Der Wirkungsgrad von Training
Energy Cost of Running

Energy Cost of Running Related to Running Intensity and Peak Oxygen Uptake

Energieverbrauch beim Laufen in Abhängigkeit von Laufintensität und Peak-Sauerstoffaufnahme


The hypotheses that a) a positive correlation between running oxygen uptake (VO2run) and peak oxygen uptake (VO2peak) is caused by a higher caloric equivalent and/or more reliance on anaerobic metabolic energy in subjects with lower VO2peak and that b) the energy cost per meter increases with relative intensity related to VO2peak were tested.
Twenty-nine males (mean±SD age: 24.4±2.7yrs; height: 179.0±5.6cm; body mass: 74.5±6.8kg; VO2peak: 51.5±5.2ml kg-1 min-1) ran at 2.6 and 3.0m s-1. VO2run, oxygen uptake per meter distance (CVO2), energy cost per meter distance based on respiratory measures and indirect calorimetry (CAER), and net increase in blood lactate concentration (CANAER), as well as total C (CTOT=CAER+CANAER) were analyzed. Ad a) VO2run and CVO2 were positively, and CANAER negatively interrelated with VO2peak (all p<0.05). CTOT was independent of VO2peak. Ad b) CVO2 was independent of relative intensity related to VO2peak, whilst CAER, CANAER and CTOT increased with relative intensity (all p<0.05).
The frequently observedpositive interrelationship between VO2run and VO2peak reflects less aerobic carbohydrate combustion and less reliance on anaerobic glycolysis in fitter subjects with higher VO2peak. Additionally, running economy decreases with increasing relative intensity.

KEY WORDS: Terrestrial Locomotion, Human, Speed, Aerobic, Anaerobic


Die Hypothesen, dass a) eine positive Korrelation zwischen Sauerstoffaufnahme beim Laufen (VO2run) und der Peak-Sauerstoffaufnahme (VO2peak) durch ein höheres kalorisches Äquivalent und/oder eine größere Abhängigkeit von anaerober metabolischer Energie bei Personen mit niedriger VO2peak verursacht wird und b) der Energieverbrauch pro Meter Laufstrecke mit relativer Intensität bezogen auf die VO2peak ansteigt, wurden getestet.
Neunundzwanzig Männer(Mittelwert±SD Alter: 24,4±2,7Jahre; Größe: 179,0±5,6cm; Körpermasse: 74,5±6,8kg; VO2peak: 51,5±5,2ml kg-1min-1) liefen mit Geschwindigkeiten von 2,6 und 3,0m s-1. VO2run, der Sauerstoffverbrauch pro Meter (CVO2), der Energieverbrauch pro Meter basierend auf respiratorischen Messgrößen und indirekter Kalorimetrie (CAER), und Nettoanstieg der Blutlaktatkonzentration (CANAER) sowie der Gesamtenergieverbrauch (CTOT=CAER+CANAER) wurden analysiert. Zu Hypothese a) VO2run und CVO2 hingen positiv und CANAER negativ mit der VO2peak zusammen (alle p<0.05). CTOT war unabhängig von der VO2peak. Zu Hypothese b) CVO2 war unabhängig von der relativen Intensität bezogen auf die VO2peak, wohingegen CAER, CANAER und CTOT mit relativer Intensität zunahmen (alle p<0.05).
Der häufig beobachtetepositive Zusammenhang zwischen VO2run und VO2peak spiegelt eine geringere aerobe Kohlenhydratverwertung und eine geringere Abhängigkeit von anaerober Glykolyse bei fitteren Personen mit höherer VO2peak wider. Zusätzlich nimmt die Laufökonomie mit zunehmender relativer Intensität ab.

SCHLÜSSELWÖRTER: Terrestrische Fortbewegung, Mensch, Geschwindigkeit,
Aerob, Anaerob


Running performance of top class endurance athletes but also of recreational and non-specifically trained physically active individuals depends on the maximum aerobic metabolic rate and the energy cost of a unit of running distance (1, 5, 6, 7, 11, 12, 16, 29). There is evidence that elite runners have better running economy (RE) than good and less capable runners (13, 26), and that in untrained individuals RE improves if they take on running exercises (2). However, some other studies reported that RE is essentially the same in sedentary and athletic subjects (9, 11, 31) and independent of running velocity if air resistance can be neglected such as during running on the treadmill (11, 14, 18, 20, 21, 27). In contrast, other studies reported an increase in oxygen uptake (CVO2) or aerobic energy expenditure (CAER) per unit of running distance with increasing relative intensity (8, 13).
Peak oxygen uptake (VO2peak) and oxygen uptake at given running velocities (VO2run) are generally accepted measures of maximum aerobic metabolic rate and economy of exercise. Frequent reports that VO2run is positively correlated with VO2peak (13, 15, 19, 23, 25, 30) are at odds with the supposedly logical conclusion that the combination of high maximum aerobic metabolic rate and high RE is the best concept for best endurance running performance (2, 11, 26, 28, 31, 33). Suggestions that among athletes with similar race performances a negative correlation between VO2peak and RE always reflects compensation of lower aerobic power by lower energy cost of a given task, is not fully convincing. Being the best athlete more likely combines high economy with high VO2peak. Attempts to explain the positive interrelationship between VO2run and VO2peak based on muscle fiber type and mitochondrial factors remained inconclusive (15, 25).
In previous studies analyzing the VO2run to VO2peak relationship in subjects with highly different VO2peak, the applied velocities in walking and running tests reflected substantially different relative intensities in the magnitude of between approximately 20 and above 90% of VO2peak (13, 25, 30). A consistent pattern of the VO2run to VO2peak interrelationship is that the correlation and regression coefficients become higher at velocities which reflect higher relative intensities (13, 30). Partly this may mirror differences in the respiratory exchange ratio (RER) between subjects with different VO2peak, which may vary the caloric equivalent and thus the metabolic energy per ml O2 in favor of the less fit subjects (13). Additionally, the relative intensity at which anaerobic energy is required to perform, as indicated by an increase in the blood lactate concentration (BLC), is highly variable between subjects and independent of the absolute metabolic rate. Consequently, the higher correlation and regression coefficients of the VO2run to VO2peak relationship at the higher velocities may indicate that the less fit subjects had to rely on a higher fraction of carbohydrate combustion than the fitter subjects. Additionally, less fit subjects may rely to some extent on anaerobic energy shifting the metabolic demand from aerobic to partly anaerobic metabolism. These factors may result in utilizing less oxygen at a given metabolic cost of locomotion in less fitter subjects.
Therefore, we tested the hypotheses that a) the positive correlation between VO2run and VO2peak is caused by a higher reliance on carbohydrate utilization and/or anaerobic metabolic energy in subjects with lower VO2peak and b) that RE in terms of the energy cost of a unit of running distance decreases with relative intensity if intensity related effects on carbohydrate and fat combustion, and anaerobic energy are considered


Twenty-nine male subjects (mean±SD age: 24.4±2.7yrs; height: 179.0±5.6cm; body mass: 74.5±6.8kg; VO2peak: 51.5±5.2ml kg-1min-1) signed informed consent conforming to internationally accepted policy statements on the use of human subjects as approved by the local ethics committee. All participants were healthy and physically active but not specifically trained.
They performed an incremental running test on an electronically driven treadmill (Ergo XELG2, Woodway, Germany) in an air-conditioned room (21°C, 60% humidity). The test started with a 3min resting reference phase. The initial running velocity was set 2.2m s-1. The speed was increased stepwise by 0.4m s-1 every 3min until exhaustion occurred. After each stage the running was interrupted for 30s for capillary blood sampling.
Respiratory gas exchange measures were taken continuously during the entire protocol (Oxycon Gamma, Mijnhard, Netherlands). The metabolic cart was calibrated using gases of known concentration and a syringe prior to each test. The breath-bybreath oxygen uptake data were reduced to stationary averages of the final 30s of each stage. Immediately before the start of a test, during each 30s break and after test-termination 20µl capillary blood was drawn from the hyperemic earlobe (Finalgon®,Thomae, Boehringer, Ingelheim, Germany) for the analysis of the BLC (Ebio plus, Eppendorf, Hamburg, Germany). The net lactate concentration (∆BLC) was calculated as the difference between pre- and post-run at each running velocity.
RE at 2.6 and 3.0m s-1 were analyzed in terms of: VO2run, CVO2, CAER, total energy cost above rest (CTOT) calculated via CAER plus anaerobic glycolytic energy per meter running distance (CANAER). Resting VO2 can hardly be measured correctly pre-testing which never fulfills conditions of true rest. These are defined as measurements in the morning, fasting and at indifferent temperature. Pre-testing conditions always include pre-test activities and excitement. Therefore, resting was defined as standing still and set as a VO2 of 4.5ml kg-1 min-1, which reflects standing still in males (4).

CAER, CANAER and CTOT were calculated by Eq.1, 2 and 3:

Eq.1: CAER[J kg-1m-1]=(VO2run-resting VO2) [ml s-1] caloric equivalent [J ml-1] body mass-1[kg-1] speed-1[s m-1], where the caloric equivalent was adjusted to the RER (32). A RER above 1.0, which clearly indicate respiratory compensation of a metabolic acidosis, was set 1.0 (2).

Eq.2: CANAER[J kg-1m-1] = ∆BLC [mmol l-1] O2-lactate equivalent [ml mmol-1l kg-1] 21.131 [J ml-1] running time-1[s-1] speed-1[s m-1], where an O2-lactate equivalent of 3.0ml mmol-1l kg-1was used, which is compatible to a distribution space of lactate of approximately 45% of the body mass (5, 10); the factor 21.131 reflects the caloric equivalent of carbohydrate oxidation (2).

Eq.3: CTOT[J kg-1m-1] = CAER[J kg-1m-1] + CANAER[J kg-1m-1].

All results are described as mean±SD. Differences between running velocities were tested via repeated measure ANOVA and Bonferroni post hoc test and effect sizes, in the form of partial eta squared (η2), were calculated. Linear and non-linear regression models were used to identify significant interrelationships between measures of running economy at given running velocities and VO2peak or relative intensity. The goodness of fits of different regression models were compared using the F-test (24). For all statistics the significance was set at P<0.05.


The VO2run at 2.6 and 3.0m s-1 running reflected 71±8% and 80±8% of the VO2peak, respectively. There were significant main effects and medium to large effect sizes for running velocity in VO2run, BLC, RER, CVO2, CAER, CANAER and CTOT. Significant pair differences were confirmed in VO2run, BLC, RER and CANAER in terms of increases from 2.6 to 3.0m s-1. Furthermore, VO2run, BLC, RER and CANAER were lower at 2.6 and 3.0m s-1 than at peak velocity whilst CVO2 was lower at peak velocity than at 2.6 and 3.0m s-1 (p<0.05; Tab.1).
At 3.0m s-1, VO2run (r=0.41, p<0.05, y=0.21x+30.0) as well as CVO2 were positively interrelated with VO2peak (Fig. 1). Irrespective of running speed, CAER was independent of VO2peak (Fig. 2). CANAER was negatively interrelated with VO2peak at both given velocities (Fig. 3). CTOT was independent of VO2peak (Fig. 4).

CVO2 was independent of relative intensity related to VO2peak (Fig. 5). A quadratic function best described (p<0.01) the positive interrelationship between CAER and CANAER, and relative intensity which flattens in CAER and increases progressively in CANAER (Fig. 6 and 7). The latter results in an interrelation between CTOT and relative intensity which was best described by a linear fit (Fig. 8).


The main findings of the present study were a) that below peak running velocity, CTOT is independent of VO2peak and b) that CTOT increases with relative intensity related to VO2peak. Consequently the results support the hypothesis a) that the positive correlation between VO2run and VO2peak at given speeds is caused by the higher reliance on carbohydrate oxidation plus anaerobic metabolic energy in subjects with lower VO2peak. They also support the hypothesis b) that RE decreases with increasing relative intensity.
The present results clearly support but also extent recent suggestions that measuring RE in terms of CAER or CTOT is more appropriate than expressing RE as VO2run or CVO2 (13). They also indicate that previous reports that VO2run was positively correlated with VO2peak, which becomes more obvious at higher running velocities and/or mechanical power and thus relative intensity (13, 15, 23, 25, 30), do not contradict the conclusion that a combination of high maximum aerobic metabolic rate and high RE is the best concept for a most successful endurance running performance (2, 3, 11, 26, 28, 31). However, they challenge suggestions that among athletes with similar race performances a positive correlation between VO2peak and VO2run always reflects compensation of lower aerobic power by less energy cost of a given task. The previously reported higher VO2run in athletes with higher VO2peak does not necessarily reflect lower RE. Alternatively it may indicate lower reliance on carbohydrate utilization and/or anaerobic energy. The positive interrelationship between VO2run and VO2peak does not require specific muscle fiber compositions or mitochondrial factors (15, 25). However, differences in the metabolic profile of slow- and fast-twitch muscle fibers may strongly contribute to a higher VO2peak in athletes with extremely high fractions of slow-twitch fibers. This allows less reliance on carbohydrate combustion and anaerobic energy including a reduced maximum glycolytic rate. These factors increase VO2run and CVO2 (Fig. 1). No corresponding interrelation between CAER and VO2peak may reflect that higher VO2run and CVO2 indicate either a higher lipid oxidation rate at a given CAER (identical RE) or even lower CAER (higher RE) and/or an increased CAER (lower RE). Irrespective of RE the combination of higher VO2peak and reduced maximum glycolytic rate results in a lower CANAER (Fig. 3). On aggregate, CTOT, the combination of all above effects, is independent of VO2peak and highly variable between individual subjects (Fig. 4).
Whilst considering for effects of aerobic substrate utilization, a potential contribution of anaerobic energy remains undetected when calculating CAER based on respiratory data only. The present findings clearly demonstrate that within the given range of the RER the aerobic fitness related effect on relative reliance of carbohydrate does not fully compensate for the underestimation of the energy cost of running based on VO2run and CVO2 related to VO2peak in less fit subjects (Fig. 1). The effect of changes of the caloric equivalent on CTOT was less than half the magnitude of that of CANAER.
No interrelationship between CVO2 and relative intensity (Fig. 5) was highly confirmatory of previous findings (11, 14, 18, 20, 21, 22, 27). However, in conjunction with growing evidence of an interrelationship between both CAER and CTOT and relative intensity (2, 9, 13; Fig. 6, 7 and 8) the independence between CVO2 and relative intensity also indicates that purely VO2run based estimations of the energy cost of running may substantially underestimate the real metabolic cost at higher exercise intensities (Fig. 8).

This underestimation gets even more evident at peak velocity where CVO2 suggests a slight but significant increase in RE in the magnitude of 2 to 3% which at RER <1.0 is similar to that in CAER. The latter clearly indicates that analyses of RE do not only require consideration of changes in the caloric equivalent but also the potential use of anaerobic energy to prevent potentially substantial overestimations of RE at running velocities causing an increase in the BLC. In the present study at peak velocity the increase of CANAER is approximately 2 and 2.5 times as high as the above mentioned decreases in CVO2 and CAER compared to 2.6 and 3.0m s-1, respectively. The latter furthermore stresses that estimations of the anaerobic fraction of energy required for a given performance based on estimations of the accumulated oxygen deficit (17) is likely to result in substantial underestimations of CANAER.
In conclusion, CTOT including CAER plus CANAER is independent of VO2peak. The consistently observed positive interrelationships between VO2run and VO2peak as well as CVO2 and VO2peak reflect less reliance on aerobic carbohydrate combustion and anaerobic glycolysis in fitter subjects with a higher VO2peak. Frequent observations of no interrelationship between CVO2 and relative intensity do not support independence of RE and running velocity. The non-linear increases of CAE Rand CANAER with the relative intensity are degressive (Fig. 6) and progressive (Fig. 7), respectively. In combination, these nonlinear increases resulted in a linear intensity related increase of CTOT (Fig. 8).

Conflict of Interest
The authors have no conflict of interest.


  1. ABE D, YANAGAWA K, YAMANOBE K, TAMURA K. Assessment of middledistance running performance in sub-elite young runners usingenergy cost of running. Eur J Appl Physiol Occup Physiol. 1998;77: 320-325.
  2. BENEKE R, HÜTLER M. The effect of training on running economyand performance in recreational athletes. Med Sci Sports Exerc.2005; 37: 1794-1799.
  3. BILLAT V, LEPRETRE PM, HEUGAS AM, LAURENCE MH, SALIM D,KORALSZTEIN JP. Training and bioenergetic characteristics in elitemale and female Kenyan runners. Med Sci Sports Exerc. 2003; 35:297-304.
  4. GEIGY C. Wissenschaftliche Tabellen Geigy TeilbandKörperflüssigkeiten. (Scientific Tables Geigy, Volume BodyFluids.) Ciba Geigy, Basel; 1985: 225-28.
  5. COETZER P, NOAKES TD, SANDERS B, LAMBERT MI, BOSCH AN, WIGGINS T,DENNIS SC. Superior fatigue resistance of elite black South Africandistance runners. J Appl Physiol. 1993; 75: 1822-1827.
  6. CONLEY DL, KRAHENBUHL GS. Running economy and distancerunning performance of highly trained athletes. Med Sci SportsExerc. 1980; 12: 357-360.
  7. COSTILL DL, WINROW E. Maximal oxygen intake among marathonrunners. Arch Phys Med Rehabil. 1970; 51: 317-320.
  8. DANIELS J, DANIELS N. Running economy of elite male and elitefemale runners. Med Sci Sports Exerc. 1992; 24: 483-489.
  9. DILL DB. Oxygen used in horizontal and grade walking andrunning on the treadmill. J Appl Physiol. 1965; 20: 19-22.
  10. DI PRAMPERO PE. Energetics of muscular exercise. Rev PhysiolBiochem Pharmacol. 1981; 89: 143-222.
  11. DI PRAMPERO PE. The energy cost of human locomotionon land and in water. Int J Sports Med. 1986; 07: 55-72.
  12. DI PRAMPERO PE, CAPELLI C, PAGLIARO P, ANTONUTTO G, GIRARDIS M,ZAMPARO P, SOULE RG. Energetics of best performances inmiddledistance running. J Appl Physiol. 1993; 74: 2318-2324.
  13. FLETCHER JR, ESAU SP, MACINTOSH BR. Economy of running: beyondthe measurement of oxygen uptake. J Appl Physiol. 2009; 107:1918–22. First published October 15, 2009
  14. HAGBERG JM, COYLE EF. Physiological comparison of competitiverace walking and running. Int J Sports Med. 1984; 05: 74-77.
  15. HUNTER GR, BAMMAN MM, LARSON-MEYER DE, JOANISSE DR,MCCARTHY JP, BLAUDEAU TE, NEWCOMER BR. Inverse relationshipbetween exercise economy and oxidative capacity in muscle. EurJ Appl Physiol. 2005; 94: 558-568.
  16. JONES AM, CARTER H. The effect of endurance training onparameters of aerobic fitness. Sports Med. 2000; 29: 373-386.
  17. KEIR DA, ZORY R, BOUDREAU-LARIVIÈRE C, SERRESSE O. Runningexercise: Effect of anaerobic-energy contribution at variousspeeds. Int J Sports Physiol Perform. 2012; 7: 382-389.
  18. KNUTTGEN HG. Oxygen uptake and pulse rate while runningwith undertermined and determined stride lengths atdifferent speeds. Acta Physiol Scand. 1961; 52: 366-371.
  19. LUCIA A, ESTEVE-LANAO J, OLIVAN J, GOMEZ-GALLEGO F, SAN JUAN AF,SANTIAGO C, PEREZ M, CHAMORRO-VINA C, FOSTER C. Physiologicalcharacteristics of the best Eritrean runners-exceptionalrunning economy. Appl Physiol Nutr Metab. 2006; 31: 530-540.
  20. MARGARIA R, CERRETELLI P, AGHEMO P, SASSI G. Energy cost ofrunning. J Appl Physiol. 1963; 18: 367-370.
  21. MCMIKEN DF, DANIELS JT. Aerobic requirements and maximumaerobic power in treadmill and track running. Med Sci Sports.1976; 8: 14-17.
  22. MENIER DR, PUGH LGCE. The relation of oxygen intake and velocityof walking and running in competition walkers. J Physiol. 1968;197: 717-721.
  23. MORGAN DW, DANIELS JT. Relationship between VO2max and theaerobic demand of running in elite distance runners. Int J SportsMed. 1994; 15: 426-429.
  24. MOTULSKY HJ, RANSNAS LA. Fitting curves to data using nonlinearregression: a practical and nonmathematical review. FASEB J.1987; 1: 365-374.
  25. PATE RR, MACERA CA, BAILEY SP, BARTOLI WP, POWELL KE. Physiological, anthropometric, and training correlates ofrunning economy. Med Sci Sports Exerc. 1992; 24: 1128-1133.
  26. POLLOCK ML. Submaximal and maximal working capacity of elitedistance runners. I. Cardiorespiratory aspects. Ann N Y AcadSci. 1977; 301: 310-322.
  27. PUGH LGCE. Oxygen intake in track and treadmill running withobservations on the effects of air resistance. J Physiol. 1970; 207:823-835.
  28. SALTIN B, LARSEN H, TERRADOS N, BANGSBO J, BAK T, KIM CK,SVEDENHAG J, ROLF CJ. Aerobic exercise capacity at sea level andat altitude in Kenyan boys, junior and senior runners comparedwith Scandinavian runners. Scand J Med Sci Sports. 1995; 5: 209-221.
  29. SAUNDERS PU, PYNE DB, TELFORD RD, HAWLEY JA. Factors affectingrunning economy in trained distance runners. Sports Med. 2004;34: 465-485.
  30. SAWYER BJ, BLESSINGER JR, IRVING BA, WELTMAN A, PATRIE JT,GAESSER GA. Walking and running economy: inverse associationwith peak oxygen uptake. Med Sci Sports Exerc. 2010; 42: 2122-2127.
  31. SLAWINSKI JS, BILLAT VL. Difference in mechanical andenergy cost between highly, well and non-trained runners.Med Sci Sports Exerc. 2004; 36: 1440-1446.
  32. STEGEMENN J. Leistungsphysiologie: physiologische Grundlagender Arbeit und des Sports. 4. überarb. Aufl. Stuttgart, New York,Thieme, 1991; pp. 65-60
  33. WESTON AR, MBAMBO Z, MYBURGH KH. Running economy of Africanand Caucasian distance runners. Med Sci Sports Exerc. 2000; 32:1130-1134.
Prof. Dr. Ralph Beneke
Department of Medicine, Training and
Health, Institute of Sports Sciences and
Motology, Philipps University Marburg
Jahnstraße 12, 35037 Marburg, Germany