The Science Journal of the American Association for Respiratory Care

2008 OPEN FORUM Abstracts

ACCURACY OF SIDESTREAM VS. MAINSTREAM END TIDAL CAPNOGRAPHY: A BLAND-ALTMAN ANALSYS

Bradley Kuch1, Al Saville2, William Vehovic2, Jesse Kirkpatrick2, Andrea S. Bumbarger2, Shekhar T. Venkataraman1,3



Background: Sidestream capnography is advocated in the pediatric population for its accuracy and reduced weight. However, the technology is limited by sampling distance, speed, and carbon dioxide (CO2) wave slurring. We hypothesized that sidestream capnography becomes increasingly inaccurate with varying rates, I:E ratios, and resistance.

Methods:
Bench model evaluating sidestream (SS) vs. mainstream (MS) capnography at varying rates, resistance, and I:E ratios. Sampling ports for each type of capnograph where place in series with breaths delivered in pressure control mode at an inspiratory pressure of 18 cmH2O and PEEP of 5 cmH2O. Carbon dioxide was bled into the test lung at 200 ml/min. Measurements were obtained at 2 breaths increments between rates of 8 through 60 with the neonatal and adult adaptors. Low rates were considered 8 - 30 and high as 32 - 60. Resistance (Raw) was added using a 2.0 endotracheal tube cut to 10 cm placed between the test lung and sampling ports. I:E ratio of 1:3, 1:2, 1:1, and 2:1 were evaluated. All measurements were recorded following 2 minutes equilibration time. Bland-Altman analysis was used to assess bias and precision (bias � precision).

Results:
The SS and MS measures correlated strongly at a steady rate (20 breaths/minute) with increasing CO2 concentrations (r2 = 0.99; p <0.001). Bland-Altman analysis reveal a baseline (1:3 I:E ratio) bias of -1.19 with a precision of ± 0.93. Low rate had greater bias then high rates (-1.88 vs. -0.63; p <0.001). Bias increased and precision decreased with increasing I:E ratio,1:2 (-2.33±0.91), 1:1 (-3.61±1.41), and 2:1 (-5.30±1.76). Addition of Raw further affected the bias and precision, 1:3 (-4.59±2.62), 1:2 (-5.15±3.0), 1:1 (-5.91±3.04), and 2:1 (-7.39±3.51).

Conclusions:
Our model revealed significant differences in the performance of SS and MS at low vs. high rates as well as increasing bias and limited precision with escalating I:E ratios and resistance. These findings may identify the cause of poor performance of SS compared to MS technology in previously published pediatric reports.