Elite cycling is continuously evolving. The combination of training, nutritional, and technological advances has resulted in records being broken on the most significant climbs of the cycling calendar, as well as a generalized increase in race speeds and, at the same time, in record power outputs during the most important races of the season. This represents a reversal of the trend observed during the years that followed the introduction of a game-changing antidoping method, such as the athlete biological passport (ABP).1,2 First developed in 2009, the ABP consists of a longitudinal screening of hematological, hormonal, and urinary biomarkers that are later subjected to Bayesian statistical analysis that highlights abnormal values that surpass previously established thresholds. Although the ABP has proven to be effective since its implementation with more than 180 antidoping rule violations identified, recent research has highlighted its limitations and the need for developing complementary antidoping methods.3,4
The International Standard for Testing and Investigations5 advocates the use of performance monitoring as an important factor within the athletes’ doping risk assessment process.5 The effectiveness of doping to improve performance has been widely proven, with studies showing rapid improvements after doping methods are implemented, and decrements when novel antidoping measures are put into practice.2,6–9 Monitoring athletes’ performance might be of utility as a complementary marker of doping particularly given the limitations associated with conventional biological testing (eg, blood, urine). Common doping practices are implemented in the weeks to months prior to the event, and thus normal testing must target these timeframes.9 In turn, performance monitoring could highlight abnormal results during the event even well after doping practices have been discontinued. Furthermore, some doping practices (eg, recombinant human erythropoietin microdosing) may be complicated to detect during biological sampling, but can still have ergogenic effects that could be potentially identified through performance monitoring.10 However, despite the potential utility of performance monitoring for the implementation of targeted testing, this strategy also has several limitations. For instance, doped athletes may overperform in terms of physiological parameters but still underperform in terms of race results or classifications, such as the Union Cycliste Internationale (UCI) ranking. Moreover, abnormal race results may occur even in the absence of doping, as cycling performance is related to several nonphysiological factors such as race strategy, technical skills, and environmental factors.11
Some of these limitations could be overcome if race results were combined with performance indicators, such as power output data, with riders being monitored on a longitudinal basis during both training and racing, preferably since the young age. Mobile power meters have been broadly implemented in the road cycling field during the last 2 decades. The analysis of power output data sets allows the longitudinal follow-up of the riders’ highest power outputs across different durations and fatigue levels.12–14 This enables the opportunity for longitudinal tracking of cyclists’ capabilities and, similarly to how the ABP is used to flag abnormal values for targeted biological testing, power output data could be used to highlight “abnormal” spikes in performance once normal intraseasonal and interseasonal variations have been accounted for.10
Given all the above, the introduction of a performance metric into the risk assessment for targeted biological testing seems appropriate.15 The assessment of power data in professional cyclists could therefore represent a valuable tool in the fight against doping.10,16 As this subject is still open to debate, this manuscript intends to provide insight and summarize the opinions and viewpoints of 15 applied sport scientists and professional cycling coaches.
Methods
To gain insight and opinion on the potential for the use of power meter data and antidoping, 3 questions were posed by the first and last authors to an expert panel of renown sport scientists and professional cycling coaches working in the field. The main criteria for participation were: (1) current coaching position in a World-Tour level cycling team and/or (2) expertise in power profiling and antidoping applied to road cycling, with relevant scientific publications during the last decade. The 3 questions were: (1) Do you believe that longitudinal monitoring of power outputs through the use of power meters could assist in doping risk assessment and be used as an additional measure for antidoping needs? Justify your previous answer; (2) If you think it would be useful, how would you implement it? If not, why don’t you see it as viable or interesting? and (3) If you answered yes to the first question and have proposed a plan for its implementation, what main limitations and difficulties do you think such a plan would encounter? To emphasize diversity of opinion, all experts were asked to explain and exemplify their choices. Consent was given on the basis that replies could be used for the purpose of this experts’ statement. In the next steps, all answers (first and last authors included) were unified and aggregated by the first and last authors, and email discussions were used to receive feedback on the subsequent versions of the manuscript.
Experts’ Opinions on the Utility and Viability of Power Data as an Antidoping Tool
Fifteen experts completed the entire feedback process and were accordingly included in panel. Two of them could be classified as sport scientists and 3 as professional cycling coaches, whereas the remaining 10 fulfilled both selection criteria. No clear trends in the experts’ opinions could be identified attending to their role as a scientist or coach. The responses provided by the experts in the field reflect 2 opposed opinions:
One subgroup of scientists and coaches declared that they could not consider power output data as a viable option to be implemented in the antidoping field given several minor and 2 major concerns. The main was the experts’ personal view on the precision, reliability, and validity of commercially available power meters, together with the problems related to calibration processes. Some of the responses read as follows: “In my experience power meters are too unreliable to perform at the level required here” or “During my coaching career I have seen variations of 30w without any good reason.” Scepticism was also related to the practical application of performance monitoring given the conflicts of interests that exist in the professional cycling world: teams and riders use several different power meters, many times provided by brands that are commercially involved in team sponsorship. Some of the responses were: “I don’t think we can bankrupt 20 power meter manufacturers and 4 teams to implement an antidoping method right now” or “For me it is not realistic to think that we can force (and enforce) the use of a single power meter for the entire peloton.” The complete list of limitations mentioned by the scientists and coaches is presented in the next section of the manuscript.
The second group of experts answered positively, affirming that power assessment is potentially implementable and would improve the antidoping fight assuming key challenges can be overcome. The most frequently highlighted challenges were again the validity of training and competition power data and system calibration. Some of the responses read as follows: “In general I think it is a good idea and can be doable but first we would have to know what is doped power vs clean power” or “In fact we are already attempting to do something similar with race power estimations calculated from climbing speeds.”17 Several responses provided general guidance on how power assessment could be potentially implemented as an antidoping tool. These ideas have been summarized and are also listed below.
Perceived Challenges of Power-Output Analysis as a Doping Marker
Power-Meter Calibration and Precision
The most reported challenge related to the implementation of power data into the antidoping fight was the ability to establish a valid tool to test intrasubject and intersubject variability across races and competitive seasons. The interdevice and intradevice precision and reliability of commercially used power meters, calibration methods, and standards is variable across manufacturers,18,19 with further research required to quantify this across the professional peloton. The experts also highlighted their personal experience with very common issues observed during training assessment: power output variations during training that were unrelated to calibration, abnormal power spikes, and especially differing accuracies among power meters from the same manufacturer and used by the same rider on different bikes. As a conclusion to these observations, several experts described power meters as “not sufficiently reliable for antidoping tasks” in their practical experience.
Conflicts of Interest, Privacy Concerns, and General Lack of Willingness to Collaborate
Several experts believed that this intervention could not be put into practice because of the conflicts of interest that would arise among teams and power meter manufacturers, which are commonly sponsoring the sport. This specific concern is that the responsible authority would need to implement one specific device that would be used by the entire peloton and would be supervised by external experts.
Furthermore, even if consensus was reached and power meter manufacturers were economically compensated for their losses, riders, coaches, and teams should all agree to share private training data, which should then by assessed and guarded by the responsible authority.
These concerns combined, although all theoretically surmountable, were stated by several of the experts as one of the main reasons to doubt the feasibility of the proposed approach.
Typical Variation in Performance
Antidoping organizations are using performance indicators such as race results as important additional information for targeted testing. Suspicious results are highlighted, and targeted testing occurs for athletes whose performance is classified as “abnormal”. One of the biggest challenges regarding the potential use of power data as an antidoping tool stated by the experts is the relative lack of knowledge of the normal seasonal variations in power output. These ranges must be established before we can attempt to determine what constitutes “abnormal” variations. In this regard, recent research has evaluated between-season variations in performance in male professional cyclists. The upper limits of variation were <12% for short efforts (≤1 min) and <8% for long efforts.20 Despite these promising findings, several experts expressed concerns about the limited knowledge available and were pessimistic about the possibility of acquiring such information in the near future.
External Factors
Power output is conditioned by several environmental factors; experts listed altitude, temperature, humidity, and road gradient as the most relevant.21–24 The longitudinal tracking of performance should account for and, ideally, factor in all these variables. Experts highlighted that up until now, scientific evidence regarding the real influence of each of these factors has been scarcely studied and has not been precisely quantified. Furthermore, some of them could not envision an implementable statistical strategy in order to properly factor in all of these particularities, although new possibilities brought by the advancement of artificial intelligence have been mentioned by several experts as a potential solution for this challenge.
Internal Factors
According to several experts, mean maximal power outputs registered from races only show what the cyclists did and not necessarily what they were capable of doing.25 Thus, if the antidoping authorities wanted to collect data reflective of maximal performances, the cyclists would also need to provide training data. This further complicates the topic as out of competition data are potentially falsifiable and concealable: the cyclist can train and test with a power meter outside of the responsible authority knowledge.
Furthermore, the already mentioned unreliability of commercial power meters is compounded by the fact that cyclists use several different bikes, each one with its own power meter, resulting in interdevice variability. Riders often change equipment between seasons, adding further potential for variance which complicates longitudinal tracking of performance.
Finally, factors unrelated to doping may influence power output: fatigue, nutritional and motivational status, changes in body mass, changes in team dynamics, legal drug use, injuries, and training tendencies were listed as the most frequent. These factors should be recorded and taken into account when assessing potentially abnormal power outputs.
A Proposal for the Implementation of Power Data Into the ABP
Some of the experts were optimistic about the potential practical implementation of this measure. Consequently, each one of them provided ideas that were summarized and organized as suggestions. All experts agreed that power data should only be used as a complimentary factor for aiding targeting of riders for follow-up biological testing via the ABP or direct doping tests for specific substances.9 In this context, an example of implementation is proposed below as a general framework upon which efforts should be directed in future years.
Measuring Device
To avoid the variability reported across different power meter manufacturers, the responsible authority, teams, staff, and riders should preferably agree upon the implementation of the same model of power measuring device, at least in the professional categories and preferably also in the development ranks. Another potential possibility would be the integration of a secondary power measuring device by the responsible authority, which would be used to recover data for antidoping testing without affecting negatively commercial and sponsorship responsibilities of each team.
Performance Tracking
Longitudinal tracking of training and competition record power outputs should preferably start prior to their entry into the professional peloton. Longitudinal data from the same rider would be used to detect abnormal intraseasonal and interseasonal variations in performance, along with changes in body mass. Data obtained from all riders taking part in the same race at the same time would also be used for comparison of individual rider performances against the rest of the peloton.
Data Analysis
Training and competition data would be integrated into a private and secure database administered by the antidoping governing bodies. Data would be statistically treated in order to detect abnormal variations in performance based on scientifically established normative ranges.
Candidates for Targeted Testing
An independent expert panel would determine which performances could be categorized as abnormal, always taking into context the already mentioned internal and external factors.26 Riders categorized as abnormal performers would enter the pool of targeted biological testing together with those showing unexpected ABP readings and illogical race results.
Practical Applications
This commentary provides suggestions for the potential implementation of power data as an antidoping tool. The proposal is based on the insights of 15 experts in the field and could be used as advice for future interventions that would aim to overcome some of the challenges highlighted along these lines.
Conclusions
An expert panel of 15 applied sport scientists and professional cycling coaches shared their insights and opinions regarding the potential use of power-meter data as an antidoping tool. These opinions could serve for future studies that aim to inform the implementation of performance monitoring in professional cycling. Given what has been discussed herein, although the potential of power data as an antidoping tool has been discussed in the literature for more than a decade, this topic still raises opposing viewpoints, with no real consensus among experts.
Acknowledgment
The authors would like to thank Professor James Hopker for his valuable insight during the creation of this manuscript.
References
- 1.↑
Iljukov S, Kauppi JP, Uusitalo ALT, Peltonen JE, Schumacher YO. Association between implementation of the athlete biological passport and female elite runners’ performance. Int J Sports Physiol Perform. 2020;15(9):1231–1236. doi:
- 2.↑
Perneger T V. Speed trends of major cycling races: does slower mean cleaner? Int J Sports Med. 2010;31(4):261–264. doi:
- 3.↑
Sottas PE, Robinson N, Rabin O, Saugy M. The athlete biological passport. Clin Chem. 2011;57(7):969–976. doi:
- 4.↑
Krumm B, Botrè F, Saugy JJ, Faiss R. Future opportunities for the Athlete Biological Passport. Front Sport Act Living. 2022;4:986875. doi:
- 6.↑
Hopker J, Passfield L, Faiss R, Saugy M. Modelling of cycling power data and its application for anti-doping. J Sci Cycl. 2016;5(2):298–303.
- 7.
Schumacher YO, Pottgiesser T. Performance profiling: a role for sport science in the fight against doping? Int J Sports Physiol Perform. 2009;4(1):129–133. doi:
- 8.
Iljukov S, Schumacher YO. Performance profiling-perspectives for anti-doping and beyond. Front Physiol. 2017;8:1102. doi:
- 9.↑
Puchowicz MJ, Mizelman E, Yogev A, Koehle MS, Townsend NE, Clarke DC. The critical power model as a potential tool for anti-doping. Front Physiol. 2018;9:643. doi:
- 10.↑
Menaspà P, Abbiss CR. Considerations on the assessment and use of cycling performance metrics and their integration in the athlete’s biological passport. Front Physiol. 2017;8:912. doi:
- 11.↑
Phillips KE, Hopkins WG. Determinants of cycling performance: a review of the dimensions and features regulating performance in elite cycling competitions. Sport Med—Open. 2020;6(1):23. doi:
- 12.↑
Pinot J, Grappe F. The record power profile to assess performance in elite cyclists. Int J Sports Med. 2011;32(11):839–844. doi:
- 13.
Sanders D, van Erp T. The physical demands and power profile of professional Men’s cycling races: an updated review. Int J Sports Physiol Perform. 2021;16(1):3–12. doi:
- 14.↑
Bell PG, Furber MJW, VAN Someren KA, Anton-Solanas A, Swart J. The physiological profile of a multiple tour de France winning cyclist. Med Sci Sports Exerc. 2017;49(1):115–123. doi:
- 15.↑
Sitko S, Mateo-March M, Zabala M. Antidoping 2.0: is it time to add power data to the biological passport? J Sci Cycl. 2023;12(1):1. doi:
- 16.↑
Hopker J, Schumacher YO, Fedoruk M, et al. Athlete performance monitoring in anti-doping. Front Physiol. 2018;9:232. doi:
- 17.↑
Millet GP, Tronche C, Grappe F. Accuracy of indirect estimation of power output from uphill performance in cycling. Int J Sports Physiol Perform. 2014;9(5):777–782. doi:
- 18.↑
Passfield L, Hopker J, Jobson S, Friel D, Zabala M. Knowledge is power: issues of measuring training and performance in cycling. J Sports Sci. 2017;35(14):1426–1434. doi:
- 19.↑
Maier T, Schmid L, Müller B, Steiner T, Wehrlin J. Accuracy of cycling power meters against a mathematical model of treadmill cycling. Int J Sports Med. 2017;38(06):456–461. doi:
- 20.↑
Valenzuela PL, Mateo-March M, Muriel X, et al. Between-seasons variability of cyclists’ peak performance: a longitudinal analysis of “real-world” power output data in male professional cyclists. Int J Sports Physiol Perform. 2023;18(10):1141–1144. doi:
- 21.↑
Muriel X, Barranco-Gil D, Zabala M, et al. The record power profile of male professional cyclists: fatigue matters. Int J Sports Physiol Perform. 2022;17(6):926–931. doi:
- 22.
Garvican-Lewis LA, Clark B, Martin DT, et al. Impact of altitude on power output during cycling stage racing. PLoS One. 2015;10(12):e0143028. doi:
- 23.
Valenzuela PL, Mateo-March M, Muriel X, et al. Road gradient and cycling power: an observational study in male professional cyclists. J Sci Med Sport. 2022;25(12):1017–1022. doi:
- 24.↑
Valenzuela PL, Mateo-March M, Zabala M, et al. Ambient temperature and field-based cycling performance: insights from male and female professional cyclists. Int J Sports Physiol Perform. 2022;17(7):1025–1029. doi:
- 25.↑
Pallares JG, Hernández-Belmonte A, Valenzuela PL, et al. Field-derived maximal power output in cycling: an accurate indicator of maximal performance capacity? Int J Sports Physiol Perform. 2022;17(11):1558–1564. doi:
- 26.↑
Montagna S, Hopker J. A Bayesian approach for the use of athlete performance data within anti-doping. Front Physiol. 2018;9:884. doi: