C72-12 Adjusting Delay Time to Minimize Errors in Breath-by-Breath Measurements of Oxygen Consumption

Rationale Valid measurements of oxygen consumption (V̇O2) require accurate synchronization of gas concentration and flow signals. The gas concentration signal is typically delayed relative to flow, due to gas transport in the sampling tubing and electronic signal processing (“delay time”). Because flow rate and volume influence both, their effects vary as respiratory rate (RR) and tidal volume (Vt) change continuously during cardiopulmonary exercise testing (CPET). Since V˙O2 near maximal exercise is emphasized, understanding these errors across RR and Vt ranges typical of CPET is essential. We used a metabolic simulator to manipulate RR and Vt to determine their influence on V̇O2 measurement accuracy related to gas signal delay time. We hypothesized that higher RR and lower Vt increase the slope of the relationship between delay time (x-axis) and V̇O2 measurement accuracy (y-axis). Methods A metabolic simulator (SN17057, VacuMed) was connected to a custom breath-by-breath (BXB) gas exchange system incorporating a mass spectrometer (MATE MGA 1100, MA Tech Services) for gas analysis, screen pneumotachometers (Series 3813/4813, Hans Rudolph) for flow, and a Douglas bag system. Two trials were conducted at a set V̇O2 (2.00L/min): one with Vt clamped at 2.0L and RR increasing from 25-60 breaths/min (bpm) in 5-10bpm increments, and another with RR clamped at 25bpm and Vt increasing from 2.0-5.0L in 1.0L increments. BXB V̇O2 was calculated over periods of Douglas bag collection. Gas volumes were corrected to STPD and used for the calculation of V̇O2. Post-trials, data were re-calculated after manually adjusting system delay time from 0.100-0.310s in 0.05-0.06s increments. Associations between %V̇O2error ([bag-BXB]/bag*100) and delay time were examined across RR and Vt. An error range of ± 5% was considered “accurate” for our BXB system. Results Across changes in RR with Vt=2.0L, the %V̇O2error vs. delay-time-slope (%/s) equation ranged from -1.882x+0.562 at RR = 25bpm to -4.381x+1.238 at RR = 60bpm (see Figure). At delay times spanning 0.275-0.294s, %V̇O2error was ±5% across all RRs. Across changes in Vt with RR = 25bpm, the %V̇O2error vs. delay-time-slope equation ranged from -1.843x+0.564 at Vt=2.0L to -1.017x+0.283 at Vt=5.0L (see Figure). At delay times spanning 0.279-0.311s, %V̇O2error was ±5% across all Vt values. Conclusion Systematic manipulation of RR and Vt via a metabolic simulator determined that BXB V̇O2 measurements become increasingly sensitive to delay time at a higher RR and at a lower Vt. Given the narrow range of delay times corresponding to ± 5% V̇O2error, delay time optimization is essential for accurate V̇O2 measurement.