Effect of external dead space removal on CO2 homeostasis in mechanically ventilated adult Covid‐19 patients

Patients with Covid‐19 respiratory failure present with hypoxemia, often in combination with hypercapnia. In this prospective, observational study we examined the effect of removing external dead space (DS) on CO2‐homeostasis in mechanically ventilated Covid‐19 patients. In addition, volumetric capnography was validated for its ability to estimate external DS volume using in vitro measured DS volumes as reference.


Editorial Comment
Elimination of carbon dioxide in mechanically ventilated cases with diffuse acute lung injury can present a clinical challenge. This study assessed possible ventilatory benefit with reducing external dead space in the breathing circuit, and with the help of advanced capnographic assessment of dead space. The findings confirm that more effective clinical carbon dioxide elimination could be achieved with this step.

| INTRODUCTION
When faced with the Covid-19 pandemic, many Covid-19 patients receiving mechanical ventilation had significant problems with eliminating carbon dioxide (CO 2 ). The setup in practice with viral filters, heat and moisture exchange (HME) filters as well as associated extensions incorporated between the Y-piece and the endotracheal tube connector, offers protection against virus contamination but will also add to the apparatus dead space (DS). This increases the risk of disturbances in CO 2 homeostasis with the need to increase minute ventilation in the ventilator to maintain stable CO 2 homeostasis. 1 The situation is similar during pressure support (PS) ventilation, where added apparatus DS is expected to increase patient work of breathing to maintain stable CO 2 . 2 When lung injury is severe, such increases in respiratory efforts may however contribute to an increased risk of patient self-inflicted lung Injury (P-SILI), even if this is yet to be clearly shown in clinical trials. 3,4 With this background we decided to perform a prospective, observational study with the primary objective to determine the effect of minimizing external (apparatus) DS on CO 2 elimination in spontaneous breathing Covid-19 patients, analyzing volumetric capnography (VCap), mechanical ventilation, and arterial blood gas data to detect potentially significant changes before and after removal of external DS. A secondary objective was to investigate VCap for its accuracy in estimating external apparatus DS volume by comparing the VCap estimated apparatus DS to in vitro measurement of the same DS volumes. was connected immediately proximal to the endotracheal tube of the patients and breath-by-breath exhaled CO 2 concentration was measured. Ventilation airflow was registered via the standard flow sensor of the Servo-U ventilator. Recorded data from this set up was then used for continuous breath-by-breath calculation of the CO 2 elimination rate (VCO 2 ) and associated VCap parameters, as described below.
After an initial 15-minute stabilization phase, to ensure stable CO 2 -homeostasis, an arterial blood sample was drawn for blood gas analysis (ABL flex 80, radiometer, Denmark). Thereafter, the endotracheal tube was temporarily clamped with a hemostat while the HME and flexible connector (i.e., investigated external DS) were removed (clamping period approximately 5-10 s). The endotracheal tube was then unclamped, and the previous ventilator settings were reconvened without any alterations. After approximately 15 min, a new arterial blood gas analysis was performed and VCap as well as ventilatory parameters were recorded again. The external DS was then re-introduced to the breathing system and the protocol was stopped.
The external DS absolute volume (flexible connectors of variable volumes between different patients) was measured in vitro by filling them with water. The HME volume was obtained from the manufacturer (HME Intersurgical Inter-Guard™, 1344007S). The sum of the volume of the flexible connectors and HME was defined as VdUSm (upstream DS measured). VDFowler: defined as anatomical DS. This was calculated from the volumetric capnogram as previously described. 5 VdUS: defined as the instrumental DS. Calculated from end tidal CO 2 fraction (FetCO 2 ) and inspired CO 2 volume for each breath (VTiCO 2 ) as follows:

| Volumetric capnography recordings
Slope3: slope of phase III, calculated for the midpoint of phase III in the volumetric capnogram where the slope coefficient equals slope III (expressed as %/l).
Two recordings of VdUS after DS removal were found to be outliers, they were, however, numerically very small and excluding these two values did not change the outcome analysis of the response in VdUS seen after DS removal.

| Changes in VCap and mechanical ventilation parameters
After external DS removal, the alveolar minute ventilation improved promptly from a median of 7094 mL min À1 (5670; 8453) to 8722 mL min À1 (7290; 14,519), p = .04. This was accompanied T A B L E 1 Demographic data. Data are expressed as percent (sex, mortality outcome) and as median (25th-75th percentile   Table 3. FACO 2 and ETCO 2 decreased significantly 5 and 15 min after DS removal, compared to baseline, as shown in Figure 1.

| Accuracy in estimating apparatus dead space
The overall bias between VdUSm and VdUS was À9 mL (lower limit of agreement À40, 95% CI: À60 to À20 and upper limit of agreement 21 mL, 95% CI: 1-40) Percentage error was 48%. The corresponding Bland Altman plot is shown in Figure 2.

| Changes in VCap and mechanical ventilation parameters
The official recommendations of the Surviving Sepsis Campaign state the importance of lung-protective ventilation also in Covid-19 ARDS. 11 In line with the recommendations, removal of external DS may have an important clinical effect and should therefore be considered for patients with severe gas-exchange abnormalities. It has the potential benefit of reducing patient respiratory drive and work of breathing during assisted ventilation, thus lowering the risk of P-SILI. The work of breathing was not measured directly in the current study. However, the rapid improvement of alveolar minute ventilation, notwithstanding unchanged respiratory rate, and tidal volumes, nevertheless, caused a significant rapid reduction in PaCO 2 , despite a parallel significant reduction in VCO 2 . This can potentially be explained by a reduced CO 2 -production. The cause of this decrease in CO 2 production was not investigated in the current study but could potentially be explained by decreased work of breathing when the DS was removed in combination with decreased rebreathing of exhaled air.
Traditional ARDS management dictates lung-protective ventilation with relatively low tidal volumes to eliminate risk of ventilator induced lung injury. 12 The added apparatus DS must however be accounted for Interestingly, a small but significant reduction in slope of phase III was seen after DS removal. This normally indicates improved ventilation/perfusion conditions, however the exact cause for the improved slope was not investigated in the current study.

| Mortality rate related to Enghoff dead space
The patients in the present study proved to have relatively high levels of alveolar DS, 56% (46-68) of the tidal volume (Enghoff DS, Table 1) with added DS, 13

| Study limitations
The main study limitation is that the current study was conducted on a small sample of patients and that subjects were investigated later in the disease and not just after intubation, as in the majority of previous ARDS studies. The originally intended sample size of 40 subjects could not be obtained for practical reasons relating to the Covid-19 pandemic. The intentional sample size was however chosen from a practical approach, since no previous experience with the tested method exists prior to this study. We believe however that the studied population size was enough to show with volumetric capnography, as proof of concept, the physiologic effects observed when external DS was reduced during assisted ventilation. Nevertheless, the results should be viewed against the relatively small sample size, even if statistical measures were made to reduce this impact.
In conclusion, in Covid-19 ARDS even a moderate DS reduction may have an impact on CO 2 homeostasis and care to reduce additional DS is likely warranted in a disease characterized by increased alveolar DS. In addition, apparatus DS can be assessed with moderate accuracy using volumetric capnography.

FUNDING INFORMATION
The work was supported by grants from the regional agreement on