Development of in-airway laser absorption spectroscopy for respiratory based measurements of cardiac output

Abstract

Respiratory approaches to determining cardiac output in humans are securely rooted in mass balance and therefore potentially highly accurate. To address existing limitations in the gas analysis, we developed an in-airway analyser based on laser absorption spectroscopy to provide analyses every 10 ms. The technique for estimating cardiac output requires both a relatively soluble and insoluble tracer gas, and we employed acetylene and methane for these, respectively. A multipass cell was used to provide sufficient measurement sensitivity to enable analysis directly within the main gas stream, thus avoiding errors introduced by sidestream gas analysis. To assess performance, measurements of cardiac output were made during both rest and exercise on five successive days in each of six volunteers. The measurements were extremely repeatable (coefficient of variation ~ 7%). This new measurement technology provides a stable foundation against which the algorithm to calculate cardiac output can be further developed.

Figure 1

Implementation of multipass Herriott cell within the molecular flow sensor. (a) Schematic of the Herriott cell system. The spatially multiplexed radiation for probing CH₄ and C₂H₂ is injected into the cell via a collimator, it follows a 21-pass optical path, and it is then detected by the photodiode labelled ‘Herriott detector.’ (b) Schematic illustrating the incorporation within the Herriott cell of the simple optical v-path used to probe CO₂ and H₂O. (c,d) Absorbance spectra from the Herriott cell are shown for 3030 ppm methane (c) and 2760 ppm acetylene (d) in synthetic air. Measured absorption spectra were obtained over 10 ms and are shown as filled red circles. The fitted Voigt distributions are shown as black lines.

Figure 2

Gas fractions and cumulative gas exchange recorded before, during and after a tracer gas wash-in. (a) Gas fractions for methane (magenta) and acetylene (blue). (b) Cumulative uptake of methane (magenta) and acetylene (blue). (c) Cumulative gas exchange for oxygen (green), nitrogen (black), and carbon dioxide (red).

Figure 3

Example datasets for measurement of cardiac output at rest (left) and during exercise (right). (a,b) Cumulative gas exchange at rest (a) and during exercise (b) for oxygen (green), carbon dioxide (red), and nitrogen (black). (c,d) Cumulative uptake at rest (c) and during exercise (d) for methane (magenta) and acetylene (blue). The data illustrate the greater oxygen consumption, carbon dioxide production and tidal volumes during exercise. Total methane uptake is comparable at rest and during exercise, whilst acetylene uptake is significantly greater during exercise.

Figure 4

Calculation of cardiac output. (a) Example record of first 60 s of gas uptake for methane (magenta) and acetylene (blue). Symbols (methane, black; acetylene, blue) indicate values at the ends of modified breaths chosen so that lung volume is the same at each time point. Circles are data, and crosses are values predicted from the model. Note that there are no symbols associated with the breath ending close to 40 s as expiration was incomplete. The modified breath then becomes a double breath. (b) Average overall cardiac output (rest, black line; exercise, grey line) as a function of the number of breaths included in the analysis. (c) Mean sum of squared residuals across the 57 collected datasets as a function of lung tissue volume, expressed as a fraction of functional residual capacity. The optimal fraction was ~ 0.2 where the mean sum of squared residuals was at a minimum, showing good agreement with the value taken from the literature.

Figure 5

Intraparticipant relationships between cardiac output and oxygen consumption. (a–f), Cardiac output measurements are plotted against measured oxygen consumption for each participant in a separate panel, with intraparticipant linear fits shown with solid black lines. The quality of the fits is reflected in the coefficients of variation (CoV); below 10% of the mean cardiac output for every participant.

Figure 6

Fitted linear relationship between cardiac output and oxygen consumption. The linear fit to the measured cardiac output data is shown with a solid black line. Individual cardiac output measurements are shown with a different symbol and colour for each participant. Horizontal lines illustrate the predicted cardiac output values for the average measured oxygen consumptions at rest and during exercise from two studies using the direct Fick method. Data from Johnson et al. are shown as full lines, and those from Bevegard et al. as broken lines.