Where is my Zone 2?
In 50 amateur cyclists VT1 sat at 51.5 ± 5.9% of ramp peak power and 81% HRmax; FatMax about a quarter lower; 72-82% HRmax spans both. A measured VT1 beats any percentage. FTP not tested (B).
Main caveat: Cross-sectional, one lab, no training outcome. It shows where markers sit, not which one gives the better adaptation.
Meixner, Filipas, Holmberg and Sperlich asked whether the markers commonly used to set "Zone 2" in a five-zone model land at the same intensity across riders.
Fifty experienced amateur cyclists (30 men, 20 women; 31.2 ± 7.8 y; VO2peak 54.1 ± 8.1 mL/kg/min) did a ramp test (100 W, +25 W every 30 s) and, within 7 days, a step test (100 W, +25 W every 3 min) with gas exchange, heart rate and earlobe lactate.
In the step test the authors located nine markers: VT1, FatMax, baseline lactate + 0.5 mmol/L, fixed lactate of 1.5, 2.0 and 2.5 mmol/L, 65% of VO2peak, and 72% and 82% of HRmax. VT1 averaged 200 ± 45 W (51.5 ± 5.9% of ramp peak power; 69.8 ± 8.9% of VO2peak; 81.4% of HRmax).
Baseline + 0.5 mmol/L (199 W), 1.5 mmol/L (194 W) and 82% HRmax (202 W) did not differ from VT1 in power; FatMax (125 W) and 72% HRmax (148 W) did not differ from each other but sat well below VT1; 2.0 and 2.5 mmol/L sat above it (211 and 224 W). Between-rider CVs of absolute power ran 22.7-29.6%.
The authors conclude that fixed percentages of HRmax and fixed lactate values misplace individuals, that VT1, baseline + 0.5 mmol/L or a measured FatMax should set Zone 2, and that "Zone 2" names different metabolic states depending on the marker.
Our reading: the useful numbers here are the relative positions and their SDs, not the absolute-power CVs, which mostly reflect rider size and fitness; the ±5.9-point SD of VT1 as a share of ramp peak is the spread any %-based easy zone has to live with.
Protocol
Participants: n = 50 (30 men, 20 women) experienced cyclists, more than 3 years of regular cycling and more than 2 sessions a week, used to clipless pedals; age 31.2 ± 7.8 y (men 34.3, women 26.5), body mass 71.7 ± 12.4 kg, body fat 13.9 ± 4.8%, VO2peak 54.1 ± 8.1 mL/kg/min. Ethics: University of Würzburg (EV2024/1-1004).
Visits: two, at least 48 h apart within 7 days; usual diet repeated 24 h before (diary), a carbohydrate-rich meal at least 3 h before, no caffeine on test days; 35 g of a carbohydrate drink (IsoFast, DextroEnergy) in 500 mL water ad libitum during warm-up and recovery. Warm-up 10 min at 1.5 W/kg plus a short sprint.
Ramp (visit 1): 2 min at 100 W, then +25 W every 30 s to exhaustion or a cadence drop of more than 10 rpm; verification bout at the last completed load after 3 min at 75 W; VO2peak = highest 30-s mean; peak power = PPOramp.
Step test (visit 2): 100 W (75 W for three riders), +25 W every 3 min to exhaustion; PPOstep interpolated from the time in the last stage. Gas exchange averaged over 10 s, final 30 s of each step used (Cosmed Quark CPET, Hans Rudolph mask); heart rate by Polar H10 chest strap; earlobe lactate before the test and in the last 15 s of each step (Biosen C-Line).
Markers: VT1 (method of their ref. 62, two blinded raters, interpolated to power); FatMax (power at the highest fat oxidation in the final 30 s of a step, fat oxidation = 1.695 × VO2 − 1.701 × VCO2); baseline lactate + 0.5 mmol/L and 1.5, 2.0, 2.5 mmol/L (third-order polynomial); 65% of VO2peak; 72% and 82% of HRmax (individual linear regressions, R² ≥ 0.95). Nine riders never had lactate below 1.5 mmol/L and were excluded from that marker (n = 41).
Statistics: mixed-effects model with Tukey comparisons; between-rider CV = SD/mean of absolute values (a robust CV for FatMax, given as median and median absolute deviation); Bland-Altman; t-tests for sex.
Funding and interests: "The authors state no funding involved"; "The authors declare no conflicts of interest." The carbohydrate drink brand is named as the product used; no sponsorship is stated.
Key Results
Each "Zone 2" marker in the step test, mean ± SD (Table 2):
| Marker | Power (W) | % of ramp peak power | % of VO2peak | Heart rate (beats/min) | % of HRmax | CV of power |
|---|---|---|---|---|---|---|
| FatMax (median ± MAD for power) | 125 ± 37 | 38.7 ± 8.9 | 56.0 ± 10.7 | 130 ± 12 | 72.3 ± 5.7 | 29.6% |
| 72% HRmax | 148 ± 41 | 38.0 ± 6.7 | 55.2 ± 8.8 | — | 72 by definition | 27.6% |
| 65% of VO2peak | 183 ± 49 | 47.1 ± 6.5 | 65 by definition | 140 ± 15 | 78.2 ± 6.0 | 26.0% |
| Lactate 1.5 mmol/L (n = 41) | 194 ± 56 | 46.0 ± 13.6 | 65.1 ± 13.2 | 141 ± 10 | 78.1 ± 6.9 | 29.0% |
| Baseline lactate + 0.5 mmol/L | 199 ± 45 | 51.2 ± 4.8 | 69.4 ± 6.7 | 146 ± 10 | 81.2 ± 3.0 | 22.8% |
| VT1 | 200 ± 45 | 51.5 ± 5.9 | 69.8 ± 8.9 | 146 ± 11 | 81.4 ± 3.9 | 22.8% |
| 82% HRmax | 202 ± 46 | 51.9 ± 5.6 | 70.3 ± 8.2 | — | 82 by definition | 22.7% |
| Lactate 2.0 mmol/L | 211 ± 53 | 54.1 ± 6.7 | 72.5 ± 8.3 | 150 ± 9 | 83.4 ± 3.3 | 24.9% |
| Lactate 2.5 mmol/L | 224 ± 52 | 57.8 ± 6.1 | 76.4 ± 7.7 | 154 ± 9 | 86.0 ± 3.0 | 23.3% |
| Maximum | ramp 387 ± 73; step 300 ± 56 | — | VO2peak 3866 ± 840 mL/min | 179 ± 10 | — | 18.9% (ramp) |
Which markers coincide (power): no significant differences among baseline + 0.5 mmol/L, 1.5 mmol/L, VT1 and 82% HRmax, nor between 72% HRmax and FatMax (p > 0.05); all other pairs differed (p < 0.01).
Bland-Altman: VT1 and baseline + 0.5 mmol/L had small bias and narrow limits; VT1 against 82% HRmax had more bias and wider limits; FatMax and 72% HRmax had no mean difference but wide limits (values shown in figures only).
Women and men: relative to step-test peak power, women sat lower than men for baseline + 0.5 mmol/L (p = 0.0023), 72% HRmax (p = 0.0197), 82% HRmax (p = 0.0105), 2.0 mmol/L (p = 0.0116) and 2.5 mmol/L (p = 0.0167); women's heart rate at FatMax was a higher share of HRmax (p = 0.0188). VT1 is not among the markers reported to differ by sex.
Positions from the group means (our arithmetic): FatMax power is about 63% of VT1 power (median against mean); the 72-82% HRmax band spans about 74-101% of VT1 power; 65% of VO2peak is about 92% of VT1 power; VT1 is about 67% of step-test peak power.
Mechanisms
Measured: VT1 and the rider's own lactate baseline + 0.5 mmol/L fell at the same power; the fixed 2.0-mmol/L point sat above VT1 and FatMax well below it. Heart rate at VT1 was a fairly tight share of HRmax (81.4 ± 3.9%), power at VT1 a looser share of ramp peak (51.5 ± 5.9%) and of VO2peak (69.8 ± 8.9%).
Authors' interpretation: the spread comes from individual differences in mitochondrial density, lactate production and clearance, oxygen transport and fibre-type mix; VT1 marks the shift from fat towards carbohydrate as the main fuel, while FatMax, about 25% lower than VT1 as a share of peak power, is a distinct intensity.
Inference: the between-rider CVs of absolute power (22.7-29.6%) are dominated by body size, sex and fitness (body mass 71.7 ± 12.4 kg, VO2peak SD 8.1), so they overstate how imprecise a marker is for prescription; the SD of each marker's relative position is the relevant spread. The heart-rate markers' "low variation" is partly built in: 72% and 82% of HRmax vary only as much as HRmax does.
Inference: a steep ramp (25 W per 30 s) inflates peak power relative to a slower ramp (387 vs 300 W at the step test's end), so every "% of peak power" here is ramp-specific.
Practical Application
Where is my Zone 2? In 50 amateur cyclists VT1 sat at 51.5 ± 5.9% of ramp peak power and 81% HRmax; FatMax about a quarter lower; 72-82% HRmax spans both. A measured VT1 beats any percentage. FTP not tested (B).
In % of FTP, as our arithmetic only: if FTP were 70-75% of this ramp's peak (a platform convention for slower ramps; no entry validates a ramp-to-FTP factor, SYNTH-53; on this steeper 50 W/min ramp the factor probably overstates FTP, so the %-of-FTP figures that follow are likely low), VT1 would average about 69-74% of FTP with an SD of about 8 points; assuming a normal spread, about one rider in seven to one in three would have VT1 at or below 65% of FTP, and 50% of FTP would sit at or just below the FatMax and 72%-HRmax level for the average rider.
Heart rate: the common 72-82% HRmax "Zone 2" covered everything from FatMax to VT1 on average; 82% of HRmax matched VT1 as a group mean but not reliably for one rider. With VT1 at 81.4 ± 3.9% of HRmax, a ceiling of 82% of a measured HRmax would put roughly half of riders at or above their VT1 (our inference from the group spread) — treat the top of that band as a ceiling, not a target (SYNTH-38).
Lactate: use the rider's own baseline + 0.5 mmol/L, not a fixed 2.0 mmol/L (which sat above VT1 here, at 83% HRmax); 1.5 mmol/L could not be found in 9 of 50 riders.
FatMax is not the top of Zone 2: in these riders it sat about 50-75 W below VT1 (our arithmetic: 75 W from FatMax's median power against VT1's mean, about 50 W from the mean shares of ramp peak, 38.7% vs 51.5% of 387 W); riding at FatMax is genuinely easy, and riding at VT1 is not "fat-burning pace".The claim that Zone 2 equals FatMax (REF-236, a blog post, C) does not fit these data; the base's reviews find Zone 2 not superior for fat oxidation or mitochondria (REF-1092).
Limitations
Cross-sectional, one lab, no training outcome. It shows where markers sit, not which one gives the better adaptation; the authors list longitudinal comparisons and FTP as future work.
Two protocols mixed: markers came from the step test and are expressed against the ramp's peak, which the authors themselves flag as a possible source of skew.
CVs of absolute values mix men and women of very different size; the relative SDs in Table 2 are the more useful spread. FatMax came from the final 30 s of 3-min steps by a stoichiometric equation, a method the authors say may add variability, and its power is a median with median absolute deviation.
Reporting (things that do not add up): the abstract says VT1 and FatMax "showed strong alignment", but Table 2 (200 vs 125 W) and the Discussion ("FatMax is generally ∼25% lower than VT1") say the opposite; the Results say 72% HRmax, 82% HRmax and 2.5 mmol/L showed the lowest variation; in Table 2, 82% HRmax has the lowest power CV (22.7%) and 2.5 mmol/L the lowest heart-rate CV (6.0%), but 72% HRmax's power CV (27.6%) is among the highest, and the two HRmax markers have no heart-rate CV (they are fixed fractions of HRmax); two p-value statements are inverted ("no significant differences, p < 0.05" and "significant differences (p > 0.01)"); Table 2 prints VT1's heart-rate share as "81.4 ±0 .3.9".The paper attributes a 1.5-2.5 mmol/L Zone 2 to the Norwegian Olympic federation and cites [10] (a 2014 comparison of intensity-analysis methods), while the base's Olympiatoppen guideline entry gives I-2 as about 1.0-2.0 mmol/L (REF-1077).
Funding and interests: no funding; no conflicts declared; no device maker, federation or data platform named as a funder.
How this entry was checked and changed
Why this level
cross-sectional measurement study with n = 50 (the rubric's B+ for this design needs 100 or more); one lab, no training outcome. Read from the full text (open access, CC BY 4.0)
Full text (checked 2026-10-28)
A cohort of n = 50 (n = 30 males and n = 20 females) experienced cyclists with more than three years of regular cycling exercise (> 2 sessions per week); relative oxygen uptake 54.1 ± 8.1 mL/min/kg (full text)
VT1: oxygen uptake 2690 ± 612 mL/min, 69.8 ± 8.9% of peak, heart rate 146 ± 11 b/min, power 200 ± 45 W, 51.5 ± 5.9% of PPOramp, CV 22.8% (full text)
FatMax: 2155 ± 407 mL/min, 56.0 ± 10.7% of peak, 130 ± 12 b/min, 72.3 ± 5.7% of HRmax, 125 ± 37 W, 38.7 ± 8.9% of PPOramp (full text)
HR72%: 148 ± 41 W, 38.0 ± 6.7% of PPOramp; HR82%: 202 ± 46 W, 51.9 ± 5.6%; BLamin+0.5: 199 ± 45 W, 51.2 ± 4.8%; BLa2.0: 211 ± 53 W, 54.1 ± 6.7% (full text)
Maximum value: 3866 ± 840 mL/min, 179 ± 10 b/min, 387 ± 73 W (Ramp), 300 ± 56 W (Step) (full text)
Table 2 shows that FatMax is generally ∼25% lower than VT1 — indicating that these markers represent distinct intensity domains (full text)
Nine participants did not reach a lactate value lower than 1.5 mmol/L and were excluded from this analysis (full text)
The participants began cycling at 100 W for 2 min, after which the load increased by 25 W every 30 s (full text)
The authors state no funding involved; the authors declare no conflicts of interest (full text)