Portable metabolic systems are frequently used to explore various physiological ventilatory variables in field tests (Crouter et al., 2006; King et al., 1999; Lampard et al., 2000; McLaughlin et al., 1999; 2001; Parr et al., 2001; Hodges et al., 2005; Macfarlane, 2001), collecting the same volume of data of the best automated metabolic systems (Macfarlane, 2001).

The results of validation of the Jaeger Oxycon Mobile (OM) are controversial (Perret and Mueller, 2006; Rosdahl and Gullstrand, 2004). We therefore compared the OM to the automated metabolic system Jaeger Oxycon Pro (OP) (Carter and Jeukendrup, 2002; Foss and Hallen, 2005; Rietjens et al., 2001). We also developed an equation which allows to correct the results of the OM to reduce the systematic differences between both analyzers.

Six moderately trained subjects (age 25.5 ± 7.8 years; weight 69.6 ± 4.3 kg; height 1.72 ± 0.06 m) volunteered to participate in the study. All were informed of the risks of the study, and signed an informed consent according to the recommendations of the declaration of Helsinki for investigation with human subjects (World Medical Association 2004). The Local Ethics Committee approved the study.

The gas analyzer OP (Erich Jaeger, Viasys Healthcare, Germany) is an automated metabolic system that measures O₂ by the differential paramagnetic principle, and CO₂ by infrared absorption method. This analyzer has been compared with the Douglas' bag technique (Foss and Hallen, 2005; Pedersen et al., 2002; Rietjens et al., 2001). Volumes are measured using a Triple turbine V^® with low resistance and dead space (Miller et al., 2005; Quanjer et al., 1993).

The portable metabolic system OM (Erich Jaeger, Viasys Healthcare, Germany) measures air volumes and air composition in a breath-by-breath fashion. It is composed of two small modules that can be attached to the subject’s chest or back using a harness, for a total weight of less than two kilograms. O₂ and the CO₂ are derived from an electrochemical cell and from thermal conductivity, respectively. Air volume is measured in the same way as in the OP system.

Each subject performed two incremental tests on a treadmill (H/P Cosmos Pulsar, H/P/COSMOS 3P 4.0^®, H/P/COSMOS Sports & Medical, Nussdorf-Traunstein, Germany) using one or the other metabolic system in a random order. After a warm up of 3 min at 9 Km·h^-1, the speed increased to 11 Km·h^-1. Thereafter, speed increased 1 Km·h^-1 every 2 minutes until exhaustion. The tests were performed with one rest day between each other, at the same hour of the day. During the experimental phase subjects, did not compete and did not undertake any physical training. At each speed, VO₂, VCO₂, VE and RER data were averaged every 15 seconds for later analysis.

The differences in the measures for the VO₂, VCO₂, VE and RER were analyses with a t-Student test for related samples. To check the validity of the OM, graphics were drawn for bias, following the procedure described by Bland and Altman, 1986. Lastly, an analysis of linear regression following the steps method was used to correct the values obtained by the OM. The coefficient of determination (R²), and the values of the t coefficients were used to check the viability of the pattern in the regression equation proposed. The residuals of this regression were analyzed to demonstrate the feasibility of the procedure. Statistical analysis was performed using SPSS 12.0 for Windows (SPSS Worldwide Headquarters, Chicago, IL). Significance level was fixed at p < 0.05.

All subjects reached a maximum speed of 17 Km·h^-1 in both tests. There were significant differences in the values of VO₂ at 12 Km·h^-1 and 17 Km·h^-1, but no statistical significance (p = 0.09) at 9 Km·h^-1 and 11 Km·h^-1. Significant differences were observed for RER in all the speeds except at 11 Km·h^-1. VCO₂ showed a tendency to be overestimated by OM, but no significant differences were observed. Likewise, V^E did not show significant differences at any speed (Table 1).

Figure 1 shows Bland and Altman plots for the variables studied at the various speeds. The bias for VO₂ was of 411.65 ± 267.5 ml·min^-1, while for RER it was of -0.12 ± 0.06, a mean error of 8.9 and 12% respectively.

Using linear regression analysis, the equation 1 explained 94% of the variance for VO₂. Equation 2 explained 65% of the variance for RER.

When the proposed regression equations were applied to the values for VO₂ and RER, the significant differences disappeared, and systematic errors were reduced to -0.89 ± 204.4 ml·min^-1 for VO₂ and 0.00015 ± 0.04 for RER. In practice, the mean errors decreased to 0.02% for VO₂ and 0.01% for RER (Figure 2).

This study showed a systematic error of OM in the measurement of VO₂ and RER. When the values were corrected with the proposed equations, the data from both analyzers were comparable, although the OM overestimated the VCO₂ in a constant, not statistically significant way.

Previous studies have shown errors in VO₂ measurement with portable metabolic systems. Some studies show relatively large errors for VO₂ (McLaughlin et al., 2001; Wideman et al., 1996), others smaller values (Crandall et al., 1994; Lothian et al., 1993; Peel and Utsey, 1993), and other investigations still report similar results between fixed and portable metabolic systems (Hausswirth et al., 1997; Schulz et al., 1997). Since VE did not show significant differences, it appears that the Triple V^® is capable of valid and reliable measures, and the differences found in the measurements of VO₂ could be related to the different procedures used in the two devices (electrochemical in OM vs paramagnetic in OP) (Perret and Mueller, 2006).

The results of this study agree with previous investigations (Perret and Mueller, 2006) showing significantly lower values for VO₂. However, the systematic errors are different: Perret and Mueller (2006) report a value of -110 ± 127 ml·min^-1 (OM - OP), while in our study the value increases to 411.65 ± 267.5 ml·min^-1 (OP - OM). This difference can arise from the differences in the ergometer used: we used a treadmill instead of a cyclergometer.

VCO₂ showed a tendency to be overestimated by the OM. Again, this is in agreement with previous data (Perret and Mueller, 2006). As a consequence of the systematic errors in VO₂ and VCO₂ measurement, RER is overestimated by the OM, and, although the systematic errors are slightly larger in our study compared with that of Perret and Mueller (2006) (-0.12 ± 0.06 vs 0.05 ± 0.03), the results are similar. Therefore, presumably by correcting VO₂ values, the differences in RER would be reduced.

When the regression equations were applied to our data, the systematic error decreased significantly for VO₂ (411.65 ± 267.5 ml·min^-1 vs -0.89 ± 204.4 ml·min^-1) and RER (-0.12 ± 0.06 vs 0.00015 ± 0.04).

The OM produces systematic errors when measuring VO₂ and RER. These errors that can be corrected with simple equations, making this portable metabolic system easy to use, and a valid measurement tool for metabolic expenditure in athletes. Given the lack of manufacturers´ information about the procedures used to measure the different variables calculated (Hodges, Brodie et al., 2005), we recommend to use regression equations to obtain comparable data between portable and automated metabolic systems.