Temporal change, reproducibility, and interobserver variability in pressure-volume curves in adults with acute lung injury and acute respiratory distress syndrome

Abstract

Objectives: To assess the reproducibility of the static pressure-volume curve of the respiratory system by using a "mini-syringe" technique; to assess the temporal change in upper (UIP) and lower inflection points (LIP) measured from pressure-volume curves of the respiratory system; to assess the inter- and intraobserver variability in detection of the UIP and LIP in patients with acute lung injury (ALI)/acute respiratory distress syndrome (ARDS); and to compare the syringe and multiple occlusion techniques for determining LIP and UIP.

Design: Prospective observational study.

Setting: Academic medical-surgical critical care unit.

Patients: Consecutive patients with ALI or ARDS.

Interventions: Static inspiratory pressure-volume curves of the respiratory system were determined twice on day 1 of diagnosis of ALI/ARDS and then once daily for up to 6 days by using the syringe technique. Pressure-volume curves were determined from zero positive end-expiratory pressure. At each time point, three separate measurements of the pressure-volume curve were made to determine reproducibility. A 100-mL graduated syringe was used to inflate patients' lungs with 50- to 100-mL increments up to an airway pressure of 45 cm H2O or a total volume of 2 L; each volume step was maintained for 2-3 secs until a plateau airway pressure was recorded. On day 1, the static pressure-volume curve also was determined by using the multiple occlusion technique. In a random and blinded sequence, the pressure-volume curves were examined visually by three critical care physicians on three different occasions, to determine the intra- and interobserver variability in visual detection of the LIP and UIP. Observers were given objective instructions to visually identify LIP and UIP.

Measurements and main results: Eleven patients were enrolled, with a total of 134 pressure-volume curves generated. LIP and UIP could be detected in 90-94% and 61-68% of curves, respectively. When the three successive pressure-volume curves were compared, both the LIP and UIP were within 3 cm H2O in >65% of curves. The index of reliability (intraclass correlation coefficient) in LIP and UIP was 0.92 and 0.89 for interobserver variability and 0.90 and 0.88 for intraobserver variability. Daily variability was as high as 7 cm H2O for LIP and 5 cm H2O for UIP. When pressure-volume curves obtained by using the multiple occlusion and syringe techniques were compared, LIP was within 2 cm H2O, and UIP was within 4 cm H2O with the two techniques.

Conclusions: The static pressure-volume curve of the respiratory system is reasonably reproducible, thus avoiding the need for multiple measurements at a single time. We found excellent interobserver and intraobserver correlation in manual identification of the LIP and UIP. Both LIP and UIP show appreciable daily variability in patients with ALI/ARDS. The multiple occlusion and syringe techniques generate similar values for LIP and UIP.