Hemodynamic Effects of Synchronized High-frequency Jet Ventilation Compared with Low-frequency Intermittent Positive-pressure Ventilation after Myocardial Revascularization

The study was approved by our institutional ethical committee and all patients gave written consent the day before surgery. Twenty-two nonconsecutive patients scheduled for coronary bypass graft surgery were prospectively studied. Exclusion criteria for eligibility for the study were other associated cardiac surgery or history of chronic obstructive pulmonary disease (defined as a ratio of forced expiratory volume in 1 s to forced vital capacity < 0.5 of predicted values). To account for the importance of the underlying myocardial function, we included equal numbers of patients with normal and decreased (< 45%) preoperative LV ejection fraction, which was assessed before surgery by echocardiography or isotopic or angiographic ventriculography.

Antianginal therapy was discontinued the evening before surgery. One hour before surgery diazepam and morphine premedication were administrated. Oral metoprolol was also given to patients receiving chronic β-adrenergic blocker therapy. A radial artery cannula and a flow-directed modified pulmonary artery catheter (Vigilance Monitor, Baxter Edwards, Irvine, CA) capable of continuous monitoring of mixed venous oxygen saturation (Sv^O2) were inserted. Anesthesia and muscle relaxation were induced and maintained with midazolam, fentanyl, and pancuronium bromide. The patient’s trachea was intubated with a cuffed Hi-Lo Jet endotracheal tube (Mallinckrodt, St Louis, MO), equipped with a principal (9.0-mm inner diameter) and two auxiliary lumens (2.4-mm inner diameter each). The proximal (jet-ventilation) port opens 5 cm from the distal orifice of the endotracheal tube. The second port (for airway pressure monitoring) opens at the distal end of the endotracheal tube. A nasogastric tube and a Foley urinary catheter were also inserted. Hypothermic (28°C) cardiopulmonary bypass and cold cardioplegic–induced cardiac arrest were used. The cardiopulmonary bypass flow was set at 2 l · min⁻¹· m⁻²and a mean arterial pressure was maintained at 60 mm Hg using vasopressive drugs. During cardiopulmonary bypass the lungs were not ventilated but inflated with a continuous positive airway pressure of 2–5 cm H²O. Weaning from cardiopulmonary bypass was achieved with inotropic or vasopressor drug support if required. After surgery patients were transferred to the surgical intensive care unit. Upon arrival, sedation and analgesia were provided by continuous infusion of midazolam and morphine titrated for a Ramsay score of 5. 6The patients were placed on mechanical ventilation (Veolar ventilator; Hamilton, Rhäzüns, Switzerland). Initial ventilator settings in controlled mechanical ventilation (CMV) mode were tidal volume, 8–10 ml/kg of body weight, and respiratory rate, 12 breaths/min). Both were adjusted to maintain arterial carbon dioxide tension at 40 ± 5 mmHg. The inspired oxygen fraction was adjusted for arterial oxygen tension > 90 mmHg. Positive end-expiratory pressure of 5 cm H²O was applied if required to obtain arterial oxygen saturation > 93%. Continuous infusion of NaCl 0.9% at a rate of 5 ml · kg⁻¹body weight · h⁻¹was given, and additional boluses of 250 ml of the same solution were given if required to maintain pulmonary artery occlusion pressure around 10 mmHg and a cardiac output index > 2 l · min⁻¹· m⁻². Body temperature, lead II electrocardiogram (ECG) and urine output were monitored throughout the postoperative period. To prevent cooling, patients were wrapped with warm blankets. The patients were observed for at least 2 h to confirm a hemodynamic steady state, which was defined as less than 10% change in hemodynamics (heart rate, cardiac output, arterial pressure), no clinically relevant bleeding, and normal body temperature. A chest radiograph was obtained and assessed before data collection to define the correct position of the endotracheal tube, pulmonary arterial and central venous catheters, surgical drains and also to confirm the absence of cardiopulmonary abnormalities (grossly enlarged mediastinal silhouette, pleural effusions, or pneumothorax). A 12-lead ECG was also recorded to document a regular sinus rhythm. Body-position changes, patient care, and endotracheal suctioning, if required, were allowed during this period but were not conducted during the study protocol.

All studies were performed on deeply sedated patients in supine position who were not spontaneously breathing, and all pressure transducers were referenced to midchest. Sv^O2was calibrated in vivo  according to manufacturer’s instruction by simultaneous measurements of mixed venous blood on a cooxymeter (Statprofile Ultra, Nova Biomedical, Waltham, MA). For the first 10 patients hemodynamic variables and airway pressure were continuously recorded on a strip-chart paper recorder. For the last 10 patients hemodynamic data were recorded intermittently using a central monitoring station (Hewlett-Packard VECTRA 712/60, Palo Alto, CA).

The study protocol consisted of five sequential ventilatory steps of 30 min each: CMV¹, sync-HFJV¹, CMV², sync-HFJV², and CMV³. CMV¹was set to maintain adequate arterial blood gases. CMV²and CMV³settings were kept identical to CMV¹and served as time and sequence control steps. Peak airway pressure was kept below 25 cm H²O during CMV by changing first inspiratory-to-expiratory ratio from 1:2 to 1:1, then by decreasing tidal volume. Sync-HFJV was delivered with no entrainment of bias flow and with passive exhalation using an Acutronic (Jona, Switzerland) AMS-1000 ventilator (driving pressure 1.8 psi; inspiratory-to-expiratory ratio 30:70, inspired oxygen fraction for arterial oxygen tension > 90 mm Hg). Synch-HFJV¹and sync-HFJV²were generated by synchronizing the HFJV inspiration phase with the QRS complex of the ECG, as previously described. 4,5Briefly, the QRS signal from the ECG monitor was used to trigger the ventilator. A variable-delay circuit between the ECG signal and the ventilator enabled timing of the jet inspiratory pulse from the early systolic to the late diastolic phase of the arterial pressure pulse. The optimal timing of the jet inspiratory pulse delivery was established at the beginning of each sync-HFJV run by delaying by 10% the ECG R-wave trigger jet pulse delivery through the cardiac cycle and observing the resultant steady state Sv^O2value (fig. 1). Sync-HFJV timing corresponding to the highest Sv^O2value was used for the sync-HFJV¹and sync-HFJV²steps. Sv^O2variation was chosen instead of the continuous cardiac output value because the former varies more rapidly than the second and represents a real online continuous monitoring.

At the end of each period, body temperature Sv^O2, heart rate, systolic, diastolic, and mean arterial and pulmonary arterial pressures, pulmonary artery occlusion, and central venous pressures were recorded from the bedside monitor (M1166A; Hewlett-Packard). Mean cardiac output was estimated by averaging triplicate injections of 10 ml ambient-temperature 0.9% NaCl delivered randomly during the respiratory cycle. At the end of each study period arterial and mixed venous blood gases were simultaneously measured (Statprofile Ultra, Nova Biomedical), and standard calculated variables were obtained from blood-gas data. Ventilator settings (respiratory rate, tidal volume, inspiratory-to-expiratory ratio, inspired oxygen fraction, positive end-expiratory pressure) and peak and mean airway pressures were also recorded.

The data were analyzed using Graph Pad Instat software (version 2.05a; Graph Pad software, San Diego, CA) for a personal computer. In order to confirm the steady-state condition, a one-way repeated-measures analysis of variances was performed to compare data for the three CMV steps. Data for CMV¹and CMV²and respectively sync-HFJV¹and sync-HFJV²were then pooled, and a two-tailed paired t  test was used. Comparisons between sync-HFJV and the preceding CMV run in two subsets of patients stratified by baseline level of LV contractility according to their preoperative LV ejection fraction (impaired LV function group, LV ejection fraction < 45%; normal LV function group, LV ejection fraction ≥ 45%) were performed by two-tailed paired t  test. All values are expressed as mean ± SD, and P < 0.05 was considered statistically significant.

Among patients who gave informed consent, two were excluded before starting the study protocol. The first one required reoperation for surgical hemostasis because of ongoing hemorrhage, and the second patient had unstable hemodynamics. Twenty patients were included in the final analysis, and all of them tolerated the experimental protocol well. Demographic and preoperative characteristics are presented in table 1. Overall LV ejection fraction was 45 ± 11%. Patients in the low LV ejection fraction group (n = 10) had an average ejection fraction of 35 ± 6%; in the normal LV ejection fraction group (n = 10) the fraction was 54 ± 6%. An average of three (range, two to four) coronary artery bypasses were performed, with all patients but one having an internal mammary artery graft. The mean cardiopulmonary bypass duration was 140 ± 32 min, and the aortic cross-clamping time was 84 ± 20 min. Catecholamine infusion was required in 14 patients for cardiopulmonary bypass weaning (dobutamine, n = 13, and/or epinephrine, n = 2). At SICU arrival, two patients were receiving continuous intravenous infusions of dobutamine (5 μg · kg⁻¹· min⁻¹), and one patient of epinephrine (0.1 μg · kg⁻¹· min⁻¹). Continuous intravenous infusion of nifedipine at a rate of 5 mg/h was given to an additional patient. In these four patients, vasoactive infusion rates were kept constant during the study protocol.

Ventilatory settings, airway pressures, and gas-exchange values are shown in table 2, and hemodynamic variables are presented in table 3. The steady-state assessment of the three different CMV runs showed a significant difference in mean arterial pressure between CMV¹and both other CMV runs and in systemic vascular resistance between CMV¹and CMV³. Other differences among the three CMV runs were minimal and not clinically relevant. During the sync-HFJV runs (pooling data from both runs) the optimal Sv^O2was obtained during the early diastolic period in 7 runs, during the late diastolic in 5, during the early systolic in 17, and during the late systolic period in 11. No significant difference was observed between the best Sv^O2obtained by delaying the jet pulse of sync-HFJV runs in patients with preoperative normal or impaired systolic LV function. Tables 4 and 5present the hemodynamic data separated by the 45% LV ejection fraction cut-off. Figure 2shows individual changes in cardiac index induced by the ventilatory mode switch from CMV (CMV¹and CMV², respectively) to HFJV (sync-HFJV¹and sync-HFJV², respectively) and after return to CMV (CMV²and CMV³, respectively). A wide variability of the response was observed among patients, and separating patients into two groups according to their preoperative LV ejection fraction did not modify individual variability.

Our data show that, unlike patients with severe congestive cardiomyopathy awaiting cardiac transplantation, patients with less severely compromised global myocardial function after cardiac revascularization surgery did not demonstrate hemodynamic improvement if airway pressure was synchronized with the cardiac cycle and normal blood-exchange values were targeted, avoiding severe hypocapnia. Indeed, no differences in cardiac output were observed if sync-HFJV was compared with CMV in our study, even if the timing of the jet pulse delivery was adjusted according to the individual best obtained value of mixed venous oxygen saturation. The absence of hemodynamic improvement was observed even though the majority of patients exhibited the highest mixed venous oxygen saturation value during the systolic or at the end of the diastolic phase (33/40 sync-HFJV runs), when minimal effect of the increase in intrathoracic pressure is expected on the venous return during the cardiac cycle.

Different results have been reported concerning the hemodynamic effects of sync-HFJV both in experimental and human studies. 7–20The discrepancy in the response is probably related to the different basal conditions of heart function and circulation state rather than changes in heart–lung interactions. However, our results are in accordance with those of Bayly et al.,  21who did not find a significant difference between low-frequency intermittent positive-pressure ventilation and sync-HFJV in 10 stable patients without heart failure, but they contrast with findings from Pinsky et al.  5and Angus et al.  22Different hypotheses may be advanced to explain the absence of improvement in our patients. In our study we targeted a normal blood-gas exchange guided by standard clinical practice criteria, avoiding the marked hypocapnia obtained in the study by Pinsky et al.  5This may have two different simultaneous consequences. The driving pressure used to deliver the HFJV tidal volume was kept at the lowest value sufficient to provide efficient gas exchange, possibly mitigating the advantage of cardiac cycle–related increases in intrathoracic pressure on the pressure gradient for LV ejection. Furthermore, a relatively low tidal volume was delivered during the CMV runs (around 8 ml/kg body weight) to obtain normal blood exchanges during conventional ventilation. The beneficial effect of synchronized HFJV might have been minimized by the reduced tidal volume ventilation during CMV. Thus, the low peak and mean airway pressures generated during all steps of the protocol may have contributed to minimize the negative effect of intrathoracic pressure swings on LV output. Another possible explanation for the lack of a differential effect of synchronized HFJV could be related to the surgical procedure. Indeed, the pericardium was left open anteriorly at the end of surgery. This may have modified the interaction between airway pressure oscillation transmission and the effective pericardial pressure. The relationship between pleural and pericardial pressures may have been further affected by a decreased lung compliance secondary to fluid accumulation in the lungs after cardiopulmonary bypass. Finally, a transient positive effect of sync-HFJV could have occurred initially but been attenuated by the intact feedback mechanisms provided by arterial baroreceptors of the great arteries. We doubt that this is the case, because we would have expected mixed venous oxygen saturation to initially decrease with the initiation of sync-HFJV and then increase toward pre–sync-HFJV values. We did not observe such temporal behavior in mixed venous oxygen saturation.

In conclusion, after coronary artery bypass graft surgery, synchronized HFJV provides no hemodynamic advantage compared with CMV.

The authors thank the physicians and nurses who cared for the patients studied. Baxter AG (Dietlikon, Switzerland) provided the Vigilance Monitor.