Dead space and CO₂ elimination related to pattern of inspiratory gas delivery in ARDS patients

Abstract

Introduction: The inspiratory flow pattern influences CO₂ elimination by affecting the time the tidal volume remains resident in alveoli. This time is expressed in terms of mean distribution time (MDT), which is the time available for distribution and diffusion of inspired tidal gas within resident alveolar gas. In healthy and sick pigs, abrupt cessation of inspiratory flow (that is, high end-inspiratory flow (EIF)), enhances CO₂ elimination. The objective was to test the hypothesis that effects of inspiratory gas delivery pattern on CO₂ exchange can be comprehensively described from the effects of MDT and EIF in patients with acute respiratory distress syndrome (ARDS).

Methods: In a medical intensive care unit of a university hospital, ARDS patients were studied during sequences of breaths with varying inspiratory flow patterns. Patients were ventilated with a computer-controlled ventilator allowing single breaths to be modified with respect to durations of inspiratory flow and postinspiratory pause (TP), as well as the shape of the inspiratory flow wave. From the single-breath test for CO₂, the volume of CO₂ eliminated by each tidal breath was derived.

Results: A long MDT, caused primarily by a long TP, led to importantly enhanced CO₂ elimination. So did a high EIF. Effects of MDT and EIF were comprehensively described with a simple equation. Typically, an efficient and a less-efficient pattern of inspiration could result in ± 10% variation of CO₂ elimination, and in individuals, up to 35%.

Conclusions: In ARDS, CO₂ elimination is importantly enhanced by an inspiratory flow pattern with long MDT and high EIF. An optimal inspiratory pattern allows a reduction of tidal volume and may be part of lung-protective ventilation.

Figures

Figure 1

Mean distribution time (MDT). MDT is the mean time during which consecutive fractions of inspired tidal volume remain in the respiratory zone of the lung (that is, the time available for distribution and mixing by diffusion of inspired gas with resident alveolar gas. The graph shows flow (black) and volume change (blue) of a breath against time. Until airway dead space (VDaw) has been inhaled (shaded area), no fresh gas arrives to alveoli, and this volume does not contribute to MDT. The following fractions of inhaled volume, N° 1 to N° n, (vertically striped area) have different distribution times in alveoli. For N° 1 distribution time is marked DT1 and for N° n DTn. MDT is the volume-weighted mean of DT1 to DTn.

Figure 2

Flow patterns. Different flow patterns studied, all at similar VT, PEEP, and expiratory time. Inspiratory flow rate is positive. (a) Only TP modified. (b) Only TI modified. (c) TI and TP modified, maintaining constant MDT. (d) TI modified at shortest possible TP (1%). (e) Shape and TP modified. The definition of end-inspiratory flow (EIF) is illustrated for increasing flow and constant flow in (e). For decreasing flow, EIF is zero, because flow rate ceases during inspiration.

Figure 3

Single-breath test for CO2 at ordinary and long postinspiratory pause in subject 4, depicting fraction of CO2 at airway opening, FCO2, against expired volume. The blue loop shows the SBT-CO2 from an ordinary breath, and the magenta loop, a breath with a prolonged TP. The blue area corresponds to VTCO2 of an ordinary breath. The additional volume of CO2 eliminated at the longer TP, ΔVTCO2, indicated by hatched area, is caused partly by a lower-airway dead space (indicated by interrupted lines) and partly by a higher level of the alveolar plateau.

Figure 4

Effects of TP variation. When only TP was modified, ΔVTCO2% increased with MDT, shown by different colors for each patient. Each dot represents one breath. To illustrate the close correlation between ΔVTCO2% and lnMDT, lines for each subject represent the relation: ΔVTCO2% = m × lnMDT + n.

Figure 5

ΔVTCO2% plotted against lnMDT in subject 4. Groups of breaths with EIF within specified ranges are indicated by separate colors. For each range of EIF, a linear relation between lnMDT and ΔVTCO2% was observed.