Electronic Supplementary Material

Bench Test Evaluation of Volume Delivered by Modern ICU Ventilators during Volume-Controlled Ventilation

Aissam Lyazidi, Arnaud W. Thille, Guillaume Carteaux, Fabrice Galia, Laurent Brochard, and Jean-Christophe M. Richard in collaboration with the “Groupe de Travail sur les Respirateurs”

Contents:

Methods

Humidifier effect
Methods

Design of the experiment

We tested ICU ventilators on a Michigan test lung (Training Test Lung, Michigan Instruments, Grand Rapids, MI) that can be used to simulate various resistive and elastic conditions. Each ventilator tested was connected to the single-lung test lung via an endotracheal tube measuring 8 mm in internal diameter with a resistance of 5 cmH2O/L/s at 1 L/s of constant flow and a ventilator circuit with a measured compliance of 2 ml/cmH2O (Intersurgical, Berkshire, UK).

Ventilators were tested in volume-controlled mode with a square inspiratory flow of 60 L/min and three preset VT values of 300, 500, and 800 ml, respectively. To evaluate the impact of a pause on the compensation algorithm, we tested the three conditions with and without a 1-sec end-inspiratory pause.

Three combinations of compliance (C) and resistance (R) were used to simulate normal, obstructive, and restrictive conditions: normal with a resistance of 5 cmH2O/L/s and a compliance of 60 ml/cm H2O, chronic obstructive pulmonary disease (COPD) with a resistance of 20 cmH2O/L/s and a compliance of 60 ml/cm H2O, and acute respiratory distress syndrome (ARDS) with a resistance of 5 cmH2O/L/s and a compliance of 30 ml/cmH2O. Under these conditions and with a VT of 500 ml, peak and plateau pressure were 18±1 cmH2O and 12±1, respectively, in the normal condition; 43±4 cmH2O and 37±3, respectively, in the restrictive condition (ARDS); and 33±5 cmH2O and 13±1, respectively, in the obstructive condition (COPD).

A low respiratory frequency (14 cycles/min) was used to minimise the risk of gas trapping. All measurements were performed using an FiO2 of 21% and a positive end-expiratory pressure (PEEP) of 5 cm H2O.

Before the measurements, each ventilator was tested according to the procedure described in the manufacturer’s user manual. To be included in the study, ventilators equipped with the circuit had to successfully pass the starting test procedures, during which the ventilator automatically checked leaks and measured circuit compliance.

Compensation for tidal volume compressed in the circuit

To evaluate compensation algorithms, we measured the volume insufflated by the ventilator at the point of exit from the ventilator (beginning of the circuit) and the VT delivered to the test lung at the Y-piece (end of the circuit).

For this purpose, we used two Fleisch No. 2 pneumotachographs, one at the beginning of the inspiratory circuit (at the point of exit from the ventilator) to measure the volume insufflated by the ventilator and the other at the end of inspiratory circuit (Y piece) to measure the VT delivered to the patient (Fig. 1). The volume error was calculated as the difference between the preset VT and the VT delivered to the test lung, expressed in percentage ([preset volume– volume delivered/preset volume] ·100).

We compared the volume insufflated by the ventilator, the VT delivered to the test lung under ATPD conditions, the actually VT delivered to a patient under BTPS conditions, and the VT preset on the ventilator. Because compensation algorithm needs some breaths before reaching a steady state, each measurement was recorded after stabilization and averaged over three cycles. In order to achieve signal stabilization for each condition, the tested ventilator was always connected to the lung model for about 2 min. In this condition in controlled ventilation, variability of the delivered volume is very low (< 4%).

The differential pressure across the pneumotachograph was measured and integrated to obtain the volume (Validyne MP45, ±2.5 cm H2O, Northridge, CA). Proximal airway pressure was measured using a differential pressure transducer (Validyne MP45, ±80 cm H2O). Signals were acquired online using an analogue-digital converter (MP100; Biopac systems, Goleta, CA), sampled at 200 Hz, and stored in a laptop computer for subsequent analysis (Acqknowledge software, Biopac systems). The volume was calibrated by integration of the flow signal using a previously calibrated 1-l syringe, in order to make integration of 1L equal to 1 Volt. The calibrating maneuver was performed at least three times keeping the maximal flow in the linearity range of the pneumotachograph, moving the piston at different speeds.

Pressure transducers were calibrated at 10 cmH2O using a water column.

Ventilators

We evaluated new-generation ICU ventilators equipped with algorithms that compensated for the compressed volume in the circuit, namely, Avea (Viasys Healthcare, Conshohocken, PA), Elisée 350 (Resmed-Saime, North Ryde, Australia), Engström (General Electric, Fairfield, CO), Esprit (Respironics, Murrysville, PA), Extend (Taema, Antony, France), Evita XL (Dräger, Lübeck, Germany), Galileo (Hamilton, Rhäzuns, Switzerland), PB 840 (TYCO, Carlsbad, CA), and Servo I (Maquet, Solna, Sweden). Two of these nine ventilators could be used with and without activation of the compensation algorithm (Avea and Servo I) and were tested in the two conditions. We compared these ventilators with two old-generation ICU ventilators, namely the Puritan-Bennett 7200 (TYCO, Carlsbad, CA), one of the first ventilators equipped with a compensation algorithm; and the Bird 8400 (Viasys Healthcare, Conshohocken, PA), which does not have a compensation algorithm. It must be noted that the Galileo ventilator measures VT via a proximal flow sensor at the Y piece and was therefore included in the group of ventilators equipped with compensation algorithms.

ATPD/BTPS

The pneumotachograph at the Y-piece measured VT under ATPD conditions, since no humidification device was used. To estimate the VT actually delivered to the lungs under BTPS conditions, we used the psychrometric method to measure gas temperature and hygrometry at the Y-piece [1]. Absolute humidity was 2.9 mg H2O/L with ventilators using dry compressed gas and 7.2 mg H2O/L with turbine ventilators using room air. At the time of the measurements, ambient air temperature was 24° C and absolute humidity was 8.4 mg H2O/L.

After measuring temperature and hygrometry, we estimated VT under BTPS conditions by converting the VT under ATPD conditions, using the following formula:

Gas volume at BTPS = gas volume at ATPD (barometric pressure – water vapour pressure / barometric pressure – 47) * 310 / 273 + Temperature in °C)

where 47 is the vapour pressure of water at 37 °C, 273 is degrees Kelvin at 0° Celsius, and 310 is degrees Kelvin at 37 °C (body temperature).

The ATPD-to-BTPS correction increased the mean delivered VT by 11.6 % for compressed gas ventilators and by 10.4 % for turbine ventilators.

Humidification

All measurements were performed without humidification devices for reasons explained below.

With the three ventilators allowing different humidification settings (Avea, Evita XL, and Engström), we measured delivered VT using a preset VT of 500 ml under normal condition of respiratory mechanic with the two settings available on the ventilator, namely, heat-and-moisture exchanger (HME) or heated humidifier (HH). This allowed estimating the difference in compensation algorithm according these two available settings.


Humidifier effect

HH or HME both introduce an additional volume in the circuit. This volume is approximately 200 mL for an HH filled with water and 65 mL for the current “small” HMEs. Given the rigid nature of these systems, they influence the compliance of the circuit only by their internal volume. Given air compressibility, also referred to as adiabatic compliance, of approximately 1 mL/cmH2O per L of air, the increase in circuit compliance with an HME will be of 0.065 mL/cmH2O, and of 0.2 mL/cmH2O for an HH. Therefore, the difference in volume compressed into the circuit will not exceed 5 ml for a peak airway pressure of 50 cmH2O which could be considered as negligible. We measured circuit compliance in vitro with no humidification system, with an HME and with an HH and found respectively values of 2 mL/cmH2O, 2 mL/cmH2O and 2.2 mL/cmH2O, thus confirming the theoretical results. When measured by the ventilator during the self-test, however, these different “circuit” compliances are taken into account by the ventilator. Therefore, although there is a slight difference in baseline circuit compliance with an HH compared to an HME, once taken into account into the compensation algorithm of the ventilator the difference becomes really negligible in terms of clinical relevance.

Second, the inspiratory resistance generated by the humidifier devices measured at a flow of 1 L/s was 2.1 cmH2O/L/s for the HME. It is not negligible for the heated circuits used with HH (because of the internal wire), however, and we found values of 1.6 cmH2O/L/s for HH circuits. The difference was 0.5 cmH2O/L/s of supplementary resistance with HME. In other terms, for a flow of 1L/s, 0.5 cmH2O of extrapressure will result from the use of an HME compared to an HH. Given the circuit compliance (2 mL/cmH2O), this means that there is one additional mL of compressed volume. Again this can be considered as negligible.


REFERENCES

1. Lellouche F, Taille S, Maggiore SM, Qader S, L'Her E, Deye N, Brochard L (2004) Influence of ambient and ventilator output temperatures on performance of heated-wire humidifiers. Am J Respir Crit Care Med 170:1073-1079

- 6 -