Comparison of Four Methods of Lung Volume Recruitment During High Frequency Oscillatory

COMPARISON OF FOUR METHODS OF LUNG VOLUME RECRUITMENT DURING HIGH FREQUENCY OSCILLATORY VENTILATION

Anastasia Pellicano, David G. Tingay, John F. Mills, Stephen Fasulakis,

Colin J. Morley, Peter A. Dargaville

ELECTRONIC SUPPLEMENTARY MATERIAL

Pellicano et al ESM Page 1

ELECTRONIC SUPPLEMENTARY MATERIAL TEXT

METHODS:

Animal model preparation

All techniques and procedures were approved by the Animal Experimentation Ethics Committee at the Royal Children’s Hospital, Melbourne; the care and handling of the animals were conducted in accordance with the guidelines of the National Health and Medical Research Council of Australia. Neonatal piglets aged between 3 – 4 weeks were anesthetized with halothane and intubated (size 4.5 mm cuffed endotracheal tube). Intravenous and intra-arterial catheters were inserted, and continuous infusions of morphine (600 µg/kg/hr), propofol (10 mg/kg/hr), pancuronium (0.22 mg/kg/hr) and 4% dextrose (3 ml/kg/hr) were commenced. Mean arterial blood pressure, heart rate, transcutaneous oxygen saturation and rectal temperature were continuously measured and displayed (Hewlett-Packard HP48S). Body temperature was maintained between 37-38° C using a warming mattress and overhead heating. The animals were conventionally ventilated using a time-cycled, pressure-limited ventilator (Bear Cub BP2001, Bear Medical, Palm Springs, CA).

Lung injury was produced by repeated saline lavage, with the endpoint being a PaO2 < 80 mm Hg whilst ventilated with an FiO2 of 1.0. CMV was continued at standardized settings (positive end-expiratory pressure (PEEP) 6 cm H2O, respiratory rate 30 breaths/minute, Ti at 0.6 sec and FiO2 1.0) for 1-2 hours to potentiate the lung injury. PAW, tidal volume and compliance of the respiratory system (CRS) were monitored continuously whilst on CMV (Florian Respiratory Monitor, Acutronic Medical Systems, Hirzel, Switzerland), and peak inspiratory pressure was adjusted to maintain tidal volume at 8-10 mL/kg. The PAW at these settings was noted (PCMV), and from it the basal PAW (Pbasal) used on HFOV was determined, where Pbasal = PCMV + 8 cm H2O. This pressure increment above PCMV was chosen on the basis of previous studies in our laboratory in which lesser increases in PAW after initiation of HFOV in the lavaged piglet were associated with poor oxygenation and cardiorespiratory instability [E1].

The animal was transferred to the CT scanner, placed supine and thereafter remained stationary throughout the experiment. During an expiratory hold at a PEEP of 6 cm H2O, a scout CT image of the chest was taken (Prospeed CT scanner, GE Healthcare, Rydalmere, Australia), and a single tomogram 1 cm above the level of the diaphragm was chosen, and used for all subsequent imaging. The CT images were acquired at 120 kV, 130 mA, 1 sec exposure, and with a slice thickness of 3 mm. CT values were calibrated against lung phantoms of air and water.

Four different HFOV lung recruitment methods were tested on each animal in random order (computer-generated random allocation). The experimental protocol is shown in Figure 1 of the manuscript. Prior to each method, the animal was disconnected to ambient pressure for 15 seconds, followed by CMV at the above settings for 10 minutes. A baseline CT image was taken with the lung held at PEEP. HFOV was then commenced at Pbasal, with a frequency of 6 Hz and an oscillatory amplitude sufficient to produce chest wiggle (around 2 x PAW). One of the four recruitment methods was then applied, followed by a 15 minute consolidation period at Pbasal. Three of the methods involved active recruitment after connection to HFOV, these being: Escalating – step-wise pressure increments of 2 cm H2O per minute for 6 min to a PAW 12 cm H2O above Pbasal (Ppeak); Sustained DI – a single inflation sustained for 20 sec at Ppeak with HFOV continuing; and Repeated DI – a short series of six 1 sec inflations to Ppeak. The value of Ppeak (12 cm H2O above Pbasal) reflects previous observations in the lavaged lung on HFOV demonstrating adequate recruitment in response to pressures of 10-15 cm H2O above the background PAW [E2, E3]. In each case PAW was returned to Pbasal immediately after the manoeuvre was completed. The fourth recruitment method, Standard recruitment, involved setting PAW immediately at Pbasal. Further CT images were taken on HFOV prior to active recruitment, and after 1, 5 and 15 minutes of the consolidation period. Arterial blood gas analyses were performed at baseline after 10 minutes on CMV, and at the end of the consolidation period. HFOV frequency and amplitude remained unchanged throughout each recruitment method.

Once all recruitment methods had been evaluated, a pressure-volume relationship of the lung was mapped with HFOV continuing. The lung was inflated in a stepwise manner (2 cm H2O per minute), held for 5 minutes at 24 cm H2O above PCMV to identify total lung capacity (TLC), and deflated in 2 cm H2O decrements to 8 cm H2O below PCMV. CT images were taken after each change in PAW. The animal was then euthanized with intravenous pentobarbitone.

CT image analysis

CT images were stored in DICOM format and analysed using digital imaging software (3-D DoctorTM, Able Software Corp., Lexington, MA). The lung area was demarcated by manual tracing around the margins of each lung, excluding the heart and great vessels. The total lung area (mm2), slice volume (mm3) and mean density (Hounsfield units, HU) were calculated. The gas volume of all voxels in the slice was calculated (voxel count x voxel volume x), this being a validated measure of thoracic gas volume (TGV) [E4]. For each animal, the TGV value thus obtained for each image was expressed as a percentage of the corresponding value at TLC [E4].

A histogram of the density distribution was produced from each CT image (ESM Figure E1). For each recruitment episode, voxel counts in the pre-recruitment image were then subtracted from those of the image 15 mins post-recruitment to produce a subtraction histogram demonstrating the effect of recruitment (Figure E1 panel E). Data from all subtraction histograms for each recruitment method were pooled. Finally, the voxels within each slice were separated into different aeration compartments using previously reported definitions [E5]. A HU value less than –900 was deemed to be overdistended lung; between –900 and –500 normally aerated lung; between –500 and –100 HU poorly aerated lung; and between –100 and +100 HU non-aerated lung [E5]. Total volume of each aeration compartment within the slice was calculated as the cumulative volume of all voxels within the given density range, expressed in mL. Gas volume of each aeration compartment was calculated as the cumulative gas volume of all voxels within the density range, expressed in mL. Change in both total volume and gas volume within each aeration compartment were compared between recruitment methods.

Data analysis

All comparisons between the four recruitment methods were made using one-way ANOVA, with differences between individual methods identified post hoc using Duncan’s multiple range test. Longitudinal data were compared within each method using a paired t-test. A P value < 0.05 was considered statistically significant.

RESULTS:

Subtraction histograms

Subtraction of the prerecruitment CT histogram from that at 15 minutes post-recruitment showed the redistribution of voxel densities that occurred after each recruitment method (ESM Figure E2). These subtraction histograms give a visual appreciation of the differences between recruitment methods, with, for example, the most extensive loss of non-aerated lung (+100 to –100 HU) recognizable after Escalating recruitment. There were also apparent differences in the distribution of newly aerated lung, being largely centered around –500 to –450 HU in the Escalating, Sustained DI and Standard methods, and around –400 HU for Repeated DI recruitment.

REFERENCES

E1.Mills JF, Davis CE, Mazzolini A, Dargaville PA (2000) Relationship between chest wall movement and tidal volume during high frequency oscillatory ventilation. Pediatr Res 47:369A.

E2.Byford LJ, Finkler JH, Froese AB (1988) Lung volume recruitment during high-frequency oscillation in atelectasis-prone rabbits. J Appl Physiol 64:1607-1614.

E3.Walsh MC, Carlo WA (1988) Sustained inflation during HFOV improves pulmonary mechanics and oxygenation. J Appl Physiol 65:368-372.

E4. Pelosi P, Goldner M, McKibben A, Adams A, Eccher G, Caironi P, Losappio S, Gattinoni L, Marini JJ (2001) Recruitment and derecruitment during acute respiratory failure: an experimental study. Am J Respir Crit Care Med 164:122-130.

E5. Vieira SR, Puybasset L, Lu Q, Richecoeur J, Cluzel P, Coriat P, Rouby JJ (1999) A scanographic assessment of pulmonary morphology in acute lung injury. Significance of the lower inflection point detected on the lung pressure-volume curve. Am J Respir Crit Care Med 159:1612-1623.

ESM Table E1 Mean blood pressure with each recruitment method

Escalating / Sustained DI / Repeated DI / Standard
At baseline on CMV / 75±17 / 71±17 / 77±15 / 79±9.4
After transition to HFOV at Pbasal / 83±8.0 / 61±21 / 74±20 / 75±14
During active recruitment at Ppeak / 73±19 / 57±23 / 68±19* / -
Immediately after active recruitment at Pbasal / 72±19 / 65±16 / 75±13 / -
After 15 min consolidation period at Pbasal / 69±31 / 74±14 / 82±18 / 69±29

All values in mm Hg. Mean±SD. No significant differences were noted between recruitment methods at any experimental time point. *Lower mean blood pressure during active recruitment compared to that at Pbasal, P = 0.0040, paired t-test). Corresponding P values for Escalating and Sustained DI methods were 0.17 and 0.089, respectively.

Pellicano et al ESM Figure E1

ESM Figure E1 CT histograms

Panel A: CT image from piglet 7 on HFOV at Pbasal immediately prior to recruitment; Panel B: Corresponding image at Pbasal 15 min after Sustained DI recruitment; note the significant improvement in aeration between A and B. Panel C: CT histogram showing the density distribution of all CT voxels in panel A. Panel D: Corresponding CT histogram for Panel B. Panel E: Subtraction histogram showing the change in voxel count between A and B for the range of CT densities. A decrease in non-aerated lung (–100 to 100 HU), and a concomitant increase in the amount of poorly aerated lung (–100 to –500 HU) and, to a lesser extent, normally aerated lung (–500 to –900 HU) is noted after recruitment. Note that, due to increase in slice cross-sectional area and therefore voxel number in the post-recruitment image, the cumulative sum of values in the subtraction histogram is greater than zero.

ESM Figure E2 Topography of recruitment

Subtraction histograms showing change in voxel count in CT images immediately prior to, and 15 min after recruitment. Both images taken on HFOV at basal PAW. Mean and SEM for each 10 HU division. Aeration compartments are indicated in Panel A, and their boundaries shown in all panels. Panel A: Escalating; panel B: Sustained DI; panel C: Repeated DI; panel D: Standard.