REF-1762 · Hunter, Fujii, Maunder & Muniz-Pumares, 2026 · Scandinavian Journal of Medicine & Science in Sports, · Paper · PubMed · Free full text
After 90 min of running just above the first ventilatory threshold, on water only, speed at VT1 fell 8.6% at 19 °C and 11.1% at 30 °C and VO2peak fell 4.0-4.8%, with no detectable difference in that fall between conditions; but rested sVT1 was already about 7% lower in the warm, so after the run it sat 17.1% below the rested temperate value, and end-test speed fell more in the warm (12.1% vs 6.8%); a randomised crossover of 10 trained runners, read in full
Main caveat: Ten completers out of 15, and two of the dropouts could not finish the warm run, so the warm results describe the more heat-tolerant runners (the authors).
Hunter and colleagues asked whether running in the heat speeds up the loss of the physiological determinants of distance-running performance during a long run.
Ten trained runners did incremental treadmill tests in temperate (19 °C) and warm (30 °C) chambers, both at 40% relative humidity, first fresh and then after 90 min of running in the same condition at a speed set from that condition's own fresh test (5% of the way from sVT1 to end-test speed, about 103% of VO2 at VT1).
Fresh, sVT1 was about 0.8 km/h (about 7%) lower in the warm. After 90 min it fell in both conditions (8.6% temperate, 11.1% warm) and VO2peak fell too (4.8% and 4.0% relative to body mass), with no condition × time interaction for either.
Running economy worsened more in the warm across the run (+5.4% vs +1.7% from 15 to 90 min, significant interaction but no single time point significant after correction), and end-test speed fell more in the warm (12.1% vs 6.8%). Because the warm run started from a lower baseline, post-run sVT1 in the warm averaged 17.1% below the fresh temperate value (9.6 vs 11.5 km/h).
The authors conclude that heat did not detectably change how much the threshold and VO2peak fell when the run was matched to condition-specific capacity, but that runners in warm weather start from a lower physiological baseline and lose more economy, so pace should be set from the conditions of the day.
Participants
15 non-heat-acclimated volunteers started; 10 completed (7 M / 3 F; age 38 ± 11 years; 1.74 ± 0.11 m; 68.2 ± 13.8 kg; VO2peak 56.0 ± 5.7 mL/kg/min). Two could not finish the warm 90-min run (both said they felt unwell with heavy legs), one withdrew injured, one ill, one data set was lost to a technical fault.
Entry: at least 3 h of running a week; a recent 10 km under 45 min or 5 km under 22 min for men (women's criteria relaxed during recruitment from under 50:00 / 24:00 to under 52:30 / 25:00); at least one run of 90 min or more in the past month; no injury in the last 3 months; no hot-climate travel, heat acclimation or sauna protocol in the previous 4 weeks. Testing in the UK summer, so some seasonal acclimatisation cannot be excluded (the authors).
Menstrual cycle phase and hormonal contraception were not controlled. Carbon-plated shoes were not allowed.
Protocol
Design: four visits within 4 weeks at about the same time of day. Block 1: fresh incremental tests in TEMP (19 °C) and WARM (30 °C), both 40% RH, random order, at least 24 h apart. Block 2: the 90-min run plus a repeat incremental test in each condition, random order, at least 72 h apart. Block 1 always preceded Block 2.
Incremental test: started within 2 min of entering the chamber, at 8.5-9.1 km/h on average, +1 km/h every 3 min until the respiratory exchange ratio reached 1.0 with a clear rise in VE/VO2, then +1 km/h every minute to failure. VT1 by ventilatory equivalents and end-tidal PO2, agreed by two researchers; sVT1 from VO2 at the end of each 3-min step; VO2peak as the highest 30-s mean; end-test speed as the last completed stage plus the completed fraction of the next.
90-min run: at 5% Δ between sVT1 and end-test speed for that condition: 11.8 ± 0.7 km/h (TEMP) vs 11.0 ± 1.1 km/h (WARM), 17.7 ± 1.1 vs 16.5 ± 1.6 km; relative intensity at 15 min 103.4 ± 4.1% vs 102.1 ± 3.0% of VO2 at VT1 (no difference). Water ad libitum at about 37 °C; no carbohydrate given. Then 5 min rest in the chamber and the repeat incremental test. Core temperature cut-off 39.5 °C (nobody reached it).
Measures: gas exchange (MetaLyzer 3B), running economy as mL/kg/km corrected for sweat-driven body-mass loss, energy cost, substrate oxidation (Jeukendrup and Wallis equations), heart rate, gastrointestinal and skin temperature, RPE (legs and overall), dyspnoea, thermal comfort and sensation, every 15 min.
Power: a priori 7 participants for main effects of time and heat; the authors state it was not powered for the condition × time interaction.
Funding and interests: the first author's research visit was funded by the Great Britain Sasakawa Foundation; no conflicts of interest declared.
Key Results
Fresh (PRE) and after 90 min (POST), mean ± SD (Table 2):
| Measure | TEMP PRE → POST | WARM PRE → POST | Condition × time |
|---|---|---|---|
| Speed at VT1, km/h | 11.5 ± 0.7 → 10.5 ± 1.1 (−8.6%) | 10.7 ± 1.1 → 9.6 ± 1.3 (−11.1%) | p = 0.312 |
| VO2 at VT1, mL/kg/min | 41.9 ± 3.5 → 40.3 ± 4.0 | 39.8 ± 4.1 → 38.5 ± 4.7 | p = 0.706 (time p = 0.058) |
| Heart rate at VT1, beats/min | 148 ± 16 → 145 ± 15 | 151 ± 17 → 149 ± 18 | no effect of time (p = 0.291) or condition |
| Energy expenditure at VT1, kcal/min | 14.1 ± 2.6 → 12.7 ± 2.4 | 13.1 ± 2.1 → 11.8 ± 2.1 | p = 0.789 |
| VO2peak, mL/kg/min | 56.0 ± 5.7 → 53.2 ± 6.2 (−4.8%) | 56.8 ± 6.6 → 54.4 ± 6.4 (−4.0%) | p = 0.750 |
| VO2peak, L/min | 3.81 ± 0.82 → 3.44 ± 0.49 (−8.3%) | 3.85 ± 0.82 → 3.55 ± 0.65 (−7.4%) | p = 0.505 |
| End-test speed, km/h | 17.0 ± 1.4 → 15.8 ± 1.7 (−6.8 ± 5.0%) | 16.7 ± 1.4 → 14.7 ± 2.0 (−12.1 ± 6.9%) | p = 0.030 |
| Maximal heart rate, beats/min | 180 ± 16 → 175 ± 14 | 183 ± 16 → 180 ± 13 | p = 0.700 |
Warm vs temperate, rested: sVT1 about 0.8 km/h (about 7%) lower in WARM (main effect of condition p < 0.001); VO2 at VT1 about 5% and energy expenditure at VT1 about 7% lower; VO2peak not different.
Extra sVT1 loss in WARM: −0.21 km/h (95% CI −0.65 to +0.23); the smallest difference this sample could detect with 80% power was 0.61 km/h. The authors: the data neither show extra loss in the warm nor establish equal durability.
Combined effect: post-run sVT1 in WARM was 17.1% below fresh sVT1 in TEMP (9.6 vs 11.5 km/h).
Running economy over the run: +1.7 ± 3.1% (TEMP) vs +5.4 ± 6.4% (WARM) from 15 to 90 min; time × condition p = 0.042 (energy cost p = 0.049), but no time point differed from 15 min after Bonferroni correction.
During the run: peak core temperature 38.9 ± 0.5 vs 38.3 ± 0.3 °C, peak skin temperature 35.0 ± 0.5 vs 31.6 ± 1.9 °C (WARM vs TEMP); body-mass loss 3.5 ± 0.9% vs 2.7 ± 0.8%; heart rate rose steadily from 30 min in WARM (interaction p < 0.001); total energy 1228 ± 196 vs 1300 ± 223 kcal; carbohydrate oxidised 131 ± 45 vs 146 ± 50 g (p = 0.311); overall RPE at 90 min 13.6 ± 1.6 vs 13.5 ± 1.6, with no condition effect on RPE or dyspnoea, while thermal comfort and sensation were worse in WARM.
Incremental test duration: shorter after the run in both conditions (TEMP 21.9 → 18.3 min; WARM 21.6 → 17.5 min).
Practical Application
After 90 min of steady running the first threshold had moved down by about 9-11% in speed, and in the warm it started about 7% lower: the warm, fatigued runner's threshold speed sat on average 17.1% below the value from a rested, cool-weather test (10 trained runners, 19 vs 30 °C, water only). For long runs and races in warm weather, set the threshold-based pace from the day's conditions and expect it to have slid by the 90-min mark.
What heat changed was the starting point and the economy, not the threshold's rate of slide. With the run matched to each condition's own capacity, the drop in sVT1 and VO2peak did not differ detectably between 19 and 30 °C, but the study could only detect a gap of 0.61 km/h or more, so a smaller extra loss in the heat is not ruled out.
Top-end speed suffered most in the warm: end-test speed fell 12.1% against 6.8%, with VO2peak falling alike, which the authors tie to the larger loss of running economy. Inference, not tested here: a late surge or finishing kick costs more on a warm day than the rested numbers suggest.
Heart rate at the threshold did not rise after the run (148 → 145 and 151 → 149 beats/min; no effect of time, p = 0.291), while speed at it fell 8.6-11.1%.This is heart rate at VT1 re-located in a new incremental test begun after 5 min of rest, not heart rate at a fixed speed during the run. SYNTH-10 and SYNTH-32 cite REF-179, in which heart rate at the re-measured moderate-to-heavy boundary rose 9-10 beats/min after 2 h of cycling, and SYNTH-10 infers from it that a fresh heart-rate cap is too low late in long efforts.In these 10 runners after 90 min, a fresh VT1 heart rate still sat at the re-measured threshold, while a fresh VT1 speed would have been 8.6-11.1% too fast. This does not overturn the guides' cycling inference: different sport, 90 min against 2 h, one small study.Inference, not tested here: heart rate during the run itself drifted upward, more in the warm, so a heart-rate cap held through continuous running was not tested.
Agrees with [SYNTH-10]'s heat rule ("lower the target for heat"; the base prescribes no per-degree correction): this study gives one measured pair, about 7% lower sVT1 at 30 °C than at 19 °C at 40% humidity, in non-acclimated runners; it is not a correction factor for other conditions.
Limitations
Ten completers out of 15, and two of the dropouts could not finish the warm run, so the warm results describe the more heat-tolerant runners (the authors).
Not powered for the condition × time interaction; the durability measure (percentage change in sVT1) has poor test-retest reliability in earlier work (ICC about 0.07-0.25, typical error about 3.6-3.9 percentage points, from a different protocol), so modest differences between conditions could be missed.
The run was matched to each condition's capacity, so in the warm the runners ran slower, covered 1.2 km less and spent less energy; a pace held fixed in the heat was not tested.
Moderate heat only (30 °C, 40% RH); water only, given at about 37 °C, which may have limited drinking; hydration status was not formally measured and body-mass loss was larger in the warm, so heat and fluid loss cannot be separated (the authors).
No carbohydrate, no blood lactate, no direct measures of cardiovascular, neuromuscular or muscle-metabolic function; chamber oxygen was not logged continuously.
Menstrual phase and contraception not controlled in the 3 women; not pre-registered; 90 min, not marathon or ultra duration.
How this entry was checked and changed
Why this level
a crossover of 10 (the rubric's B row for crossovers under 15); speed at VT1, VO2peak, end-test speed and running economy are capacity outcomes, not markers, so no step down