REF-1611 · Grummt et al., 2024 · Sports Medicine - Open · Paper · PubMed · Free full text

What Borg numbers usually match on a graded test: in 6311 people (mean age 37) tested on a bike or treadmill at a Berlin sports-medicine clinic, the median rating was 13 at 2 mmol/l lactate, 15 at 3 mmol/l, 16 at 4 mmol/l and 15 at the individual threshold LT2, with the middle half of people spread over 2-3 points; sex, age, fitness and bike vs treadmill shifted ratings mainly at 2 mmol/l, and longer tests went with higher ratings at every anchor. Read in full

Main caveat: Retrospective, single centre, routine clinic data; no information on participants' conditions; user and equipment error cannot be excluded (the authors).
moderate evidenceMeasured on non-athletes to trained athletesRead in full

Grummt and colleagues took each person's first graded exercise test from the 2015-2022 records of a Berlin university sports-medicine clinic and asked what Borg 6-20 rating went with fixed lactate values and with the lactate threshold LT2.

Lactate and RPE were taken in the last 30 s of each 3-min stage; for each person a spline through the stage values gave the rating at 2, 3 and 4 mmol/l and at LT2, defined as the first rise above baseline plus 1.5 mmol/l. Of 10,934 tests, 6311 qualified.

Median ratings were 13 at 2 mmol/l, 15 at 3 mmol/l, 16 at 4 mmol/l and 15 at LT2, each with an interquartile range of 2-3 points. In ordinal models, women, older and fitter people and bike tests gave higher ratings at 2 mmol/l; at higher lactate most of these effects faded, while the number of stages raised ratings at every anchor.

The authors propose RPE 11 or less for light, 12-14 for moderate and 15-17 for vigorous intensity, slightly above the ACSM bands, and RPE 15 as the boundary between the heavy and severe domains.

Participants

n = 6311 (3949 men, 2362 women), 36.65 ± 16.68 years; bike 3672, treadmill 2639; BMI 18.5-24.9 in 60.9%; VO2max 38.64 ± 12.68 ml/kg/min (men 41.88, women 33.24; treadmill testers 48.72, bike testers 31.40). Only each person's first test; tests with fewer than three stages, missing lactate or RPE, or implausible data were excluded.

Protocol

Test: bike starting at 10-100 W and rising 10-50 W every 3 min, or treadmill starting at 3-10 km/h and rising 1-2 km/h every 3 min, chosen individually; to volitional exhaustion; room 19-20.5 °C.
Rating: the Borg 6-20 scale explained before the test; overall exertion rated in the last 30 s of each stage, defined as how hard, heavy and strenuous the task feels.
Lactate: earlobe sample each stage; LT1 as the first rise above baseline, LT2 as LT1 + 1.5 mmol/l.
Analysis: rating at each anchor from a per-person monotone cubic spline (at least three stage pairs), rounded; four multivariable cumulative ordinal models with sex, age, ergometer, VO2max and number of stages.

Key Results

AnchorMedian RPE [IQR]
2 mmol/l lactate13 [11; 14]
3 mmol/l lactate15 [13; 16]
4 mmol/l lactate16 [15; 17]
LT2 (mean lactate 3.03 ± 0.66 mmol/l)15 [14; 16]
Odds ratio for a higher rating (95% CI)2 mmol/l3 mmol/l4 mmol/lLT2
Male sex0.65 (0.587-0.719)0.762 (0.689-0.844)not significantnot significant
Treadmill vs bike0.754 (0.641-0.886)0.851 (0.723-1.002), not significantnot significant0.615 (0.522-0.725)
Number of stages (each 3 min)1.345 (1.300-1.394)1.290 (1.244-1.336)1.229 (1.187-1.274)1.155 (1.115-1.197)
Age, per year1.015 (1.012-1.018)1.006 (1.003-1.010)not significantnot significant
VO2max, per ml/kg/min1.023 (1.015-1.030)1.012 (1.004-1.019)not significantnot significant
Size of the shifts: at 2 mmol/l the model predicts RPE 13 for a 20-year-old man on a treadmill and 15 for a 75-year-old woman on a bike; at 3 mmol/l the authors find no practically relevant shift (RPE 15 for everyone).
Test length: the only factor at every anchor; the authors note that intensity also rose with each stage, so stage count is not a clean measure of duration.
Comparison: ratings at 3 and 4 mmol/l and LT2 were 1-2 points higher than in the previous largest study (2560 people), which the authors put down to their per-person interpolation and broader sample.

Practical Application

On graded bike and treadmill tests in 6311 clinic patients (mean age 37), a Borg 6-20 rating of about 13 usually went with 2 mmol/l lactate, about 15 with the threshold LT2 and 3 mmol/l, and about 16 with 4 mmol/l, but the middle half of people spread over 2-3 points at each. For a rider learning the scale this gives a first map ("around 13 is still below threshold, around 15 is threshold"), and the spread is the reason to check it against one's own threshold test, heart rate or power (inference: checking was not studied here).
Ratings at every anchor were higher in tests with more stages, a comparison between people in which, the authors note, intensity also rose with each stage. Inference, not tested here: in a long ride, a rating taken late overstates the intensity relative to the same rating early on.
At lower intensity women, older and fitter riders rated the same lactate higher, men and treadmill tests lower; the authors see a practical difference only at 2 mmol/l.
The authors propose RPE 12-14 for moderate and 15-17 for vigorous work, one point above the ACSM's 12-13 and 14-17, and RPE 15 as the heavy-to-severe boundary; that is their proposal for a broad population, not a tested training prescription.
The base's own practice: the app logs a 1-10 rating, and no conversion from these Borg values is tested (SYNTH-40).

Limitations

Retrospective, single centre, routine clinic data; no information on participants' conditions; user and equipment error cannot be excluded (the authors).
Ratings come from a ramp test with 3-min stages, so intensity and time rose together.
LT2 from one formula only; other threshold methods were not compared.
Spline interpolation and rounding to whole ratings add error; one rating point is hard to place exactly.
No menstrual-cycle information, and lactate is only one of many signals behind a rating.
Two models each of bike and treadmill; tests at varying times of day.
How this entry was checked and changed
Why this level

the rubric's cross-sectional measurement study with ≥100 (n = 6311); not higher because it is retrospective clinic data from one centre, the ratings come from the last 30 s of 3-min stages of a graded test rather than from steady training, and LT2 rests on one formula (first lactate rise + 1.5 mmol/l)