ARDS is a common reason for critical illness and respiratory failure with high mortality [
16], making routinely collected measurements that accurately predict outcomes of utmost importance. They may assist in clinical management and prognostication as well as stratification of patients enrolled in clinical trials. Several physiologic parameters available at the bedside have been described to serve these purposes including
P/
F ratio [
4], oxygenation index [
17], driving pressure [
3], and dead space [
5]. Dead space can be measured by the Bohr method [
18], but this requires collection of expired gas and measurement of expired PCO
2 and arterial PCO
2. Here, we present that the estimation of dead space by either the arterial-ET difference or the simplified Enghoff modification equation is independently predictive of mortality in critically ill patients with ARDS.
A potential alternative to the direct measurement of dead space is the use of the arterial to end-tidal PCO
2 difference, which theoretically should be a reliable estimate of dead space. To date, there are limited data supporting this inference. Very early studies by Severinghaus, et al. [
19] and others [
20,
21] showed that this difference tracked with changes in dead space and the degree of ventilation/perfusion (V/Q) mismatch in
healthy animal models. Nunn et al. [
22] obtained arterial-ET differences in 12
healthy anesthetized patients and proposed this measurement as the simplest method of demonstrating the existence of V/Q mismatch. Shetty et al. evaluated 215 patients presenting to the emergency department (ED) and the arterial-ET difference modestly predicted adverse outcomes in patients presenting with suspected sepsis due to
non-respiratory causes. Those with normal arterial-ET differences were noted to have much lower risk for hospital mortality and prolonged ICU length of stay [
23]. Yamanaka et al. [
24] studied 17 patients requiring endotracheal intubation and mechanical ventilation using an average of exhaled PCO
2 at the end of several breaths over a duration of 30 s and found that the difference between arterial and exhaled CO
2 correlated closely with physiological dead space (
r = 0.80,
P < 0.05). Similarly, a prospective study in 106 trauma patients requiring emergency surgery noted that the arterial-ET difference was lower during all phases of surgery in survivors (5.8 ± 4.5 vs. 16.5 ± 14.7 mm Hg) (
P < 0.001) [
25]. In another study that included 412 patients presenting to the ED with shortness of breath, ETCO
2 was measured with a sampling cannula. A difference > 10 mm Hg was strongly predictive of the need for positive pressure ventilation via face mask or endotracheal tube (AUC 0.91 [95% CI 0.87–0.94]) [
7]. In our current study, high estimations of deadspace measured within the first 24 h of ARDS onset were significantly associated with in-hospital mortality. These statistically significant associations persisted with multiple cohorts and when using either the artieral-ET difference or simplified Enghoff equation. Accordingly, we believe that the arterial-ET difference and the simplified Enghoff equation can be useful for early prognostication in ARDS patients in a highly generalizable manner.
Our study has strengths and limitations. Our validation cohort was derived from a single-center analysis. Mortality in the University of Chicago cohort group was significantly higher than the ARDS Network group. This difference is likely explained by the University of Chicago cohort being significantly older and having dramatically worse gas exchange by P/F ratio. Both derivation and validation cohorts found the arterial-ET difference or simplified Enghoff equation to be an independent predictor of mortality, which speaks to the generalizability of our findings. This is a retrospective study with no precise timing between difference measurements. In order to minimize changes in the difference between arterial and end-tidal CO2, we required that the two measurements occurred no more than 1 h apart and with no modifications of ventilator settings. Given that arterial blood-gas and ETCO2 measurements are largely dependent on temporal hemodynamic and respiratory factors, our study is limited by potential disparities due to the rapid, time-dependent fluctuations of the measured variables. The use of volumetric capnometry may be a more accurate measurement than ETCO2, given that it allows the separation of physiologic dead space from apparent changes in dead space due to shunt and thus can give a more precise indication of physiological mechanism; however, it is often not practical in the ICU setting. Optimization of future investigation of the association between dead space fraction and mortality could include measuring the arterial blood gas and ETCO2 at the same time to reduce the risk of hemodynamic changes potentially skewing the data.