Journal of Surgical Research 166, e71–e81 (2011) doi:10.1016/j.jss.2010.10.022
Comparison of ‘‘Open Lung’’ Modes with Low Tidal Volumes in a Porcine Lung Injury Model Scott Albert, M.D.,* Brian D. Kubiak, M.D.,* Christopher J. Vieau, B.A.,* Shreyas K. Roy, M.D.,*,1 Joseph DiRocco, M.D.,* Louis A. Gatto, Ph.D.,† Jennifer L. Young, Ph.D.,‡ Sudipta Tripathi, Ph.D.,* Girish Trikha, M.D.,§ Carlos Lopez, M.D.,k and Gary F. Nieman, B.A.* *Department of Surgery, Upstate Medical University, Syracuse, New York; †Department of Biological Sciences, SUNY Cortland, Cortland, New York; ‡Department of Pediatrics, University of Rochester, Rochester, New York; §Division of Pulmonary and Critical Care, Upstate Medical University, Syracuse, New York; and kDepartment of Anesthesiology, Upstate Medical University, Syracuse, New York Submitted for publication July 26, 2010
Background. Ventilator strategies that maintain an ‘‘open lung’’ have shown promise in treating hypoxemic patients. We compared three ‘‘open lung’’ strategies with standard of care low tidal volume ventilation and hypothesized that each would diminish physiologic and histopathologic evidence of ventilator induced lung injury (VILI). Materials and Methods. Acute lung injury (ALI) was induced in 22 pigs via 5% Tween and 30-min of injurious ventilation. Animals were separated into four groups: (1) low tidal volume ventilation (LowVt -6 mL/ kg); (2) high-frequency oscillatory ventilation (HFOV); (3) airway pressure release ventilation (APRV); or (4) recruitment and decremental positiveend expiratory pressure (PEEP) titration (RMDOP) and followed for 6 h. Lung and hemodynamic function was assessed on the half-hour. Bronchoalveolar lavage fluid (BALF) was analyzed for cytokines. Lung tissue was harvested for histologic analysis. Results. APRV and HFOV increased PaO2/FiO2 ratio and improved ventilation. APRV reduced BALF TNF-a and IL-8. HFOV caused an increase in airway hemorrhage. RMDOP decreased SvO2, increased PaCO2, with increased inflammation of lung tissue. Conclusion. None of the ‘‘open lung’’ techniques were definitively superior to LowVt with respect to VILI; however, APRV oxygenated and ventilated more effectively and reduced cytokine concentration compared with LowVt with nearly indistinguishable histopathology. These data suggest that APRV may be of potential benefit to critically ill patients but other
1
To whom correspondence and reprint requests should be addressed at Department of Surgery, Upstate Medical University, 750 East Adams Street, Syracuse, NY 13210. E-mail:
[email protected].
‘‘open lung’’ strategies may exacerbate injury.
Ó 2011
Elsevier Inc. All rights reserved.
Key Words: ventilator induced lung injury; porcine; acute lung injury; airway pressure release ventilation; recruitment maneuver; high frequency oscillatory ventilation.
INTRODUCTION
Mechanical ventilation is necessary to sustain lung function in critically ill patients of diverse etiologies, however improper mechanical ventilation may cause ventilator induced lung injury (VILI), a mechanical and inflammatory injury to the lung, and exacerbate the underlying pulmonary pathophysiology. Continued insult and unchecked inflammation of the lung may progress to acute respiratory distress syndrome (ARDS), which carries an extremely high mortality rate. An epidemiologic study by Rubenfeld and Herridge proposed that an estimated 17,000–43,000 lives could be saved per year if VILI were eliminated [1]. The current standard of care for mechanically ventilated patients is low tidal volume ventilation (LowVt) with moderate positive-end expiratory pressure (PEEP) per the ARDSnet protocol. By recommending a tidal volume of 6 mL/kg predicted body weight, this strategy is based on the assumption that low tidal volumes and plateau pressures below 30 cmH2O [2] prevent overdistension of heterogeneously injured alveoli. In clinical trials and animal studies, the low tidal volume approach has met with success in terms of a reduction in markers of inflammation and, most importantly, overall mortality compared with ventilation
e71
0022-4804/$36.00 Ó 2011 Elsevier Inc. All rights reserved.
e72
JOURNAL OF SURGICAL RESEARCH: VOL. 166, NO. 1, MARCH 2011
with high tidal volumes [3, 4]. The ARDSnet clinical trial demonstrated that lung protective approaches improved patient outcomes as shown by the significant reduction in mortality from 40% to 31% with the use of low tidal volume ventilation [2]. Despite the success of ARDSnet, alternative ventilation regimes have been proposed, claiming superior oxygenation, ventilation, and lung protection [5–7]. The term ‘‘open-lung’’ has been coined as an umbrella for strategies that promote and maintain full lung inflation by first increasing distending pressures to ‘‘recruit’’ collapsed units, then maintaining adequate positive end expiratory pressure (PEEP) to prevent alveolar collapse. In theory, open lung modes create a more homogenous distribution of airflow, reducing the stress/strain associated with cyclic alveolar opening and collapse. This study focuses on three open lung strategies: high frequency oscillatory ventilation (HFOV), airway pressure release ventilation (APRV), and a strategy utilizing aggressive recruitment maneuvers (RM) with a decremental titration of PEEP to determine the minimum PEEP necessary to prevent lung collapse that we have designated as the ‘‘optimal PEEP’’ (OP). Small animal models have corroborated theoretical reports by showing HFOV, APRV, and recruitment maneuvers to be beneficial to the injured lung [8–10], however translation to the clinical setting has been limited. Therefore, further investigation is required to elucidate the benefits and disadvantages between different ventilation strategies in clinically relevant models. In this large animal model of acute lung injury (ALI), we hypothesized that each open lung ventilation mode (HFOV, APRV, and RMþOP) would diminish the physiologic and histologic evidence of VILI compared with low tidal volume ventilation, thus demonstrating that they offer superior lung protection. METHODS Vertebrate Animals Experiments described in this study were performed in accordance with the National Institutes of Health guidelines for the use of experimental animals in research. The Committee for the Humane Use of Animals at Upstate Medical University approved the experimental protocol (approval no. 040).
Surgical Preparation Yorkshire pigs (n ¼ 22) were pretreated with glycopyrrolate (0.01 mg/kg, intramuscular), telazol (tiletamine hydrochloride and zolazepam hydrochloride (5 mg/kg, intramuscular), and xylazine (2 mg/kg, intramuscular). A ketamine (3 mg/1mL)/xylazine (0.003/1 mL) continuous infusion via an infusion pump (3M model 3000) was used to maintain anesthesia throughout the experiment. A Galileo ventilator (Hamilton Medical, Reno, NV) was used. Baseline ventilator settings were as follows: conventional volume-controlled ventilation (CMV) with a tidal volume 12 cc/kg; PEEP 3 cmH20; inspiration-to-
expiration ratio 1:2, FiO2 of 21% and respiratory rate 15 breaths/ min. A left carotid artery, left internal jugular, and left external jugular catheter were placed for hemodynamic monitoring, blood sampling, and administration of fluids. Blood gas analysis was performed (model ABL5’ Radiometer Inc., Copenhagen, Denmark) to ensure pulmonary health. A right internal jugular pediatric Swan-Ganz catheter was placed for cardiac output and filling pressures. An open tracheostomy was performed with placement of a 7.5 French cuffed endotracheal tube. ECG monitoring, pulse oximetry, central venous pressure, pulmonary artery pressure, and arterial pressure were continuously monitored (CMS-2001; Agilent, Santa Clara, CA, USA). A Foley catheter was placed through a cystotomy for urine output monitoring.
Fluid Administration Each pig received a 1 L lactated ringers fluid bolus during surgical preparation to maintain end-organ perfusion. A slow infusion 2cc/kg/h was maintained throughout the experiment. Additional fluid boluses were given in response to hypotension (MAP < 65 mmHg).
Lung Injury Pigs were placed in a right decubitus position for instillation of 5% Tween-20 in normal saline (1.5 cc/kg/lung) through a 2 mm i.d. catheter threaded down the endotracheal tube. During instillation, the ventilator settings were changed to pressure controlled ventilation (PCV) with a Pcontrol 25 cm H2O, PEEP 10 cmH2O, and FiO2 1.0 for 10 min. This was repeated with the pig in a left decubitus position. Immediately following Tween instillation, the pig was placed supine for 30 min of injurious ventilation (baseline settings, see above). Acute lung injury caused by Tween instillation has been characterized by previous work in our laboratory [11].
Experimental Groups Pigs were separated into four groups immediately following lung injury as follows. LowVt: (n ¼ 6, 24.4 6 2.7 kg) pigs were placed on volume-controlled ventilation with a Vt of 6 mL/kg, PEEP 14 cmH2O, and FiO2 of 0.8. Settings were adjusted according to the ARDSnet sliding scale FiO2 and PEEP protocol [2]. Briefly, blood gases were obtained every 30 min and the FiO2 and PEEP were adjusted to maintain a PaO2 between 55 and 80 mmHg [2]. Plateau pressure (Pplat) was maintained 30cmH2O, the tidal volume was reduced by 1 cc/kg. If pH fell below 7.30, respiratory rate was increased until pH > 7.30 or PaCO2 < 25 mmHg. Recruitment Maneuver þ Optimal PEEP Group (RMþOP): (n ¼ 5, 21.4 6 1.9 kg); this strategy was based on the method described by Borges et al. [12]. Briefly, an aggressive recruitment maneuver (RM) was performed followed by a decremental PEEP trial to find the amount of PEEP necessary to keep the newly recruited alveoli open. For a detailed description of this procedure, please refer to Figure 1. HFOV: (n ¼ 5, 33.7 6 1.9 kg). Animals were switched to HFOV (Sensormedics 3100B oscillator, San Diego, CA, USA) and a RM was performed by increasing mean airway pressure (Pmean) to 40 cmH2O for 40 s with a FiO2 of 1.0. The initial oscillator settings were: Pmean 30 cmH2O, amplitude (DP) ¼ 50 cmH2O if pH > 7.25 or DP ¼ 70 cmH2O if pH < 7.25 cm H2O. Frequency ¼ 7 Hz (if animal < 25 kg) or 6 Hz (if animal > 25kg), percent inspiratory time 33%, and a power of 4.3. Adjustments were made to DP (minimum DP of 3) to achieve a pH target of 7.30 to 7.45. APRV: (n ¼ 6, 23.17 6 0.5 kg). The ventilator was switched to the DuoPAP-Plus mode (Galileo Ventilator; Hamilton Medical, Reno, NV) with a FiO2 of 1.0 and baseline ventilator settings determined by the adult recommended values: High pressure (Phigh) 30 cmH2O, low pressure (Plow) 0 cmH2O, the time at high pressure (Thigh) set to 5 s, and the time at low pressure (Tlow) 0.2 s. A RM was performed immediately after switching to APRV. Recruitment
ALBERT ET AL.: HFOV, APRV, AND RM+OP VERSUS LOWVT
e73
removed and the bronchus to the right middle lobe was exposed, cannulated with a small endotracheal tube, and the balloon was inflated to secure the tube. The lobe was lavaged with 60 mL normal saline (three 20 mL injections) and collected. The bronchoalveolar lavage fluid (BALF) was centrifuged and frozen for measurement of inflammatory mediators.
Wet/Dry Weight Ratio Representative samples of lung tissue were minced, placed in preweighed tin dishes, weighed, and placed in an oven to dry. The dishes were weighed daily until there was no weight change for 24 h, at which time the tissue was determined to be completely dry and water free. The wet to dry weight ratio (W/D) was calculated by dividing the wet weight by the dry after subtracting the weight of the dishes.
Histology
FIG. 1. RMþOP protocol. Immediately following lung injury, the lung was recruited by increasing PEEP to 25 cmH2O for four minutes and an arterial blood gas was taken (arrows). If the PaO2 þ PaCO2 was less than 400 mmHg, the PEEP was increased by 5 cmH2O for 2 min, then returned to 25 for 2 min, and an arterial blood gas taken. This was repeated until the PaO2 þ PaCO2 was greater then 400 mmHg (asterisk arrow). The maximum PEEP used was termed the ‘‘recruited’’ PEEP. After recruitment, a decremental PEEP trial was performed by lowering the PEEP in 2 cmH2O increments every 2 min until the PaO2 þ PaCO2 was less than 380 mmHg (arrowhead). The PEEP 2 cmH2O greater than the PEEP where PaO2 þ PaCO2 was less than 380 mmHg was termed the ‘‘optimal’’ PEEP. Following the PEEP trial, the lung was re-recruited using the ‘‘recruited PEEP’’ for 4 min and then the ventilator was set using low tidal volume (6 mL/kg) with the ‘‘optimal’’ PEEP for the duration of the experiment.
was performed by increasing the Phigh to 40 cmH2O with a Thigh of 40 s. After the recruitment maneuver was completed, the peak expiratory flow rate (PEFR) was measured at Tlow. Expiratory flow was analyzed using the freeze function and the tracer on the Galileo. Tlow was adjusted to resume CPAP phase before the lung fully exhaled (i.e., between 50% and 75% PEFR). Animals were allowed to spontaneously breathe while on APRV.
Hemodynamic and Lung Function Measurements Arterial and venous blood gases (ABG), heart rate (HR), mean arterial blood pressure (MAP), central venous pressure (CVP), cardiac output (CO), and urine output (UOP) were recorded at baseline (after surgical preparation), immediately post-injury (T0) and every 30 min for 6 h (T360). CO was measured using a thermodilution technique (5F catheter, Edwards Lifesciences, Irvine, CA, USA) and recorded as the average of three measurements. Pulmonary parameters (respiratory rate (RR), peak airway pressure (Ppeak), mean airway pressure (Pmean), plateau pressure (Pplat), tidal volume (Vt), PEEP, static compliance (Cstat), were measured and calculated inline by the Galileo ventilator at each 30 min time point.
Necropsy Euthanasia was performed by barbiturate overdose (150 mg/kg). The heart and lungs were removed en bloc. The heart was then
The lung was immersed in formalin for a minimum of 48 h prior to histologic sectioning. A constant airway pressure of 25 cmH2O was used to fix the lung. Samples were obtained from the lower left lobe and were scored by a blinded histologist using a 0–4 scale, with 4 representing the most profound injury. Samples were evaluated for evidence of fibrinous deposits, red blood cells in the alveolar space (airspace hemorrhage), vessel congestion, alveolar wall thickness, and leukocyte infiltration. A detailed description of the histologic methodology and lesion descriptions can be found in previous work by our lab [13].
BALF TNF-a and IL-8 Bronchoalveolar lavage fluid (BALF) levels of TNF-a and IL-8 were determined by porcine-specific enzyme linked immunosorbent assay (ELISA) kits (R and D Systems, Minneapolis, MN) using the manufacturer’s specified protocol.
Apoptosis A terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay was used to measure the number of apoptotic cells in the lung tissue on the surface of each histologic slice. Eightmicron sections were cut, de-parinized, and then stained using with ChemiCon’s ApopTag Red in situ. The counts were made using a CY3 filter at 203 magnification.
Statistics Physiologic variables were analyzed following group assignment using a repeated measures ANOVA (RM ANOVA) to determine differences between groups and within groups. To determine differences at specific time points, Dunnet’s post-hoc test was performed. Graded histologic data was analyzed to determine a c2 value via a likelihood ratio calculation. Comparisons were limited to the aims of this study, which were to compare the respective open lung strategies to low tidal volume ventilation, and were not extended to compare open lung groups with each other. Analyses were performed on ver. 5.0.1.2 of JMP (SAS Institute, Cary, NC). All data are presented as mean 6 standard error mean. An a level of 0.05 or less was considered statistically significant.
RESULTS
There were no significant differences between groups in any of the hemodynamic or respiratory parameters at baseline or immediately following lung injury. All
e74
JOURNAL OF SURGICAL RESEARCH: VOL. 166, NO. 1, MARCH 2011
animals developed a severe lung injury (PaO2/FiO2 ratio 151.8 6 25.5) following instillation of Tween-20 and nonprotective ventilation (Fig. 2A) regardless of group assignment. Respiratory Parameters
Average Vt was similar in the RMþOP and LowVt groups (6.6 6 0.1 mL/kg versus 6.3 6 0.5 mL/kg) while the experimentally derived ‘‘optimal’’ PEEP in the RMþOP group was significantly greater (15.7 6 0.2 cmH2O versus 9.0 6 0.4 cmH2O, P > 0.0001). Although PEEP was significantly different in the RMþOP (Fig. 3C), there was no difference in static compliance compared with the LowVt (Table 1). Tidal volume was significantly greater with APRV (15.4 6 0.3) than
with LowVt. Per protocol, the mean airway pressure (mPaw) was lowest in the LowVt, and maintained at a constant pressure of 30 cmH2O in the HFOV and APRV groups; mPaw in the RMþOP group was significantly greater than LowVt (Fig. 3D). Following injury, the PaO2/FiO2 ratio improved significantly in all of the ‘‘open lung’’ groups (RMþOP, HFOV, APRV) compared with LowVt (Fig. 2A). Systemic oxygen utilization measured by SvO2 was significantly greater with APRV and HFOV versus LowVt; with RMþOP it was significantly reduced (Fig. 2B). CO2 clearance was significantly improved in the HFOV and APRV groups compared with LowVt, whereas the RMþOP displayed hypercapnia more severe than LowVt, albeit not statistically significant (Fig. 2C). These changes in PaCO2 were reflected in the arterial pH, which was in the
FIG. 2. Blood gas analysis between LowVt (filled circles), APRV (filled triangles), HFOV (open circles), and RMþOP (open triangles): (A) PaO2/FiO2: lung injury was uniform between all groups following lung injury and subsequently improved with the application of HFOV (P ¼ 0.0567 by group; 0.0060 by time*group), APRV (P ¼ 0.0246 by group) and RMþOP (P ¼ 0.0260 by group). (B) SvO2: venous oxygen saturation was significantly greater with HFOV (P ¼ 0.0118 by group; 0.0229 by time*group) and APRV (P ¼ 0.0399 by group; 0.0019 by time*group) than LowVt. SvO2 was significantly lower in the RMþOP group over the duration of the experiment (P ¼ 0.0278 by time*group). (C) PaCO2: arterial CO2 content was effectively maintained at normal levels with application of HFOV (P ¼ 0.029 by group;
