Estimation of the transpulmonary pressure from the central venous pressure in mechanically ventilated patients

Abstract

Transpulmonary pressure (P_L) calculation requires esophageal pressure (P_ES) as a surrogate of pleural pressure (Ppl), but its calibration is a cumbersome technique. Central venous pressure (CVP) swings may reflect tidal variations in Ppl and could be used instead of P_ES, but the interpretation of CVP waveforms could be difficult due to superposition of heartbeat-induced pressure changes. Thus, we developed a digital filter able to remove the cardiac noise to obtain a filtered CVP (f-CVP). The aim of the study was to evaluate the accuracy of CVP and filtered CVP swings (ΔCVP and Δf-CVP, respectively) in estimating esophageal respiratory swings (ΔP_ES) and compare P_L calculated with CVP, f-CVP and P_ES; then we tested the diagnostic accuracy of the f-CVP method to identify unsafe high P_L levels, defined as P_L>10 cmH₂O. Twenty patients with acute respiratory failure (defined as PaO₂/FiO₂ ratio below 200 mmHg) treated with invasive mechanical ventilation and monitored with an esophageal balloon and central venous catheter were enrolled prospectively. For each patient a recording session at baseline was performed, repeated if a modification in ventilatory settings occurred. P_ES, CVP and airway pressure during an end-inspiratory and -expiratory pause were simultaneously recorded; CVP, f-CVP and P_ES waveforms were analyzed off-line and used to calculate transpulmonary pressure (P_LCVP, P_Lf-CVP, P_LP_ES, respectively). Δf-CVP correlated better than ΔCVP with ΔP_ES (r = 0.8, p = 0.001 vs. r = 0.08, p = 0.73), with a lower bias in Bland Altman analysis in favor of P_Lf-CVP (mean bias - 0.16, Limits of Agreement (LoA) -1.31, 0.98 cmH₂O vs. mean bias - 0.79, LoA - 3.14, 1.55 cmH₂O). Both P_Lf-CVP and P_LCVP correlated well with P_LP_ES (r = 0.98, p < 0.001 vs. r = 0.94, p < 0.001), again with a lower bias in Bland Altman analysis in favor of P_Lf-CVP (0.15, LoA - 0.95, 1.26 cmH₂O vs. 0.80, LoA - 1.51, 3.12, cmH₂O). P_Lf-CVP discriminated high P_L value with an area under the receiver operating characteristic curve 0.99 (standard deviation, SD, 0.02) (AUC difference = 0.01 [-0.024; 0.05], p = 0.48). In mechanically ventilated patients with acute respiratory failure, the digital filtered CVP estimated ΔP_ES and P_L obtained from digital filtered CVP represented a reliable value of standard P_L measured with the esophageal method and could identify patients with non-protective ventilation settings.

Keywords: ARDS; Central venous pressure; Digital filtering analysis; Esophageal pressure; Mechanical ventilation; Pleural pressure; Respiratory failure; Transpulmonary pressure.

Fig. 1

A showed the CVP curve (in green) and P_ES curve (in blue). ΔCVP and ΔP_ES were calculated as differences between values obtained during end-inspiratory and end-expiratory pauses. For each pause, CVP values were identified at the base of the “c” wave (the dotted horizontal line in CVP wave), while for P_ES positive peak values were considered (the dotted horizontal line in Pes wave). The double headed arrows in the CVP and P_ES curves identified the magnitude of ΔCVP and ΔP_ES, respectively. In the red boxes the inspiratory and expiratory pauses were enlarged to appreciate the marker location. X axis, time as seconds (s); y axis, amplitude in cmH₂O. B displayed the same curves in Fig. 1A with the addition of f-CVP curve (in black). Δf-CVP was calculated as differences between values obtained during end-inspiratory and end-expiratory pauses. For each pause, f-CVP values were identified at the plateau level (the dotted horizontal line in f-CVP wave). The double headed arrow in the f-CVP curve identified the magnitude of Δf-CVP. In the red boxes the inspiratory and expiratory pauses were enlarged to appreciate the marker location. X axis, time as seconds (s); y axis, amplitude in cmH₂O

Fig. 2

Correlation between Δf-CVP and ΔP_ES. Δf-CVP, respiratory changes in filtered central venous pressure; ΔP_ES, respiratory changes in esophageal pressure. Solid line, linear regression line

Fig. 3

Bland-Altman analysis for the agreement between Δf-CVP and ΔP_Es, and between ΔCVP and ΔP_ES, box A and B, respectively. Bland-Altman analysis for the agreement between P_Lf-CVP and P_LP_ES, and between P_LCVP and P_LP_ES, box C and D, respectively. Broken lines, bias; dotted lines, ± 1.96 SD of the bias. ΔCVP, respiratory changes in central venous pressure; Δf-CVP, respiratory changes in filtered central venous pressure; ΔP_ES, respiratory changes in esophageal pressure; P_Lf-CVP, transpulmonary pressure obtained using f-CVP; P_LP_ES, transpulmonary pressure obtained using esophageal pressure; P_LCVP, transpulmonary pressure obtained using CVP

Fig. 4

Correlation between P_Lf-CVP and P_ES. P_Lf-CVP, transpulmonary pressure calculated with filtered central venous pressure; P_ES, transpulmonary pressure calculated with esophageal pressure. Solid line, linear regression line

Fig. 5

Four-quadrant trend plot of changes of P_LP_ES and P_Lf-CVP from baseline to second measurements. Four-quadrant trend plot representing the relationship between changes (∆) in transpulmonary pressure (P_L) estimated by esophageal pressure (∆P_LP_ES) and filtered central venous pressure (∆P_Lf-CVP). ∆ was calculated by subtracting the first P_L (baseline) from the second (after an adjustment in ventilator setting). Twelve ∆P_L pairs of measurements were considered because four were excluded from the analysis as at least one ∆P_L value was zero. Solid line, line of regression