1 vs MLSS −55.2 ± 30.3 W; MLSS vs LT + 0.5 −4.5 ± 23.2 W (r = 0.78); VT2 vs LT + 2.0 2.8 ± 24.0 W; VT2 vs OBLA −1.2 ± 28.5 W.
Individual spread (our arithmetic, bias ± 1.96 SD): VT1 minus MLSS about −115 to +4 W; VT1 minus LT about −81 to +54 W; MLSS minus LT + 0.5 about −50 to +41 W.
Reliability (two identical tests): VT1 CV 3.5% (3.6% in the Results text), ICC 0.95, bias −2.9 ± 13.3 W; VT2 CV 2.1%, ICC 0.96, bias −2.7 ± 11.4 W; LT CV 3.7%, ICC 0.85; DMAX CV 10.3%, ICC 0.57, bias 15.4 ± 42.7 W; RER = 1 CV 6.4%; V-slope 5.0%.
Heart rate and RPE at each event (Table 3, 95% CI of the group mean): VT1 (LT) 71-78% HRmax, 62-71% HRR, RPE 11-12; MLSS (LT + 0.5) 81-87% HRmax, 76-83% HRR, RPE 13; VT2 (LT + 2.0) 87-92% HRmax, 83-89% HRR, RPE 14-15.
The authors' zones (Table 4, their "personal approach", not tested): R0 = 70-90% of VT1 or LT (52-67% HRmax, RPE 8-9); R1 = 90-110% of VT1 or LT (67-82% HRmax, RPE 10-11); R2 = 90-100% of MLSS (82-87% HRmax, RPE 12-14); R3 = 95-105% of VT2 (87-95% HRmax, RPE 15-16).
Ratios from the group means (our arithmetic): VT1 = 78% of MLSS power, 66% of VT2 power and 52% of peak power; MLSS = 66% of peak power.
Mechanisms
Measured: MLSS fell between VT1 and VT2 (255 W, against 200 and 304 W), not at VT2; the authors stress that MLSS corresponds to neither threshold. Lactate at MLSS after 30 min (4.5 ± 0.9 mmol/L) was higher than at VT2 in the ramp (4.2 ± 1.0), at a lower power.
Authors' interpretation: VT1 is where CO2 from bicarbonate buffering of lactate first drives ventilation up, and VT2 (the respiratory compensation point) is the highest metabolic rate with an elevated but stable acidosis; training in each band produces different adaptations. Ventilation responds more sensitively to muscle acidosis than blood lactate does.
Inference: the band between VT1 and MLSS — about 55 W wide in these riders — is the heavy domain, where "tempo" and "sweet spot" rides sit; a ride is easy only below VT1. Because a ramp's lactate lags behind a constant-load ride's, a fixed mmol/L value from a ramp does not mark MLSS: the 4-mmol/L point matched VT2 (304 W), not MLSS (255 W, where 30-min lactate was 4.5 mmol/L); the marker that matched MLSS was relative to the rider's own baseline (LT + 0.5).
Inference: the authors' lowest zone (R0, 70-90% of VT1) is the band most easy and recovery rides in amateur plans fall into; the paper names it but does not test what it does.
Practical Application
Where does my first threshold sit against my threshold power? In 14 well-trained men VT1 was 200 W against 255 W at MLSS (78%, our ratio), at 71-78% HRmax and RPE 11-12 (ramp; CIs of the mean, not one rider's band). One lab, men (B).
From % of FTP to % of VT1, as arithmetic across studies: FTP was not measured here. If FTP sat about 7% above MLSS, as in trained cyclists in REF-1325 (MLSS = 93.1% of FTP), these riders' VT1 would be near 73% of FTP; a ride at 65% of FTP would be about 89% of VT1 (the "10-15% below VT1" band of REF-705) and one at 50% about 68% of VT1, just under the authors' R0 band (our arithmetic on group means of different riders). Treat it as the average case only.
The individual spread is larger than the gap between zones: the 95% limits of VT1 minus MLSS were about −115 to +4 W (our arithmetic from the Bland-Altman SD), so in one rider "easy" ends a few watts below threshold and in another 100 W below. A measured first threshold — lab VT1, lactate first rise, or a validated field estimate (SYNTH-53) — replaces any ratio.
A lactate test can stand in for gas exchange, on average: the first rise above baseline matched VT1 within 14 W as a group, baseline + 0.5 mmol/L matched MLSS and baseline + 2.0 mmol/L matched VT2; for one rider LT and VT1 could differ by about 50-80 W (our arithmetic). DMAX and RER = 1 repeated poorly (CV 10.3% and 6.4%); avoid them when other markers are available (the authors' advice).
Heart rate and RPE as rough checks: VT1 fell at 71-78% of HRmax and RPE 11-12 (between "light" and "somewhat hard") as group-mean CIs taken in the 25 W/min ramp, the only protocol the authors apply them to (heart rate lags power in a fast ramp, so a steady ride at VT1 power will likely show a higher share of HRmax; inference); across 100 adults the lactate threshold ranged far wider in % HRmax (REF-038), so a heart-rate ceiling from a percentage is a starting point, not a measurement (SYNTH-38).
Not from this paper; the base's own practice: the base caps easy riding at about 70% of FTP and a recovery spin at about 55% (SYNTH-70); by the arithmetic above, the 70% cap sits a little below an average trained man's VT1 and the 55% spin in the authors' R0 band.
Limitations
Small, single-lab, men only. Fourteen well-trained young men, one ergometer, one ramp protocol; the authors' own limitation is that other ramp designs, step tests or other sports may change these relationships.
Group means, not individual bands. The heart-rate and RPE values are 95% CIs of the group mean; individual riders spread much wider. MLSS was determined once, so its own repeatability is not reported.
The zone table is the authors' scheme (Table 4, "personal author's approach"), derived from an earlier publication of theirs; its heart-rate bands are not tested here.
Reporting (things that do not add up): in Table 4 the lowest zone gives 53-62% of heart-rate reserve against 52-67% of HRmax, but at any heart rate the reserve percentage is lower than the HRmax percentage, so the lower bounds cannot both hold; the Discussion describes the protocol as "75 W every 4 min" while 25 W per minute makes 100 W in 4 min (our arithmetic); VT1's CV is 3.5% in Table 2 and 3.6% in the Results text; the MLSS row of Table 1 lists the LT + 2.5 difference as 66 W where the means give 56 W (311 − 255) and the bias is −55.7 W; lactate at LT (1.6 ± 0.3 mmol/L) is higher than at VT1 (1.1 ± 0.4) although LT is defined as the last workload before lactate rises above baseline; the training intensity is written "nVT1" without definition.
No FTP, CP or performance test. Translating to % of FTP needs other studies and is arithmetic, not a finding.
Funding and interests: no specific funding; no competing interests declared; no device maker, federation or data platform funded the work.
How this entry was checked and changed
Why this level
a test validation study in one laboratory with n = 14 (the rubric's B for validation studies; single-lab, small n). Read from the full text (open access, CC BY 4.0)
Full text (checked 2026-10-28)
VT1 workload (200 ± 36 W) was different to the workload for the rest of the CBL thresholds except for the LT threshold (214 ± 33 W, p = 0.147) (full text)
The workload at MLSS (255 ± 32 W); VT2 workload (304 ± 39 W); V˙O2max workload 388 ± 32 W in Table 2 (full text)
Bland-Altman: VT1 vs LT bias -13.6 ± 34.3; VT1 vs MLSS bias -55.2, SD of bias 30.3; MLSS vs LT+0.5 bias -4.5 ± 23.2 (full text)
Table 3, 95% confidence interval: VT1 (LT) 71%—78% HRMax, 62%—71% HRR, RPE 11—12; MLSS (LT+0.5) 81%—87%, 76%—83%, RPE 13—13; VT2 (LT+2.0) 87%—92%, 83%—89%, 14—15 (full text)
Table 4 zone R0: 70%—90% VT1 or LT, 52%—67% HRMax, 53%—62% HRR, RPE 8—9 (full text)
Reliability: VT1 and VT2 revealed low CV (3.6% and 2.1%), high ICC (r = 0.95–0.96); Table 2 CV for VT1 3.5%, DMAX 10.3% (full text)
Fourteen trained men cyclists, V˙O2max 62.1 ± 4.6 ml·kg·min^-1, endurance training experience 10.9 ± 4.9 yr (full text)
Training during the study: cycling sessions of 90 minutes every 48 hours at the individual intensity of nVT1 interspersed with efforts of 5–7 min at 90–95% intensity of VT2 each 20 min (full text)
The authors received no specific funding for this work; the authors have declared that no competing interests exist (full text)