ARTICLE REVIEW

The use of pulse pressure variation (ΔPV) as a predictor of fluid responsiveness – that is at least a 15% increase in cardiac output following administration of a fluid challenge – is well established.  However, the accuracy of this is significantly higher at tidal volumes above 8 ml/kg [1].  This presents a challenge in utilising ΔPV in patients with Acute Respiratory Distress Syndrome (ARDS), in whom lung-protective ventilation strategies with lower tidal volumes normalised for ideal body weight are adopted. Further, although a liberal fluid strategy may be needed in initial resuscitation in patients with sepsis, in established ARDS a conservative approach is likely to be favoured [2], driving the need to assess these patients’ fluid responder status.

It has previously been observed that correcting ΔPV for either the airway driving pressure (ΔPaw) or tidal volume does not improve its accuracy [3]. Liu et al. hypothesised that ΔPV corrected for respiratory changes in pleural pressure (ΔPpl) may establish a tool useful in guiding clinicians’ decisions regarding fluid administration in ARDS [4].

In this single-centre prospective observational study in a Chinese University General Intensive Care Unit, 96 patients with ARDS, circulatory failure and signs of organ hypoperfusion were recruited between July 2013 and July 2014. Patients were ventilated in a volume control mode with tidal volume set to 5-8 ml/kg ideal body weight. Oesophageal manometry was used to measure end-inspiratory and end-expiratory oesophageal pressure, and the difference used as a surrogate for ΔPpl [5]. 54% were fluid responders as assessed by either cardiac output or stroke volume, and the baseline characteristics of both responders and non-responders were similar, with an average PaO2:FiO2 of 135 mmHg and 138 mmHg respectively.

Haemodynamic measurements including heart rate, mean arterial pressure, central venous pressure, stroke volume, ΔPP and cardiac output measured by thermodilution were recorded before and after fluid challenge with 500 ml saline over 20 minutes. Repeated-measures analysis of variance was used to evaluate differences in haemodynamic variables between the responder and non-responder groups, linear regression models used to correlate the variables and a multivariate logistic regression model used to identify which variables were significantly associated with correct classification of fluid responder status.

The results show:
•    ΔPP with a cut-off of 12% had a sensitivity of 42.3% and a specificity of 88.6% in predicting fluid responsiveness
•    ΔPP with a cut-off of 10% had a sensitivity of 67.3% and a specificity of 84%
•    ΔPP/ΔPpl with a cut-off of 2% had a sensitivity of 92.4% and a specificity of 93.2%
•    Area under the receiver operator curve (AUC) for ΔPP adjusted for ΔPpl, chest wall elastance (ECW), and ECW/total respiratory system elastance (ERS) significantly larger (p < 0.01) than AUC for ΔPP alone
•    No difference adjusting for plateau pressure, ΔPaw, tidal volume, ERS, heart rate/respiratory rate, and lung elastance

Applying a grey zone approach:
•    45% of patients were within a ΔPP grey zone of 7-12%
•    3.1% were within a ΔPP/ΔPpl grey zone of 1.94-2.1% with a normal fluid strategy
•    11.4% were within a ΔPP/ΔPpl grey zone of 2.02-2.45% with a restrictive fluid strategy

There are a number of limitations. This is a single centre study, which should be followed up in further centres to ensure that results are generalisable. Patients were ventilated at tidal volumes of 5-8 ml/kg ideal body weight, rather than by following a specific ARDS ventilation strategy, such as the ARDSnet protocol or an Open Lung Ventilation approach, and ΔPP/ΔPpl should be validated for the approach used in individual centres prior to adoption. Further, there is no subgroup analysis comparing patients with extra-pulmonary ARDS and those with pulmonary ARDS, in whom pleural pressures and respiratory system elastance may be expected to behave differently. The average PaO2:FiO2 reflects moderate ARDS as per the Berlin Definition [6], and it would be interesting to see if the results are consistent in subgroups with severe ARDS. Finally, it should be studied whether applying ΔPP/ΔPpl to determine fluid management has a morbidity or mortality benefit.

In conclusion, this study does identify a tool in ΔPP/ΔPpl, which has the potential to better predict fluid responsiveness compared to ΔPP alone in patients with ARDS, and which with further evaluation could be applicable and useful to everyday clinical practice.

Article review was prepared and submitted by ESICM Journal Review Club member David Woods.

REFERENCES
1. De Backer D, Heenen S, Piagnerelli M, et al: Pulse pressure variations to predict fluid responsiveness: Influence of tidal volumes. Intensive Care Med 2005; 31:1004-1010
2. Roch A, Guervilly C, Papazian L: Fluid management in acute lung injury and ARDS. Ann Intensive Care 2011; 1:16
3. Vallée F, Richard JC, Mari A, et al: Pulse pressure variations adjusted by alveolar driving pressure to assess fluid responsiveness. Intensive Care Med 2009; 35:1004-1010
4. Liu Y, Wei L, Li G, et al: Pulse pressure variation adjusted by respiratory changes in pleural pressure, rather than by tidal volume, reliably predicts fluid responsiveness in patients with Acute Respiratory Distress Syndrome. Crit Care Med 2016; 44:342-351
5. Chiumello D, Carlesso E, Cadringher P, et al: Lung stress and strain during mechanical ventilation for acute respiratory distress syndrome. Am J Respir Crit Care Med 2008; 178:346-355
6. The ARDS Definition Task Force: Acute Respiratory Distress Syndrome: The Berlin Definition. JAMA 2012;307:2526-2533