Relationship between esophageal and central venous pressure variation during respiratory activity: the best choice among four different methods to measure the amplitude of central venous pressure changes

Changes in lung and pleural pressures during mechanical ventilation are transmitted to intrathoracic structures, including the venous system. As a result, respiratory swings in central venous pressure (CVP) may reflect variations in pleural pressure, typically estimated using esophageal pressure (P es ). However, the point on the CVP waveform used for measurement may influence this relationship. Our aim is to find the best association between ∆P es and each ∆CVP considering different points in the CVP waveform to calculate ∆CVP. In this prospective study, we analysed CVP and Pes waveforms in 22 mechanically ventilated patients. For each cardiac cycle, three CVP values were identified: the minimum (CVPmin), the maximum (CVPmax), and the average (CVPavg). From these, four types of respiratory CVP variation (∆CVP) were calculated per breath: ∆CVPmin = max-min of CVPmin values; ∆CVPmax = max-min of CVPmax values; ∆CVPavg = max-min of CVPavg values; ∆CVPtot = absolute difference between the highest CVPmax and the lowest CVPmin within each respiratory cycle. The primary objective was to assess the association between ∆Pes and each ∆CVP variant using linear mixed-effects models. A total of 2,286 breaths (median 104 per patient) was analysed. All ∆CVP measures showed a statistically significant association with ∆Pes. The strongest correlation was observed for ∆CVPtot, with a marginal R² of 0.50 and a conditional R² of 0.78. The association further improved when using the median of multiple consecutive breaths. Specifically, the marginal R² increased from 0.60 (median of 3 breaths) to 0.70 (median of 23 breaths). Among the different measurement strategies, ∆CVPtot—defined as the difference between the highest and lowest CVP values within a respiratory cycle—showed the strongest association with ∆Pes. Averaging over at least three consecutive breaths further enhanced this relationship. These data are useful for understanding how ∆CVP can be used to estimate ∆Pes at the bedside.