Cardiorespiratory effects of conventional and high frequency ventilation in rabbits with bilateral pneumothoraces and surfactant depleted lungs

We compared high frequency ventilation (HFV) to conventional mechanical ventilation (CMV) under normoxic and normocapnic condition in surfactant depleted rabbits with bilateral pneumothoraces. We hypothesized that lower airway pressures would be required with HFV under these conditions. We applied CMV and HFV in 8 anaesthetized rabbits with a prototype ventilator at frequencies of 30, 100, 200, and 300 cycles/min. A positive end-expiratory pressure (PEEP) just below the pressure sufficient to open the air leak from the pneumothoraces was applied at all frequencies. Airway pressures, gas exchange, heart rate, and mean arterial pressure were recorded. Peak airway pressure decreased significantly from 2.50 to 2.10 kPa when the frequency of ventilation was increased from 30 to 300 cycles/min. There were no significant changes in mean airway pressure, PaO2, arterial pH, heart rate, and mean arterial pressure when HFV was compared to CMV. In conclusion, during HFV peak airway pressures measured at the mouth were decreased. Our ability to maintain adequate gas exchange in the face of ongoing pulmonary air leaks may reflect lower alveolar pressures.     

Acute respiratory distress syndrome. Liquid ventilation

In summary, the use of partial liquid ventilation with perfluorocarbon seems to be safe in neonates, children, and adults with ALI or ARDS. Partial liquid ventilation can be used for the following goals: recruitment of lung volume; enhancement or uniformity of lung inflation; improved oxygenation; improved lung compliance; and prevention or minimization of lung damage induced by mechanical ventilation. Partial liquid ventilation improved gas exchange and lung mechanics in test subjects. Cardiopulmonary interaction and long-term safety require further study. Extrapolations from animal data, however, suggest that there are no long-term undesired effects. The current clinical data strongly suggest that close monitoring and adjustment of ventilator parameters during drug instillation are necessary to reduce hypoxia, bradycardia, and pneumothorax. Further studies are required to explore the efficacy of partial liquid ventilation with perfluorocarbon in ALI or ARDS subjects, in particular, to evaluate the safety and efficacy of smaller doses; mechanical ventilation strategies; and outcomes. Additional research including large randomized studies is clearly required to answer these and other questions.     

Adaptive lung ventilation (AVL). Evaluation of new closed loop regulated respiration algorithm for operation in the hyperextended lateral position

The lateral decubitus position is the standard position for nephrectomies. There is a lack of data about the effects of this extreme position upon respiratory mechanics and gas exchange. In 20 patients undergoing surgery in the nephrectomy position, we compared a new closed-loop-controlled ventilation algorithm, adaptive lung ventilation (ALV), which adapts the breathing pattern automatically, to the respiratory mechanics with conventionally controlled mandatory ventilation (CMV). The aims of our study were (1) to describe positioning effects on respiratory mechanics and gas exchange, (2) to compare ventilatory parameters selected by the ALV controller with traditional settings of CMV, and (3) to assess the individual adaptation of the ventilatory parameters by the ALV controller. The respirator used was a modified Amadeus ventilator, which is controlled by an external computer and possesses an integrated lung function analyzer. In a first set of measurements, we compared parameters of respiratory mechanics and gas exchange in the horizontal supine position and 20 min after changing to the nephrectomy position. In a second set of measurements, patients were ventilated with ALV and CMV using a randomized crossover design. The CMV settings were a tidal volume of 10 ml/kg body weight, a respiratory rate of 10 breaths/min, an I:E ratio of 1:1.5, and an end-inspiratory pause of 30% of inspiratory time. With both ventilation modes F1O2 was set to 0.5 and PEEP to 3 cm H2O. During ALV a desired alveolar ventilation of 70 ml/ kg KG.min was preset. All other ventilatory parameters were determined by the ALV controller according to the instantaneously measured respiratory parameters. Positioning induced a reduction of compliance from 61.6 to 47.9 ml/cm H2O; the respiratory time constant shortened from 1.2 to 1.08 s, whereas physiological dead space increased from 158.9 to 207.5 ml. On average, the ventilatory parameters selected by the ALV controller resembled very closely those used with CMV. However, an adaptation to individual respiratory mechanics was clearly evident with ALV. In conclusion, we found that the effects of positioning for nephrectomy are minor and may give rise to problems only in patients with restrictive lung disease. The novel ALV controller automatically selects ventilatory parameters that are clinically sound and are better adapted to the respiratory mechanics of ventilated patients than the standardized settings of CMV are.     

Mechanical ventilation of patients with acute lung injury

Ventilatory management of patients with acute lung injury (ALI), particularly its most severe subset, acute respiratory distress syndrome (ARDS), is complex. Newer lung protective strategies emphasize measures to enhance alveolar recruitment and avoid alveolar overdistention, thus minimizing the risk of ventilator-induced lung injury (VILI). Key components of such strategies include the use of smaller-than-conventional tidal volumes which maintain peak transpulmonary pressure below the pressure associated with overdistention, and titration of positive end-expiratory pressure to promote maximal alveolar recruitment. Novel techniques, including prone positioning, inverse ratio ventilation, tracheal gas insufflation, and high frequency ventilation, are considerations in severe ARDS. No single approach is best for all patients; adjustment of ventilatory parameters to individual characteristics, such as lung mechanics and gas exchange, is required.     

Long-term mechanical ventilation in amyotrophic lateral sclerosis

BACKGROUND: Acute respiratory insufficiency (ARI) with alveolar hypoventilation or incapacitating dyspnoea but without peripheral muscle involvement can be an early manifestation of respiratory involvement in amyotrophic lateral sclerosis (ALS). Some of these patients benefit from assisted ventilation. The object of this study was to analyse the results of long-term mechanical ventilation (LTMV) in ten patients with ALS. METHODS: A retrospective analysis of intensive care unit (ICU) or ambulant patients with ALS who underwent LTMV in a conventional hospital ward was performed. Erect and supine spirometry, blood gas analysis and pulse oximetry were performed before the start and during the course of ventilation. RESULTS: Ten patients on LTMV were included. Four from the ICU were ventilated via tracheostomy, and six ambulant patients had non-invasive (nasal) ventilation. In all cases, ventilation was performed in a conventional hospital ward. The ambulant patients improved symptomatically during ventilation, confirmed by measurement of gas exchange and of SaO2 by continuous pulse oximetry. Three of the ten patients survive in long-term care--two with nasal and one with tracheostomy ventilation. CONCLUSIONS: LTMV outside ICU was possible in ten patients, seven of whom returned home. Returning home is very difficult for patients dependent on a ventilator who lack family support.     

Artificial ventilation in acute respiratory distress syndrome in adults. Towards an individual optimization

Artificial ventilation plays a key role in the treatment of acute respiratory distress syndrome (ARDS). Initially, the goal is to normalize gas exchange compromised by the lung disease. Positive pressure ventilation can however aggravate prior lesions of the pulmonary parenchyma, at least in areas of the lung accessible to ventilation. Computed tomography of the lung has given us a better understanding of the pathogenesis of these ventilation-induced lesions, leading to new ventilatory strategies aimed at assuring adequate oxygenation without damaging the parenchyma. These ventilatory modes may tolerate a certain degree of hypercapnia to avoid lung injury. Improved oxygenation relies on optimizing the ventilation/perfusion ratio, either with inhaled nitric oxide or a supine position to improve alveolar recruitment. In the most severe cases, extra-corporal gas exchange systems have shown their efficacy for patients whose lungs cannot be ventilated. Thus ventilation should be carefully adapted to each patient based on the severity of the ARDS and its clinical course. We present a practical protocol based on a hierarchy rationale for each ventilation mode and indicate the explorations required to adapt each mode to a specific patient.     

Current aspects of acute respiratory distress syndrome in children

Acute respiratory distress syndrome (ARDS) is a frequent condition in pediatric intensive care units. The mortality remains high despite advances in conventional mechanical ventilation and aetiological treatment. Several animal studies have documented lung injury during mechanical ventilation with high tidal volume, and clinical investigations have shown that in human ARDS, most ventilation is distributed to the small areas of remaining aerated lung resulting in overdistension of these areas and lung injury ("baby lung" theory). Nevertheless the usefulness of extrapulmonary gas exchange remains much debated. New ventilatory strategies have been developed in order to reduce ventilator-induced lung injury and to improve systemic oxygenation but multicentric randomized clinical trials are needed before these strategies can be validated.     

The effects of long-term prone positioning in patients with trauma-induced adult respiratory distress syndrome

Prone positioning improves gas exchange in some patients with adult respiratory distress syndrome (ARDS), but the effects of repeated, long-term prone positioning (20 h duration) have never been evaluated systemically. We therefore investigated 20 patients with ARDS after multiple trauma (Injury Severity Score [ISS] 27.3 +/- 10, ARDS score 2.84 +/- 0.42). Patients who fulfilled the entry criteria (bilateral diffuse infiltrates, severe hypoxemia, pulmonary artery occlusion pressure [PAOP] < 18 mm Hg, and PaO2/fraction of inspired oxygen [FIO2] < 200 mm Hg at inverse ratio ventilation with positive end-expiratory pressure [PEEP] > 8 mm Hg for more than 24 h) were turned to the prone position at noon and were turned back to the supine position at 8:00 AM on the next day. Thus only two turns per day were necessary, and the risk of disconnecting airways or medical lines was minimized. Prone positioning was repeated for another 20 h if the patients fulfilled the entry criteria. Except for FIO2, the ventilator settings remained unchanged during the study period. All patients were sedated and, if needed paralyzed to minimize patient discomfort. One hour before and after each position change, ventilator settings and pulmonary and systemic hemodynamics were recorded and blood was obtained for blood gas analysis. Derived cardiopulmonary and ventilatory variables were calculated using standard formulas. Overall mortality was 10%. Oxygenation variables improved significantly each time the patients were placed prone. Immediately after the first turn from the supine to the prone position the following changes were observed: PaO2 increased from 97 +/- 4 to 152 +/- 15 mm Hg, intrapulmonary shunt (Qva/Qt) decreased from 30.3 +/- 2.3 to 25.5 +/- 1.8, and the alveolar-arterial oxygen difference decreased from 424 +/- 24 to 339 +/- 25 mm Hg. All these changes were statistically significant. Most of these improvements were lost when the patients were turned supine, but could be reproduced when prone positioning was repeated after a short period (4 h) in the supine position. Short periods in the supine position were necessary to allow for nursing care, medical evaluation, and interventions such as placement of central lines. No position-dependent changes of systemic hemodynamic variables were observed. We conclude that, in trauma patients with ARDS undergoing long-term positioning treatment, lung function improves significantly during prone position compared to short phases of conventional supine position during which the beneficial effects are partly lost.     

Lung aeration and pulmonary gas exchange during lumbar epidural anaesthesia and in the lithotomy position in elderly patients

We investigated a total of 36 subjects with a mean (SD) age of 65 (13) years, during baseline conditions (supine, before any anaesthesia), and then during one of the following protocols: (1) lithotomy positioning (n = 12), (2) epidural anaesthesia (n = 12), (3) general anaesthesia in the supine position (n = 12). Lung aeration, ventilation/perfusion matching, gas exchange and functional residual capacity were measured. Lung aeration was normal during baseline assessment with almost no regions with poor aeration and no substantial dependent densities. Shunt and perfusion of poorly ventilated regions were minor. Lithotomy positioning did not reduce functional residual capacity and did not affect aeration of the lung or ventilation/perfusion matching. Epidural anaesthesia, in general, had no effect on aeration, ventilation/perfusion matching or gas exchange, regardless of whether the patient was in the supine or lithotomy position. General anaesthesia, however, caused significant increases in poorly aerated lung regions and in dependent densities (interpreted as atelectasis). In conclusion, no or little impairment of lung aeration and ventilation/perfusion matching was caused by the lithotomy position and/or epidural anaesthesia, contrary to the effects seen during general anaesthesia. However, our findings also suggest that being overweight is a factor that may cause impairment of lung aeration.     

Erwin Stransky and intrapsychic ataxy

A high-frequency ventilator was developed, consisting of a single-phase induction motor, an unbalanced mass and a mechanical vibration system. Intermittent positive pressure respiration was combined with high-frequency ventilation to measure end-tidal pCO2. Hysteresis was observed between the rotational frequency of the high-frequency ventilator and end-tidal pCO2. A fuzzy proportional plus integral control system, designed on the basis of the static characteristics of the controlled system and a knowledge of respiratory physiology, successfully regulated end-tidal pCO2. The characteristics of gas exchange under high-frequency ventilation was approximated by a first-order linear model. A conventional PI control system, designed on the basis of the approximated model, regulated end-tidal pCO2 with a performance similar to that of the fuzzy PI control system. The design of the fuzzy control system required less knowledge about the controlled system than that of the conventional control system.     

Adjuncts to mechanical ventilation

Adjunctive ventilatory strategies have been developed to improve oxygenation and carbon dioxide (CO2) removal during mechanical ventilation of critically ill patients. These techniques allow clinicians to attain their clinical goals at lower levels of ventilatory support. In this article, the authors discuss extracorporeal CO2 removal, venovenous intravena caval oxygenator, and tracheal gas insufflation as adjuncts to CO2 removal and nitric oxide, surfactant replacement therapy, perfluorocarbon-associated gas exchange, and prone positioning as adjuncts to oxygenation.     

Evaluation of lung function after intratracheal perfluorocarbon administration in healthy animals

OBJECTIVES: To investigate the effects of partial liquid ventilation (i.e., mechanical ventilation in combination with intratracheal administration of perfluorocarbon) on lung function, with particular attention to the integrity of the alveolocapillary membrane in healthy adult animals. DESIGN: Prospective, randomized, controlled study. SETTING: Laboratory at the Department of Experimental Anesthesiology, Erasmus University Rotterdam. SUBJECTS: Ten adult male New Zealand rabbits. INTERVENTIONS: Five rabbits were intratracheally treated with 12 mL/kg of perfluorocarbon while conventional mechanical ventilation (volume-controlled, tidal volume of 12 mL/kg, respiratory rate of 30 breaths/min, inspiration/expiration ratio of 1:2, positive end-expiratory pressure of 2 cm H2O, and an FIO2 of 1.0) was applied for 3 hrs. To assess the permeability of the alveolocapillary membrane, pulmonary clearance of inhaled technetium-99m-labeled diethylenetriamine pentaacetic acid (99mTc-DTPA) measurements were performed at 3 hrs and compared with data from the control group (n = 5) treated with mechanical ventilation only, using the same ventilatory parameters. MEASUREMENTS AND MAIN RESULTS: Pulmonary gas exchange and lung mechanical parameters were measured in both groups at 30-min intervals. Mean values for PaO2 in the perfluorocarbon group, although at adequate levels, were less than those values of the control group during the 3-hr study period (370 +/- 44 vs. 503 +/- 44 torr at 3 hrs [49.3 +/- 5.9 vs. 67.1 +/- 5.9 kPa]). Peak and mean airway pressures were higher in the perfluorocarbon group (ranging from 1.9 to 3.4 cm H2O and 0.7 to 1.3 cm H2O, respectively) compared with the control group, while end-inspiratory airway pressure was similar in both groups. The half-life of 99mTc-DTPA was 83.7 +/- 24.5 mins in the control group, which was significantly longer (p < .01) than in the perfluorocarbon group (49.8 +/- 6.1 mins). CONCLUSIONS: These findings suggest that partial liquid ventilation with perfluorocarbons lowers pulmonary gas exchange in healthy animals, and the increased pulmonary clearance of 99mTc-DTPA after 3 hrs of this type of ventilatory support may reflect minimal reversible changes in the lung surfactant system.     

Perfluorochemical liquid ventilation: from the animal laboratory to the intensive care unit

Perfluorochemical or perfluorocarbon liquids have an enormous gas-carrying capacity. During tidal liquid ventilation the respiratory medium of both functional residual capacity and tidal volume is replaced by neat perfluorocarbon liquid. Tidal liquid ventilation is characterized by convective and diffusive limitations, but offers the advantage of preserved functional residual capacity, high compliance and improved ventilation-perfusion matching. During partial liquid ventilation only the functional residual capacity is replaced by perfluorocarbon liquid. Both tidal and partial liquid ventilation improve gas exchange and lung mechanics in hyaline membrane disease, adult respiratory distress models and meconium aspiration. Compared to gas ventilation, there is less histologic evidence of barotrauma after liquid ventilation. Cardio-pulmonary interaction, inherent to the high density of liquid, and long term safety need further study. However, extrapolating from animal data, and taking into account promising human pilot studies, liquid ventilation has the desired properties to occupy an important place in the therapy of restrictive lung disease in man.     

Pulmonary gas-exchange analysis by using simultaneous deposition of aerosolized and injected microspheres

Numerical methods for determining end-capillary gas contents for ventilation-to-perfusion ratios were first developed in the late 1960s. In the 1970s these methods were applied to validate distributions of ventilation-to-perfusion ratios measured by the multiple inert-gas-elimination technique. We combined numerical gas analysis and fluorescent-microsphere measurements of ventilation and perfusion to predict gas exchange at a resolution of approximately 2.0-cm3 lung volume in pigs. Oxygen, carbon dioxide, and inert gas exchange were calculated in 551-845 compartments/animal before and after pulmonary embolization with 780-micrometers beads. Whole lung gas exchange was estimated from the perfusion- and ventilation-weighted end-capillary gas contents. Before lung injury, no significant difference existed between microsphere-estimated arterial PO2 and PCO2 and measured values. After lung injury, the microsphere method predicted a decrease in arterial PO2 but consistently underestimated its magnitude. Correlation between predicted and measured inert gas retentions was 0.99. Overestimation of low-solubility inert gas retentions suggests underestimation of areas with low ventilation-to-perfusion ratios by microspheres after lung injury. Regional deposition of aerosolized and injected microspheres is a valid method for investigating regional gas exchange with high spatial resolution.     

Computer-controlled minute ventilation in preterm infants undergoing mechanical ventilation

INTRODUCTION: Computer-controlled minute ventilation (CCMV) continuously adjusts the ventilator rate to changes in spontaneous respiratory drive and pulmonary mechanics to maintain a preset total minute ventilation. HYPOTHESIS: We hypothesized that CCMV would maintain ventilation and oxygenation with fewer mechanical breaths than conventional intermittent mandatory ventilation in very low birth weight infants. METHODS: Very low birth weight infants in clinically stable condition who were undergoing mechanical ventilation were enrolled. The number of mechanical breaths, total and mechanical expiratory minute ventilation, mean airway pressure, oxygen hemoglobin saturation by pulse oximetry, and transcutaneous partial carbon dioxide and partial oxygen tensions were obtained during intermittent mandatory ventilation and CCMV (45 to 60 minutes) and compared by paired t test. RESULTS: Fifteen infants were studied. Birth weight (median, range) was 700 gm (550 to 1205 gm), gestational age 26 weeks (23 to 34 weeks), age 21 days (3 to 50 days). When switched from intermittent mandatory ventilation to CCMV, the number of mechanical breaths was reduced (15 +/- 2.8 to 8.6 +/- 2.9 breaths per minute, p < 0.001), leading to lower airway pressure (3.97 +/- 1.00 to 3.45 +/- 1.00 cm H2O, p < 0.001) and lower expiratory minute ventilation generated by the mechanical ventilator (116 +/- 31 to 65 +/- 28 ml/min per kilogram, p < 0.001), while total expiratory minute ventilation remained unchanged. Mean transcutaneous partial carbon dioxide and oxygen tensions, oxygen hemoglobin saturation, and the time spent within different oxygen hemoglobin saturation ranges did not differ between both ventilatory modes. CONCLUSION: CCMV maintained adequate ventilation and oxygenation with lower mechanical ventilatory support than IMV. CCMV may reduce barotrauma and chronic lung disease during long-term use.     

Noninvasive positive pressure ventilation: what is its role in treating acute respiratory failure?

Noninvasive positive pressure ventilation (NIPPV) is a viable option in treating appropriately selected patients with acute respiratory failure. It is often well tolerated, and it avoids endotracheal intubation with its potential complications. Moreover, gas exchange is reportedly improved. Several issues relating to the use of NIPPV are unresolved, however. The optimal interface, best ventilator mode, and patient selection criteria have not been firmly established. Also, studies are needed to compare the efficacy, safety, and cost-effectiveness of NIPPV and standard endotracheal ventilation. Despite these unresolved issues, NIPPV clearly represents an important addition to the techniques available in managing acute respiratory failure. Except in situations in which immediate endotracheal intubation is required, it may become first-line therapy in elderly patients in whom resuscitation status is unsettled.     

Ventilation of patients with asthma and obstructive lung disease

Mechanical ventilation in a patient with obstructive airway disease may be a lifesaving measure; however, it may also be associated with significant morbidity and mortality. It is important for a physician to be familiar with the potential complications of mechanical ventilation in this group of patients and to know how to avoid them by carefully applying safe ventilator strategies. The cornerstone of such strategies is to minimize minute ventilation, maximize time for expiration, and avoid hyperinflation of the lung. Several bedside parameters (iPEEP, VEI, Pplat) that reflect presence of gas trapping and potential hyperinflation may be measured. In addition to mechanical ventilation, management should include inhaled bronchodilators and systemic corticosteroid therapies. In the event controlled hypoventilation is necessary, sedation with or without the use of muscle relaxants may be required. Unconventional therapies such as the use of Heliox, magnesium sulfate, ketamine, and inhalational anesthetics may be attempted in severe cases that do not respond to conventional management. With appropriate use of ventilator strategies, a reduction in the mortality and morbidity of patients with obstructive airway disease requiring mechanical ventilation has recently been noted.     

Growth hormone treatment prevents the decrease in insulin-like growth factor I gene expression in patients undergoing abdominal surgery

Gas exchange is improved during partial liquid ventilation with perfluorocarbon in animal models of acute lung injury. The specific mechanisms are unproved. We measured end-expiratory lung volume (EELV) by null-point body plethysmography in anesthetized sheep. Measurements of gas exchange and EELV were made before and after acute lung injury was induced with intravenous oleic acid to decrease EELV and worsen gas exchange. Measurements of gas exchange and EELV were again performed after partial liquid ventilation with 30 ml/kg of perfluorocarbon and compared with gas-ventilated controls. Oxygenation was significantly improved during partial liquid ventilation, and EELV (composite of gas and liquid) was significantly increased, compared with preliquid ventilation values and gas-ventilated controls. We conclude that partial liquid ventilation may directly recruit consolidated alveoli in the lung-injured sheep and that this may be one mechanism whereby gas exchange is improved.     

Out-of-hospital ventilation: bag--valve device vs transport ventilator

OBJECTIVE: To examine the patterns of out-of-hospital airway management and to compare the efficacy of bag-valve ventilation with that of the use of a transport ventilator for intubated patients. METHODS: A prospective, nonrandomized, convenience sample of 160 patients requiring airway management in the out-of-hospital urban setting was analyzed. A survey inquiring about airway and ventilatory management was completed by emergency medical services (EMS) personnel, and arterial blood gas (ABG) samples were obtained within 5 minutes of patient arrival in the ED. The ABG parameters were compared for patients grouped by different airway techniques and presence or absence of cardiac arrest (systolic blood pressure < 50 mm Hg) upon ED presentation. RESULTS: Over a one-year period, 160 surveys were returned. The majority (62%) of the patients were men; the population mean age was 61 +/- 19 years. Presenting ABGs were obtained for 76 patients; 17% (13/76) had systemic perfusion and 83% (63/76) were in cardiac arrest. There was no difference in ABG parameters between the intubated cardiac arrest patients ventilated with a transport ventilator (pH 7.17 +/- 0.17, PaCO2 37 +/- 20 torr, and PaO2 257 +/- 142 torr) and those ventilated with a bag-valve device (pH 7.20 +/- 0.16, PaCO2 42 +/- 21 torr, and PaO2 217 +/- 138 torr). The patients ventilated via an esophageal obturator airway (EOA) device had impaired gas exchange, compared with the groups who had endotracheal (ET) intubation (pH 7.09 +/- 0.13, PaCO2 76 +/- 30 torr, and PaO2 75 +/- 35 torr). The intubated patients not in cardiac arrest had similar ABG parameters whether ventilated manually with a bag-valve device or with a transport ventilator. Endotracheal intubation was successfully accomplished in 93% (123/132) of attempted cases. CONCLUSIONS: In this sample, ET intubation was the most frequently used airway by EMS providers. When ET intubation was accomplished, adequate ventilation could be achieved using either bag-valve ventilation or a transport ventilator. Ventilation via the EOA proved inadequate.     

High-frequency partial liquid ventilation in respiratory distress syndrome: hemodynamics and gas exchange

Partial liquid ventilation using conventional ventilatory schemes improves lung function in animal models of respiratory failure. We examined the feasibility of high-frequency partial liquid ventilation in the preterm lamb with respiratory distress syndrome and evaluated its effect on pulmonary and systemic hemodynamics. Seventeen lambs were studied in three groups: high-frequency gas ventilation (Gas group), high-frequency partial liquid ventilation (Liquid group), and high-frequency partial liquid ventilation with hypoxia-hypercarbia (Liquid-Hypoxia group). High-frequency partial liquid ventilation increased oxygenation compared with high-frequency gas ventilation over 5 h (arterial oxygen tension 253 +/- 21.3 vs. 17 +/- 1.8 Torr; P < 0.001). Pulmonary vascular resistance decreased 78% (P < 0.001), pulmonary blood flow increased fivefold (P < 0.001), and aortic pressure was maintained (P < 0.01) in the Liquid group, in contrast to progressive hypoxemia, hypercarbia, and shock in the Gas group. Central venous pressure did not change. The Liquid-Hypoxia group was similar to the Gas group. We conclude that high-frequency partial liquid ventilation improves gas exchange and stabilizes pulmonary and systemic hemodynamics compared with high-frequency gas ventilation. The stabilization appears to be due in large part to improvement in gas exchange.