The use of microsoft excel as a user interface for biological simulations

We used Microsoft Excel 4.0 for Windows running on a PC-486 to develop a user interface for two biological simulation models: a lung fluid balance model and a fractal model of the pulmonary circulation. The simulation programs were written in the C programming language, while the user interface was written in the macro language of Excel. The interface builds input data files for the simulation programs and provides a mechanism for displaying relevant information from output files produced from the simulations. Input fields are partially protected so that the user cannot modify certain portions of the spreadsheet. The Excel interface is used to build models from different available components and to select appropriate parameters for these models. The developed interface was also useful for running models in the batch mode. After selecting changes in lung fluid balance parameters, the interface allows users to find new steady state values by automatically running the model and adjusting initial conditions. Several different graphical options allow users to easily investigate the effects of selecting particular models and parameters. Techniques used in developing our user interface can be extended to most biological simulation programs which manipulate input and output data files.

Parameter identification in coronary pressure flow models: a graphical approach

The confident identification of parameters is important in the practical application of physiological models. However, the task of parameter identification is often complicated by interactions among parameters and by the fact that the sensitivity of the model to changes in a given parameter is generally a function of all the other parameters. Here we illustrate a graphical approach to parameter identification that allows the modeler to visualize the behavior of the model, the sensitivity functions, and certain functions characteristic of parameter interdependence. The visual display can be generated over any desired portion of parameter space. The technique is applied to a simple, four-parameter, myocardial pump model of the coronary circulation. The results indicate that over specified ranges of parameters, it is possible to distinguish among the four parameters of the model: the ratio of proximal-to-distal resistance, alpha; the overall resistance of the vascular bed, R; the compliance of the vascular bed, C; and a parameter, kappa, relating tissue pressure to left ventricular pressure. It was found that in order to identify all parameters uniquely, it was necessary to regress upon both coronary inflow and outflow.

Use of scaling theory to relate measurements of lung endothelial barrier permeability

This work examined the relationships between lung microvascular permeability-surface area products (PS) for small solutes in animals of different size and for columns of endothelial-covered microcarrier beads. We assembled PS data (humans, sheep, lambs, and rabbits) for labeled sucrose, mannitol, urea, 1,2-propanediol, 1,3-propanediol, and 1,4-butanediol. In addition, PS for cell columns using sucrose, mannitol, and sodium fluorescein were evaluated. A new mathematical model for the analysis of cell columns that accounts for transit time variations was derived and compared with models neglecting this variation. Allometric relationships between PS and body weight or exchange surface (S) were examined. Permeability relative to diffusivity (P/D) correlated inversely with S for all animals. In addition, P/D for the cell columns fell near this regression line. The results suggest either that permeability for hydrophilic tracers is higher for smaller animals or that the indicator-dilution measurement is a fractal process dependent on scale. Furthermore, the P/D-S correlations may help relate cell column experiments to animal studies.

Effects of pulmonary blood flow on the fractal nature of flow heterogeneity in sheep lungs

The spatial heterogeneity of pulmonary blood flow can be described by the relative dispersion (RD) of weight-flow histograms (RD = SD/mean). Glenny and Robertson (J. Appl. Physiol. 69: 532-545, 1990) showed that RD of flow in the lung is fractal in nature, characterized by the fractal dimension (D) and RD for the smallest realizable volume element (RDref). We studied the effects of increasing total pulmonary blood flow on D and RDref. In eight in situ perfused sheep lung preparations, 15-microns radio-labeled microspheres were injected into the pulmonary artery at five different blood flows ranging, in random order, from 1.5 to 5.0 l/m. The lungs were in zone 2 at the lower flows and in zone 3 at the higher flows. The lungs were removed, dried, cut into 2 x 2 x 2-cm3 pieces, weighed, and then counted for microsphere radioactivity. Fractal plots of log(weight) vs. log(RD) were constructed by iteratively combining neighboring pieces and then calculating RD with the increasingly larger portion size. D, which is one minus the slope of the fit through this plot, was 1.14 +/- 0.09 and did not change as blood flow increased. However, RDref decreased significantly (P < 0.01) as total flow increased. We conclude that the fractal nature of pulmonary blood flow distribution is not altered by changes in overall flow.

Pulmonary vascular resistance distribution and recruitment of microvascular surface area

To test the hypothesis that the distribution of hemodynamic resistance is involved in the control of pulmonary capillary surface area, we measured permeability-surface area product (PS) and longitudinal resistance distribution (LRD) as functions of perfusion rate in isolated rabbit lungs under zone II conditions (n = 10) and through the zone II-III transition (n = 4). PS, considered to be indicative of functioning capillary surface area, was measured with the aid of the diffusion-limited tracer [14C]propanediol, whereas LRD was determined using a viscous bolus technique. LRD was seen to change character with increasing flow and increasing PS/surface area, becoming bimodal with low central resistance as full capillary recruitment was approached in zone III. Effects of hypoxic ventilation were studied in zone II in five lungs; it was found that hypoxia altered the LRD and eradicated the normoxic dependence of PS/surface area on perfusion rate. It was concluded that LRD is involved in the determination of functioning capillary surface area.

Sulfhemoglobinated erythrocytes as an optical intravascular tracer in the lung

Sulfhemoglobinated erythrocytes (SHb-RBC's) were examined for utility as an optical multiple indicator dilution tracer in lung studies. A device was developed to measure this tracer optically in flowing blood. Arterial blood was sampled from cannulated, anesthetized dogs and pumped through the device that measured the optical density (OD) of blood at 620 nm. This system was calibrated for increasing SHb-RBC concentrations using an unsteady-state indicator dilution procedure. Areas under optical density (delta OD) profiles were well correlated with injected SHb-RBC volumes using linear regression (r2 > 0.9). This linearity was independent of blood oxygenation, hematocrit, or pH. In vivo lung indicator dilution studies in the intact dog were performed and compared to radioisotope indicator studies using 51Cr labeled erythrocytes. Coefficient of variation (CV) between the two curves was 0.065 under baseline conditions, 0.085 for studies performed during hypoxia, and 0.073 after pH was lowered. We conclude that this device linearly measured SHb-RBC content in whole blood and that SHb-RBC is as accurate a lung indicator dilution tracer as 51Cr-erythrocytes.

Reliability of retrospective self-reports of alcohol consumption among women: data from a U.S. national sample

Drinking histories (retrospective self-reports) can be a valuable resource for time-ordered analyses of causes and consequences of drinking. However, there is a scarcity of data on the reliability of drinking histories from general population samples. We report here on the reliability and consistency of reported ages of onset and typical drinking frequencies, quantities and volume, from drinking histories provided in 1981 and 1986 by national samples of women drinkers with and without drinking problems. Statistical reliability was generally modest, yet large percentages of women gave exactly the same reports 5 years apart. Reliability was apparently reduced by limited response options, and was lower among younger drinkers, whose drinking was more changeable between 1981 and 1986. We discuss ways to improve reliability and to make best use of drinking histories.

Deduction of pulmonary microvascular hematocrit from indicator dilution curves

We have developed a new model describing the relationship between plasma and red cell tracers flowing through the lung. The model is the result of an analysis of the transport of radiolabeled plasma albumin between two flowing phases and shows that differences between red cell and plasma tracer curves are related to microvascular hematocrit. The model was tested in an isolated, blood-perfused dog lung preparation in which we injected 51Cr-labeled red cells and 125I-labeled plasma albumin into the pulmonary artery. From the tracer concentration-time curves at the venous outflow, we calculated hr, the ratio of microvascular hematocrit to large-vessel hematocrit. In 18 baseline experiments, hr = 0.92 +/- 0.01 (mn +/- sem) at a blood flow rate of 10.7 +/- 0.3 ml s-1. We determined the effects of (a) glass bead embolization, (b) alloxan, and (c) lobe ligation on hr. Embolization attenuated the separation between plasma and red cells (increased hr), probably as a consequence of passive vasodilation. Alloxan enhanced separation of plasma and red cells (decreased hr), possibly as a result of arteriolar vasoconstriction. Ligation of a fraction of the perfused tissue at constant flow did not cause significant change in hr in the remaining perfused tissue. The model assumes that large-vessel transit times are uniform and that all dispersion occurs in the microvasculature. A theoretical analysis apportioning dispersion between large and small vessels disclosed that the error associated with these assumptions is likely to be less than 15% of the measured hr. We conclude from this study that the microvascular hematocrit model describes experimental plasma and red cell curves. The results imply that hr can be readily deduced from tagged red cells and plasma and can be accounted for in calculating permeability-surface area in diffusing tracer experiments.

Perilla ketone increases endothelial cell monolayer permeability in vitro

Perilla ketone (PK) is a potent lung toxin that causes increased microvascular permeability pulmonary edema in grazing animals. Because the mechanism of action of PK is not know, we investigated whether PK directly affects endothelial cells. Bovine aortic endothelial cells were grown to confluence on Cytodex-3 microcarrier beads and placed in a chromatographic cell column. Monolayer permeability was evaluated from the elution profiles of three optical tracers: blue dextran (2 x 10(6) mol wt), sodium fluorescein (NaF, 342 mol wt), and cyanocobalamin (B12, 1,355 mol wt). Perfusion with 1.2 mM PK increased permeability within 15 min to NaF and B12 by 51 +/- 6 and 54 +/- 11%, respectively. Permeability returned to baseline after PK removal. These in vitro results suggest that PK produces a rapid and reversible increase in endothelial permeability directly. Staining of fixed cells with rhodamine-phalloidin revealed a major disruption of actin microfilaments after PK treatment. Because previous reports suggested that PK may be activated via cytochrome P-450, we attempted to block this using the cytochrome P-450 inhibitor ketoconazole. Ketoconazole alone did not significantly affect permeability, and the combination of PK and ketoconazole resulted in permeability increases similar to those measured for PK alone. This suggests that PK may not require cytochrome P-450 to increase vascular permeability.

A fast track to IAIMS: the Vanderbilt University strategy

In July 1991, Vanderbilt University Medical Center (VUMC) initiated a fast track approach to the implementation of an Integrated Academic Information Management System (IAIMS). The fast track approach has four elements: 1) an integrated organizational structure combining various operational information management units and the academic informatics program into a single entity to enhance efficiency; 2) technology transfer and network access to remote resources in preference to de novo development; 3) parallel IAIMS planning and infrastructure construction; 4) restriction of the scope of the initial IAIMS to permit a manageable implementation project. The fast track approach is intended to provide a truly functional IAIMS within a time period (7 years) associated with other major construction projects such as the building of a replacement hospital.

Serving two masters: conflicts in practice

Accustomed to putting patients' needs first, internists now face daily pressure to contain health care costs. With a view from the trenches, a practicing internist and former ASIM president enumerates the dilemmas physicians confront in trying to serve patients and payers.

Surface area-independent assessment of lung microvascular permeability with an amphipathic tracer

A combination of an amphipathic-indicator-dilution (ID) diffusing tracer 1,4[14C]butanediol (B) and a hydrophilic tracer ([14C]urea) (U) was hypothesized to provide a capillary surface area- (S) independent assessment of lung microvascular permeability (P). We performed ID studies on isolated perfused dog lungs and administered randomly two interventions, increasing P by alloxan infusion and reduction in S by lobar ligation. The ratio of PS product of U (PSU) to that for butanediol (PSB) was sensitive to changes in P yet insensitive to changes in S. We performed ID studies in which the dependence of PSU and PSB on flow, hematocrit, and plasma protein binding were examined. Measurements of PSU and PSB after flow and hematocrit were changed suggested that these factors have no significant independent effects. From ID and in vitro studies we also found that no significant binding of B to plasma proteins (albumin) occurred. We concluded that ID techniques using B and U provide a consistent measure of P, despite changes in S, hematocrit, plasma protein concentration, and recruitment.

Effects of ligation and embolization on Kf and multiple tracer measurements in dog lungs

In isolated blood-perfused dog lungs, the capillary filtration coefficient (Kf) and the permeability-surface area product of urea (PS) were measured to determine their responses to two different methods of altering filtration area: lobe ligation (LL, n = 5) and glass bead embolization (GBE, n = 4) during constant perfusion rates (700 +/- 45 ml/min). When two of three lobes were ligated, Kf decreased (1.36 +/- 0.13 to 0.58 +/- 0.23 g.min-1.cmH2O-1; P less than 0.05), but PS did not change (2.02 +/- 0.4 to 1.71 +/- 0.3 ml/s). Kf per gram of perfused blood-free dry lung weight was unchanged by LL (0.051 +/- 0.17 to 0.052 +/- 0.18 g.min-1.cmH2O-1), indicating that surface area per gram measured by Kf remained the same. However, PS per gram dry lung doubled (0.07 +/- 0.016 to 0.146 +/- 0.06 ml/s; P less than 0.05) after LL, suggesting that recruitment occurred in the remaining lobe. When three lobes were embolized with 200-microns glass beads (0.48 +/- 0.01 g beads/kg body wt), PS decreased (2.1 +/- 0.22 to 0.94 +/- 0.09 ml/s; P less than 0.05), but Kf was not altered (1.01 +/- 0.17 to 1.04 +/- 0.18 g.min-1.cmH2O-1). The constancy of Kf after GBE implies that the vascular pressure increase during the Kf measurement was transmitted to both blocked and flowing vessels and thereby measured the same filtration area before and after GBE. PS decreased significantly after GBE because of a loss of perfused surface area by the beads blocking flow in small arterial vessels.(ABSTRACT TRUNCATED AT 250 WORDS)

Lung microvascular transport properties measured by multiple indicator dilution methods in patients with adult respiratory distress syndrome. A comparison between patients reversing respiratory failure and those failing to reverse

We conducted indicator dilution studies on the lungs of patients in the early phases of adult respiratory distress syndrome (ARDS) to test the hypothesis that capillary permeability was increased in patients with respiratory failure. Indicator dilution studies were performed using 51Cr-erythrocytes, 125I-albumin, 14C-urea, and 3H-water as tracers. The injectate was infused as a bolus into a central venous line. Peripheral arterial blood was collected and counted for radioactivity. Mathematical analysis of the indicator curves yielded cardiac output, measures of the product of capillary permeability and surface area for urea (PS and D1/2S), the intravascular lung volume (Vv), and the extravascular lung water volume (Ve). Permeability was separated from surface area by normalizing PS and D1/2S to Vv. Patients could be divided into 16 in whom blood gas determinations and radiologic criteria for ARDS were reversed and 23 in whom they were not. We examined indicator dilution and other measures of lung function in the two groups to determine whether significant differences in microvascular function existed. PS and PS/Vv were significantly higher in the nonreversal patients. Ve was above normal, but not different between groups. Linear regression analysis showed significant correlations for all of the following in the nonreversal group: Ve and all measures of permeability, pulmonary vascular resistance (PVR), and the inverse of permeability-surface area measures and AaDO2 and PVR. Only measures of Ve and PS correlated in the reversal group. These results support the hypothesis that capillary permeability is increased in patients with early ARDS and continuing respiratory failure.(ABSTRACT TRUNCATED AT 250 WORDS)

Measurement of lung fluid volumes and albumin exclusion in sheep

A radioactive tracer technique was used to determine interstitial diethylenetriaminepentaacetic acid (DTPA) and albumin distribution volume in sheep lungs. 125I- and/or 131I-labeled albumin were injected intravenously and allowed to equilibrate for 24 h. 99mTc-labeled DTPA and 51Cr-labeled erythrocytes were injected and allowed to equilibrate (2 h and 15 min, respectively) before a lethal dose of thiamylal sodium. Two biopsies (1-3 g) were taken from each lung and the remaining tissue was homogenized for wet-to-dry lung weight and volume calculations. Estimates of distribution volumes from whole lung homogenized samples were statistically smaller than biopsy samples for extravascular water, interstitial 99mTc-DTPA, and interstitial albumin. The mean fraction of the interstitium (Fe), which excludes albumin, was 0.68 +/- 0.04 for whole lung samples compared with 0.62 +/- 0.03 for biopsy samples. Hematocrit may explain the consistent difference. To make the Fe for biopsy samples match that for homogenized samples, a mean hematocrit, which was 82% of large vessel hematocrit, was required. Excluded volume fraction for exogenous sheep albumin was compared with that of exogenous human albumin in two sheep, and no difference was found at 24 h.

The optical measurement of 1,2-propanediol for the determination of lung capillary permeability surface area

A technique has been developed which allows for the optical measurement of the concentration-time relationship for a diffusion-limited material in indicator dilution studies. The material, 1-2 propanediol, is used as a probe of the permeability of capillaries in the lung. Comparisons between standard radioisotope measurements and the optical measurements are provided and show excellent agreement. The optical method represents an improvement over the standard radioisotope method in that it provides the same data at lower cost, lower risk, and without the delay required by the radiographic methods.

Effects of endotoxemia on the sheep lung microvascular membrane: a two-pore theory

We analyzed the effects of Escherichia coli endotoxin infusion on pulmonary microvessels in sheep by using a two-pore mathematical model of the microvascular barrier. Five sheep were prepared with lung lymph fistulas and instrumented to measure pulmonary arterial and left atrial pressures. Multiple indicator-dilution curves (with 125I-labeled albumin, 51Cr-labeled erythrocytes, [14C]urea, and 3H2O) were measured at base line and during phases 1 and 2 of the endotoxin response. Alterations in the membrane integrity in response to endotoxin infusion were quantified by using a two-pore theory of the microvascular barrier that incorporated lymph, protein, pressure, and multiple indicator measurements. The modeling results showed a slight change in the size of the pores during phase 1 but a 56% decrease in the number of small pores and a twofold increase in the number of large pores with respect to base-line values. During phase 2 the large pore size increased by 40%, and the total number of pores returned to base-line values. The analysis showed that endotoxin effects on fluid and protein exchange in the lung cannot be explained by hemodynamic and surface area changes alone. An apparent increase in lung microvascular permeability occurs during phases 1 and 2 of the endotoxin reaction, with a substantial decrease in perfused microvascular surface area during phase 1.

Lung water and vascular permeability-surface area in premature newborn lambs with hyaline membrane disease

Extravascular lung water and vascular permeability-surface area products were measured with a multiple indicator dilution method in 6 premature lambs with hyaline membrane disease 1-5 hours following delivery by cesarean section. The indicators used were 51Cr-labelled erythrocytes, 125I-albumin, 3H-water, and 14C-urea. Results were compared with previously obtained data in newborn lambs without hyaline membrane disease also delivered by cesarean section. Extravascular lung water was significantly higher in lambs with hyaline membrane disease [23.2 +/- 1.0 (SEM) vs. 10.7 +/- 1.4 ml/kg body wt]. Vascular permeability-surface area products for 14C-urea were significantly lower in lambs with hyaline membrane disease (0.30 +/- 0.10 vs 0.78 +/- 0.11 ml/s per kg). It is concluded that extravascular lung water is high in lambs with hyaline membrane disease. Permeability-surface area products for 14C-urea is low in lambs with hyaline membrane disease, which probably indicates a decrease in detectable surface area for exchange due to derecruitment or hypoperfusion of pulmonary exchange vessels in edematous and hypoxic areas of the lungs.

Comparison of labeled propanediol and urea as markers of lung vascular injury

The purpose of these studies was a comparison of [14C]urea (U) and 1,3-[14C]propanediol (Pr) as measures of lung vascular permeability-surface area (PS) under base-line conditions and after lung injury caused by alloxan infusion in isolated perfused dog lungs. Indicator mixtures of 125I-albumin, 51Cr-red blood cells, 3HOH, and U or Pr were injected under base-line conditions, after 1.2 g of alloxan, and after an additional 0.8 g of alloxan. Indicator-dilution curves were analyzed from sampled outflow blood to provide PS, the square root of effective extravascular diffusivity multiplied by exchange surface area (D1/2S), and extravascular lung water (EVLW) from the tracer mean transit times (VW). Results show that alloxan increases PS and D1/2S for U, D1/2S for Pr, and VW and EVLW by desiccation. All indicator-dilution parameters correlate significantly with alloxan dose. Interpretation of Pr transport suggests that materials with lipid and hydrophilic pathways might be used in conjunction with U to minimize the effects of surface area changes and increase the sensitivity of these tracers to permeability alteration. In addition Pr may be a useful alternative to U as a marker of vascular damage.

Models of lung transvascular fluid and protein transport

Transport theory has been applied to lymph flow (QL), protein lymph to plasma concentration ratios (L/P), and permeability surface area for urea (PSu) in unanesthetized sheep. Three models of the plasma-interstitial barrier have been used: a single pathway fiber matrix model, a continuous cylindrical-pore model with log normal distribution of filtration coefficients, and a cylindrical two-pore model. The fiber matrix model was unable to match measured PSu, QL, and L/P. The continuous-pore model was capable of describing the data, but the fitted median pore size was inconsistent with a continuum theory. The two-pore model described steady-state data and was used in additional model applications. We explored the 90% confidence limits for the fitted structural parameters of the two-pore theory. We found that many sets of model parameters were capable of fitting the available experimental data. We therefore sought combinations of parameters that might characterize the microvascular barrier under baseline and altered permeability situations. One combination that looks promising is the ratio of large-pore to small-pore radius raised to the sixth power and multiplied by the large-pore frequency. This value remains relatively constant following elevations in microvascular pressure, saline infusions, and plasma infusions but increases dramatically after endotoxin infusion.

Effect of perfusate hematocrit on urea permeability-surface area in isolated dog lung

Seven dog lower left lung lobes were statically inflated and perfused at a constant rate for each lobe with a perfusate in which the hematocrit was altered over a wide range. The permeability-surface area of urea was calculated from multiple indicator dilution curves using two separate injectates for each hematocrit level. One injectate contained only 125I-albumin as the vascular reference tracer and the other contained both 51Cr-erythrocytes and 125I-albumin as the vascular reference tracers; both contained [14C]urea as the permeating tracer. The results strongly indicate that the phenomenon of "erythrocyte trapping" of urea does not affect the calculation of urea permeability-surface area product provided the appropriate albumin-erythrocyte composite reference tracer is utilized in its calculation.

Effect of sodium bicarbonate and sodium bentonite on digestion, solid and liquid flow, and ruminal fermentation characteristics of forage sorghum silage-based diets fed to steers

Six ruminally cannulated steers, five Holsteins and one Hereford (250 to 295 kg), were fed 84% forage sorghum silage plus 16% supplement or 50% forage sorghum silage plus concentrate and supplement diets containing either no addition (controls), 1% sodium bicarbonate (NaHCO3) or 2% sodium bentonite in a 2 X 3 factorial arrangement of treatments in a 6 X 6 Latin-square experiment with 3-wk periods. Sodium bicarbonate increased dry matter (DM) intake when concentrate was included, but neither compound affected intake of the 84% silage diet. Bentonite lowered DM, neutral detergent fiber (NDF), and acid detergent fiber (ADF) digestibilities, but NDF disappearance from nylon bags was unchanged. Ruminal pH, osmolality and L(+) and D(-) lactate were not affected by treatment. Both NaHCO3 and bentonite tended to lower ruminal NH3-N concentrations. Bentonite lowered the molar proportion of isobutyrate in ruminal fluid relative to controls, but proportions of other volatile fatty acids (VFA) and total VFA concentrations were unchanged. Neither NaHCO3 nor bentonite affected ruminal liquid or solid volumes, dilution rate constants or ruminal outflow rates. Markers overestimated volumes, but correction with measured volumes did not change interpretation of treatment effects. It was concluded that control diets had sufficiently high baseline values of pH, dilution rate and acetate proportion to preclude changes induced by either compound, especially at 1 or 2% of DM intake. An effect on palatibility through neutralization of silage acids may have been responsible for the intake response to NaHCO3.

Effects of hypoproteinemia on lung microvascular protein sieving and lung lymph flow

Experiments were conducted on five chronically instrumented unanesthetized sheep to determine the effects of sustained hypoproteinemia on lung fluid balance. Plasma total protein concentration was decreased from a control value of 6.17 +/- 0.019 to 3.97 +/- 0.17 g/dl (mean +/- SE) by acute plasmapheresis and maintained at this level by chronic thoracic lymph duct drainage. We measured pulmonary arterial pressure, left atrial pressure, aortic pressure, central venous pressure, cardiac output, oncotic pressures of both plasma and lung lymph, lung lymph flow rate, and lung lymph-to-plasma ratio of total proteins and six protein fractions for both control base-line conditions and hypoproteinemia base-line conditions. Moreover, we estimated the average osmotic reflection coefficient for total proteins and the solvent drag reflection coefficients for the six protein fractions during hypoproteinemia. Hypoproteinemia caused significant decreases in lung lymph total protein concentration, lung lymph-to-plasma total protein concentration ratio, and oncotic pressures of plasma and lung lymph. There were no significant alterations in the vascular pressures, lung lymph flow rate, cardiac output, or oncotic pressure gradient. The osmotic reflection coefficient for total proteins was found to be 0.900 +/- 0.004 for hypoproteinemia conditions, which is equal to that found in a previous investigation for sheep with a normal plasma protein concentration. Our results suggest that hypoproteinemia does not alter the lung filtration coefficient nor the reflection coefficients for plasma proteins. Possible explanations for the reported increase in the lung filtration coefficient during hypoproteinemia by other investigators are also made.