ArticlePDF Available

Is the FTP Test a Reliable, Reproducible and Functional Assessment Tool in Highly-Trained Athletes?


Abstract and Figures

The aim of the current study was to assess reliability of the Functional Threshold Power test (FTP) and the corresponding intensity sustainable for 1-hour in a "quasi-steady state". Highly-trained athletes (n = 19) completed four non-randomized tests over successive weeks on a Wattbike; a 3-min incremental test (GxT) to exhaustion, two 20-min FTP tests and a 60-min test at computed FTP (cFTP). Power at cFTP was calculated by reducing 20-min FTP data by 5% and was compared with power at Dmax and lactate threshold (TLac). Ventilatory and blood lactate (BLa) responses to cFTP were measured to determine whether cFTP was quasi-steady state. Agreement between consecutive FTP tests was quantified using a Bland-Altman plot with 95% limits of agreement (95% LoA) set at ± 20 W. Satisfactory agreement between FTP tests was detected (95% LoA = +13 and-17 W, bias +2 W). The 60-min effort at cFTP was successfully completed by 17 participants, and BLa and ventilatory data at cFTP were classified as quasi-steady state. A 5% increase in power above cFTP destabilized BLa data (p < 0.05) and prompted VO2 to increase to peak GxT rates. The FTP test is therefore deemed representative of the uppermost power a highly-trained athlete can maintain in a quasi-steady state for 60-min. Agreement between repeated 20-min FTP tests was judged acceptable.
Content may be subject to copyright.
Is the FTP Test a Reliable, Reproducible and Functional Assessment Tool in
Highly-Trained Athletes?
Human Performance Laboratory, Departments of Anatomy and Physiology, School of Medicine,
Trinity College Dublin, IRELAND
Denotes graduate student author, Denotes professional author
International Journal of Exercise Science 12(4): 1334-1345, 2019. The aim of the current study was to
assess reliability of the Functional Threshold Power test (FTP) and the corresponding intensity sustainable for 1-hour
in a “quasi-steady state”. Highly-trained athletes (n = 19) completed four non-randomized tests over successive
weeks on a Wattbike; a 3-min incremental test (GxT) to exhaustion, two 20-min FTP tests and a 60-min test at
computed FTP (cFTP). Power at cFTP was calculated by reducing 20-min FTP data by 5% and was compared with
power at Dmax and lactate threshold (TLac). Ventilatory and blood lactate (BLa) responses to cFTP were measured to
determine whether cFTP was quasi-steady state. Agreement between consecutive FTP tests was quantified using a
Bland-Altman plot with 95% limits of agreement (95% LoA) set at ± 20 W. Satisfactory agreement between FTP tests
was detected (95% LoA = +13 and -17 W, bias +2 W). The 60-min effort at cFTP was successfully completed by 17
participants, and BLa and ventilatory data at cFTP were classified as quasi-steady state. A 5% increase in power
above cFTP destabilized BLa data (p < 0.05) and prompted VO2 to increase to peak GxT rates. The FTP test is therefore
deemed representative of the uppermost power a highly-trained athlete can maintain in a quasi-steady state for 60-
min. Agreement between repeated 20-min FTP tests was judged acceptable.
KEY WORDS: Functional threshold power, incremental exercise test, critical power, maximum
lactate steady state.
Functional threshold power (cFTP) is defined as the uppermost power sustainable for 60-min in a
quasi-steady state (1). Data for cFTP is calculated as 95% of the mean power output during a
maximal 20-min effort, the sole outcome measure assessed is power, see Table 1. This 20-min
maximum effort does not reflect a specific energy system, rather a performance, derived from all
available energetics. Dissimilarly, exercise physiologists commonly break-down the elements of
whole-body energetics to enhance our understanding of how a competitive performance is
derived, such as blood lactate (BLa) and ventilation to determine physiological thresholds;
namely, load at lactate threshold (TLac), load at ventilatory threshold (Tvent) and load at Dmax (4).
Ordinarily these results are used to appraise the effectiveness of the preceding training and to
direct future preparation. However, BLa and ventilation measures are impacted upon by test
Int J Exerc Sci 12(4): 1334-1345, 2109
International Journal of Exercise Science
format; namely, increment duration and magnitude (4, 24). A new power-based test may add
insight into these already established physiological test indices.
The authors of the FTP test (1) state that the test can be trialed satisfactorily in the field.
Irrespective, we felt it prudent, at this juncture, to appraise the FTP test in a controlled laboratory
setting. The complexities of controlling the multitude of variables required for ecological
validation were beyond the scope of the current investigation.
Table 1. Test protocol for assessing FTP(1)
Endurance pace
3 by 1-min
with 1-min recoveries
Fast pedaling,
100 revˑmin-1
Not applicable
Easy riding
Maximum effort
Easy riding
FTP test
Maximum effort
Cool down
10 to 15-min.
Easy riding
(FTP) Functional threshold power, (min) minute, (revˑmin-1) revolutions per minute
The first hypothesis of the current study was that, in a highly-trained athletic cohort, repeated 20-
min FTP tests would be reliable and reproducible. The second hypothesis was that the FTP test
was sustainable for 60-min. The prefatory description “quasi-steady state” BLa and VO2 was
assessed using already established definitions of steady state BLa (20) and VO2 (26). The
interchangeability of cFTP with the following mathematical calculations of a threshold intensity;
Tvent, TLac and Dmax were examined. The aim was to find a comparable intensity to cFTP. The
variation in each calculation (Tvent, TLac and Dmax) was foreseen to add a range of threshold
powers and therefore enhance the likelihood of a possible match. The computation of Tvent, TLac
and Dmax is shown in the data reduction section below. For statistical reasons, the null
hypothesis was that no significant differences (p > 0.05) would be detected between the load at
cFTP and TLac, Dmax and/or Tvent. Where the null hypothesis was accepted, agreement between
tests was assessed using statistical measures of reliability and reproducibility.
The current study obtained ethical approval from the Faculty of Health Sciences Research Ethics
Committee in Trinity College Dublin and was performed in accordance the ethics standards of
the International Journal of Exercise Science (16). Participants completed a consent form prior to
beginning any trials. Nineteen (12 male, 7 female) highly-trained cyclists and triathletes, mean age
28-yr (± 6) participated. The mean VO2peak, body mass, body mass index (BMI) and % body fat of
enlisted participants are presented in Table 2.
Int J Exerc Sci 12(4): 1334-1345, 2109
International Journal of Exercise Science
Table 2. Mean (± SD) VO2peak, mass, BMI and % body fat data for participants.
VO2peak FTP
Body fat
57.2 ± 3.3
57.0 ± 4.1
20.1 ± 1.2
16.3 ± 1.4
28 ± 6
65.1 ± 5.5
66.1 ± 4.9
22.0 ± 1.0
10.5 ± 1.3
28 ± 7
(SD) standard deviation, (BMI) Body mass index, (GxT VO2peak) Peak maximum oxygen uptake during
graded incremental test, (VO2peak FTP) peak maximum oxygen uptake during functional threshold power test,
(mLˑkg-1ˑmin-1) -milliliter of Oxygen consumed per kilogram body mass per minute, (kg) kilogram, (kgˑm-2)
kilogram divided by meter squared, (yr) year.
Participants attended the laboratory on four occasions, in a rested, carbohydrate loaded and well-
hydrated state, having abstained from alcohol and caffeine in the 24-h prior to testing. Trials were
separated by a minimum of 6 and maximum of 10-day. Training load was agreed with both
athletes and coaches prior to commencing the study, with weekly training load and stimulus
remaining as constant as possible for the testing period. Exercise was limited to aerobic work for
48-h prior to each test. A 24-h food diary was completed prior to the first trial. A copy of the food
diary was retained by both the researcher and the participant. Each participant was requested to
replicate their food intake prior to all tests, or if different, to consume comparable carbohydrate
quantities. When necessary, participants were helped to plan their meals. Pre-test hydration
status was assessed as urine specific gravity (USG) by refractometry (Bellingham & Stanley, Kent,
UK) prior to all tests, euhydration was defined as USG < 1.020. In addition, a full blood count was
performed to screen for sub-clinical anaemia and infection.
Visit one involved a detailed medical assessment and a 3-min graded incremental test (GxT) to
volitional exhaustion. Participants completed a 10-min warm-up on an air-braked cycle ergometer
(WattBike, Nottingham, UK) below the initial power output required for the GxT (female < 90 W,
male < 120 W). This time interval allowed participants to make minor adjustments to their cycling
position (saddle height and fore-aft position, handlebar height and reach); these individual
positions were noted and replicated during subsequent visits. The increments for the female and
male cohort were 20 and 25 W, respectively, and cadence was controlled between 75 and 110
revˑmin-1. Finger-tip capillary blood samples for lactate analysis were collected at 2-min into each
3-min increment and assessed using a calibrated YSI 1500 Sports Lactate analyzer (YSI, OH, USA).
Breath-by-breath ventilatory (VE, VO2 and VCO2) and heart rate (HR) data, recorded using a
calibrated cardio-metabolic cart (CPET; Cosmed, Rome, Italy), were averaged across the final
minute of each increment. Each participant’s VO2peak was determined as the highest consecutive
five breath average during the final test increment.
All participants had previously completed an FTP test prior to participation in the current study.
During visits two and three, 20-min FTP tests were completed in the laboratory and projected 60-
min FTP (cFTP) was computed (cFTP = 95% of the 20-min trial). Five participants were available
for 3 visits only due to race commitments; therefore, they completed the GxT, one 20-min FTP test
and the 60-min cFTP test. Data for these five participants were omitted from all analyses
pertaining to test-retest reliability and repeatability. However, their data were included in the
assessment of the validity of the assertion that cFTP could be sustained in a quasi-steady state for
Int J Exerc Sci 12(4): 1334-1345, 2109
International Journal of Exercise Science
60-min. Endeavoring to adhere as closely as possible to the construct of the FTP warm-up, the
first FTP test warm-up used 65% Dmax as a proxy for 65% cFTP, the 5-min time-trial effort during
the second FTP test was controlled at the same power attained during the initial FTP test and
cadence during the warm-up was un-controlled. Table 3 outlines the timelines for capillary blood
sample collection and measurement of ventilatory variables during all test sessions.
During visit four, participants (n = 19) were tasked with sustaining their cFTP power for 60-min.
The pre-test warm-up was standardized as 7-min at 55% of Dmax, 5-min at 65% of Dmax, 3-min
at cFTP and 3-min stationary rest. On-line VO2 and HR data were recorded during the 60-min
cFTP test at the start and end (0 to 10 and 50 to 60-min). The facemask was removed from 11 to
49-min to facilitate participants drinking carbohydrate beverage ad libitum. Participants were
asked to bring their normal race beverage, with the only stipulation being that the drink
contained carbohydrate. Fingertip capillary blood samples were collected for lactate analysis at
10-min intervals throughout the 60-min trial, see Table 3.
Table 3. Time intervals during tests when BLa and ventilatory data were measured.
Lactate sample
Ventilatory data
2-min into each workload
FTP1 and FTP2
Post 5-min time-trial
Pre 20-min time-trial
Post 20-min time-trial
From 31 to 66-min
60-min at cFTP
10, 20, 30, 40, 50 and 60-min
Start (0 to 10-min)
End (50 to 60-min)
(BLa) Blood lactate, (GxT) - graded incremental test, (FTP) functional threshold power, (min) - minute
Threshold loads (W) for Tvent, TLac and Dmax were determined from the recorded GxT data. VO2
was used as the dependent variable; the associated power output was used to estimate Tvent and
Dmax (4). Load at Tvent data were identified using a segmented regression model minimizing the
squared sum of the residuals following logarithmic transformation of VO2 and load data
(SigmaPlot 13.0; Systat, IL, USA). Load at TLac and Dmax was identified using “Lactate E”
software (17). The line of best fit for the Dmax calculation is calculated using third order
curvilinear regression using VO2 and BLa data at each workload. Thereafter, the maximum
perpendicular distance to the straight line between the lowest and highest BLa data identified
load at Dmax (4). The calculation of TLac is accomplished using a linear spline minimizing the sum
of the squared differences between workload and BLa, the workload preceding a departure from
the straight line is identified as TLac (12). A single factor repeated measures ANOVA, quantified
using post-hoc Bonferroni tests with p < 0.05 inferring significance, compared threshold data and
was used to cull variables on the basis of there being a statistically significant difference
compared to cFTP (2).
Statistical Analysis
An a priori power test was conducted for expected outcomes with a Type 1 error probability of
0.05, a power of 0.85 and a projected effect size 0.3. This analysis indicated that n = 19 would
provide a statistical power of 85% (G*Power v3.0.10 free software; Institute of Experimental
Psychology, Heinrich Heine University, Dusseldorf, Germany). All data were checked for
Int J Exerc Sci 12(4): 1334-1345, 2109
International Journal of Exercise Science
normality using both the D'Agostino and Pearson, and the Shapiro-Wilk tests. The reliability of
the FTP test was assessed using intra-class correlation (ICC), standard deviation of the residuals
(Sy.x), coefficient of determination (r2) and computing 95% confidence intervals of r (95% CI). The
reproducibility of the FTP test was assessed using relative technical error of the measure (% TEM)
and Bland Altman 95% limits of agreement (95% LoA). These analyses were performed using
Prism 7 (Graph Pad, CA, USA).
Group data are presented as mean and standard deviation (SD). VO2 measured during the 60-min
trial was classified as steady-state if it remained significantly lower (p < 0.05, paired Student’s T
test) than VO2peak, a specific demarcation used in identifying critical power (CP), a metric
described as the boundary between steady- and non-steady-state cycling (26). Steady- state BLa
data were assessed using maximal lactate steady state (MLSS) criteria, that is, an increase in BLa <
0.05 mmolˑL-1ˑmin-1 (8). We hypothesized that peak 60-min BLa would be significantly (p < 0.05)
lower than GxT peak BLa. This might be sufficient to affirm Allen and Coggan’s (1) assertion that
cFTP elicits a steady-state VO2 response. However, this would not elucidate whether the intensity
was the upper-limit intensity. MLSS can be identified using increases of 0.2 Wˑkg-1 until the load
destabilized BLa data (19). The difference in power output between the 20-min time-trial and
cFTP was 5%.Consequently, we reasoned that if the 5% difference in power equated to £ 0.2
Wˑkg-1 and marked the boundary between steady and non-steady state, then the 60-min FTP
power was the upper limit of steady-state.
For the purpose of comparing cFTP data with alternate threshold tests, acceptable 95% LoA were
set as ± 20 W. These LoA were considered appropriate for clinical use (2); namely, to discern
biological change whilst mitigating against the risk of erroneous or misleading results.
“Performance VO2 was calculated by expressing VO2 data during the 60-min test at cFTP as a
percent of VO2peak (5). During the 60-min test, HR data were recorded during the initial and final
10-min of exercise, individual changes in mean HR data are reported. Peak HR measures were
assessed for significant differences (p < 0.05, paired Student’s T test) with GxT HR maxima.
ANOVA identified no significant (p > 0.05) difference in power output between cFTP (259 ± 40 W)
and Dmax or TLac (246 ± 38 and 244 ± 47 W, respectively). However, the load at Tvent (197 ± 43 W)
was significantly lower (p < 0.01) than cFTP, consequently, Tvent data were excluded from further
data analysis, see Figure 1. Mean group and gender specific power, normalized to body mass
(Wˑkg-1), at Dmax, TLac, cFTP and last completed GxT stage (peak GxT) are listed in Table 4. The
95% LoA, the mean bias, r2, Sy.x and 95% CI associated with; cFTP1 vs. cFTP2, cFTP vs. Dmax and
cFTP vs. TLac are detailed in Table 5.
Int J Exerc Sci 12(4): 1334-1345, 2109
International Journal of Exercise Science
Figure 1. Bar graph of load at computed thresholds, bars denote SD, n = 15. Asterisk symbol (*) infers
significantly higher than Tvent, ** infers p < 0.01.
Table 4. Mean power (Wˑkg-1) associated with load at Dmax, TLac, cFTP and peak GxT load. Columns indicate the
whole population sample, female only and male only.
Dmax (Wˑkg-1)
3.5 ± 0.4
3.2 ± 0.3
3.7 ± 0.4
TLac (Wˑkg-1)
3.6 ± 0.6
3.2 ± 0.3
3.8 ± 0.6
cFTP (Wˑkg-1)
3.9 ± 0.5
3.5 ± 0.5
4.1 ± 0.5
Peak GxT (Wˑkg-1)
4.8 ± 0.5
4.3 ± 0.4
5.0 ± 0.4
(Wˑkg-1) Watt per kilogram of body mass, (Dmax) Load at maximum displacement, (TLac) Lactate threshold,
(cFTP) Functional threshold power , (GxT) graded incremental test.
Table 5. Results of comparisons for cFTP, Dmax and TLac computing % TEM, ICC, r2 and 95% CI.
cFTP1 vs. cFTP2
cFTP vs. Dmax
cFTP vs. TLac
95% LoA (W)
+ 13 to -17
+ 43 to - 7
+ 68 to 37
Mean bias (W)
95% CI
0.87 to 1.11
0.72 to 1.10
0.51 to 1.33
Sy.x (W)
(cFTP) Functional threshold power, (Dmax) Load at maximum displacement, (TLac) Lactate threshold, (%
TEM) Relative technical error of measurement, (ICC) - Inter-class correlation coefficient, (r2) Coefficient of
determination, (CI) confidence interval, (LoA) limits of agreement, (Sy.x) Standard deviation of the
residuals, (W) Watt.
No significant difference (p > 0.05) was detected comparing GxT peak BLa data with 20-min FTP
BLa maxima (6.9 ± 1.3 vs. 6.9 ± 1.9 mmolˑL-1, respectively). However, peak BLa data were
significantly lower (p < 0.05) during the 60-min cFTP test than peak GxT BLa (4.0 ± 1.1 vs. 6.9 ± 1.3
mmolˑL-1, respectively). During the 60-min test, mean HR after 10-min at cFTP (158 ± 14
beatsˑmin-1) increased across time by 9% (14 ± 7 beatsˑmin-1), however, mean HR between 50 and
60-min at cFTP remained significantly (p < 0.05) lower (178 ± 11 beatsˑmin-1) than peak HR data
cFTP1 cFTP2 Dmax TLac Tvent
Mean power (W)
Int J Exerc Sci 12(4): 1334-1345, 2109
International Journal of Exercise Science
recorded during the GxT (186 ± 12 beatsˑmin-1). In addition, mean VO2 data increased by 6% from
start to end of the cFTP test (52.0 ± 6.4 vs. 55.0 ± 5.7 mLˑkg-1ˑmin-1, respectively), but similarly to
HR data, remained significantly (p < 0.05) lower than peak GxT data (62.9 ± 6.2 mLˑkg-1ˑmin-1).
The 5% power output difference between the 20 and 60-min trials equated to an average of 0.19 ±
0.02 Wˑkg-1. “Performance VO2 increased from 84% at 10-min to 88% of VO2peak during the last 10-
min of the 60-min cFTP trial.
In the current study, individual power at cFTP was sustained for 60-min by seventeen (89 %) of
nineteen athletes assessed; the two athletes unable to complete the cFTP trial aborted at 35 and 52-
min, respectively. Analysis of their BLa kinetics from 10 to 60-min indicated that the rate of
change in BLa data from 10-min onwards remained below 0.05 mmol.L-1.min-1 (Figure 2). The two
athletes that could not complete the 60-min trial had their BLa data assessed at test termination.
Neither participant’s BLa had increased by > 0.05 mmolˑL-1ˑmin-1 upon termination. As the
facemask was removed to facilitate drinking, determination as to whether their VO2 was also in a
steady-state was not possible. Individual BLa data measured at 10-min intervals during the 60-
min trial at cFTP are presented in Figure 2.
Figure 2. Individual BLa data at 10-min intervals during the 60-min trial at cFTP (n = 19).
The results of the current study demonstrate that FTP is a reliable test. Both ICC and r2 were high,
while the corresponding residuals had a low standard deviation (ICC = 0.98 r2 = 0.96, Sy.x = 9 W).
The FTP test was also found to be repeatable with satisfactory limits of agreement (+13 to -17 W;
mean bias -2 W) and a low relative TEM of 2.3%.Consequently, the reliability and repeatability of
the FTP test supports the inclusion of the 5 participants who completed a single FTP prior to the
validity trial.
The FTP test was shown to identify a cycling power output that could be sustained for 1-hour in
seventeen of nineteen (89%) athletes. Participants were asked to stop cycling after 1-hr as this met
10 20 30 40 50 60
Time (min)
BLa (mmolL-1)
Int J Exerc Sci 12(4): 1334-1345, 2109
International Journal of Exercise Science
the criteria specified as FTP (1). The purpose of the trial was not to determine a time to
exhaustion. This eliminates the establishment of an absolute 60-min limit of tolerance. However,
the FTP test itself provides a 20-min limit of tolerance, supporting the contention that 60-min
maximum power is < 5 % above cFTP. Of note, this study group were highly trained, mean
fraction of VO2max sustained during the 60-min trial was 84 % (using the first 10-min rather than
the last 10-min VO2 data as the numerator), a fraction reflective of elite endurance athletes (8). The
high sustained fraction of VO2 at cFTP is particularly relevant if percentages of FTP are used to
formulate exercise prescription, as currently recommended (1). Were a trained and untrained
cyclist to train at the same percentage of their FTP, each would likely be cycling at different
percentages of absolute maximum 60-min power. In the context of training at percentages of
maximum, overtraining and / or a poor response to training zones has previously been described
(18). The cause of volition exhaustion appears beyond the scope of this investigation. The two
unsuccessful participants were euhydrated, reportedly carbohydrate loaded and adhered to all
pre-trial exercise controls. The two participants also met the definitions of steady state BLa. As a
consequence of the face mask being removed to allow drinking, it cannot be discerned whether
VO2 was steady state on cessation. However, both participants maintained steady-state VO2 for in
excess of 30-min. Both cyclists were observed to be under a lot of stress on volitional exhaustion
and consequently motivation did not appear to be the limiting factor. These investigators would
recommend future research include an examination of the reliability of the 60-min trial.
Additionally, future research into the trained state of the cyclist may prove insightful. The
distinction between an index for a 35-min versus a 60-min limit of tolerance is worth considering.
In the context of exercise prescription, where the ideal intensity is to be steady-state, one might
suggest the FTP metric adequate.
Strategies taken from investigations into MLSS and CP were used to measure whether quasi-
steady state VO2 and BLa data were attained, not to expound whether FTP was interchangeable
with MLSS or CP. The BLa and VO2 responses in this investigation support the concept of a
steady state being reached at cFTP. Furthermore, the £ 0.2 Wˑkg-1 increment required to
destabilize BLa and VO2 support the claim that cFTP was the uppermost intensity whereby a
quasi-steady state could be maintained.
The strength of the relationship between cFTP and Dmax was stronger than the relationship
between cFTP and TLac (r2 = 0.89, Sy.x = 14 W, 95% CI = 0.72 to 1.10 vs. r2 = 0.73, Sy.x = 26 W, 95%
CI = 0.64 to 1.38, respectively). Nonetheless, the LoA associated with cFTP and Dmax were
greater (+ 43 to – 7 W) than the LoA set a priori (+20 to –20 W). As such, we recommend that cFTP
should not be used interchangeably with Dmax or TLac. Dmax and TLac are commonly used to
determine physiological thresholds, but do not have associated limits of tolerance; namely, a
defined upper limit duration (time) associated with a specific power output. This potentially
highlights another likely function for the FTP test. A uniform racing pace tactic has been reported
to be the fastest strategy where terrain is even, conditions still and duration extensive (7). Under
these circumstances, cFTP is most likely a superior index to TLac or Dmax as a uniform power
output pacing strategy is possible, unlike pacing using physiological indices, which are
susceptible at some level to temporal changes (9).
Int J Exerc Sci 12(4): 1334-1345, 2109
International Journal of Exercise Science
An investigation into the interchangeability of cFTP with CP may be worthwhile. CP “is
mathematically defined as the power-asymptote of the hyperbolic relationship between power
output and time to exhaustion” (26). However, of note, CP is reported sustainable for < 30-min
(22), although this disparity may be explained in part by the untrained status of the participants
in investigations (21). The landmark paper by Monod and Scherrer (13) stated that exercise at CP
could be sustained indefinitely, although the context was in respect to local muscle work,
illustrated as being equivalent to “one leg” exercises, independent of respiratory and
cardiovascular factors. The determination of CP also furnishes a metric named W¢ which
represents the fixed amount of work that can be performed above CP (21). The energy provided
by W¢ is likely to include anaerobic substrates and energy derived from oxygen stores, although,
the exact physiological underpinnings of W¢ are unknown (22). Were CP and cFTP are found to
be interchangeable, one might posit that the difference between quasi-steady state (cFTP) and
maximum power to be correlated with the size of W¢. The FTP test has been appraised elsewhere,
whereby participants were tasked with completing a 60-min time-trial as fast as possible, and the
mean power achieved said to reflect cFTP (3). This methodology appears to consider cFTP as the
maximum intensity a participant can maintain, rather than that of a quasi-steady state. Vanhatalo
et al. (26) categorized a W¢ of 25 kJ as “large”, however, extrapolated across a 1-h period a W¢ of 25
kJ equates to 7 W, this is equivalent to a 2.3 % increase in power above cFTP for an elite rider
weighing 60-kg with a cFTP of 5 Wˑkg-1. The reproducibility in common performance cycling tests
is typically 2 to 3% (11) and meaningful change in elite sport is suggested as being approximately
0.4 to 1.5 % (11, 23). The practical and or experimental significance of a 2.3% differential between
cFTP and maximum power should perhaps be considered in the context of the purpose of the test.
The current investigation used the upper-limit of BLa change in MLSS to establish a tenet criteria
to assess the claim that cFTP was a quasi-steady state intensity. Currently, there is a scarcity of
peer-reviewed investigations into FTP. A single investigation into the interchangeability of FTP
and MLSS reported “trivial” differences between cFTP and MLSS (3). The associated LoA in “well
trained” and “trained” populations (11.8 and 6.4 %, respectively) were reported in percentage
format (3). The determination of absolute differences (W) versus percentages of differences likely
requires deliberation, as a higher cFTP would result in larger absolute LoA.
Researchers have indicated that maximum sustainable power can only be achieved with the
concession of aerodynamic position (6, 19, 25). If the FTP test data were derived from a power
maximized position, consideration would need to be given when applying cFTP to a less
powerful drag-optimized position. Beyond the cycling position adopted by the cyclist, part of our
reluctance to appraise FTP in the field was that the average power might be affected negatively
by the topography of the testing environment. Allen and Coggan (1) introduce a non-validated
algorithm termed “normalized power” (NP), which purportedly overcomes underestimates of
mean power as a consequence of stochastic cycling. If validated, this weighted average could
potentially facilitate the appraisal of the FTP test in the field. To compute NP the power data are
smoothed using a 30-s moving average, raised to the fourth power, averaged over time and lastly
the fourth root computed (10). The adopted pacing strategy in the Borszcz et al. (3) investigation
was open and variable. This strategy is very clear but may overlook the negating effect of a
variable versus a uniform pacing strategy (7). The current study required position and power
Int J Exerc Sci 12(4): 1334-1345, 2109
International Journal of Exercise Science
output to be held constant, and as such, our findings only reflect similar pacing strategies and
conditions. Our focus was to determine whether it was possible physiologically to maintain FTP
power for 60-min rather than to factor in “all” pacing strategies and / or positions.
In conclusion, the FTP test was shown to be a reliable and repeatable test. The current
investigation demonstrated that the FTP test successfully determined 60-min power in 89% of
assessed participants and prompted a quasi-steady state BLa and VO2 in all assessed participants.
These findings are however limited to a similarly highly trained, highly motivated cohort.
Extrapolation to lesser-trained groups would require fresh investigation. The findings herein
support the utility of FTP as part of a pacing strategy in competition. Proceeding forwards, it
appears that research into the interchangeability of FTP with MLSS and or CP may be
worthwhile. In the instance that these tests were found interchangeable with FTP, the benefits
would include more affordable and less time-consuming testing. Noteworthy, the disadvantages
of using the FTP test in lieu of physiological data is the lack of insight as to where performance
gains might be made. A performance is the sum of whole-body energetics that does not provide
information such as fuel use, efficiency, aerobic versus anaerobic capacity. The principle
limitation of the current study is that FTP was not assessed outside; where for the most part the
majority of cycling is performed. However, the current investigation provides a starting point for
further ecological assessment, given the demonstrated efficacy of the test. Inevitably, as Morton
stated, “the single most remarkable conclusion is on the one hand such a plethora of models fit
the real world so well, yet there is so much more to discover” (14).
The authors disclose no conflicts of interest or financial arrangements related to this research.
We would like to thank all athletes for their gracious participation in this research.
1. Allen H, Coggan A. Training and racing with a power meter. VeloPress: CO, USA; 2010.
2. Bland J, Altman D. Statistical methods for assessing agreement between two methods of clinical
measurement. Lancet 8: 307-310, 1986.
3. Borszcz F, Tramontin A, Costa V. Is the functional threshold power interchangeable with the maximal
lactate steady state in trained cyclists? Int J Sports Physiol Perform, 2019; doi: 10.1123/ijspp.2018-0572. [Epub
ahead of print]
4. Cheng B, Kuipers H, Snyder A, Keizer H, Jeukendrup A, Hesselink M. A new approach for the
determination of ventilatory and lactate thresholds. Int J Sports Med 13: 518-522, 1992.
5. Costill D. Metabolic responses during distance running. J Appl Physiol 28: 251-255, 1970.
6. Fintelman D, Sterling M, Hemida H, Li F.-X. Optimal cycling time-trial position models; aerodynamics
versus power output and metabolic energy. J Biomech 47: 1894-1898, 2014.
7. Fukuba Y, Whipp B. A metabolic limit on the ability to make up for lost time in endurance events. J Appl
Physiol 87: 853-861, 1999.
Int J Exerc Sci 12(4): 1334-1345, 2109
International Journal of Exercise Science
8. Heck H, Mader G, Hess G, Mucke S, Muller R, Hollmann W. Justification of the 4-mmolˑL-1 lactate
threshold. Int J Sports Med 6: 117-130, 1985.
9. Jeukendrup A, van Diemen A. Heart rate monitoring during training and competition in cyclists. J Sports
Sci 16: S91-S99, 1988.
10. Jobson S, Passfield L, Atkinson G, Barton G, Scarf P. The Analysis and Utilization of Cycling Training Data.
Sports Med 39: 833-844, 2009.
11. Lamberts R, Swart J, Woolrich R, Noakes T, Lambert M. Measurement error associated with performance
testing in well-trained cyclists: Application to the precision of monitoring changes in training status. Int Sports
Med J 10: 33-44, 2009.
12. Lundberg M, Hughson R, Weisiger K, Jones R, Swanson G. Computerised estimation of lactate threshold.
Comput Biomed Res 19: 481-486, 1986.
13. Monod H, Sherrer J. The work capacity of a synergic muscular group. Ergonomics 8: 329-338, 1965.
14. Morton R. The critical power and related whole-body bioenergetic models. Eur J Appl Physiol 96: 339-354,
15. Naimark A, Wasserman K, McIlroy M. Continuous measurement of ventilatory exchange ratio during
exercise. J Appl Physiol 19: 644-652, 1964.
16. Navalta J, Stone W, Lyons T. Ethical issues relating to scientific discovery in exercise science. Int J Exerc Sci
12: 1-8, 2019.
17. Newell J, Higgins D, Madden N, Cruickshank J, Einbeck J, Mc Millan K, Mc Donald R. Software for
calculating blood lactate endurance markers. J Sports Sci 25: 1403-1409, 2007.
18. Olbrecht J. The science of winning. Planning, periodizing and optimizing swim training. F & G Partners,
Belgium; 2007. pp. 115-117.
19. Padilla S, Mujika I, Angulo F, Goirene J. Scientific approach to the 1-h cycling world record: A case study. J
Appl Physiol 89: 1522-1527, 2000.
20. Palleres J, Moran-Navarro R, Ortega J, Fernandez-Elias V, Mora-Rodriguez R. Validity and reliability of
ventilatory and blood lactate thresholds in well-trained cyclists. PLoS ONE 11: e0163389.
21. Poole D, Ward S, Garner G, Whipp B. Metabolic and respiratory profile of the upper limit for prolonged
exercise in man. Ergonomics 31: 1265-1279, 1988
22. Poole D, Burnley M, Vanhatalo A, Rossiter H, Jones A. Critical power: An important fatigue threshold in
exercise physiology. Med Sci Sports Exerc 48: 2320-2334, 2016.
23. Pyne D. Physiological tests for elite athletes. Second edition. Australian Institute of Sport. Human Kinetics.
Champaign, IL, USA. pp. 35-42.
24. Stockhausen W, Grathwohl D, Burklin C, Spranz P, Keul J. Stage duration and increase of work load in
incremental testing on a cycle ergometer. Eur J Appl Physiol 76: 295-301, 1997.
Int J Exerc Sci 12(4): 1334-1345, 2109
International Journal of Exercise Science
25. Underwood L, Schumacher J, Burette-Pommay J, Jermy M. Aerodynamic drag and biomechanical power of
a track cyclist as a function of shoulder and torso angles. Sports Eng 14: 147-154, 2011.
26. Vanhatalo A, Jones A, Burnley M. Application of critical power in sport. Int J Sports Physiol Perform 6:
128-136, 2011.
ResearchGate has not been able to resolve any citations for this publication.
Full-text available
This work aims to present concepts related to ethical issues in conducting and reporting scientific research in a clear and straightforward manner. Considerations around research design including authorship, sound research practices, non-discrimination in subject recruitment, objectivity, respect for intellectual property, and financial interests are detailed. Further, concepts relating to the conducting of research including the competency of the researcher, conflicts of interest, accurately representing data, and ethical practices in human and animal research are presented. Attention pertaining to the dissemination of research including plagiarism, duplicate submission, redundant publication, and figure manipulation is offered. Other considerations including responsible mentoring, respect for colleagues, and social responsibility are set forth. The International Journal of Exercise Science will now require a statement in all subsequent published manuscripts that the authors have complied with each of the ethics statements contained in this work.
Full-text available
Purpose: Functional Threshold Power (FTP), determined as 95% of the average power during a 20-minute time-trial test, is suggested as a practical test for the determination of the maximal lactate steady state (MLSS) in cycling. Therefore, the objective of the present study was to determine the validity of FTP in predicting MLSS. Method: Fifteen cyclists, 7 classified as trained and 8 as well-trained (mean ± standard deviation; maximal oxygen uptake = 62.3 ± 6.4 mL/kg/min, maximal aerobic power = 329 ± 30 Watts), performed an incremental test to exhaustion, an FTP test, and several constant load tests to determine the MLSS. The bias ± 95% limits of agreement (LoA), typical error of the estimate (TEE), and Pearson´s coefficient of correlation (r) were calculated to assess validity. Results: For the power output measures, FTP presented a bias ± 95% LoA of 1.4 ± 9.2%, a moderate TEE (4.7%), and nearly perfect correlation (r = 0.91) with MLSS in all cyclists together. When divided by the training level, the bias ± 95% LoA and TEE were higher in the trained group (1.4 ± 11.8% and 6.4%, respectively) than in the well-trained group (1.3 ± 7.4% and 3.0%, respectively). For the heart rate measurement, FTP presented a bias ± 95% LoA of −1.4 ± 8.2%, TEE of 4.0%, and very -large correlation (r = 0.80) with MLSS. Conclusion: Therefore, trained and well-trained cyclists can use FTP as a noninvasive and practical alternative to estimate MLSS
Full-text available
Purpose The purpose of this study was to determine, i) the reliability of blood lactate and ventilatory-based thresholds, ii) the lactate threshold that corresponds with each ventilatory threshold (VT1 and VT2) and with maximal lactate steady state test (MLSS) as a proxy of cycling performance. Methods Fourteen aerobically-trained male cyclists (V˙O2max 62.1±4.6 ml·kg⁻¹·min⁻¹) performed two graded exercise tests (50 W warm-up followed by 25 W·min⁻¹) to exhaustion. Blood lactate, V˙O2 and V˙CO2 data were collected at every stage. Workloads at VT1 (rise in V˙E/V˙O2;) and VT2 (rise in V˙E/V˙CO2) were compared with workloads at lactate thresholds. Several continuous tests were needed to detect the MLSS workload. Agreement and differences among tests were assessed with ANOVA, ICC and Bland-Altman. Reliability of each test was evaluated using ICC, CV and Bland-Altman plots. Results Workloads at lactate threshold (LT) and LT+2.0 mMol·L⁻¹ matched the ones for VT1 and VT2, respectively (p = 0.147 and 0.539; r = 0.72 and 0.80; Bias = -13.6 and 2.8, respectively). Furthermore, workload at LT+0.5 mMol·L⁻¹ coincided with MLSS workload (p = 0.449; r = 0.78; Bias = -4.5). Lactate threshold tests had high reliability (CV = 3.4–3.7%; r = 0.85–0.89; Bias = -2.1–3.0) except for DMAX method (CV = 10.3%; r = 0.57; Bias = 15.4). Ventilatory thresholds show high reliability (CV = 1.6%–3.5%; r = 0.90–0.96; Bias = -1.8–2.9) except for RER = 1 and V-Slope (CV = 5.0–6.4%; r = 0.79; Bias = -5.6–12.4). Conclusions Lactate threshold tests can be a valid and reliable alternative to ventilatory thresholds to identify the workloads at the transition from aerobic to anaerobic metabolism.
Full-text available
The hyperbolic form of the power-duration relationship is rigorous and highly conserved across species, forms of exercise and individual muscles/muscle groups. For modalities such as cycling, the relationship resolves to two parameters, the asymptote for power (critical power, CP) and the so-called W' (work doable above CP), which together predict the tolerable duration of exercise above CP. Crucially, the CP concept integrates sentinel physiological profiles - respiratory, metabolic and contractile - within a coherent framework that has great scientific and practical utility. Rather than calibrating equivalent exercise intensities relative to metabolically distant parameters such as the lactate threshold or V˙O2 max, setting the exercise intensity relative to CP unifies the profile of systemic and intramuscular responses and, if greater than CP, predicts the tolerable duration of exercise until W' is expended, V˙O2 max is attained, and intolerance is manifested. CP may be regarded as a 'fatigue threshold' in the sense that it separates exercise intensity domains within which the physiological responses to exercise can (<CP) or cannot (>CP) be stabilized. The CP concept therefore enables important insights into 1) the principal loci of fatigue development (central vs. peripheral) at different intensities of exercise, and 2) mechanisms of cardiovascular and metabolic control and their modulation by factors such as O2 delivery. Practically, the CP concept has great potential application in optimizing athletic training programs and performance as well as improving the life quality for individuals enduring chronic disease.
Full-text available
Small changes in performance, as low as 1%, are regarded as meaningful in well-trained cyclists. Being able to detect these changes is necessary to fine tune training and optimise performance. The typical error of measurement (TEM) in common performance cycle tests is about 2-3%. It is not known whether this TEM is lower in well-trained cyclists and therefore whether small changes in performance parameters are detectable. In this research, after familiarisation, 17 well-trained cyclists each completed three Peak Power Output (PPO) tests (including VO2max) and three 40km time trials (40km TT). All tests were performed after a standardised warm-up at the same relative intensity and under a strict testing-protocol. TEM within the PPO-test was 2.2% for VO2max and 0.9% for PPO, while TEM for the 40km TT was 0.9%. In conclusion, measurement of PPO and 40km TT time, after a standardised warm-up, has sufficient precision in well-trained cyclists to detect small meaningful changes.
The aerodynamic drag of a cyclist in time trial (TT) position is strongly influenced by the torso angle. While decreasing the torso angle reduces the drag, it limits the physiological functioning of the cyclist. Therefore the aims of this study were to predict the optimal TT cycling position as function of the cycling speed and to determine at which speed the aerodynamic power losses start to dominate. Two models were developed to determine the optimal torso angle: a ‘Metabolic Energy Model' and a ‘Power Output Model'. The Metabolic Energy Model minimized the required cycling energy expenditure, while the Power Output Model maximized the cyclists' power output. The input parameters were experimentally collected from 19 TT cyclists at different torso angle positions (0–24°). The results showed that for both models, the optimal torso angle depends strongly on the cycling speed, with decreasing torso angles at increasing speeds. The aerodynamic losses outweigh the power losses at cycling speeds above 46 km/h. However, a fully horizontal torso is not optimal. For speeds below 30 km/h, it is beneficial to ride in a more upright TT position. The two model outputs were not completely similar, due to the different model approaches. The Metabolic Energy Model could be applied for endurance events, while the Power Output Model is more suitable in sprinting or in variable conditions (wind, undulating course, etc.). It is suggested that despite some limitations, the models give valuable information about improving the cycling performance by optimizing the TT cycling position.
The speed attained by a track cyclist is strongly influenced by aerodynamic drag, being the major retarding force in track events of more than 200 m. The aims of this study were to determine the effect of changes in shoulder and torso angles on the aerodynamic drag and power output of a track cyclist. The drag of three competitive track cyclists was measured in a wind tunnel at 40 kph. Changes in shoulder and torso angles were made using a custom adjustable handlebar setup. The power output was measured for each position using an SRM Power Meter. The power required by each athlete to maintain a specific speed in each position was calculated, which enabled the surplus power in each position to be determined. The results showed that torso angle influenced the drag area and shoulder angle influenced the power output, and that a low torso angle and middle shoulder angle optimised the surplus power. However, the lowest possible torso angle was not always the best position. Although differences between individual riders was seen, there was a strong correlation between torso angle and drag area.
A new conception of dynamic or static muscular work tests is presented. The authors define the critical power of a muscular work from the notions of maximum work and maximum time of work. The work capacity is then considered in the case of dynamic work, and of continuous or intermittent static work. From the data presented it is possible to define the maximum amount of work that can be performed in a given time as well as the conditions of work performed without fatigue. (French & German summaries) (22 ref.) (PsycINFO Database Record (c) 2012 APA, all rights reserved)
To obtain optimal training effects and avoid overtraining, it is necessary to monitor the intensity of training. In cycling, speed is not an accurate indicator of exercise intensity, and therefore alternatives have to be found to monitor exercise intensity during training and competition. Power output may be the most direct indicator, but heart rate is easier to monitor and measure. There are, however, limitations that have to be taken into account when using a heart rate monitor. For example, the position on the bicycle may change heart rate at a given exercise intensity. More important, however, is the increase in heart rate over time, a phenomenon described as 'cardiac drift'. Cardiac drift can change the heart rate-power output relationship drastically, especially in hot environments or at altitude. It is important to determine whether one is interested in monitoring exercise intensity per se or measuring whole-body stress. Power output may be a better indicator of the former and heart rate may, under many conditions, be a better indicator of the latter. Heart rate can be used to evaluate a cyclist after training or competition, or to determine the exercise intensity during training. Heart rate monitoring is very useful in the detection of early overtraining, especially in combination with lactate curves and questionnaires. During overtraining, maximal heart rates as well as submaximal heart rates may be decreased, while resting and, in particular, sleeping - heart rates may be increased.