Effectiveness of individualized lung recruitment strategies at birth: an experimental study in preterm lambs

Transcription

1 Am J Physiol Lung Cell Mol Physiol 312: L32 L41, First published November 23, 2016; doi: /ajplung RESEARCH ARTICLE Macro Real-Time Visualization of Lung Function: From Micro to Effectiveness of individualized lung recruitment strategies at birth: an experimental study in preterm lambs X David G. Tingay, 1,2,3,4 Anushi Rajapaksa, 1,4 Emanuela Zannin, 5 Prue M. Pereira-Fantini, 1,4 Raffaele L. Dellaca, 5 Elizabeth J. Perkins, 1,2 Cornelis E. E. Zonneveld, 1 Andy Adler, 6 Don Black, 1 Inéz Frerichs, 7 Anna Lavizzari, 1,8 Magdy Sourial, 1 Bartłomiej Grychtol, 9 Fabio Mosca, 8 and Peter G. Davis 1,3,10 1 Neonatal Research, Murdoch Children s Research Institute, Parkville, Australia; 2 Neonatology, The Royal Children s Hospital, Parkville, Australia; 3 Neonatal Research, The Royal Women s Hospital, Parkville, Australia; 4 Department of Paediatrics, University of Melbourne, Melbourne, Australia; 5 TBM Laboratory, Dipartimento di Elettronica, Informazione e Ingegneria Biomedica-DEIB, Politecnico di Milano University, Milano, Italy; 6 Systems and Computer Engineering, Carleton University, Ottawa, Canada; 7 Department of Anaesthesiology and Intensive Care Medicine, University Medical Centre Schleswig-Holstein, Campus Kiel, Kiel, Germany; 8 NICU, Fondazione IRCCS Ca Granda, Ospedale Maggiore Policlinico-Università degli Studi di Milano, Milano, Italy; 9 Fraunhofer Project Group for Automation in Medicine and Biotechnology, Mannheim, Germany; and 10 Department of Obstetrics and Gynaecology, University of Melbourne, Melbourne, Australia Submitted 16 September 2016; accepted in final form 20 November 2016 Tingay DG, Rajapaksa A, Zannin E, Pereira-Fantini PM, Dellaca RL, Perkins EJ, Zonneveld CE, Adler A, Black D, Frerichs I, Lavizzari A, Sourial M, Grychtol B, Mosca F, Davis PG. Effectiveness of individualized lung recruitment strategies at birth: an experimental study in preterm lambs. Am J Physiol Lung Cell Mol Physiol 312: L32 L41, First published November 23, 2016; doi: /ajplung Respiratory transition at birth involves rapidly clearing fetal lung liquid and preventing efflux back into the lung while aeration is established. We have developed a sustained inflation (SI OPT ) individualized to volume response and a dynamic tidal positive end-expiratory pressure (PEEP) (open lung volume, OLV) strategy that both enhance this process. We aimed to compare the effect of each with a group managed with PEEP of 8 cmh 2O and no recruitment maneuver (No-RM), on gas exchange, lung mechanics, spatiotemporal aeration, and lung injury in day preterm lambs. Forty-eight fetal-instrumented lambs exposed to antenatal steroids were ventilated for 60 min after application of the allocated strategy. Spatiotemporal aeration and lung mechanics were measured with electrical impedance tomography and forced-oscillation, respectively. At study completion, molecular and histological markers of lung injury were analyzed. Mean (SD) aeration at the end of the SI OPT and OLV groups was 32 (22) and 38 (15) ml/kg, compared with 17 (10) ml/kg (180 s) in the No-RM (P 0.024, 1-way ANOVA). This translated into better oxygenation at 60 min (P 0.047; 2-way ANOVA) resulting from better distal lung tissue aeration in SI OPT and OLV. There was no difference in lung injury. Neither SI OPT nor OLV achieved homogeneous aeration. Histological injury and mrna biomarker upregulation were more likely in the regions with better initial aeration, suggesting volutrauma. Tidal ventilation or an SI achieves similar aeration if optimized, suggesting that preventing fluid efflux after lung liquid clearance is at least as important as fluid clearance during the initial inflation at birth. Address for reprint requests and other correspondence: D. Tingay, Neonatal Research, Murdoch Children s Research Institute, Royal Children s Hospital, Flemington Rd., Parkville 3052 Victoria, Australia ( david.tingay@rch.org.au). preterm; infant; open lung ventilation; sustained inflation; lung mechanics; regional lung injury THE INITIAL TRANSITION from in utero life poses many problems for the preterm infant (14). Effective tidal ventilation requires the underdeveloped lung to rapidly transition from a fluid-filled to aerated state (12, 14). This complex process requires rapid clearance of fetal lung fluid from the airway and alveoli, establishing a functional residual capacity and then prevention of fluid efflux back into the alveoli spaces during expiration (13). Inability to successfully complete this process exposes the preterm lung to heterogeneous states of aeration and ventilation, both increasing the risk of injury (14, 40). Both positive end-expiratory pressure (PEEP) and an initial sustained lung inflation (SI) have been suggested as methods of optimizing respiratory transition (13, 18). In rabbit pups, an initial SI has been shown to assist clearance of lung fluid, with PEEP then preventing fluid efflux back into alveolar spaces during subsequent tidal ventilation (31). Subsequent clinical and preclinical data regarding an SI have been conflicting (16, 18, 28, 31), with some studies showing no substantive benefit over tidal ventilation with sufficient PEEP (11, 39). The interpretation of these studies has been hampered by inconsistencies in both the SI and PEEP strategies used and the use of proven lung protective therapies, such as antenatal corticosteroids and exogenous surfactant (38). We have previously shown that tidal ventilation at birth with a dynamic PEEP, or open lung ventilation (OLV), approach that exploited hysteresis resulted in more homogenous spatiotemporal aeration and better outcomes than ventilation with a fixed PEEP of 8 cmh 2 O or a 30-s SI (33, 39). The advantage of an OLV approach is that it uses real-time feedback on the mechanical properties of an individual s lung (3, 15). In contrast, SI strategies have traditionally L /17 Copyright 2017 the American Physiological Society

2 used a predetermined pressure, duration, or delivered volume (24, 30, 31, 33, 37 39). To be effective, an SI must overcome the long time constants of the fluid-filled respiratory system (30), which are highly variable even in standardized preterm animal models (34, 40). Recently, we demonstrated that a volumetric SI, in which the SI duration was individualized to achieving stable aeration optimized outcomes compared with a predefined 30-s SI (34). It is possible that the SI strategy used in our previous OLV studies, and by many other groups, was inappropriate given the variable mechanical characteristics of the preterm lung. Notwithstanding this, the observation that an OLV strategy achieved lung aeration without the need for an initial SI to rapidly clear lung fluid is intriguing. To date, optimal SI and OLV approaches have not been compared in the preterm lung exposed to antenatal steroids. We hypothesized that optimizing lung recruitment at birth using an SI or OLV would create different pathways of spatiotemporal aeration during respiratory transition. The specific aims of the study were to compare the effects of an individualized SI, dynamic PEEP OLV, and a static PEEP approach on gas exchange, lung mechanics, spatiotemporal aeration, and lung injury in preterm lambs exposed to antenatal steroids. METHODS All techniques and procedures were approved by the Animal Ethics Committee of the Murdoch Children s Research Institute, Melbourne, Australia in accordance with National Health and Medical Research Committee guidelines. Experimental instrumentation. Surfactant-deficient Border-Leicester/Suffolk day preterm lambs (term 145 days; n per group) born to date-mated ewes who received 11.7 mg of betamethasone 24 and 48 h before delivery were delivered via Caesarean section under general anesthesia and instrumented as described in detail previously (33, 34, 38, 40). Twin pregnancies were preferred to minimize maternal variability. After we exposed the fetal head and neck, the carotid artery and external jugular vein were cannulated, an ultrasonic flow probe placed around the contralateral carotid artery, airway intubated with a 4.0 cuffed endotracheal tube (ETT), and lung fluid passively drained. Electrical impedance tomography (EIT) (Goe-MF II EIT system; CareFusion, Hoechberg, Germany) needles were then positioned around the chest (20, 33, 34, 37, 39, 40). At delivery, the lambs were weighed, placed supine, and commenced on intravenous infusions of ketamine and midazolam. Measurements. Peripheral oxygen saturation (SpO 2), heart rate (HR), arterial blood pressure (HP48S monitor, Hewlett Packard, Andover, MA), carotid blood flow (TS420 Perivascular Flow Module; Transonic Systems, Ithaca, NY), airway pressure (P AO), gas flow, and tidal volume (V T) at the airway opening (Florian; Acutronic Medical Systems, Hirzel, Switzerland) were measured continuously from birth. Global and regional lung volume changes were acquired by EIT at 25 scans/s (24, 33, 34, 39), and the unfiltered global lung volume change was displayed in real-time using the Thorascan software package (CareFusion). Arterial blood gases were measured at 5 min of life and every 15 min from birth. The respiratory system resistance (R RS) and reactance (X RS) were computed using the forced oscillation technique (FOT) (4) from the pressure and flow signals measured at the inlet of the tracheal tube during a small-amplitude (5 cmh 2O) 5-Hz oscillatory pressure superimposed onto the ventilation waveform for 10 s. These were conducted on completion of the ventilation strategy at birth and immediately after every arterial blood gas sample (5, 6, 41). Ventilation strategies. Lambs were randomly assigned before delivery to receive one of the following strategies at birth using a L33 computer-generated system that ensured equal matching of group permutations and no repetition of the same strategy within twined ewes. 1) No intentional recruitment maneuver (No-RM) group: positive-pressure ventilation (PPV; SLE5000; SLE, South Croydon, UK) in volume-targeted ventilation (VTV) mode at PEEP 8 cmh 2O, inspiratory time 0.4 s, rate 60 inflations/min and set V T of 7 ml/kg [inspiratory pressure (PIP) 40 cmh 2O] (39). 2) Optimized SI (SI OPT) group: an optimal aeration SI at 40 cmh 2O individualized to each lamb, defined as 10 s after aeration plateau visually determined by two investigators (CEZ, DT) in the global EIT volume signal on the Thorascan display (34). The SI was administered with a Neopuff Infant T-Piece Resuscitator (Fisher & Paykel Healthcare, Auckland, New Zealand) at 8 l/min flow. On completion of the SI, the lung was held at a PEEP of 8 cmh 2O for 5 s before the clamping of the ETT and transferring the lamb to the SLE5000 ventilator and PPV VTV per No-RM (34, 39). 3) OLV group: step-wise PEEP strategy (OLV) using PPV VTV (V T 7 ml/kg) (33, 39). PEEP was increased by 2 cmh 2O every 10 inflations from an initial PEEP of 6 cmh 2O until 20 cmh 2O (PEEP MAX) and then similarly decreased to 6 cmh 2O. PPV VTV was then commenced at PEEP 8 cmh 2O after a transient 10-inflation rerecruitment at PEEP 20 cmh 2O ( s total duration) followed by PPV VTV per No-RM. Ventilation strategy and general management after birth. Additional 10-s FOT measurements were made at PEEP MAX (OLV), on completion of the SI OPT and OLV, and 90 s and 180 s (No-RM and SI OPT). All lambs were initially supported in 0.21 fraction-inspired oxygen (FI O2 ). PPV VTV was applied for 60 min using a standardized strategy, including titration of FI O2 and VTV to maintain SpO % and Pa CO mmhg after the first arterial gas. At the end of the study period, all lambs received a lethal dose of pentobarbitone. The ETT was then disconnected to atmosphere until lung collapse. A static in vivo super-syringe pressure-volume curve was generated (maximum pressure 40 cmh 2O) to determine the static mechanical properties of the respiratory system and calibrate the EIT signal. An additional (1, 34) 10 age-matched fetuses (unventilated controls, UVC group) were euthanized at delivery for injury analysis comparison. Data acquisition and analysis. Physiological parameters were recorded at 1,000 Hz (LabChart V7; AD Instruments, Sydney, Australia) and analyzed at key time points (34). Together with these data, EIT data were monitored in real-time and recorded (Thorascan, Carefusion) continuously for the first 15 min, and subsequently for 2 min with each arterial blood gas. EIT signals were also recorded during the static in vivo PV curve. PIP, PEEP, and inspiratory change in pressure ( P) were determined from the P AO data, and dynamic compliance (C dyn) was calculated from the P and V T data. X RS (indicator of lung recruitment) and R RS were computed from FOT recordings (34, 42). Static respiratory system compliance (C RS) was determined from the PV curve. The alveolar-arterial difference in oxygen (AaDO 2) was calculated from the arterial blood gases. Time-course EIT image data were reconstructed using an anatomically correct custom-built GREIT algorithm (2) with non-lung regions excluded during postprocessing and filtering to the respiratory domain, as we have described previously (39). The trough values of the EIT signal were used to determine global and regional change in end-expiratory volume (aeration; EEV) and the trough to peak amplitudes V T. The global EEV was calibrated (ml/kg) to the change in impedance during the static PV curve (34, 39). Functional EIT images of aeration throughout the imaged cross-sectional slice were also computed to determine the spatiotemporal patterns of aeration (39). The fraction of V T within each gravity-dependent third of the lung weighted to equal lung pixel area was determined (39) and calibrated to ml/kg from the corresponding Florian V T value. Lung injury analysis. Immediately after euthanasia, the lung was removed. The right upper lobe was fixed at 20 cmh 2O with 4% paraformaldehyde and five hematoxylin and eosin-stained 5- m sections from each of the upper, middle, and lower gravity-dependent

3 L34 sections scored for lung injury (n 15 total/lamb) on the following criteria: 1) alveolar wall thickness, 2) detached epithelial cells, 3) hyaline membranes, and 4) alveolar collapse/atelectasis by an investigator blinded to treatment allocation (34, 39). Bronchoalveolar lavage with saline was performed on the left lung and total protein concentration determined using the Lowry method (17). Lung tissue samples were collected from the gravity-dependent (lower) and nondependent (upper) zones of the right lower lobe and immediately snap frozen in liquid nitrogen. RNA was extracted from lung tissue using TRIzol, and 0.1 kg RNA was reverse-transcribed into complementary DNA. Primers of early biomarkers of lung injury [connective tissue growth factor (CTGF), cysteine-rich 61 (CYR61), early growth response protein 1 (EGR1), and the interleukins-1b, 6, and 8] were designed using the Roche Universal ProbeLibrary Assay Design Center. All reactions were performed in triplicate on the Light-Cycler 480 System (Roche, Mannheim, Germany). The 2 Ct method was used to calculate relative changes in gene expression, determined from qrt-pcr experiments using GAPDH as a housekeeping gene and relative to the UVC group (22). Statistical analysis. On the basis of our previous studies with this model (34), a sample size of 12 lambs per group would detect a clinically meaningful difference (SD) in C dyn of 0.08 (0.067) ml kg 1 cmh 2O 1 at 60 min (power of 0.8 and error 0.05). Data were first tested for normality and analyzed with t-tests, one-way ANOVA, Kruskal-Wallis test, or two-way repeated-measures ANOVA (using time and ventilation strategy as factors) and Tukey s, Dunn s, or Dunnett s (against UVC) posttests as appropriate. Statistical analysis was performed with GraphPad PRISM 6 (GraphPad Software, San Diego, CA), and P 0.05 was considered significant. RESULTS Forty-eight lambs were studied. The groups were well matched for weight, gestation, drained lung fluid, and fetal wellbeing (Table 1). Four lambs were excluded because of severe fetal hypoxia, two in the fetal UVC group, one attributable to unrecognized esophageal intubation (No-RM), and one severe intrauterine growth restriction (OLV). Four lambs developed pneumothoraces after inflation to 40 cmh 2 O during the static PV curve, three in the No-RM and one in the SI OPT groups. Initial aeration during respiratory transition. The time needed to achieve plateau aeration during the SI was a median (range) 72 (36 132) s and resulted in a mean (SD) 32 (22) ml/kg EEV from birth (Fig. 1A). This was comparable to the EEV of 38 (15) ml/kg at PEEP MAX (90 s, OLV), mean (95% CI) difference 6 ( 13, 24) ml/kg, Tukey s posttest (2-way repeated-measure ANOVA). The OLV maneuver resulted in hysteresis with a similar EEV at completion (PEEP 8 cmh 2 O) and PEEP MAX, and both greater than the EEV of 19 (13) ml/kg immediately before the OLV (10-s life) (P , 1-way ANOVA). The EEV at the end of the OLV maneuver was similar to the SI OPT group at 180 s. In contrast, the No-RM group demonstrated less time-based recruitment although this was still significant from birth, increasing from 10 (6) ml/kg at 10 s to 17 (10) ml/kg by 180 s (P 0.004, 1-way ANOVA). EEV was significantly higher in the OLV compared with No-RM group from 90 s onward (P 0.033, 1-way ANOVA). SI OPT and OLV had similar regional aeration patterns on completion of the recruitment maneuvers (Fig. 1, B and C), with lower relative aeration in the most gravity-dependent third of the chest compared with the least gravity-dependent and middle regions (P and P , respectively, 1-way ANOVA). The No-RM group had more heterogeneous aeration with both the most and least gravity-dependent third of the chest having lower relative aeration at 90 s and 180 s (both P ). Within each region, the OLV group had relatively greater aeration in the least dependent third at 90 s/peep MAX and 180 s compared with No-RM group (P and P , respectively). Although not statistically significant, the SI OPT group had the most homogenous aeration at 180 s of life. There were no differences within each region of interest over time for all strategies. Lung mechanics. No-RM and OLV groups showed timedependent improvement in C dyn immediately after birth (Fig. 2A). C dyn was higher in the No-RM group at 90 s (P ) compared with 10 s. At the end of the SI, C dyn was similar to that of the OLV group at the same time point (PEEP MAX ) and greater than No-RM group (P 0.16, Tukey s posttest). At 180 s, C dyn was lower in the SI OPT and No-RM groups than the OLV group (end of RM) (P and P , respectively). Thereafter, SI OPT and OLV groups behaved similarly and with better C dyn than No-RM although this was only significant at 5 min of life (P , OLV vs. No-RM; Tukey s posttest). This translated to better PV curves at the end of the study (Fig. 2B). No-RM and OLV groups showed improvement in X RS from 10 s and PEEP MAX (OLV) (both P , 1-way ANOVA) (Fig. 2C). X RS did not change over time in the SI OPT group, but the post-si X RS was higher than the X RS at 10 s for No-RM and OLV groups. X RS was higher at the end of the OLV RM compared with No-RM group at 180 s (P 0.025, 2-way repeated-measures ANOVA, Tukey s posttest) but not SI OPT. By 5 min of life, there was no difference among the groups. Table 1. Study group characteristics No-RM SIOPT OLV Fetal UVC Number Gestational age, days (1.0) (0.8) (0.8) (0.7) Female, n (%) 5 (42%) 7 (58%) 7 (54%) 6 (60%) Singleton, n (%) 1 (8%) 0 (0%) 0 (0%) 1 (10%) First born, n (%) 7 (64%) 5 (42%) 5 (38%) 3 (30%) Birth weight, g 3,169 (519) 3,042 (420) 3,120 (724) 2,765 (432) Fetal lung fluid, m/kg 15.9 (8.6) 16.5 (10.0) 20.1 (5.4) N/A Arterial cord ph 7.33 (0.04) 7.33 (0.06) 7.36 (0.06) 7.20 (0.19) Arterial cord Pa o2, mmhg 22.0 (5.1) 22.7 (3.7) 22.3 (6.2) 21.4 (14.7) Static C RS, ml kg 1 cmh 2O (0.26) 1.20 (0.35) 1.16 (0.39) N/A All applicable data are means (SD). No differences between groups (1-way ANOVA or 2 test as appropriate). No-RM, no recruitment maneuver; SI OPT, optimized sustained inflation; OLV, open lung ventilation; UVC, unventilated controls; C RS, respiratory system compliance.

4 Fig. 1. Change in global end-expiratory volume ( EEV) from birth (ml/kg) (A) following the no recruitment maneuver (No-RM) (black circles), open lung ventilation (OLV) (gray circles), and optimized sustained inflation (SI OPT) (open diamonds) strategy. *OLV vs. No-RM; P way repeatedmeasures ANOVA with Tukey s posttest. OLV vs. 10-s life EEV, No-RM and OLV vs. 10-s life EEV; P way ANOVA with Tukey s posttest. B: relative regional aeration at 90 s, end of SI or maximum positive endexpiratory pressure (PEEP MAX) in the least gravity-dependent (black), middle (gray), and most gravity-dependent (white) third of the cross section of the thorax, expressed as a percentage of total aeration. C: relative regional aeration at 180 s or end of OLV maneuver. *P 0.05, **P 0.01, ***P 0.001; 1-way ANOVA. All data are means SD. L35 Overall, R RS behaved similarly between all groups (Fig. 2D). R RS was significantly lower at PEEP MAX and at the end of OLV compared with 10 s (P 0005, repeated-measures 1-way ANOVA). Ventilator parameters. All groups demonstrated a reduction in P requirements with time (all groups P , RM 1-way ANOVA), with OLV having the greatest decrease in P during early respiratory transition (Fig. 3A). P was higher in the No-RM group throughout the study, reaching significance against OLV, but not SI OPT, from 180 s (P 0.001, Tukey s posttest) to 10 min of life (P 0.049). No strategy was within target V T range before 180 s, with No-RM and OLV groups being below range and SI OPT group above range (Fig. 3B). Measured V T was lower in the No-RM group at 10 s compared with OLV group (P ). Both No-RM and OLV groups had a lower V T at 90 s compared with the SI OPT post-si (P ). Thereafter, V T was the same in all groups. Figure 4 shows V T within the middle, least, and most gravity-dependent thirds of the lung. At 5 min, the OLV and SI OPT groups had higher V T in the least dependent regions compared with middle (both) and most dependent (SI OPT )(P and P 0.011, respectively, 1-way ANOVA). By 60 min, all groups had significantly higher V T in the least dependent third of the lung compared with all other regions [P (No-RM), P (SI OPT ), P (OLV)]. Only the No-RM groups demonstrated increased spatiotemporal redistribution between 5 and 60 min (Fig. 4C), with significant changes in V T toward the least dependent lung (P ). Oxygenation. OLV and SI OPT attained target SpO 2 (in FI O2 0.21) by 180 s and 5 min, respectively (Fig. 3C), compared with 15 min for the No-RM group [P against OLV (90 s to 5 min) and SI OPT (3 and 5 min), Tukey s posttest]. Required FI O2 was similar for SI OPT and OLV (Fig. 3D) and lower than No-RM. By 60 min, FI O2 was a mean (95% CI) 0.26 (0.03, 0.42) higher than OLV in the No-RM group (Tukey s posttest). AaDO 2 was higher in the No-RM group (Fig. 3E) from 30 min (P 0.014, OLV) and 45 min (P 0.036, SI OPT ). All groups could be ventilated within the target Pa CO2 range (Fig. 3F) although the initial 5-min Pa CO2 was higher in the No-RM group (P , OLV and P , SI OPT ). EEV. OLV and SI OPT had similar EEV beyond the initial period of respiratory transition (Fig. 1A). The No-RM group had a lower EEV at all time points of the study, and this was significant compared with OLV [with mean (95% CI) differences ranging between 19 (1, 38) ml/kg (30 min) and 20 (1, 39) ml/kg (10 min), Tukey s posttest]. All groups showed redistribution in aeration from the most dependent lung regions to the middle regions over the duration of the study (Fig. 5). This was most apparent in the No-RM group. Both RM groups behaved similarly. Lung histology. All strategies had similar total lung injury scores on histology, and all were greater than UVC overall (P , 1-way ANOVA). Regionally there was no difference in lung injury between all groups in the middle regions (Fig. 6A). Only SI OPT had higher injury in the most gravity-dependent lung regions compared with UVC (P , Tukey s posttest), and all strategies demonstrated significantly more injury in the least gravity-dependent (upper) region compared with UVC (P , No-RM and SI OPT ; P 0.07, OLV). Within strategies, regional injury distribution differed. The No-RM and SI OPT strategies had higher scores in

5 L36 Fig. 2. Dynamic compliance (C dyn) (A), static pressure-volume curves (B), respiratory system reactance (X RS) (C) and respiratory system resistance (R RS) (D). Symbols per Fig. 1A. P 0.05 *SI OPT vs. No-RM (90 s), OLV vs. No-RM and SI OPT, OLV vs. No-RM; 2-way repeatedmeasures ANOVA and **OLV and No- RM, OLV vs. 10 s or SI OPT post-si vs. OLV and No-RM at 10 s; 1-way ANOVA. All data are means SD. the least gravity-dependent regions compared with the most (No-RM, P 0.030) and middle (SI OPT, P 0.011). Injury scores were higher in the middle compared with least dependent regions in the OLV strategy (P 0.001). There was no difference in total lung protein (Fig. 6B). Molecular evidence of injury. Gene expression of markers of lung injury (Fig. 6C) was increased in all interventional groups when compared with UVC in the nondependent lung (all markers P 0.004, Kruskal-Wallis test) although SI OPT (CTGF) did not differ on subgroup analysis (P respectively, Dunnett s posttest). In the dependent lung, all strategies exhibited increased IL1, IL6, IL8, CYR61, and EGR1 (P ) gene expression compared with the UVC group but not CTGF (P 0.90). DISCUSSION This is the first study to compare the mechanical and injury responses after two different active lung recruitment maneuvers individualized to the mechanical response of the lung and designed to optimize aeration at birth. One strategy focused on rapid lung liquid clearance (SI OPT ) and the other on preventing fluid efflux via PEEP and gradual tidal aeration (OLV). Both strategies produced similar benefits over a control group using tidal ventilation with no active recruitment maneuver in steroid-exposed preterm lambs that would represent clinically meaningful short-term differences. This has implications for clinical practice, as optimal approaches to PEEP and SI have yet to be determined in human infants; our study suggests that both could be equally effective. Birth involves the rapid transition from a fluid-filled to aerated lung state, a process that is essential for successful respiratory function (14). Moving lung fluid from the airways and into the interstitium can only happen when a driving pressure gradient is applied to the lung (12). Thereafter, sufficient end-expiratory pressure is needed to prevent fluid efflux back into the alveolar spaces during tidal ventilation (12, 13). This mechanical explanation of aeration at birth emphasizes the need for sufficient applied pressure. The SI OPT and OLV strategies we employed both aimed to optimize aeration at birth but focused on pressure during different components of the respiratory transition, the initial inflation pressure (SI OPT ) and dynamic PEEP during tidal inflations. The median SI OPT duration was similar to the time to OLV PEEP MAX, and both achieved similar absolute lung volumes that were greater than No-RM. The improved aeration was associated with beneficial mechanical changes in the lung. Thus it is not unexpected that the subsequent clinical outcomes were similar. How best to support the preterm lung at birth remains unknown (18). SIs have been extensively investigated (11, 16, 24, 28, 33, 34, 37 40), with large clinical trials ongoing (7), but investigation of the role of PEEP has been limited (32). Our study reiterates the need for active maneuvers to facilitate lung aeration at birth, but, unlike previous studies, ours emphasizes that more than one effective option exists. Whichever maneuver is chosen, it is critical that it is applied optimally. A striking finding of our study was the high variability in the aeration duration and subsequent volumes achieved irrespective of strategy; between 36 and 132 s was needed to optimize the SI response. We have previously demonstrated that lung aeration at birth depends on both the strategy employed and the mechanical properties of the recipient s lung. For this reason, our sample size was considerably greater than previous physiological studies (24, 28, 30, 31, 33, 40). Most of these studies have used predefined SI durations of 60 s (11, 24, 28, 33, 38 40) and/or static PEEP (27, 31). Our study reinforces that such an approach to both SI and PEEP is very unlikely to create

6 L37 Fig. 3. Change in pressure ( P) (A), tidal volume (V T)(B), SpO 2 (C), FI O2 (D), alveolar-arterial difference in oxygen (AaDO 2) (E) and Pa CO2 (F). Symbols per Fig. 1A. P 0.05 *SI OPT vs. No-RM, OLV vs. No-RM, No-RM vs. OLV and SI OPT (2-way repeated-measures ANOVA), SI OPT post SI vs. OLV and No-RM 90 s (1-way ANOVA). All data are means SD. uniform outcomes across an intervention group. This may explain the conflicting and inconclusive outcomes of previous human (26) and animal studies (11, 24, 28, 30, 31, 38, 39). Our dynamic stepwise PEEP strategy is based on the OLV concept (33). This patient-defined approach aims to place ventilation on the deflation limb of the pressure-volume relationship, the region known to optimize lung mechanics in the already aerated lung (9, 35, 36). Our results suggest that future clinical studies should also focus on providing meaningful physiological feedback mechanisms at the bedside. In this context, EIT offers promise as a direct measure of aeration that is robust in the presence of poor face mask leak and can demonstrate upper airway obstruction (8). The EIT data showed that the benefits seen with both SI OPT and OLV were due to improved distal lung aeration, particularly within the least gravity-dependent lung. In contrast, aeration was mainly limited to regions associated with the major airways and adjacent alveoli (middle third of chest) in the Fig. 4. Regional V T (ml/kg) in the least-gravity dependent (black), middle (gray), and most-gravity dependent (white) third of the cross section of the thorax at 5 min (A) and 60 min (B). C: change in V T in each region between 5 and 60 min. *P 0.05, **P 0.01, ***P 0.001; 1-way ANOVA. All data are means SD.

7 L38 Fig. 5. Functional electrical impedance tomography (feit) graphs of the regional gravity-dependent distribution of relative aeration (end-expiratory volume, EEV) within 22 nondependent (top) to dependent (bottom) slices of the thorax immediately after the recruitment maneuver (SI OPT and OLV) or at 180 s of life (No-RM) and 60 min (right) following No-RM, SI OPT, and OLV strategies. feit images divided into the least-gravity dependent (dotted), middle (nonfilled) and most-gravity dependent (dashed) thirds shown in Fig. 1B and C. All bars are means SD. There was no spatial difference between 2 time points within each group. No-RM group. Despite this, both SI OPT and OLV failed to achieve truly uniform aeration in our surfactant-deficient lambs, mainly attributable to poorer recruitment of the most gravity-dependent lung. This is not unexpected, as these lung regions are the hardest to recruit in poorly compliant diseased lungs (3, 10, 21). EIT measures relative aeration and ventilation differences, so it cannot determine whether these lung regions were anatomically atelectatic. Interesting, across all groups, lung injury markers were greater in those regions easiest to engage in ventilation and aeration. The two recruitment strategies resulted in the greatest heterogeneity of V T by 5 min. This suggests that volutrauma is an important component of preterm lung injury in early life, potentially more so than atelectasis, and that there may be a risk from excessive recruitment. That both recruitment approaches already had evidence of significant lung injury by 60 min and failed to demonstrate any meaningful benefit in injury over No-RM despite more uniform aeration is intriguing and not simply explained by the potential effects of antenatal steroids. This finding challenges the hypothesis that fluid/aeration inhomogeneity in the lung at birth increases injury (13). It is possible that protecting lung regions that are very poorly compliant, and thus hardest to aerate, from the initial high driving pressures placed on the lung during aeration may be beneficial (37). The use of VTV with a threshold maximum PIP allows this to be achieved, as the volume exposure is primarily defined by lung mechanics. The V T during the first 3 min of life in the two PPV groups was much less than the intended 7 ml/kg, and changes then mirrored the temporal increases in C dyn, R RS, and X RS as more Fig. 6. Hematoxylin and eosin lung injury score in the most dependent (open), middle (shaded), and least (solid) dependent regions of lung (A). P 0.05, *between regions, against all other strategies, against SI OPT (1-way ANOVA). B: total bronchoalveolar fluid protein by strategy. C: IL1B, IL6, IL8, CTGF, CYR61, and EGR1 gene expression in the most and least gravity-dependent lung regions. P 0.05, *between regions, against all other strategies (Kruskal-Wallis test with Dunn s posttest), against unventilated controls (UVC) (Kruskal-Wallis test with Dunnett s posttest). All bars are means SD.

8 lung units engaged in tidal ventilation. The simultaneous assessment of breath-to-breath regional aeration, ventilation, and FOT-derived mechanics, unique to our study, allows the lung behavior at birth to be defined both spatially and temporally. This approach clearly demonstrates that focusing on immediate and uniform aeration is too simplistic a model of lung protection in the preterm lung at birth. A lack of consistency in animal models, SI strategies, and PEEP levels has hampered interpretation of preclinical studies (including our own) of respiratory transition. We have previously found that a PEEP of 8 cmh 2 O optimized aeration in our lambs, a finding supported by data from other groups (25). However, we acknowledge that this is higher than in other studies (11, 28). We contend that the focus should be on the intent of the strategy. The SI OPT focused on achieving aeration through a prolonged but rapid inspiratory driving pressure upon the highly resistive fluid-filled lung. In contrast, the OLV uses a slower approach with a higher transient PEEP applied to overcome the viscoelastic forces, allowing fluid efflux and alveoli collapse during expiration between ongoing tidal fluid clearance and recruitment via exploiting hysteresis. This is based on the observation that recruitment is an ongoing process in the diseased lung and differentiates the OLV from both the No-RM and SI OPT strategies (15). OLV strategies are used in atelectatic lung conditions (19) but outside our group have not been reported at birth. On the basis of observations in 40-g rabbit pups (30), it was recently proposed that the flow dynamics and tissue mechanics of the lung make it difficult to ventilate at birth using high pressures and short inspiratory times (tidal ventilation) without overinflating and injuring aerated lung regions (13). Lung injury was not reported in these studies (30, 31). Our study, using a larger sample size and lambs with different chest mechanics, refutes this hypothesis, as similar regional aeration and injury characteristics were reported using both tidal ventilation and an SI, which all resulted in similar reductions in R RS. This suggests that preventing fluid efflux over time and understanding of the dynamic mechanical responses of the lung may be more important than how the airways are cleared of fluid. In our steroid-exposed lambs without exogenous surfactant therapy, the initial benefits of EEV were universally lost by 60 min, associated with a gradual redistribution of aeration toward the least dependent lung (39) and were not prevented using antenatal steroids. This complicates the interpretation of the regional injury results. It is possible that the complex heterogeneous injury pattern at study completion was a response to one, or both, of either the initial aeration event or the subsequent spatiotemporal changes. Spatiotemporal changes are most evident in the regional V T pattern of the No-RM group. We have now observed this in both the steroid- and nonsteroidexposed surfactant-deficient preterm lamb lung by 60 min of life (39). The spatiotemporal redistribution of aeration after birth is likely to be a major contributor to early development of lung injury but is poorly understood. Clearly long-term evaluation is needed, as the spatiotemporal behavior beyond 60 min of life may differ again, perpetuating new injury patterns. We postulate that the beneficial effects of surfactant relate to stabilizing spatiotemporal injury pathways (29, 40), and our study suggests that there may be an important clinical interaction between recruitment strategies at birth and surfactant efficacy. The breath-to-breath imaging by EIT, coupled with L39 newer methods of injury evaluation with proteomics (23), offers potential to detail injury mechanisms at a high temporal resolution. This study has some other limitations. Similar to other translational studies in this field (11, 12, 24, 25, 27, 28), the lambs were anaesthetized, intubated, and managed with a cuffed ETT, inconsistent with clinical practice and likely to result in less variable aeration than in humans managed with a face mask, but necessary to ensure the accuracy of the mechanical measures. In spontaneously breathing human infants, the variability we observed is likely to be greater still, and future work should consider developing spontaneously breathing animal models or using our EIT methods in the delivery room (8). Most previous animal studies of preterm ventilation at birth (18, 24, 25, 28, 30, 31, 33, 37, 39, 40) have not exposed fetal lambs to corticosteroids, which significantly alters the mechanical benefits of the ventilation strategies (38). It is for this reason that we chose to expose lambs to steroids before birth, but we did not use surfactant after birth, despite also having proven benefit (29). Our sample size was almost double that of most similar studies, but injury outcomes were still underpowered. Future studies should be larger still if lung injury is a primary outcome. We focused on alveolar injury only, but Hillman and coworkers (11) have recently shown that ventilation strategy at birth also influences airway injury. Furthermore, lung inflammation was limited to mrna expression in fetuses without overt infection. Immunohistochemistry and examination of other pathways, such as surfactant catabolism and fibrosis, may have yielded different patterns (11) and may not translate to fetuses exposed to chorioamnionitis. Finally, the limitations of EIT are well described elsewhere (8). Specifically, EIT is only able to measure regional volume characteristics through a single cross-sectional plane, and assumptions regarding whole lung patterns should not be made. In the preterm lamb, similar outcomes can be achieved using an SI and dynamic PEEP recruitment maneuver (OLV) as long as both are applied using a strategy that accounts for the individual variability of the mechanical state of the preterm lung at birth. We found that similar aeration and injury outcomes can be achieved using tidal ventilation alone as long as a dynamic PEEP approach was used, suggesting that preventing fluid efflux after lung liquid clearance may be the critical step in successful respiratory transition at birth. ACKNOWLEDGMENTS The authors acknowledge Georgina Huan for preparing and scoring the hematoxylin and eosin data, Dan Pavlic for assisting in PCR analysis, and Sarah White and Rebecca Sutton for assistance in preparation of the ewes. GRANTS This study is supported by a National Health and Medical Research Council Project Grant (Grant ID ) and the Victorian Government Operational Infrastructure Support Program (Melbourne, Australia). D. Tingay is supported by a National Health and Medical Research Council Clinical Career Development Fellowship (Grant ID ). P. Davis is supported by a National Health and Medical Research Council Program Grant (Grant ID ). P. Davis is supported by a National Health and Medical Research Council Practitioner Fellowship (Grant ID ). DISCLOSURES Politecnico di Milano University, the institution of Emanuela Zannin and Raffaele L. Dellaca, owns a patent on the use of forced oscillation technique

9 L40 for the detection of lung volume recruitment/derecruitment. The other authors have no competing interests to declare. AUTHOR CONTRIBUTIONS D.G.T., A.E.R., E.Z., P.M.P.-F., E.P., C.E.E.Z., D.B., A.L., and M.S. performed experiments; D.G.T., A.E.R., E.Z., P.M.P.-F., R.D., E.P., C.E.E.Z., D.B., and B.G. analyzed data; D.G.T., A.E.R., E.Z., P.M.P.-F., R.D., C.E.E.Z., A.A., I.F., and P.G.D. interpreted results of experiments; D.G.T., A.E.R., and P.M.P.-F. prepared figures; D.G.T. and A.E.R. drafted manuscript; D.G.T., E.Z., R.D., E.P., C.E.E.Z., A.A., I.F., and P.G.D. edited and revised manuscript; D.G.T., A.E.R., E.Z., P.M.P.-F., R.D., E.P., C.E.E.Z., A.A., I.F., A.L., M.S., B.G., F.M., and P.G.D. approved final version of manuscript. REFERENCES 1. Adler A, Amyot R, Guardo R, Bates JH, Berthiaume Y. Monitoring changes in lung air and liquid volumes with electrical impedance tomography. J Appl Physiol 83: , Adler A, Arnold JH, Bayford R, Borsic A, Brown B, Dixon P, Faes TJ, Frerichs I, Gagnon H, Garber Y, Grychtol B, Hahn G, Lionheart WR, Malik A, Patterson RP, Stocks J, Tizzard A, Weiler N, Wolf GK. GREIT: a unified approach to 2D linear EIT reconstruction of lung images. Physiol Meas 30: S35 S55, Dargaville PA, Rimensberger PC, Frerichs I. Regional tidal ventilation and compliance during a stepwise vital capacity manoeuvre. Intensive Care Med 36: , Dellaca RL, Andersson Olerud M, Zannin E, Kostic P, Pompilio PP, Hedenstierna G, Pedotti A, Frykholm P. Lung recruitment assessed by total respiratory system input reactance. Intensive Care Med 35: , Dellaca RL, Zannin E, Kostic P, Olerud MA, Pompilio PP, Hedenstierna G, Pedotti A, Frykholm P. Optimisation of positive end-expiratory pressure by forced oscillation technique in a lavage model of acute lung injury. Intensive Care Med 37: , Dellaca RL, Zannin E, Ventura ML, Sancini G, Pedotti A, Tagliabue P, Miserocchi G. Assessment of dynamic mechanical properties of the respiratory system during high-frequency oscillatory ventilation. Crit Care Med 41: , Foglia EE, Owen LS, Thio M, Ratcliffe SJ, Lista G, Te Pas A, Hummler H, Nadkarni V, Ades A, Posencheg M, Keszler M, Davis P, Kirpalani H. Sustained Aeration of Infant Lungs (SAIL) trial: study protocol for a randomized controlled trial. Trials 16: 95, Frerichs I, Amato MB, van Kaam AH, Tingay DG, Zhao Z, Grychtol B, Bodenstein M, Gagnon H, Bohm SH, Teschner E, Stenqvist O, Mauri T, Torsani V, Camporota L, Schibler A, Wolf GK, Gommers D, Leonhardt S, Adler A. Chest electrical impedance tomography examination, data analysis, terminology, clinical use and recommendations: Consensus statement of the TRanslational EIT development study group. Thorax Sep 5. pii: thoraxjnl doi: /thoraxjnl [Epub ahead of print] 9. Froese AB. The incremental application of lung-protective high-frequency oscillatory ventilation. Am J Respir Crit Care Med 166: , Gattinoni L, Pesenti A, Bombino M, Baglioni S, Rivolta M, Rossi F, Rossi G, Fumagalli R, Marcolin R, Mascheroni D. Relationships between lung computed tomographic density, gas exchange, and PEEP in acute respiratory failure. Anesthesiology 69: , Hillman NH, Kemp MW, Noble PB, Kallapur SG, Jobe AH. Sustained inflation at birth did not protect preterm fetal sheep from lung injury. Am J Physiol Lung Cell Mol Physiol 305: L446 L453, Hooper SB, Kitchen MJ, Wallace MJ, Yagi N, Uesugi K, Morgan MJ, Hall C, Siu KK, Williams IM, Siew M, Irvine SC, Pavlov K, Lewis RA. Imaging lung aeration and lung liquid clearance at birth. FASEB J 21: , Hooper SB, Te Pas AB, Kitchen MJ. Respiratory transition in the newborn: a three-phase process. Arch Dis Child Fetal Neonatal Ed 101: F266 F271, Jobe AH, Hillman N, Polglase G, Kramer BW, Kallapur S, Pillow J. Injury and inflammation from resuscitation of the preterm infant. Neonatology 94: , Lachmann B. Open up the lung and keep the lung open. Intensive Care Med 18: , Lista G, Boni L, Scopesi F, Mosca F, Trevisanuto D, Messner H, Vento G, Magaldi R, Del Vecchio A, Agosti M, Gizzi C, Sandri F, Biban P, Bellettato M, Gazzolo D, Boldrini A, Dani C; SLI Trial Investigators. Sustained lung inflation at birth for preterm infants: a randomized clinical trial. Pediatrics 135: e457 e464, Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Protein measurement with the Folin phenol reagent. J Biol Chem 193: , McCall KE, Davis PG, Owen LS, Tingay DG. Sustained lung inflation at birth: What do we know, and what do we need to know? Arch Dis Child Fetal Neonatal Ed 101: F175 F180, Miedema M, de Jongh FH, Frerichs I, van Veenendaal MB, van Kaam AH. Changes in lung volume and ventilation during lung recruitment in high-frequency ventilated preterm infants with respiratory distress syndrome. J Pediatr 159: ; e192, Milesi I, Tingay DG, Zannin E, Bianco F, Tagliabue P, Mosca F, Lavizzari A, Ventura ML, Zonneveld CE, Perkins EJ, Black D, Sourial M, Dellaca RL. Intratracheal atomized surfactant provides similar outcomes as bolus surfactant in preterm lambs with respiratory distress syndrome. Pediatr Res 80: , Pellicano A, Tingay DG, Mills JF, Fasulakis S, Morley CJ, Dargaville PA. Comparison of four methods of lung volume recruitment during high frequency oscillatory ventilation. Intensive Care Med 35: , Pereira-Fantini PM, Rajapaksa AE, Oakley R, Tingay DG. Selection of reference genes for gene expression studies related to lung injury in a preterm lamb model. Sci Rep 6: 26476, Pereira-Fantini PM, Tingay DG. The proteomics of lung injury in childhood: challenges and opportunities. Clin Proteomics 13: 5, Polglase GR, Tingay DG, Bhatia R, Berry CA, Kopotic RJ, Kopotic CP, Song Y, Szyld E, Jobe AH, Pillow JJ. Pressure- versus volumelimited sustained inflations at resuscitation of premature newborn lambs. BMC Pediatr 14: 43, Probyn ME, Hooper SB, Dargaville PA, McCallion N, Crossley K, Harding R, Morley CJ. Positive end expiratory pressure during resuscitation of premature lambs rapidly improves blood gases without adversely affecting arterial pressure. Pediatr Res 56: , Schmolzer GM, Kumar M, Aziz K, Pichler G, O Reilly M, Lista G, Cheung PY. Sustained inflation versus positive pressure ventilation at birth: a systematic review and meta-analysis. Arch Dis Child Fetal Neonatal Ed 100: F361 F368, Siew ML, Te Pas AB, Wallace MJ, Kitchen MJ, Lewis RA, Fouras A, Morley CJ, Davis PG, Yagi N, Uesugi K, Hooper SB. Positive endexpiratory pressure enhances development of a functional residual capacity in preterm rabbits ventilated from birth. J Appl Physiol 106: , Sobotka KS, Hooper SB, Allison BJ, Te Pas AB, Davis PG, Morley CJ, Moss TJ. An initial sustained inflation improves the respiratory and cardiovascular transition at birth in preterm lambs. Pediatr Res 70: 56 60, Soll RF. Prophylactic natural surfactant extract for preventing morbidity and mortality in preterm infants. Cochrane Database Syst Rev 2000: CD000511, te Pas AB, Siew M, Wallace MJ, Kitchen MJ, Fouras A, Lewis RA, Yagi N, Uesugi K, Donath S, Davis PG, Morley CJ, Hooper SB. Effect of sustained inflation length on establishing functional residual capacity at birth in ventilated premature rabbits. Pediatr Res 66: , te Pas AB, Siew M, Wallace MJ, Kitchen MJ, Fouras A, Lewis RA, Yagi N, Uesugi K, Donath S, Davis PG, Morley CJ, Hooper SB. Establishing functional residual capacity at birth: the effect of sustained inflation and positive end-expiratory pressure in a preterm rabbit model. Pediatr Res 65: , te Pas AB, Walther FJ. A randomized, controlled trial of delivery-room respiratory management in very preterm infants. Pediatrics 120: , Tingay DG, Bhatia R, Schmolzer GM, Wallace MJ, Zahra VA, Davis PG. Effect of sustained inflation vs. stepwise PEEP strategy at birth on gas exchange and lung mechanics in preterm lambs. Pediatr Res 75: , Tingay DG, Lavizzari A, Zonneveld CE, Rajapaksa A, Zannin E, Perkins E, Black D, Sourial M, Dellaca RL, Mosca F, Adler A, Grychtol B, Frerichs I, Davis PG. An individualized approach to sustained inflation duration at birth improves outcomes in newborn preterm lambs. Am J Physiol Lung Cell Mol Physiol 309: L1138 L1149, 2015.

10 L Tingay DG, Mills JF, Morley CJ, Pellicano A, Dargaville PA. The deflation limb of the pressure-volume relationship in infants during highfrequency ventilation. Am J Respir Crit Care Med 173: , Tingay DG, Mills JF, Morley CJ, Pellicano A, Dargaville PA. Indicators of optimal lung volume during high-frequency oscillatory ventilation in infants. Crit Care Med 41: , Tingay DG, Polglase GR, Bhatia R, Berry CA, Kopotic RJ, Kopotic CP, Song Y, Szyld E, Jobe AH, Pillow JJ. Pressure-limited sustained inflation vs. gradual tidal inflations for resuscitation in preterm lambs. J Appl Physiol 118: , Tingay DG, Rajapaksa A, McCall K, Zonneveld CE, Black D, Perkins E, Sourial M, Lavizzari A, Davis PG. The interrelationship of recruitment manoeuvre at birth, antenatal steroids and exogenous surfactant on compliance and oxygenation in preterm lambs. Pediatr Res 79: , Tingay DG, Rajapaksa A, Zonneveld CE, Black D, Perkins EJ, Adler A, Grychtol B, Lavizzari A, Frerichs I, Zahra VA, Davis PG. Spatiotemporal aeration and lung injury patterns are influenced by the first inflation strategy at birth. Am J Respir Cell Mol Biol 54: , Tingay DG, Wallace MJ, Bhatia R, Schmolzer GM, Zahra VA, Dolan MJ, Hooper SB, Davis PG. Surfactant before the first inflation at birth improves spatial distribution of ventilation and reduces lung injury in preterm lambs. J Appl Physiol 116: , Zannin E, Dellaca RL, Kostic P, Pompilio PP, Larsson A, Pedotti A, Hedenstierna G, Frykholm P. Optimizing positive end-expiratory pressure by oscillatory mechanics minimizes tidal recruitment and distension: an experimental study in a lavage model of lung injury. Crit Care 16: R217, Zannin E, Ventura ML, Dellaca RL, Natile M, Tagliabue P, Perkins EJ, Sourial M, Bhatia R, Dargaville PA, Tingay DG. Optimal mean airway pressure during high-frequency oscillatory ventilation determined by measurement of respiratory system reactance. Pediatr Res 75: , 2014.