Changes in CO2-Derived Variables, Induced by Passive Leg Raising Test, Detect Preload Responsiveness in Mechanically Ventilated Patients: A Pilot Study

TL;DR

This study shows that changes in CO2-related variables during a passive leg raise test can help determine if ICU patients will benefit from fluid resuscitation.

Contribution

The study introduces passive leg raising-induced CO2-derived variables as a novel method to assess preload responsiveness in ventilated ICU patients.

Findings

ΔP(cv-a)CO2 and ΔP(cv-a)CO2/C(a-cv)O2 correlated with VTI changes in preload responders.

PLR-induced decrease in P(cv-a)CO2 was significantly associated with preload responsiveness.

AUC for ΔP(cv-a)CO2 to predict preload responsiveness was 0.89, indicating strong predictive value.

Abstract

Background/Objectives. Changes in CO2-derived variables during a fluid challenge have been proposed as markers of fluid responsiveness. We investigated whether, instead of fluid administration, passive leg raising (PLR)-induced changes in the CO2-derived variables, namely central venous-arterial carbon dioxide partial pressure (P(cv-a)CO2) and the ratio between P(cv-a)CO2 and the arterial-central venous oxygen content (P(cv-a)CO2/C(a-cv)O2), could detect preload responsiveness in critically ill patients. Methods. We studied 30 mechanically ventilated patients in whom a PLR test was performed due to acute circulatory failure. Routine hemodynamic variables, velocity-time integral (VTI), in the left ventricular outflow tract, and CO2-derived variables, were measured before, during, and after a PLR test. A PLR-induced increase in VTI of ≥10% defined preload responsiveness. The differences (Δ) of P(cv-a)CO2 and P(cv-a)CO2/C(a-cv)O2 between PLR and pre-PLR were calculated. The predictive values of PLR-induced changes in the CO2-derived variables was determined by receiver operating characteristic area under curves (ROC-AUCs). Results. Fifteen patients (50%) were classified as preload responsive. ΔP(cv-a)CO2 and ΔP(cv-a)CO2/C(a-cv)O2 were correlated with VTI changes and differed significantly between responders and non-responders −1.3 (−2–−0.6) vs. 0.6 (−0.1–1.1) mmHg, p < 0.001, and −0.38 (−0.97–−0.34) vs. 0.1 (−0.15–0.57) mmHg/mL O2, p < 0.001, respectively. The PLR-induced decrease in P(cv-a)CO2 was significantly associated with preload responsiveness (OR 0.48, CI 0.20–0.89, p = 0.016, bootstrap CI 0–0.85). The AUC curves for both ΔP(cv-a)CO2 and ΔP(cv-a)CO2/C(a-cv)O2 ratio to predict preload responsiveness were 0.89 (CI 0.74–1), p < 0.001, and 0.85 (CI 0.70–1), p < 0.001, respectively. Conclusions. In mechanically ventilated ICU patients with circulatory shock, PLR-induced changes in P(cv-a)CO2 and P(cv-a)CO2/C(a-cv)O2 ratio were correlated with VTI changes. The change in P(cv-a)CO2 was the only variable detecting preload responsiveness assessed by PLR; therefore, it could serve as an indirect marker, useful to guide fluid resuscitation when cardiac output measurement is not feasible.

Figures