Errors in MET estimates of physical activities using 3.5 ml x kg(-1) x min(-1) as the baseline oxygen consumption

J Phys Act Health. 2010 Jul;7(4):508-16. doi: 10.1123/jpah.7.4.508.


Purpose: To compare intensity misclassification and activity MET values using measured RMR (measMET) compared with 3.5 ml x kg(-1) x min(-1) (standMET) and corrected METs [corrMET = mean standMET x (3.5 / Harris-Benedict RMR)] in subgroups.

Methods: RMR was measured for 252 subjects following a 4-hr fast and before completion of 11 activities. VO2 was measured during activity using indirect calorimetry (n = 2555 activities). Subjects were classified by BMI category (normal-weight or overweight/obese), sex, age (decade 20, 30, 40, or 50 y), and fitness quintiles (low to high). Activities were classified into low, moderate, and vigorous intensity categories.

Results: The (mean +/- SD) measMET was 6.1 +/- 2.64 METs. StandMET [mean (95% CI)] was (0.51(0.42, 0.59) METs) less than measMET. CorrMET was not statistically different from measMET (-0.02 (-0.11, 0.06) METs). 12.2% of the activities were misclassified using standMETs compared with an 8.6% misclassification rate for METs based on predicted RMR (P < .0001). StandMET differences and misclassification rates were highest for low fit, overweight, and older individuals while there were no differences when corrMETs were used.

Conclusion: Using 3.5 ml x kg(-1) x min(-1) to calculate activity METs causes higher misclassification of activities and inaccurate point estimates of METs than a corrected baseline which considers individual height, weight, and age. These errors disproportionally impact subgroups of the population with the lowest activity levels.

Publication types

  • Comparative Study
  • Research Support, N.I.H., Extramural

MeSH terms

  • Adult
  • Body Mass Index
  • Exercise / physiology*
  • Female
  • Humans
  • Male
  • Middle Aged
  • Oxygen Consumption / physiology*
  • Physical Fitness / physiology*
  • Reference Values
  • Young Adult