High Frequency Oscillatory Ventilation for Acute Respiratory Distress Syndrome (ARDS)

BACKGROUND AND RATIONALE Recent two-center results (1) support a beneficial effect of combined high-frequency oscillation (HFO), recruitment maneuvers (RMs) and tracheal gas insufflation (TGI) on the survival of patients with severe acute respiratory distress syndrome (ARDS). The addition of TGI to HFO improves gas-exchange (1-4); however, its value with respect to outcome still remains uncertain (1). TGI is likely useful in patients with very severe oxygenation disturbances and/or poor tolerance to hypercapnia (2).

The main goals of the present study are 1) The determination of the effect of the intermittent, combined use of HFO and RMs (HFO-RMs - intervention group) on the survival as compared to the best possible strategy of lung-protective conventional mechanical ventilation (CMV - control group); and 2) the elucidation of the mechanism of action of HFO on respiratory function and the ARDS-related inflammatory response.

The consecutive hypotheses supporting the conduct of the present trial can be summarized as follows:

The use of HFO + RMs will likely augment lung recruitment, with consequent improvements in oxygenation and lung compliance The HFO-related, physiological benefits will likely be maintained during the subsequent CMV, if an adequate positive end-expiratory pressure (PEEP) level is used (1).

Therefore, the following sequence of events is expected from the intermittent use of HFO-RMs:

• Lung Recruitment \& Compliance →
• Ventilation pressures during the subsequent CMV (as compared to pre-HFO CMV)→
• Risk of ventilator-associated lung injury →
• ARDS-related inflammatory response →
• ARDS-related organ or system failure(s) →
• Survival

These hypotheses are consistent with preceding results on 125 patients (1). The present HFO-RMs protocol will be simplified relative to that of our previous study, in order to improve its generalized applicability. The use of HFO-TGI will be optional and limited to 1) rescue oxygenation procedures for both the HFO-RMs and CMV group; and 2) certain patients of the HFO-RMs-group with "very severe oxygenation disturbances" (see below for definition), or "poor control" of arterial pH (pHa)/PaCO2 (see below). The frequency of HFO-TGI use will be compared between the 2 study groups after study completion.

METHODS Patients The study protocol has received Institutional Review Board Approval. An Information Sheet detailing the potential benefits and risks of study participation will be provided to the next-of-kin of eligible patients. Following discussion of the study with one of the investigators, a written, next-of-kin consent for study participation will be requested. As soon as clinically feasible, patients will be informed of the study and of their right to withdraw.

Study participants must fulfill the eligibility criteria presented in the dedicated section. Continuous patient monitoring will include electrocardiographic lead II, inra-arterial pressure with/without cardiac index (PICCO plus, Pulsion Medical Systems, Munich, Germany), and peripheral oxygen saturation (SaO2). Anesthesia will be maintained with continuous infusions of midazolam or propofol and fentanyl or remifentanil. Neuromuscular blockade with cisatracurium will be used in concordance with standard recommendations (6), and as part of attending physician-prescribed medical treatment. During the first 48 hours post-enrollment, all patients will receive a continuous infusion of cisatracurium (7).

Randomization Patient randomization will result in assignment to the control (CMV) group (which will receive treatment with CMV alone), or to the intervention (HFO-RMs) group [which will receive treatment with an extended HFO session of at least 96 hours (HFO session can be interrupted before the 96-hour time point only if PaO2/FiO2 rises to >200 mmHg for >12 hours)], and whenever required, additional HFO sessions of at least 12-24 hours, interspersed with CMV as described below).

For each participating center, a sequence of unique random numbers from 1 to 200 will be generated apriori with the Research Randomizer (www.randomizer.org). To achieve concealment until patient study entry, each random number will be marked on a piece of paper placed in an opaque envelope, which will be subsequently sealed. Envelopes will be prepared by the department's statistician, and will be externally labelled with the patient serial number. Envelopes will be opened following the patient serial number order after the obtainment of signed informed consent. Thus, on each consecutive study entry, an envelope will be opened and the random number it contains will be assigned to the patient as his/hers unique study-number. Patients with even and odd study-numbers will be allocated to the CMV and HFO-RMs group, respectively.

CMV strategy Immediately after randomization, study participants will receive CMV with the following combinations of FiO2/PEEP: 0.5/10-12 cm H2O, 0.6/14-16 cm H2O, 0.7/14-16 cm H2O, 0.8/14-16 cm H2O, 0.9/16-18 cm H2O, 1.0/20-24 cmH2O. Whenever oxygenation improves, FiO2 will be reduced first, followed by the reduction in PEEP according to the aforementioned FiO2/PEEP combinations. Whenever oxygenation deteriorates, PEEP will be increased first, followed by the increase in FiO2 according to the aforementioned FiO2/PEEP combinations. Rationale for the use of high PEEP: According to a recent meta-analysis (8), the use of PEEP levels comparable to those proposed by the present protocol may be associated with an improved survival.

In patients with a body mass index of >27 kg/m2 and/or an urinary bladder pressure of ≥15 mmHg, the positivity of end-expiratory transpulmonary pressure (PLend-exp) will be confirmed with the esophageal balloon technique (9,10), once daily for the first 10 days post-randomization; in case of a negative PLend-exp, the PEEP level will be increased so that PLend-exp becomes positive (10); rationale: obese patients and patients with increased intra-abdominal pressure are more likely to require higher PEEP levels for the prevention of expiratory derecruitment.

Tidal volume will be within 5.5-7.5 mL/Kg predicted body weight. The maximal plateau pressure limit will be 40 cmH2O, and target plateau pressure will be ≤32 cmH2O (7); rationale: as in the study of Meade et al (11), a higher plateau pressure will be tolerated to allow for the use of a higher PEEP level. When plateau pressure exceeds 32 cmH2O for >15 min, the following adjustments will be conducted: tidal volume reduction up to 4.0 mL/kg predicted body weight, respiratory rate increase up to 35/min, and PEEP reduction by ≥2 cmH2O. These adjustments will have to concurrently result in the achievement of the below-provided gas-exchange targets.

The respiratory rate will be titrated to a pHa of 7.20-7.45. The inspiratory-to-expiratory time (Ι:Ε) ratio will be ≤1/2. The oxygenation target will be SaO2=90-95%, and/or PaO2=60-80 mmHg. At pHa<7.20, breathing circuit deadspace will be minimized by connecting the Y piece directly to the tracheal tube (7), tidal volume will be increased up to 8.0 mL/kg predicted body weight, and respiratory rate will be increased up to 35/min. If these measures fail, the criterion of "poor control of pHa/PaCO2," and the use of a bicarbonate infusion will be permitted. Additional options will include the use of TGI of 6-7 L/min, or extracorporeal CO2 removal.

In the CMV group, RMs (Continuous positive airway pressure of 45 cmH2O for 40 s at FiO2=1.0) will be used at a frequency of 3/day for the first 5 days post-randomization. RMs will start at 9:00 a.m. and will be repeated every 5 hours; this means that the second RM will be performed at 2:00 p.m., and the third RM at 7 p.m. If during an RM, SaO2 drops by 10%, or mean arterial pressure drops by 25%, the RM will be immediately aborted, and the next RM will be performed after at least 10 hours. If the first daily RM is aborted, then the second RM will be cancelled and the third RM will be administered. If the second daily RM is aborted, then the third daily RM will be cancelled. If the third daily RM of days 1, 2, 3, or 4 is aborted, the next RM will be conducted at 9:00 a.m. of the subsequent day. After day 6, there will be no protocolized use of RMs; rationale: as time from ARDS onset passes, the probability of RM-associated oxygenation improvement decreases and the risk of RM-related hypotension increases (12).

HFO-RMs strategy In the HFO-RMs group, HFO sessions will be used during days 1-5 post-randomization; rationale: in NCT00637507 HFO was used for ≤5 days in ~75% of the patients of the intervention group (1). The daily HFO sessions will start at 9 a.m. and last a minimum of 12 hours. The HFO-RMs strategy is described below.

Recently published recommendations regarding HFO use (Sensormedics 3100B ventilator, Sensormedics, Yorba Linda, CA, USA) include the following steps (13).

1. Sufficient level of sedation for the abolishment of respiratory muscles activity, with or without neuromuscular blockade; the latter will be mandatory during the first 48 hours (7).
2. Confirmation of endotracheal tube patency and placement of the tube at 3-4 cm above carina.
3. RMs: Immediately after patient-oscillator connection, an RM will be performed (increase in the circuit pressure to 45-50 cmH2O for 40 s with the oscillator's piston off). During days 1-5 post-randomization, RMs will be repeated every 5-6 hours. RM abort criteria will be the same as for the CMV group. In the HFO-RMs group, RMs will be used solely during HFO.
4. FiO2 will initially be set at 1.0, and then adjusted according to protocol (see below).
5. Bias flow will be set at 60 L/min; rationale: this maximal bias flow setting is expected to result in improved CO2 clearance from the HFO breathing circuit.
6. Initial oscillation frequency will be 4 Hz, and will be titrated to a pHa of >7.20. The minimum value of oscillation frequency will be 3.5 Hz; rationale: our HFO experience suggests that the use of frequencies of <3.5 Hz are associated with increased risk of high-frequency ventilator malfunction.
7. Oscillatory pressure amplitude (ΔP) will be initially set at 90 cmH2O and will be titrated according to a pHa of >7.20 (range=60-100 cm H2O).
8. A tracheal tube cuff leak will be placed to facilitate CO2 elimination. The associated reduction in mean airway pressure (mPaw) of 4-5 cmH2O will be immediately reversed by using the corresponding control knob. Cuff pressure will be maintained at ≥20 cmH2O.
9. If pHa<7.20, despite an oscillation frequency of 3.5 Hz and a maximal ΔP setting of 100 cmH2O, the deadspace of the breathing circuit will be minimized by connecting the Y piece directly to the tracheal tube (7). If pHa still remains <7.20, the criterion of "poor control of pHa/PaCO2" will be fulfilled, and the use of a bicarbonate infusion will be permitted. As during CMV (see above), additional options will include the use of TGI of 6-7 L/min, or extracorporeal CO2 removal.
10. The I:E ratio will be maintained at 1:2.
11. mPaw adjustments will be as follows: Α] Initial mPaw=mPaw CMV + 10-13 (maximal allowable=45) cm H2O, Β] Within the next 2 hours: mPaw titrations of ±3 cm H2O to determine the "optimal mPaw setting" that achieves the highest PaO2 at FiO2=1.0, and C] Reduction of mPaw at a rate of 1-2 cmH2O/6 hours, with each downward titration being preceded by an RM.
12. Patients will be returned to CMV after a maximum of 96 hours after HFO initiation, provided that a PaO2/FiO2 of >200 mmHg is achieved for >6 hours. Return to CMV before 96 hours after HFO initiation will be allowed whenever PaO2/FiO2 rises of >200 mmHg will be achieved and maintained for >12 hours during HFO. Within days 5-10, patients will be returned to HFO whenever PaO2/FiO2 falls below 200 mmHg for >12 hours; the criterion of re-return to CMV will again be PaO2/FiO2 >200 mmHg for >12 hours or the end of day 10, provided that at this time point PaO2/FiO2 exceeds 100 mmHg; any subsequent use of HFO will be according to the below-presented protocol of "Rescue Oxygenation."
13. TGI of 6-7 L/min will be permitted as an option if the patient fulfills the following criterion of "very severe oxygenation disturbance": During the pre-HFO CMV, the patient requires an FiO2 of 0.9-1.0 (and a PEEP level of ≥16 cmH2O) to maintain an SaO2 of 90-95% (and/or a PaO2 of 60-80 mmHg); this is virtually equivalent to a patient having a PaO2/FiO2 of <100 mmHg at a PEEP level of ≥16 cmH2O. In such cases, the rest of the above-described, protocolized adjustments and interventions of the HFO-RMs protocol will be used without any additional change; rationale: TGI may be useful in patients, who require maximal support during CMV to maintain a clinically acceptable level of oxygenation (3).

Rescue Oxygenation Patients of both groups will be eligible for rescue oxygenation if they fulfill the following criterion: Patient is on CMV with an FiO2 of 1.0 and a PEEP level of ≥20 cmH2O, and has a sustained and life-threatening hypoxemia (i.e., PaO2<60 mmHg for >30 min), not associated with a "promptly reversible" factor (e.g. pneumothorax, malpositioning or obstruction of the tracheal tube, or ventilator malfunction). Rescue oxygenation techniques may include the use of HFO-RMs and/or HFO-TGI, prone position (14), inhaled nitric oxide (Nitric Oxide - ΝΟ), intravenous almitrine, and extracorporeal membrane oxygenation. The use of one or more rescue oxygenation techniques will last at least until the achievement of the reversal of the life-threatening hypoxemia for 1 hour.

Patient follow-up Baseline patient data will be recorded within 2 hours pre-randomization. Daily recordings will include physiologic/laboratory data (days 1-28 post-randomization), intervention-associated complications (days 1-10; examples: RM-induced hypotension or desaturation), mechanical ventilation-associated barotrauma [study-independent radiologists will assess chest radiographs for pathologic gas collection(s), e.g. pneumothorax], data on organ/system failures and medication (days 1-60), episodes of failure to maintain unassisted breathing and various complications (until hospital-discharge or death; examples: infections, heparin-induced thrombocytopenia).

During days 1-10, sets of physiologic measurements will be obtained as follows: 1) CMV group: 3 measurements/day, starting at 8:30 a.m. 2) HFO-RMs group: just before, during, and 6 hours after HFO, and as in CMV group after day 5. Measurements will include arterial/central-venous blood-gas analysis, hemodynamics, and respiratory mechanics while on CMV (including respiratory compliance); also, on the morning of each one of days 1-10, we will determine and record the fluid balance of the preceding 24 hours. For between-group comparisons, we will use CMV-data obtained within 8:30-9 a.m. in both groups. Daily fluid balance will also be compared between the 2 groups.

Lastly, in patients who have received HFO-TGI during days 1-5, a brief fiberoptic inspection of the trachea will be conducted on the morning of day 6 to detect any potential, TGI-associated tracheal mucosal damage. Bronchoscopic findings will be defined as follows: Grade I: Pink and glistening tracheal mucosa; Grade II: Reddened and/or swollen mucosa with/without presence of purulent secretions; Grade IIIA: Hemorrhagic mucosa and/or presence of thrombotic material; Grade IIIB: Limited localized necrosis, especially at the carina, and/or presence of necrotic mucosal slough; and Grade IIIC: Extensive localized necrosis, especially at the carina, and/or presence of necrotic mucosal slough. Grade IIIA-IIIC findings will be considered as suggestive of TGI-related mucosal damage. A bleeding diathesis (if present) should be considered as an independent risk factor for Grade IIIA findings. Findings suggestive of TGI-related tracheal mucosal damage in a patient will result in no further use of HFO-TGI for rescue oxygenation in that particular patient.

Bronchoalveolar lavage (BAL) BAL of ≤100 mL will be performed on day 1 and 6 post-randomization in patients of both groups. Patients will be eligible for BAL if their PaO2/FiO2 has been maintained at >100 mmHg for >12 hours and they are intubated with an orotracheal tube with an internal diameter of ≥8.5 mm, or a tracheostomy tube. An (additional) RM will be performed after the fiberoptic bronchoscopic procedure. BAL fluid samples will be used for microbiological cultures, cell count, and the determination of the concentration of phospholipids, surfactant-related proteins, and inflammatory markers. The purpose of the aforementioned investigational interventions is the elucidation of the effect of HFO on the function of the surfactant and the ARDS-related inflammation (15-18). During the bronchoscopic procedures, blood samples will also be obtained for the determination of the concentrations of the same inflammatory markers in the peripheral blood.

BAL fluid studies The initial, 20-mL portion of the BAL fluid aspirate will be sent for microbiological cultures, and the rest will be stored in ice-cold tubes. Subsequently, the BAL fluid will be filtered through sterile gauze and centrifuged at 500 g for 15 min at 4 0C. The supernatant will be used for the determination of the concentrations of inflammatory markers, phospholipids, and surfactant-related proteins. The sediment will be used for total cell count, determination of cell type, and estimation of cell viability on a Neubauer plate. Both the supernatant and sediment will be stored at -700C.

Surfactant aggregates, surfactant-related proteins, and inflammatory markers The cell-free supernatant of the 500 g centrifugation will be subjected to additional consecutive centrifugation at 30,000 g και 100,000 g for 90 min at 4 0C. This will be done to separate surfactant aggregates according to their size. The Large Surfactant Aggregates (LSAs) will be obtained from the 30,000 g centrifugation sediment. LSAs are considered as major determinants of the alveolar surface tension. The less active Small Surfactant Aggregates and the Very Small Surfactant Aggregates will be respectively obtained from the sediment and supernatant of the 100,000 g centrifugation (15).

Surfactant studies will include total lipid concentration, separation of lipid classes with thin layer chromatography, determination of lipid phosphorus content (15), and surfactant-related proteins (17). In addition, the supernatant will be analyzed to determine the concentrations of tumor necrosis factor (TNF) alpha, interleukin (IL) 1-beta, IL-1 receptor antagonist, IL-6, IL-8, transforming growth factor alpha (16,18), activin alpha, and folistatin, whereas the same inflammatory markers will be determined in the peripheral blood.

POTENTIAL RISKS OF INVESTIGATIONAL INTERVENTIONS AND THEIR PREVENTION Potential risk: Barotrauma. Preventive measures: This potential risk is equally high during CMV or HFO (19,20). Prevention includes the rapid HFO-RMs-related improvement in oxygenation and lung compliance, and the consequent reduction in the ventilation pressures during the subsequent CMV. Regarding the theoretical risk of bronchoscopy-related barotrauma, the bronchoscopies will be performed by an experienced operator, and intra-procedural ventilation (comprising tidal volumes of 2-3 mL/kg predicted body weight at rates of 35/min and PEEP temporarily reduced to 0-5 cmH2O) will also be managed by an experienced physician. Potential risk: Hypotension-drop in cardiac output. Preventive measures: This potential risk is equally high during CMV or HFO (19,20). If RM-associated, RMs will be aborted for ≥10 hours (this holds also for cases of RM-induced desaturation - see above). TGI-related complications: Such complications are not expected, because there will be only short-term use of TGI (1-4). Nevertheless, TGI may cause tracheal mucosal damage, inspissation of secretions, pneumothorax, gas embolism, and hemodynamic compromise (1-4). In NCT00637507, 1 of the 61 patients of the intervention group (1.6%) may have suffered a reversible tracheal mucosal injury (1). This patient had received TGI for a total of 118.3 hours over a period of 10 days (or 240 hours); he had a prolonged but full recovery from the severe ARDS and is currently leading a normal life.