Impact of k4b2 calibration drift on respiratory measurements during walking and running

Chris Easton, C.C.F. Howe, Richard O Matzko, Fabio Piaser, Yannis P. Pitsiladis

Research output: Contribution to journalMeeting Abstractpeer-review


The K4b2 (COSMED s.r.l, Rome, Italy) is a portable metabolic analyser designed to measure breath-by-breath respiratory variables in the field. While the majority of published evidence suggests that the K4b2 is valid and reliable during rest and exercise, it has been previously demonstrated that the accuracy of the %O2 and %CO2 measurements can drift over time (McLaughlin et al. 2001. Int J Sports Med 22(4): 280-284). However, the effect of K4b2 calibration drift on respiratory variables during prolonged exercise tests is presently unclear.

PURPOSE: To examine the accuracy of K4b2 measurements of VO2, VCO2 and VE following a one hour delay between calibration and measurement.

METHODS: Ten male participants (age: 31 ± 11 years, height: 181 ± 3 cm and body mass: 84.1 ± 10.0 kg) completed three maximal discontinuous incremental exercise tests on a motorized treadmill at speeds corresponding to 4, 5, 6, 8, 10, 12, 14 and 16 km·hr-1, or until volitional exhaustion. Participants completed 3 min of exercise at each speed, followed by 3 min active recovery for all speeds above 10 km·hr-1. Respiratory variables were measured continuously using either a laboratory based metabolic cart (Oxycon Pro, Carefusion, Germany) (OP), a K4b2 calibrated immediately before use (K4b2) or a K4b2 calibrated one hour prior to the test (K4b2DEL).

RESULTS: VCO2 and VE were not different when measured by K4b2 (P=0.27, P=0.53 respectively) or K4b2DEL (P=0.14, P=0.39 respectively) compared to the OP. VO2 was not different between OP and K4b2 (P=0.19) but K4b2DEL overestimated VO2 compared to the OP at faster running speeds (P=0.05). Bland and Altman analysis indicated good agreement in the measurement of VO2 between OP and K4b2 with a mean difference of 38 ml·min-1 and limits of agreement between -285 and 208 ml·min-1. The mean difference between K4b2DEL and OP was 84 ml·min-1 with limits of agreement between -469 and 302 ml·min-1.

CONCLUSIONS: The K4b2 overestimated VO2 during faster running speeds following a one hour delay between calibration and measurement. Whilst care should be taken when using this device for longer duration exercise tests, the extent of the differences were small and newer models of the K4b2 are fitted with a periodical recalibration mechanism which may eliminate this issue.
Original languageEnglish
Pages (from-to)S434-S435
Number of pages3
JournalMedicine & Science in Sports & Exercise
Issue number5S
Publication statusPublished - May 2012


Dive into the research topics of 'Impact of k4b2 calibration drift on respiratory measurements during walking and running'. Together they form a unique fingerprint.

Cite this