Men who started a 2-h moderate ride better hydrated, after randomised hydration counselling, saw perceived exertion and thirst climb more slowly and more often rode the full 120 min before the 2-3% body-mass-loss stop (75% vs 29%) than men on their habitual intake, in a trial of 30 physically active men at 22 °C; core temperature from a skin sensor did not significantly predict RPE once group was in the model (p = 0.81), though the better-hydrated group was also fitter at baseline
Main caveat: Randomised, but the groups differed in VO2max at baseline (51.2 vs 46.1 mL/kg/min) and starting the ride depended on reaching the allocated group's hydration…
Skotniczny and colleagues asked whether core temperature, perceived exertion and thirst rise together during a long moderate ride, and whether starting hydration status changes those links.
Thirty physically active men were randomised to structured hydration counselling or to their habitual drinking for a qualification phase of up to 4 weeks; the counselled group's fasting urine specific gravity averaged 1.009 and the habitual group's 1.019 at baseline.
Each man then rode at 50% of his maximal power for up to 120 min at 22 °C without drinking, stopping at 2-3% body-mass loss or when he could no longer hold cadence; a CORE heat-flux sensor estimated core temperature and RPE, thirst and heart rate were logged every 15 min. Core temperature rose and then levelled off around 60-75 min, while RPE and thirst kept climbing.
In the adjusted model group (the authors' "hydration status") predicted RPE strongly and core temperature did not; the habitual-intake group's RPE and thirst rose faster and fewer of them reached 120 min. The authors conclude that starting euhydrated lowers perceived effort and thirst in long thermoneutral cycling, and that thirst and heart rate can serve as field warnings.
The groups were not balanced for fitness, which the adjustment can only partly handle.
Participants
n = 30 men, age 34.7 ± 6.1 years, body mass 80.6 ± 8.9 kg, VO2max 48.8 ± 5.8 mL/kg/min; inclusion age 25-60, fat mass under 30%, no acute disease.
Counselling group (EXP) n = 16: VO2max 51.2 ± 6.2 mL/kg/min, urine specific gravity 1.009 ± 0.008.
Habitual-intake group (CON) n = 14: VO2max 46.1 ± 4.0 mL/kg/min (p = 0.01 vs EXP), urine specific gravity 1.019 ± 0.005 (p = 0.001).
Sport, training volume and competition level: not reported.
Protocol
Qualification phase of up to 4 weeks: physician screening, ramp VO2max test on a Lode ergometer, three weekly fasting urine specific gravity checks; starting the trial depended on meeting the hydration expectation of the allocated group.
Counselling for EXP (applied to training outside the laboratory): 2.5 L of fluid a day; 5-7 mL/kg 4 h before exercise and a further 3-5 mL/kg 2 h before if urine colour scored 4 or more on an 8-point chart; 200-300 mL of water every 15 min during exercise; 120-150% of body-mass loss replaced afterwards. CON kept their usual habits.
Test day: 3-5 mL/kg of water on waking, a breakfast of about 4.5 kcal/kg 2 h before, 3-5 mL/kg of fluid 2 h before, then no drinks until all post-exercise measures.
Ride: 50% Wmax at 80-90 rpm, up to 120 min or until 2-3% body-mass loss, ended early if cadence could not be held; 22 °C, 45% RH, air speed under 0.2 m/s, upper body bare; paused every 15 min (up to 2 min, clock stopped) for ratings and nude weighing.
Measures every 15 min: core temperature (CORE heat-flux sensor, which the authors treat as approximate), heart rate (Polar H10), RPE (0-10 sheet with Borg anchors), thirst (0-10).
Analysis: ΔMean and ΔPeak are the mean and the largest of the change scores between consecutive 15-min points from 15 to 90 min; ANCOVA with baseline as covariate; data after 90 min were excluded from the primary analyses because dropouts were not random.
Key Results
| Outcome (15-90 min unless stated) | EXP (counselled) | CON (habitual) | Between groups |
|---|---|---|---|
| Rode the full 120 min | 75.0% (shortest 90 min) | 28.6% (most often stopped at 105 min, 57.1%; shortest 75 min) | χ2 = 10.57, p = 0.014, Cramér's V 0.55 |
| ΔRPE mean, median [IQR] | 0.50 [0.46; 0.63] | 1.00 [1.00; 1.14] | reported difference +0.23 (95% CI 0.13 to 0.34), p < 0.001, d = 0.89 |
| ΔRPE peak, median [IQR] | 1.00 [1.00; 1.00] | 2.50 [2.00; 3.00] | reported difference +0.53 (0.32 to 0.74), p < 0.001, d = 1.02 |
| ΔThirst mean, median [IQR] | 0.63 [0.63; 0.71] | 1.13 [1.00; 1.14] | reported difference +0.23 (0.14 to 0.32), p < 0.001, d = 1.11 |
| ΔThirst peak | 2.00 [2.00; 2.00] | 3.00 [3.00; 4.00] | reported difference +0.65 (0.30 to 1.01), p < 0.001, d = 0.74 |
| ΔCore temperature mean | 0.13 [0.09; 0.17] °C | 0.08 [0.04; 0.10] °C | not significant, p = 0.83 |
| ΔCore temperature peak | 0.39 [0.28; 0.58] °C | 0.36 [0.20; 0.45] °C | not significant, p = 0.12 |
| ΔHeart rate mean | 3.87 [2.43; 5.14] bpm | 3.50 [3.00; 4.43] bpm | not significant, p = 0.18 |
| RPE at 90 min, median | 3.5 | 7.0 | descriptive |
ANCOVA for RPE (overall F = 13.76, p < 0.001): group ("hydration status") partial η2 = 0.41, p < 0.001; core temperature partial η2 = 0.002, p = 0.81; VO2max partial η2 = 0.13, p = 0.058.
ANCOVA for thirst: group partial η2 = 0.27 (p = 0.005) with pre-exercise urine specific gravity as covariate (itself p = 0.35), and 0.50 (p < 0.001) with body fat as covariate; VO2max p = 0.051 and p = 0.48.
Correlations across all 30 (Spearman): ΔRPE with ΔThirst r = 0.84, with urine specific gravity 0.54, with body fat 0.39, with VO2max −0.62 and with ΔCore temperature −0.38; ΔThirst with urine specific gravity 0.62 and with VO2max −0.45.
Core temperature rose until about 60-75 min and then plateaued in CON; RPE and thirst rose at every step in both groups.
Practical Application
Starting a long steady session under-hydrated is one ordinary reason for RPE to run high at an unchanged load. In this trial men on their habitual intake (morning urine specific gravity about 1.019) saw RPE and thirst climb faster over a 2-h moderate ride and fewer rode the full 120 min before the body-mass-loss stop than counselled men (about 1.009), with no drinking during the ride; the counselled group was also fitter, so part of the gap may be fitness.
Where it meets [SYNTH-39]: the guide's list of causes for an sRPE that runs high against the power score names heat, a stale threshold, sleep and stress, illness, under-fuelling and session shape, and lists starting under-hydrated beside under-fuelling, citing this trial as RPE rated during the ride. It agrees with the guide that a high RPE says a cause is present without saying which: here core temperature from the sensor did not explain RPE in thermoneutral air.
Where it meets [SYNTH-68]: both start from arriving euhydrated. The counselling's fixed drinking during exercise (200-300 mL every 15 min) and 120-150% replacement were advice for training at home and were not what the laboratory ride tested (no drinking was allowed), so this trial gives no support for drinking on a schedule over the guide's drink-to-thirst default.
Inference, not tested here: thirst rising faster than usual during a long ride, together with a higher RPE at the usual power, is worth noting with the session; the authors suggest thirst and heart rate as field warnings, but heart-rate drift did not differ significantly between groups in this trial (p = 0.18).
Not from this paper: any effect in the heat, at higher intensity, in women, or of drinking during the session; and a hydration threshold for urine colour or specific gravity that changes performance.
Limitations
Randomised, but the groups differed in VO2max at baseline (51.2 vs 46.1 mL/kg/min) and starting the ride depended on reaching the allocated group's hydration target, so group membership carries fitness and possibly habit as well as hydration; the authors note residual confounding cannot be excluded.
Blinding is not reported; participants necessarily knew whether they were counselled, and RPE and thirst are self-rated. Pre-exercise urine specific gravity did not itself predict thirst once group was in the model (p = 0.35), so the effect belongs to the counselled group as a whole, not to a measured test-day hydration value.
Some reported between-group statistics do not match the tables: the ΔRPE medians differ by 0.5 (mean) and 1.5 (peak) and the ΔThirst mean medians by 0.5, against reported differences of +0.23, +0.53 and +0.23; the core-temperature intervals do not fit their stated estimates and t values; and the heart-rate paragraph labels one of its results 'ΔRPE peak'.For RPE and thirst the direction agrees with the tables; for core temperature the sign of the reported difference is unclear.
"Completing 120 min" is shaped by the stop rule (2-3% body-mass loss, or failing to hold cadence); the paper does not say which rule stopped each rider.
The ΔPeak values in Table 3 equal each group's change from rest to 15 min (RPE 1.00 vs 2.50; thirst 2.00 vs 3.00), although ΔPeak is defined over 15-90 min, so the peak gap reflects ratings already apart in the first 15 min. Primary analyses stop at 90 min; the time to peak core temperature and heart rate differed partly because CON riders stopped earlier.
Core temperature came from a heat-flux sensor, which the authors treat as approximate; error in it would weaken its link with RPE.
30 men, thermoneutral conditions, one moderate load; sex, heat and intensity were not varied. No external funding; the authors declare no conflicts of interest.
How this entry was checked and changed
Why this level
a parallel trial of about 15 per arm (the B+ band) held one step lower because the groups differed at baseline in VO2max (51.2 vs 46.1 mL/kg/min, p = 0.01), starting the ride depended on meeting the hydration target of the allocated group, blinding is not reported, and some between-group statistics do not agree with the paper's own tables