Table 5 Characteristics and synthesis of the conceptual studies based on the PICO scheme

Di Prampero et al. [8]

12 medium-level sprinters (from previous study by di Prampero et al. [15]) and data for Usain Bolt

Summary of theoretical aspects underlying the metabolic power approach

Practical conclusions such as implementation in GPS software, estimation of actual VO₂, comparison of actual and estimated VO₂, comparison of mechanical accelerating power of medium-level sprinters and soccer players to Usain Bolt; use of original equation for metabolic power analysis (di Prampero et al. [15])

GPS (20 Hz) derived, actual VO₂ consumed was close to VO₂ determined by portable metabolic carts

Di Prampero and Osgnach [17]

/

Theoretical assumptions to update metabolic power approach

Extension of metabolic power approach by addressing (1) air resistance and (2) differences between running and walking periods

Air resistance: equation for calculating ES was extended by addition of a second equivalent slope equation: ES_D = k*v²*g⁻¹, effects of air resistance are minor and only amount of ~ 2% of total energy expenditure;

Walking periods: new equation when locomotion is identified as walking: ECw_vES = (ECw_vLES + ΔECw_v*(ES-LES)*(HES-LES)⁻¹)*EM, effects on whole match energy expenditure are ~ 14% smaller than previously obtained

Gaudino et al. [48]

29 professional male soccer players (19 ± 1 years)

Maximum sprint (12 m) and shuttle test with 180° change of direction (12 + 12 m) on different terrains (grass, artificial turf, sand)

Energetic and biomechanical variations in sprints with and without change of direction on different terrains using a GPS (5 Hz); use of modified equation for metabolic power analysis (Osgnach et al. [3])

Modified equation was extended by multiplication of an additional constant (KT = 1.45) for calculating EC on sand; EC and metabolic power were highest, while speed and acceleration were lowest on sand (p < 0.001); no significant differences between grass and artificial turf (p > 0.5)

Gray et al. [16]

/

Theoretical alternative energetic approach

Attempt to further quantify energetic costs of team sports especially during collisions via a mechanical modeling approach

Metabolic power approach shows limitations especially in collisions-based sports; alternative approach to derive energetic demands through mechanical work (external work + internal work) which can be predicted by obtaining speed and/or acceleration data

López-Fernández et al. [49]

16 Spanish 2nd Division female soccer players (20 ± 2 years)

SSG on different terrains (ground, grass, artificial turf) and different pitch sizes (400, 600, 800 m²)

Metabolic power demands of SSG played on different terrains using a GPS (5 Hz interpolated to 15 Hz); use of modified equation for metabolic power analysis (Osgnach et al. [3])

All metabolic variables were significantly lower (p < 0.05) on ground compared to all pitch sizes on grass and all except smallest pitch size on artificial turf

Osgnach et al. [3]

399 Italian elite soccer players of unknown sex (27 ± 4 years)

Data from 56 competitive matches during one season

Match performance based on speed, acceleration, and metabolic power using a video camera system (25 Hz); EC and metabolic power were calculated via metabolic power approach; use of original equation for metabolic power analysis (di Prampero et al. [15])

Original equation was extended by multiplication of a constant (KT = 1.29) to take different terrain into account (grass vs. treadmill); mean EE during match play is 14.60 ± 1.57 kcal/kg

Osgnach and di Prampero [50]

Subjects from 2, then unpublished, studies: 1. soccer players (no further description); 2.497 Italian outfield soccer players

Theoretical approach as well as data from then unpublished studies

Estimation of corresponding time course of actual VO₂, aerobic and anaerobic energy supply as well as high and low intensity energy bouts

Equation for estimating actual VO₂ kinetics: VO₂T_n(t) = (E_n−VO₂T_n(0))*(1−e^−t/T) + VO₂T_n(0); anaerobic or aerobic energy supply is given when metabolic power requirement is greater or lower than actual VO₂; 5-step procedure to identify high and low intensity energy bouts (excess of a defined threshold, duration, peak and mean power, subsequent bouts, low intensity)

Polglaze and Hoppe [19]

/

Summary of metabolic power approach and its limitations and benefits as well as future perspectives

/

Metabolic power approach addresses energetic cost of changing speed by analyzing interaction between speed and acceleration but it is not capable to estimate overall energy expenditure or mechanical work; distinction between acceleration conducted at different starting speeds; validity of metabolic power approach: VO₂ and metabolic power cannot simply be compared as VO₂ shows aerobic, whereas metabolic power shows aerobic and anaerobic contribution; metabolic power is a sensible tool to quantify intensity in team sports with the potential to use individualized thresholds

Ponzano and Gollin [51]

12 nationally ranked male tennis players (16 ± 3 years)

Data from 24 matches with 12 matches being played on red clay and on hard court each

Analysis of speed, heart rate, acceleration, deceleration, metabolic power using a GPS (15 Hz); use of original equation for metabolic power analysis (di Prampero et al. [15])

Mean metabolic power was significantly higher (p < 0.05, effect size = 0.72) on clay (3.9 ± 0.3 W/kg) compared to hard court (3.7 ± 0.3 W/kg)

Savoia et al. [52]

Two-part study: (1) 17 Italian professional male soccer players (24 ± 3 years), (2) 13 out of the 17 players of first part of study (22 ± 6 years)

Assessment of VO_2max on a treadmill run to exhaustion; (1) 6 min aerobic-based steady-state run at 10.29 km/h on a 160 m circular course; (2) 8 min soccer-specific run at varying speeds

Determination of energy cost of running on grass as well as updating and validating the metabolic power equation using a GPS (10 Hz)

Energy cost of running on grass was 4.7 J/kg/m; converting metabolic power algorithm to a new on including energy cost of running on grass: EC = 30.4x⁴−5.0975x³ + 46.3x² + 17.696 + 4.66; correlation between metabolic power via VO₂ and via GPS with new equation as well as via GPS with old equation was 0.66 and 0.63, respectively; estimates of fixed and proportion bias were negligible in both approaches; significant difference between metabolic power via VO₂ and via GPS with old equation (p < 0.001) and no difference between metabolic power via VO₂ and via GPS with new equation (p = 0.853)

Vescovi and Falenchuk

[53]

28 professional female soccer players of unknown age

Data from official matches

Examination of impact of different surfaces (natural vs. artificial turf) on metabolic power distances using a GPS (5 Hz); use of modified equation for metabolic power analysis (Osgnach et al. [3])

High-, elevated-, and maximal-metabolic power distances were elevated on artificial turf compared to natural turf (p = 0.004, 0.097, 0.239, respectively)