Reproducibility of plasma and extracellular fluid volume measurements in critically ill patients.

Plasma and extracellular fluid (ECF) volume measurements may provide valuable complementary data to the hemodynamic measurements currently used to compare fluid infusions in critically ill patients. To assess the reproducibility of plasma and extracellular fluid volume measurements in critically ill patients, we injected 131I-labeled albumin (10 microCi) and 35S-sodium sulfate (50 microCi), respectively, into 15 stable patients on two occasions 150 min apart. Plasma was sampled at 20, 30, and 40 min after each injection and the volume of distribution of each radioisotope was calculated from the extrapolated zero time counts. We found that plasma and ECF volume did not differ significantly between the first (42.4 +/- 4.7 ml/kg and 186 +/- 39 ml/kg) and second (42.8 +/- 5.5 ml/kg and 193 +/- 48 ml/kg) measurements. Specifically, the mean difference between the two measurements was 0.4 +/- 3.2 ml/kg and 7 +/- 17 ml/kg respectively. We conclude that measurements of plasma and ECF volume are reproducible over 150 min in stable critically ill patients.

Physiologic effects and side effects of prostaglandin E1 in the adult respiratory distress syndrome.

Because PGE1 previously has been reported to increase survival of patients with ARDS, we evaluated physiologic effects and side effects of PGE1 in a prospective open-label study of patients with ARDS. Seventeen patients with ARDS who did not have significant renal or hepatic dysfunction received PGE1 by continuous central venous infusion (30 ng/kg/min). Seventeen control patients with ARDS without renal or hepatic dysfunction who had similar APACHE II and ARDS scores and causes of ARDS did not receive PGE1. Prostaglandin E1 significantly decreased the SVRI and oxygen extraction ratio. Concentrations of total and polymorphonuclear leukocytes, but not platelets, increased significantly during PGE1 infusion, but did not change in control patients. There was no change in the Do2I and Vo2I during the course of the PGE1 infusion. There were no differences in Do2I and Vo2I during PGE1 infusion between survivors and nonsurvivors. Prostaglandin E1 was infused for a mean of 5.9 +/- 1.8 days (+/- SD) and was discontinued on ten occasions in seven patients because of supraventricular dysrhythmias (n = 4), hypotension (n = 3), thrombocytopenia (n = 3), and cardiac arrest (n = 2). Nonsurvivors had PGE1 discontinued prematurely more frequently than survivors (56 percent [5/9] vs 25 percent [2/8], respectively). The prevalence of multiple-system organ failure and the in-hospital mortality of both PGE1-treated and control patients were not different. Although PGE1 causes significant systemic vasodilation and possibly decreased intrapulmonary polymorphonuclear leukocyte sequestration, PGE1 does not influence multiple-system organ failure or mortality of patients with ARDS without renal or hepatic dysfunction.

Plasma ionized calcium and blood lactate concentrations are inversely associated in human lactic acidosis.

Plasma ionized calcium [Ca++] concentrations are decreased in patients having lactic acidosis. To further investigate this observation, we prospectively studied nine critically ill patients who had lactic acidosis and measured arterial pH, PCO2, [Ca++], lactate, and albumin concentrations. We found a strong association between decreased [Ca++] and increased plasma lactate concentrations (r2 = 0.78, p less than or equal to 0.001). This unexpected association--[Ca++] usually increases with increasing acidosis--might be clinically important and the mechanism deserves further investigation.

Blood flow to the trachea and bronchi: the pulmonary contribution.

It is generally assumed that when pulmonary vascular pressures are normal the pulmonary contribution to central airway blood flow (Qp) is negligible compared with systemic blood flow (Qs). However, it has been suggested in recent reports that a substantial portion of central airway blood flow is Qp. We have attempted to confirm whether there is a pulmonary contribution to central airway blood flow and to describe how it is anatomically distributed. Measurements of Qp were made using the radioactive microsphere technique in anesthetized ventilated dogs (n = 7) and sheep (n = 6). Qs to the central airways was also measured in another group of sheep (n = 10). At the end of each study, animals were killed and the lungs and trachea were excised. Qp to the upper and lower trachea, mainstem bronchi, and lobar bronchi was calculated, and the relative distribution of Qp and Qs to mucosa, cartilage, and adventitia was determined. Results showed a progressive increase (P < 0.01) in Qp (in ml.min-1 x 100 g-1) from upper trachea to lobar bronchi in both dogs and sheep. Qp and Qs supplied mainly the airway adventitia and the mucosa, respectively. Expressed as a percentage, 89 +/- 4% (SE) of Qp was to the adventitia and 0.1 +/- 0.07% was to the mucosa (P < 0.01), whereas 60 +/- 3.2% of Qs was to the mucosa and 22 +/- 4.6% was to the adventitia. In conclusion, in dogs and sheep there is a pulmonary contribution to central airway blood flow that increases from upper trachea to lobar bronchi and that supplies mainly the peritracheal adventitia.

Distribution of blood flow and neutrophil kinetics in bronchial vasculature of sheep.

The bronchial circulation, as opposed to the pulmonary circulation, is the likely source of the edema and inflammatory cells that contribute to airflow obstruction and airway narrowing associated with asthma and pulmonary edema. The purpose of this study was to understand the mechanism of edema formation and inflammation in airway walls. Therefore, we sought first to determine the normal bronchial venous drainage pathways. In anesthetized, ventilated, open-chest sheep we measured the relative distribution of 51Cr-labeled red blood cells to the right and left ventricles after injection into the bronchial artery (n = 7). Using this information, we then studied the kinetics of leukocytes in the bronchial vascular bed. We measured the extraction of 111In-labeled neutrophils during their first pass through the microvasculature after injection into the bronchial artery or right ventricle (n = 6). In the first set of experiments, we found > 85% of the systemic blood flow to the lung returns to the left ventricle. In the second set of experiments, we found that extraction of neutrophils in the bronchial vasculature (50-60%) was less (P < 0.05) than that in the pulmonary vasculature (80%). This finding may be explained by differences in the anatomy and/or hydrodynamic dispersal forces between the pulmonary and bronchial vascular beds or may reflect sequestration of neutrophils within the pulmonary microvasculature while traversing bronchial-to-pulmonary anastomotic pathways.

Bronchial vascular occlusion does not attenuate or accentuate oleic acid lung injury in anesthetized sheep.

To determine if bronchial blood flow affects the consequences of acute pulmonary vascular injury, we studied oleic acid lung injury in 12 anesthetized sheep. In six sheep (group 1), we injected 2 ml of ethanol directly into the bronchoesophageal artery to decrease bronchial blood flow. In the control sheep (group 2), we injected 2 ml of normal saline. One hour later, oleic acid (0.1 ml/kg) was injected into the right ventricle in both groups. We measured hemodynamics and lung mechanics at baseline, 1 h after injection into the bronchoesophageal artery but just before the injection of oleic acid, and 3 h after injection of oleic acid. We measured bronchial blood flow at baseline and 3 h after injection of oleic acid and extravascular lung water at 3 h after injection of oleic acid. One hour after injection of ethanol or saline into the bronchoesophageal artery, hemodynamics and lung mechanics did not change. Three hours after injection of oleic acid, systemic arterial pressure decreased, pulmonary arterial pressure increased, cardiac output decreased, dynamic compliance decreased, pulmonary resistance increased, arterial oxygen tension decreased, and extravascular lung water was greater than normal. There were no differences in these measurements between the two groups. However, bronchial blood flow decreased only in group 1. We conclude that decreasing bronchial blood flow does not attenuate or accentuate the consequences of oleic acid lung injury.

Kinetics of radiolabeled neutrophils in swine.

The kinetics of radiolabeled neutrophils (PMNs) as they pass through the lungs of swine were evaluated and compared with those in rabbits (J. Appl. Physiol. 63: 1806-1815, 1987) and dogs (J. Appl. Physiol. 63: 1253-1261, 1987; 65: 1217-1225, 1988) previously reported from our laboratory. 111In-labeled PMNs (111In-PMNs) and 99mTc-labeled erythrocytes were simultaneously injected into the right atrium, and the 111In-PMN percent extraction on the first passage through the lung was determined by the indicator-dilution technique. After 10 min of circulation the distribution of 111In-PMNs in selected organs was determined. The extraction of 111In-PMNs in swine was 88 +/- 3%, which was significantly greater than that of rabbits (78 +/- 3%) or dogs (72 +/- 2%). The recovery of the 111In-PMNs in the lungs of swine was 60 +/- 7%, which was two to three times higher than the recovery in lungs of rabbits or dogs. These results show that radiolabeled PMNs injected intravenously are less able to pass through the pulmonary vasculature and are retained much more within the lung in swine than in rabbits or dogs. This difference could be the result of the presence of pulmonary intravascular macrophages in the lungs of swine.

Effects of beta-adrenergic agents in lungs of normal and air-embolized awake sheep.

It is unclear whether beta-adrenergic agonists or antagonists affect lung liquid and protein exchange by changing pulmonary hemodynamics or microvascular leakiness. In 23 unanesthetized, instrumented sheep with long-term lung lymph fistulas, we assessed the effect of the beta-agonist terbutaline or the beta-antagonists propranolol, nadolol, and atenolol, all infused intravenously, on lung lymph flow under base-line conditions and during the acute lung injury caused by 4 h of venous air embolism. Under base-line conditions, neither beta-stimulation nor blockade had any effect. During air embolism, terbutaline decreased pulmonary vascular resistance and lymph flow by 25%. Propranolol and nadolol (non-selective beta 1,beta 2-antagonists) but not atenolol (selective beta 1-antagonist) also decreased lymph flow by 22% on average. We favor the more conservative (hemodynamic) over the more liberal (altered permeability) explanation for our results. First, beta-stimulation clearly caused vasodilation, which lowered the pulmonary microvascular pressure at the site of injury. beta-blockade caused changes similar to alpha-stimulation (J. Appl. Physiol. 62: 2147-2153, 1987). We therefore interpret the beta-blockade as unmasking pulmonary arterial alpha-receptors stimulated by the air-embolism injury, thus allowing vasoconstriction upstream to the site of injury. We do not believe the explanation of the beta-agent effects requires any modulation of lung microvascular leakiness by beta-adrenergic agents.

Systemic blood flow to the lung after bronchial artery occlusion in anesthetized sheep.

To compare the effectiveness of different embolizing agents in reducing or redistributing bronchial arterial blood flow, we measured systemic blood flow to the right lung and trachea in anesthetized sheep by use of the radioactive microsphere method before and 1 h after occlusion of the bronchoesophageal artery (BEA) as follows: injection of 4 ml ethanol (ETOH) into BEA (group 1, n = 5), injection of approximately 0.5 g polyvinyl alcohol particles (PVA) into BEA (group 2, n = 5), or ligation of BEA (group 3, n = 5). After occlusion, angiography showed complete obstruction of the bronchial vessels. There were no changes in tracheal blood flow in any of the groups. Injection of ETOH produced a 75 +/- 14% (SD) reduction in flow to the middle lobe (P less than 0.02) and a 75 +/- 13% reduction to the caudal lobe (P less than 0.01), whereas injection of PVA produced a smaller reduction in flow to these two lobes (41 +/- 66 and 51 +/- 54%, respectively). After BEA ligation there was a 52 +/- 29% reduction in flow to the middle lobe and a 53 +/- 38% reduction to the caudal lobe (P less than 0.05). This study has significant implications both clinically and experimentally; it illustrates the importance of airway collateral circulation, in that apparently complete radiological obstruction of the BEA does not necessarily mean complete obstruction of systemic blood flow. We also conclude that, in experimental studies in which the role of the bronchial circulation in airway pathophysiology is examined, ETOH is the agent of choice.

Direct injury to the bronchial vasculature in anesthetized sheep.

Injury to the bronchial vasculature may contribute to liquid and solute leakage into the lung during noncardiac pulmonary edema. The purpose of this study was to measure changes in hemodynamics, pulmonary mechanics, extravascular lung water, and lung morphometry after selectively injuring the bronchial vasculature in anesthetized sheep. In two groups of seven sheep, we injected oleic acid (0.1 ml/kg) or normal saline directly into the bronchoesophageal artery. We measured systemic and pulmonary arterial pressures, cardiac output, oxygen saturation, pulmonary resistance and compliance, and lung volumes before and 1 and 4 h after injection. The lungs were removed for measurement of extravascular water, histology, and morphometry. Four hours after injection of oleic acid, cardiac output decreased but pulmonary arterial pressure did not change. In addition, pulmonary resistance increased and dynamic compliance and vital capacity decreased. Extravascular lung water was slightly but significantly greater in the oleic acid group. Histological examination showed interstitial edema and leukocytes in airway walls and sloughing of bronchial epithelium but little or no alveolar edema. Morphometric analysis showed significant thickening of airway walls. We conclude that direct injury to the bronchial vasculature increases lung resistance, decreases dynamic compliance, and increases extravascular lung water by the accumulation of an inflammatory infiltrate in airway walls.

Logistic regression model to predict outcome after in-hospital cardiac arrest: validation, accuracy, sensitivity and specificity.

OBJECTIVE: To develop and validate a logistic regression model to identify predictors of death before hospital discharge after in-hospital cardiac arrest. DESIGN: Retrospective derivation and validation cohorts over two 1 year periods. Data from all in-hospital cardiac arrests in 1986-87 were used to derive a logistic regression model in which the estimated probability of death before hospital discharge was a function of patient and arrest descriptors, major underlying diagnosis, initial cardiac rhythm, and time of year. This model was validated in a separate data set from 1989-90 in the same hospital. Calculated for each case was 95% confidence limits (C.L.) about the estimated probability of death. In addition, accuracy, sensitivity, and specificity of estimated probability of death and lower 95% C.L. of the estimated probability of death in the derivation and validation data sets were calculated. SETTING: 560-bed university teaching hospital. PATIENTS: The derivation data set described 270 cardiac arrests in 197 inpatients. The validation data set described 158 cardiac arrests in 120 inpatients. INTERVENTIONS: none. MEASUREMENTS AND RESULTS: Death before hospital discharge was the main outcome measure. Age, female gender, number of previous cardiac arrests, and electrical mechanical dissociation were significant variables associated with a higher probability of death. Underlying coronary artery disease or valvular heart disease, ventricular tachycardia, and cardiac arrest during the period July-September were significant variables associated with a lower probability of death. Optimal sensitivity and specificity in the validation set were achieved at a cut-off probability of 0.85. CONCLUSIONS: Performance of this logistic regression model depends on the cut-off probability chosen to discriminate between predicted survival and predicted death and on whether the estimated probability or the lower 95% C.L. of the estimated probability is used. This model may inform the development of clinical practice guidelines for patients who are at risk of or who experience in-hospital cardiac arrest.

Time-weighted nursing demand is a better predictor than midnight census of nursing supply in an intensive care unit.

PURPOSE: Labor costs are the largest fraction of operating costs in an intensive care unit (ICU). Estimation of appropriate nursing supply is frequently based on the midnight census of patients, which is a "snapshot" view of the ICU. We postulated that the midnight census would not correlate as well as time-weighted nursing demand (a calculation of need for nursing staff) with the actual number of nurses who were required to staff the ICU (nursing supply). The purpose of this study was to compare the correlation between midnight census and actual nursing supply with the correlation between time-weighted nursing demand and nursing supply. MATERIALS AND METHODS: We measured nursing activity, midnight census, and actual nursing supply for each of 77 consecutive days in a 14-bed medical-surgical ICU within a 450-bed tertiary care teaching hospital. We calculated time-weighted nursing demand based on 1:1 nursing for ICU patients, 1:2 nursing for step-down patients, 0.5 additional nurse hours for each cardiac arrest, and 0.5 additional nurse hours for each new admission to the ICU. RESULTS: There was a correlation between midnight census and nursing supply (r2 = .42, P<.0001) and between nursing demand and nursing supply (r2 = .83, P<.0001). The correlation coefficient for the relationship between nursing demand and nursing supply was significantly greater than that for the relationship between midnight census and nursing supply (P<.01). CONCLUSIONS: Time-weighted nursing demand is a better predictor than midnight census of nursing supply in an ICU.

Mind the gap: Opportunities for improving end-of-life care for patients with advanced heart failure.

BACKGROUND: Patients with advanced heart failure (HF) experience progressive symptoms, decreased quality of life, and more frequent hospitalizations as they approach the end of life (EOL). Understanding patient perspectives and preferences regarding EOL issues is necessary to identify key opportunities for improving care. OBJECTIVE: To identify, from the patient's perspective, the major opportunities for improving EOL care for patients hospitalized because of advanced HF. METHODS: A cross-sectional survey of patient perspectives regarding EOL care was administered via interview of 106 hospitalized patients who had advanced HF in five tertiary care centres across Canada. The study compared which aspects of EOL care patients rated as 'extremely important' and their level of satisfaction with these aspects of EOL care to identify key opportunities for improvement of care. RESULTS: The greatest opportunities for improvement in EOL care were reducing the emotional and physical burden on family, having an adequate plan of care following discharge, effective symptom relief and opportunities for honest communication. The three most important issues ranked by patients were avoidance of life support if there was no hope for a meaningful recovery, communication of information by the doctor and avoidance of burden for the family. CONCLUSIONS: Advanced care planning that seamlessly bridges hospital and home must be standard care for patients who have advanced HF. Components must include coordination of care, caregiver support, comprehensive symptom management, and effective communication regarding HF and EOL issues.

The role of HIV infection and drug and alcohol dependence in hospital mortality among critically ill patients.

PURPOSE: Critical care outcomes among HIV-infected patients have improved because of advances in HIV therapy and general improvements in intensive care unit (ICU) management. There is a high co-occurrence of drug and alcohol dependence among HIV-infected patients, and the independent role of drug and alcohol dependence among patients with and without HIV infection in outcomes of critical illness is unclear. MATERIALS AND METHODS: We analyzed a prospectively collected database of 7015 index ICU admissions at 2 teaching hospitals between January 1999 and January 2006. The ICU diagnoses were determined from prospective chart review and classified according to the dictionary of diagnoses developed by the Intensive Care National Audit and Research Council. We used logistic regression to determine the independent association of drug and alcohol dependence as well as HIV infection with in-hospital mortality. Covariates that were adjusted for included acute drug overdose, Acute Physiology and Chronic Health Evaluation II score, age, sex, hospital site, and socioeconomic variables. RESULTS: Of all patients, 4.4% (309 of 7015) were HIV infected; and of these, 56% (173 of 309) had a history of drug and alcohol dependence, whereas only 7.4% (502 of 6706) of the HIV-negative group had a history of drug and alcohol dependence. Drug and alcohol dependence was not independently associated with hospital mortality in either the model including all admissions (adjusted odds ratio [AOR] 0.80; 95% confidence interval [CI] 0.62-1.03) or the model including pneumonia and sepsis admissions only (AOR 0.92; 95% CI 0.59-1.41). Infection with HIV was independently associated with hospital mortality (AOR 2.16; 95% CI 1.60-2.93). CONCLUSIONS: Although HIV infection is associated with increased hospital mortality, drug and alcohol dependence is not associated with an increased hospital mortality independent of HIV infection.

Distribution of normal saline and 5% albumin infusions in septic patients.

OBJECTIVE: To determine the relative distribution of fluid within the extracellular fluid volume (ECFV) after infusing either normal saline or 5% albumin in septic, critically ill patients. DESIGN: Prospective, randomized, unblinded, interventional study. SETTING: Intensive care unit in a 450-bed, tertiary care, teaching hospital. PATIENTS: Septic, critically ill patients (n = 18). INTERVENTIONS: Infusion of either normal saline or 5% albumin to a hemodynamic end point determined by the patient's clinician. MEASUREMENTS AND MAIN RESULTS: Plasma volume (PV), ECFV, cardiac index, and arterial oxygen content were measured immediately before (baseline) and after each fluid infusion. PV and ECFV were measured by dilution of 131I-albumin and 35S sodium sulfate, respectively. Interstitial fluid volume (ISFV) was calculated as ECFV - PV. Baseline values for PV, ISFV, ECFV, and oxygen delivery index did not differ between treatment groups. Infusion of normal saline increased the ECFV by approximately the volume infused, and the expansion of the PV to ISFV was in a ratio of 1:3. Infusion of 5% albumin increased the ECFV by double the volume infused, with both the PV and ISFV expanding by approximately equal amounts. Oxygen delivery index did not increase after either infusion due to the effect of hemodilution. CONCLUSION: Expansion of the ECFV in excess of the volume of 5% albumin infused suggests that fluid may move from the intracellular fluid volume to the ECFV in septic patients who receive this fluid.

Distribution of normal saline and 5% albumin infusions in cardiac surgical patients.

OBJECTIVE: To determine the relative distribution of fluid within the extracellular fluid volume (ECFV) and the effect on oxygen delivery after infusing either normal saline or 5% albumin in cardiac surgical patients. DESIGN: Prospective, randomized, unblinded, interventional study. SETTING: Cardiac surgical intensive care unit in a 450-bed teaching hospital. PATIENTS: Postoperative cardiac surgical patients (n = 40). INTERVENTIONS: Infusion of either normal saline or 5% albumin to a hemodynamic end point determined by the patient's clinician. MEASUREMENTS AND MAIN RESULTS: Plasma volume (PV), ECFV, cardiac index, and arterial oxygen content were measured immediately before (baseline) and after each fluid infusion. PV and ECFV were measured by dilution of (131)I-albumin and [(35)S]sodium sulfate, respectively. Interstitial fluid volume (ISFV) was calculated as ECFV - PV. Baseline values for PV, ISFV, ECFV, and oxygen delivery index did not differ between treatment groups. Infusion of normal saline and 5% albumin increased PV by 9 +/- 23% and 52 +/- 84% of the volume infused, respectively (p <.05), whereas there was no significant difference between saline and albumin in the change in ISFV per volume infused. Only 5% albumin significantly increased cardiac index, although oxygen delivery did not change significantly after either infusion. CONCLUSIONS: In postoperative cardiac surgical patients, infusion of 5% albumin is approximately five times as efficient as a PV expander but has comparable effects on changes in ISFV and oxygen delivery relative to normal saline.

Systolic and diastolic learning: an analogy to the cardiac cycle.

It has been observed that the active-passive classification of adult learning can be viewed in terms of a systolic-diastolic model. This model represents an analogy to the cardiac cycle and the work done by the heart during these two phases of the cycle. The determinants of systolic and diastolic learning can be compared to the determinants of cardiac function: preload, afterload and contractility. Similarly, dysfunction in these two phases of learning can be compared to cardiac dysfunction from a pathophysiologic perspective.

One-step isolation of neutrophils using an elutriator.

To simplify the isolation of neutrophils, we developed a one-step procedure using elutriation. The perfusate (0.2% gelatin and 0.1% glucose in phosphate buffered saline) was pumped through an elutriator rotor at 4 ml/min (25 degrees C) with the rotor speed at 2370 rpm. Twenty milliliters of anticoagulated porcine venous blood were mixed with 60 ml of perfusate and loaded into the elutriator chamber. The flow rate was increased by 2 ml/min increments and 100-ml fractions of effluent were collected at each increment. Concentrations of neutrophils and mononuclear cells were measured in each fraction, and the percentage of total neutrophils or mononuclear cells was plotted against flow rate. The optimal yield (46%) and purity (95.1%) of neutrophils (n = 8) was obtained in pooled fractions at flow rates greater than 20 ml/min. Neutrophils in this preparation were round, the granules were intact, and the nuclei were lobulated. In addition, the cells produced superoxide in the presence of phorbol myristate acetate and phagocytosed zymosan particles. These characteristics were similar to those of porcine neutrophils prepared by a conventional sedimentation method. The yield (43%) and purity (94%) of human neutrophils isolated using the elutriator method was similar to that for porcine cells. This one-step method provides a moderate yield of pure neutrophils that have retained their morphology and function.

Thromboxane and prostacyclin in acute lung injury caused by venous air emboli in unanesthetized sheep.

In 4 unanesthetized sheep with lung lymph fistulas we studied the role of thromboxane A2 (TxA2) in the lung microvascular injury caused by 4 h of venous air embolism. In each sheep, we did paired air embolism experiments: control and after giving the TxA2-synthetase inhibitor, U63,557A. Inhibition of TxA2-synthetase did not affect the hemodynamic response to air emboli; lung lymph flow increased slowly but to the same final rate as with air alone. The pulmonary production of TxA2, estimated as the lymph concentration of thromboxane B2 (TxB2) X lymph flow, increased during air embolism in the control experiment but increased much less when TxA2-synthetase was inhibited. Similarly, prostacyclin (PGI2) production increased during air embolism but was reduced when TxA2-synthetase was inhibited. Our results suggest that TxA2 does not mediate the pulmonary hypertension caused by venous air emboli, but it might modulate the increase in lymph flow by acting on the lymphatic pump or affecting hemodynamics in the pulmonary microcirculation. Our results do not support that TxA2 mediates the injury caused by venous air embolism. Pulmonary production of PGI2 seems to depend on the pulmonary production of TxA2, as shown by the parallel changes of their metabolites during air embolization.

Effects of plasmapheresis and of hypoproteinemia on lung liquid conductance in awake sheep.

The Starling equation, which describes net transvascular liquid flow, does not include the possibility that a reduction in plasma protein concentration may have a direct effect on lung liquid conductance or microvascular protein permeability. Nevertheless, both effects have been reported. Since these results were not predictable, we wondered whether the changes were due to the decrease in plasma protein concentration or to the process by which protein depletion was accomplished (batch plasmapheresis which involves considerable handling of blood). To separate these factors, we did control (sham) and protein removal plasmapheresis in awake sheep by two plasmapheresis methods (batch and continuous-flow). We monitored pulmonary hemodynamics, measured lung lymph flow, and determined protein concentrations in lymph and plasma. We calculated or measured the protein osmotic pressures of plasma and lymph. After control plasmapheresis, lymph flow increased and lymph:plasma protein concentration decreased but had returned to baseline levels by 4 hours. After protein removal plasmapheresis, the changes persisted for 24 hours. However, lung microvascular conductance (filtration coefficient) was not increased, except during the first 4-hour period. The changes in lymph flow and protein concentration ratio are explained using a simple two-pore model. We conclude that, over the range studied, hypoproteinemia does not increase lung microvascular liquid conductance or protein permeability.