Validity of empirical estimates of the ratio of dead space to tidal volume in ards

Abstract

BACKGROUND: The ratio of dead space to tidal volume (VD/VT) is a clinically relevant parameter in ARDS; it has been shown to predict mortality, and it determines the extent to which extracorporeal CO2 removal reduces tidal volume (VT) and driving pressure (∆P). VD/VT can be estimated with volumetric capnography, but empirical formulas using demographic and physiological information have been proposed to estimate VD/VT without the need of additional equipment. It is unknown whether estimated and measured VD/VT produce similar estimates of the predicted effect of extracorporeal CO2 removal on ∆P. METHODS: We performed a secondary analysis of data from a previous clinical trial including subjects with ARDS in whom VD/VT and CO2 production ( _VCO2 ) were measured with volumetric capnography. The estimated ratio of dead space to tidal volume (VD,est/VT) was calculated using standard empiric formulas. Agreement between measured and estimated values was evaluated with Bland-Altman analysis. Agreement between the predicted change in ∆P with extracorporeal CO2 removal as computed using the measured ratio of alveolar dead space to tidal volume (VDalv/VT)oresti-mated VDalv/VT (VDalv,est/VT) was also evaluated. RESULTS: VD,est/VT was higher than measured VD/VT, and agreement between them was low (bias 0.05, limits of agreement –0.21 to 0.31). Differences between measured and estimated _VCO2 accounted for 57% of the error in VD,est/VT.The predicted reduction in ∆P with extracorporeal CO2 removal computed using VDalv,est/VT was in rea-sonable agreement with the expected reduction using VDalv/VT (bias –0.7 cm H2O, limits of agreement –1.87 to 0.47 cm H2O). In multivariable regression, measured VD/VT was associated with mortality (odds ratio 1.9, 95% CI 1.2–3.1, P = .01), but VD,est/VT was not (odds ratio 1.2, 95% CI 0.8–1.8, P = .3). CONCLUSIONS: VD/VT and VD,est/VT showed low levels of agreement and cannot be used interchangeably in clinical practice. Nevertheless, the predicted decrease in ∆P due to extracor-poreal CO2 removal was similar when computed from either estimated or measured VDalv/VT.