Sharing best practices in teaching biomedical engineering design.

In an effort to share best practices in undergraduate engineering design education, we describe the origin, evolution and the current status of the undergraduate biomedical engineering design team program at Johns Hopkins University. Specifically, we describe the program and judge the quality of the pedagogy by relating it to sponsor feedback, project outcomes, external recognition and student satisfaction. The general pedagogic practices, some of which are unique to Hopkins, that have worked best include: (1) having a hierarchical team structure, selecting team leaders the Spring semester prior to the academic year, and empowering them to develop and manage their teams, (2) incorporating a longitudinal component that incudes freshmen as part of the team, (3) having each team choose from among pre-screened clinical problems, (4) developing relationships and fostering medical faculty, industry and government to allow students access to engineers, clinicians and clinical environments as needed, (5) providing didactic sessions on topics related to requirements for the next presentation, (6) employing judges from engineering, medicine, industry and government to evaluate designs and provide constructive criticisms approximately once every 3-4 weeks and (7) requiring students to test the efficacy of their designs. Institutional support and resources are crucial for the design program to flourish. Most importantly, our willingness and flexibility to change the program each year based on feedback from students, sponsors, outcomes and judges provides a mechanism for us to test new approaches and continue or modify those that work well, and eliminate those that did not.

Hindlimb unweighting affects rat vascular capacitance function.

Microgravity is associated with an impaired stroke volume and, therefore, cardiac output response to orthostatic stress. We hypothesized that a decreased venous filling pressure due to increased venous compliance may be an important contributing factor in this response. We used a constant flow, constant right atrial pressure cardiopulmonary bypass procedure to measure total systemic vascular compliance (C(T)), arterial compliance (C(A)), and venous compliance (C(V)) in seven control and seven 21-day hindlimb unweighted (HLU) rats. These compliance values were calculated under baseline conditions and during an infusion of 0.2 microg*kg(-1)*min(-1) norepinephrine (NE). The change in reservoir volume, which reflects changes in unstressed vascular volume (DeltaV(0)) that occurred upon infusion of NE, was also measured. C(T) and C(V) were larger in HLU rats both at baseline and during the NE infusion (P < 0.05). Infusion of NE decreased C(T) and C(V) by ~20% in both HLU and control rats (P < 0.01). C(A) was also significantly decreased in both groups of rats by NE (P < 0.01), but values of C(A) were similar between HLU and control rats both at baseline and during the NE infusion. Additionally, the NE-induced DeltaV(0) was attenuated by 53% in HLU rats compared with control rats (P < 0.05). The larger C(V) and attenuated DeltaV(0) in HLU rats could contribute to a decreased filling pressure during orthostasis and thus may partially underlie the mechanism leading to the exaggerated fall in stroke volume and cardiac output seen in astronauts during an orthostatic stress after exposure to microgravity.

The effects of hindlimb unweighting on the capacitance of rat small mesenteric veins.

Microgravity is associated with an impaired cardiac output response to orthostatic stress. Mesenteric veins are critical in modulating cardiac filling through venoconstriction. The purpose of this study was to determine the effects of simulated microgravity on the capacitance of rat mesenteric small veins. We constructed pressure-diameter relationships from vessels of 21-day hindlimb-unweighted (HLU) rats and control rats by changing the internal pressure and measuring the external diameter. Pressure-diameter relationships were obtained both before and after stimulation with norepinephrine (NE). The pressure-diameter curves of HLU vessels were shifted to larger diameters than control vessels. NE (10(-4) M) constricted veins from control animals such that the pressure-diameter relationship was significantly shifted downward (i.e., to smaller diameters at equal pressure). NE had no effect on vessels from HLU animals. These results indicate that, after HLU, unstressed vascular volume may be increased and can no longer decrease in response to sympathetic stimulation. This may partially underlie the mechanism leading to the exaggerated fall in cardiac output and stroke volume seen in astronauts during an orthostatic stress after exposure to microgravity.

Chronic isolation of carotid sinus baroreceptor region in conscious normotensive and hypertensive rats.

We have developed a chronic technique to isolate the carotid sinus baroreceptor region in the conscious rat model. Our technique, when used in conjunction with other methods, allows for the study of the control of arterial pressure, heart rate, and cardiac output by the carotid sinus baroreceptor reflex in conscious, unrestrained rats. The performance of our technique was evaluated in two strains: normotensive Sprague-Dawley (SD) rats and spontaneously hypertensive rats (SHR). Each rat was instrumented with an aortic flow probe and a catheter placed in the right femoral artery to monitor cardiac output and arterial pressure, respectively. The cervical sympathetic trunk and aortic depressor nerve were ligated and cut bilaterally, leaving vagus nerves intact. The right and left carotid sinuses were isolated using our new technique. We tested the open-loop function of the carotid sinus baroreceptor reflex system in the conscious rat after recovery from the isolation surgery. We found that changes in nonpulsatile carotid sinus pressure caused significant changes in arterial pressure, heart rate, and total peripheral resistance in both rat strains. However, the cardiac output responses differed dramatically between strains. Significant changes were seen in the cardiac output response of SHR, whereas no significant changes were observed in normotensive SD rats. We have found this technique to be a highly reliable tool for the study of the carotid sinus baroreceptor reflex system in the conscious rat.

Reduction in arterial compliance alters carotid baroreflex control of cardiac output in a model of hypertension.

Baroreflex regulation of cardiac output is determined by the performance of the heart as well as the available blood flow returning to the heart (i.e., venous return). We hypothesized that a decrease in arterial compliance (C(a)) would affect carotid baroreflex control of cardiac output by altering the slope of the venous return curve (VR curve). Baroreflex control of systemic arterial pressure (Pa), central venous pressure (Pv), heart rate, cardiac output (CO), and peripheral vascular resistance (R) were determined during bilateral carotid occlusion (BCO) in spontaneously hypertensive (hypertensive, HT) and Sprague-Dawley (normotensive, NT) rats. C(a) was determined from the rate of arterial pressure decay when CO was transiently stopped, and the VR curve was obtained during graded inflation of a vascular balloon positioned in the right atrium. The inverse slope of the VR curve was used as an index of the resistance to venous return (RVR). The baseline slope of the VR curve was -50.5 +/- 3.3 vs. -35.5 +/- 2.6 ml.kg-1.min-1.mmHg-1 in NT vs. HT, respectively (P < 0.05). Control values of Pa (96 +/- 5 vs. 124 +/- 8 mmHg) and R [0.43 +/- 0.04 vs. 0.80 +/- 0.07 peripheral resistance units (PRU)] were reduced in NT, whereas Ca (0.062 +/- 0010 vs. 0.036 +/- 0.003 ml.kg-1.mmHg-1) was elevated in NT vs. HT, respectively (P < 0.05). Analysis of the pressure dependence of C(a) demonstrated that C(a) was a nonlinear function of Pa, and the exponential decay constant for the C(a)-Pa relationship was reduced in HT (0.0055 +/- 0.0012 vs. 0.0012 +/- 0.0002 min, NT vs. HT, P < 0.05). Baroreflex activation by BCO significantly increased Pa (delta Pa, 20 +/- 4 vs. 28 +/- 3 mmHg) and R (delta R, 0.16 +/- 0.04 vs. 0.24 +/- 0.06 PRU) in NT vs. HT, respectively. However, BCO significantly decreased CO in NT but not HT (delta CO, -24 +/- 5 vs. -4 +/- 6 ml.kg-1.min-1, P < 0.05). In NT, RVR was increased 39 +/- 9% during BCO (P < 0.05), whereas RVR increased 8 +/- 3% in HT (P = NS). From these findings, we conclude that the difference in baroreflex control of CO is mediated, in part, by the reduction in C(a), which minimized the baroreflex-evoked increase in RVR.

Invariance of the resistance to venous return to carotid sinus baroreflex control.

Despite the well-established fact that the carotid sinus baroreflex system has profound control over the physical properties of the systemic circulation, the resistance to venous return (RVR) seems to be invariant of such control. We hypothesized that this apparent paradox may be explained from the baroreflex changes in systemic arterial compliance. In 12 pentobarbital-anesthetized mongrel dogs, RVR was measured at controlled carotid sinus pressures (CSP) of 50 and 200 mmHg with normal and artificially increased arterial compliance. Arterial compliance was determined from the arterial pressure decay when systemic blood flow was stopped with total vena caval occlusion. Changing CSP between 50 and 200 mmHg changed RVR significantly only under the condition of artificially increased arterial compliance. A four-parameter lumped model of the systemic circulation revealed that the baroreflex changes in arterial compliance and arterial resistance, which occurred in opposite directions, prevented a change in RVR when CSP was changed. The data also suggested that approximately 75% of RVR was attributed to large and conduit veins, the resistances along which were insensitive to baroreflex control. We concluded that the invariance of RVR results from a combination of 1) baroreflex change in the arterial compliance, 2) baroreflex insensitivity of the resistance along large and conduit veins, and 3) spatially distinct location between the major site of reflex change in capacitance and the major site of compliance.

Effect of arterial compliance on carotid sinus baroreceptor reflex control of the circulation.

Capacitive properties of the arterial and venous segments of the peripheral circulation are important in the regulation of cardiac output and arterial blood pressure. We examined whether an acute increase in arterial compliance C(a) would alter carotid sinus baroreflex control of the circulation. Eight mongrel dogs were anesthetized with pentobarbital sodium, and the carotid sinus regions were isolated and perfused with nonpulsatile pressures. Open-loop baroreflex response curves for systemic arterial pressure (SAP), heart rate (HR), aortic blood flow (ABF), peripheral vascular resistance (PVR), and left ventricular (LV) contractility were obtained when carotid sinus pressure (CSP) was changed in 25-mmHg steps between 50 and 200 mmHg under a control condition and when C(a) was increased by including two hydraulic compliant chambers to the arterial circulation (CS 1.72 ml/mmHg and CL 5.05 ml/mmHg). The compliant chambers significantly increased C(a) and altered the ratio of arterial to venous compliance C(a)/Cv). Changes in C(a)/Cv significantly decreased the maximal open-loop baroreflex gain (Gmax) for SAP (-2.3 +/- 0.5, -1.6 +/- 0.3, and -1.1 +/- 0.2 mmHg/mmHg, control vs. CS vs. CL, P < 0.05). Gmax for ABF was decreased by CS (-0.9 +/- 0.2 vs. -0.3 +/- 0.1 ml.kg-1.min-1, control vs. CS, P < 0.05), and CL reversed the reflex changes in ABF (Gmax: +0.6 +/- 0.3 ml.kg-1.min-1). Gmax for HR, PVR, and LV contractility was not altered when C(a) was increased (P > 0.05). These findings indicate that an increase in C(a) changes C(a)/Cv and alters carotid baroreflex control of SAP by modifying the ABF response. We conclude that a change in C(a)/Cv affects the reflex control of the circulation by altering the distribution of blood volume between the arterial and venous circulations.

Alpha- and beta-adrenergic mechanisms in the control of vascular capacitance by the carotid sinus baroreflex system.

We examined the active and passive contributions of the alpha- and beta-adrenergic receptor mechanisms to the changes in systemic vascular capacitance caused by the carotid sinus baroreflex system in anesthetized, vagotomized dogs. The carotid sinuses were isolated from the systemic circulation and perfused with controlled pressures. To determine the changes in vascular capacitance, a constant flow, constant venous pressure cardiopulmonary bypass was used. The changes in unstressed vascular volume were calculated when carotid sinus pressure was reduced from 200 to 50 mmHg without any adrenergic receptor antagonist, with either an alpha- (phentolamine) or a beta- (propranolol) antagonist and then with both. The reflex change in unstressed vascular volume in the systemic circulation (22.6 +/- 9.0 ml/kg without any antagonist) was reduced by 72% with phentolamine, by 35% with propranolol, and by 73% with both antagonists. Our results suggest that the alpha-adrenergic mechanisms contribute significantly to active changes in systemic venous capacity. In addition, the beta-adrenergic system has very little effect on active changes in venous vessels but does contribute to the overall capacity changes by dilating the hepatic outflow resistance when the carotid sinus baroreflex system is activated.

Two-port analysis of systemic venous and arterial impedances.

Hemodynamic properties of the systemic vasculature were measured in eight anesthetized dogs using two-port impedance analysis. Blood pressures and flows were measured at the aortic root and the caval-atrial junction. Impedances were computed from 0.05 to 20 Hz to characterize the systemic vasculature. Pseudorandom variations in flow were produced with an extracorporeal perfusion system. Impedance measurements were made at carotid baroreceptor pressures of 50, 125, and 200 mmHg. A six-parameter lumped-element model best fitted the measured impedance spectra. At 125 mmHg, the mean parameter values were venous inertance, 13.5 g.kg.cm-4; venous and arterial compliances, 0.769 and 0.0214 ml.mmHg-1.kg-1; venous and arterial characteristic impedances, 0.028 and 0.084 mmHg.kg.min.ml-1; and arterial-to-venous small-vessel resistance, 0.706 mmHg.kg.min.ml-1. Regression analysis showed significant dependence of small-vessel resistance on baroreceptor pressure. The other parameters were not dependent on carotid sinus pressure, which is consistent with baroreflex control of venous unstressed volume but not compliance. We conclude that two-port impedance analysis is a useful tool for studying venous hemodynamics and the dynamic coupling between the veins and the right heart.

Pial microvascular hemodynamics in anemia.

Isovolemic hemodilution and subsequent anemia increase cerebral blood flow (CBF). We hypothesized that pial microvascular pressure also increases with hemodilution and that arteriolar diameter varies concurrently as a myogenic autoregulatory response. First- and second-order arterioles (31-92 microns, n = 29) and large venules (65-215 microns, n = 17) were studied in thiopental-anesthetized rats. Microvascular pressure was determined using the servo-null technique, and vessel diameters were obtained directly from a video monitoring system. We measured the increase in CBF (radiolabeled microspheres) that accompanies hemodilution in a separate group of animals (n = 20). Hematocrit was reduced to 16-36% with homologous plasma (hemodilution group, n = 13) or held constant with homologous whole blood (control group, n = 4). In control animals, arteriolar and venular diameter varied +/- 1-2 microns from baseline values, and microvascular pressure remained unchanged from baseline. In the hemodilution group, CBF increased, but there was no systematic pial vasodilation. Furthermore, intraluminal pressure did not increase in pial microvessels, suggesting that proximal vasodilation was negligible even at the lowest hematocrit studied. Vascular resistance fell proportionately in both large vessel and microvascular segments. We conclude that experimental anemia does not produce alterations in microvascular pressure in rats, and the hyperemia accompanying hemodilution is largely viscosity mediated.

Blood volume changes in microcirculation of rat intestine caused by carotid sinus baroreceptor reflex.

This series of experiments quantified the role of the rat small intestinal arterioles in capacity changes during bilateral carotid occlusion (BCO) and compared them with previous venular studies. We also determined the role of autoregulation in arteriolar constriction during BCO. First-order, second-order, and third-order arteriolar diameter changes were measured during changes in arteriolar pressure and/or sympathetic activity (via BCO). The results indicated that an arteriolar pressure drop caused an immediate significant transient diameter decrease of 14% (P < 0.001), followed by an average steady-state diameter increase of 11% (P < 0.001) over control because of autoregulation. During BCO, the arterioles demonstrated an initial 8-22% decrease in diameter (P < 0.05). The largest first-order vessels were the least responsive to changes in pressure and BCO, while the smallest third-order vessels were the most responsive. Apparently autoregulation, and not sympathetic activity, was responsible for most of the arteriolar constriction during BCO. We also developed an anatomic model of the rat intestinal vasculature which revealed that venules hold 70% of the microcirculatory volume and are responsible for 80% of the total blood shift during BCO. Venular constriction and, to a minor degree, arteriolar constriction result in a 14% decrease in microcirculatory intestinal blood volume during BCO.

Systemic responses to carotid occlusion in the anesthetized rat.

We evaluated the effects of four standard anesthetization regimens on the systemic cardiovascular responses to bilateral common carotid artery occlusion in 28 adult male rats. Rats were randomly assigned to anesthesia groups: thiopental sodium (PT; 100 mg/kg ip), alpha-chloralose (CH; 100 mg/kg iv), ketamine hydrochloride plus acepromazine (KA; 135 mg/kg and 1.5 mg/kg sc), and pentobarbital sodium (PB; 50 mg/kg ip). PT and PB animals had similar baseline heart rates (HR; 333 and 345 beats/min, respectively) and arterial pressures (MAP; 126 and 118 mmHg, respectively), whereas both were lower in CH and KA (314 and 288 beats/min, 92 and 85 mmHg). During bilateral carotid occlusion, PT demonstrated the largest change in MAP (dMAP; +27 mmHg) but the smallest change in HR (dHR; +8 beats/min). CH and PB demonstrated similar dHR (+24 and +16 beats/min) and dMAP (+20 and +19 mmHg). KA demonstrated a significant dHR (+14 beats/min), but the average dMAP was not statistically significant (+3 mmHg). Therefore, carotid occlusion in rats anesthetized with PT, PB, or CH consistently elicits a systemic arterial pressor response comparable with that reported for conscious animals. When the magnitude and stability of baseline HR and MAP are also considered, PT and PB anesthetization seem to be the most reliable for evaluation of the carotid occlusion pressor response in rats.

Modeling the circulation with three-terminal electrical networks containing special nonlinear capacitors.

Development, first of analog and later of digital computers, as well as algorithms for analysis of electrical circuits, stimulated the use of electrical circuits for modeling the circulation. The networks used as building blocks for electrical models can provide accurate representation of the hydrodynamic equations relating the inflow and outflow of individual segments of the circulation. These networks, however, can contain connections in which voltages and currents have no analogues in the circulation. Problems arise because (a) electrical current must flow in closed loops, whereas no such constraints exist for hydraulic models; and (b) electrical capacitors have a number of characteristics that are not analogous to those of hydraulic compliant chambers. Disregarding these differences can lead to erroneous results and misinterpretation of phenomena. To ensure against these errors, we introduce an imaginary electrical element, the nonlinear residual-charge capacitor (NRCC), with characteristics equivalent to those of a compliant chamber. If one uses appropriate circuit connections and incorporates the residual-charge capacitor, then all voltages and currents in the model are proper analogues of pressures and flows in the circulation. It is shown that the capacitive current represents the rate of change of volume of blood inside the vessel, as well as the rate of the corresponding displacement of volume of the surrounding tissue.

Pressure dependence of baroreceptor-mediated vasoconstriction in rat skeletal muscle.

To determine whether microvessels in resting or contracting skeletal muscle constrict during baroreceptor activation, vascular diameters were measured in the spinotrapezius muscle of adult rats (n = 12) during occlusion of the common carotid arteries. Neural and myogenic components were distinguished using two types of occlusion: 1) "normal" (arterial pressure was allowed to increase with baroreceptor activation) and 2) "isobaric" (arterial pressure was maintained constant by decreasing blood volume). During normal occlusions, intermediate and small arteriolar diameters decreased in resting and contracting muscle (10-15% and 25-30%, respectively). Large arterioles and all-sized venules distended slightly (approximately 5%) in resting muscle, but diameters were maintained or decreased in contracting muscle. When arterial pressure was maintained constant (isobaric), the microvascular responses to baroreceptor activation in both resting and contracting muscle were essentially eliminated. We conclude that nearly all the arteriolar constriction observed in the spinotrapezius muscle during normal carotid artery occlusion is myogenic in origin, secondary to increased arterial pressure. This pressure-dependent constriction is augmented during skeletal muscle contraction and functional vasodilation.

Carotid sinus baroreceptor reflex control of venular pressure-diameter relations in rat intestine.

We tested the hypothesis that the venules of the small intestinal muscle are responsible for decreases in vascular capacitance during bilateral carotid artery occlusion. We measured microvascular venular pressure and diameter relations in 135 vessels during both control and baroreflexive conditions (bilateral carotid occlusion). Microvascular pressure was measured using a servo-null pressure system, and diameters were obtained from a video-monitoring system with a total magnification of X1,000. First-, second-, and fourth-order microvenules were studied in rat small intestinal muscle. The vessels showed an average diameter decrease of 11-12% and an increased stiffness or pressure-diameter slope of 31-53% during bilateral occlusion. We also tested whether the observed constriction during bilateral occlusion was caused by an increase in the sympathetic nerve activity to the venules and/or increased hormonal release via the baroreflex system. We studied an additional 22 microvenules before and after denervation of the preparation. Denervation eliminated any significant change in diameter or stiffness during bilateral occlusion. Based on our data, we conclude that the changes in the venular properties observed during bilateral occlusion are due to the increased sympathetic nerve activity resulting from decreased carotid sinus pressure. Intestinal venules can actively constrict to change vascular capacitance during bilateral carotid occlusion.

New technique to completely isolate carotid sinus baroreceptor regions in rats.

We developed a method by which we can completely isolate the carotid sinus baroreceptor regions in the rat. The carotid sinus baroreceptor region is exposed and, with the use of extra-fine forceps, a human hair is placed around and tied at the root of the bifurcation. This procedure occludes the external carotid artery and blood flow to the carotid body. An injector is then attached to a catheter in the common carotid artery. We introduce a cylindrical rubber plug into either the palentine or internal carotid artery. A second plug is introduced to occlude the other artery. In six of the eight rats studied, these procedures completely isolated the carotid sinus region. In those cases where a small leak persisted at a carotid sinus pressure of 180 mmHg, we introduced a small particle of the animal's own previously clotted blood. Carotid sinus pressure was either randomly changed between 40 and 180 mmHg in 20-mmHg increments or in sequential 20-mmHg steps from 40 to 180 mmHg while measuring the animal's pulsatile and mean blood pressures. Arterial pressure-carotid sinus pressure relationship indicates that there is a highly sigmoidal relationship between the two pressures. The peak gain of the carotid sinus reflex system had a range from 1.5 to 4.0 and a mean value of 2.07 +/- 0.08. Our data indicate that the rat exhibits a significant carotid sinus baroreceptor reflex response. This technique combined with other techniques will allow for the study of neural control of cardiovascular function in the rat.

Rat venular pressure-diameter relationships are regulated by sympathetic activity.

The hypothesis that the pressure-diameter relationship of intestinal venules in rats is primarily determined by sympathetic nervous system activity was tested. The pressure-diameter relationship of the smallest to largest diameter (20-100 microns) intestinal venules of the rat was measured at rest, during hemorrhage to increase sympathetic neural activity, and during saline volume expansion to decrease sympathetic activity. During hemorrhage, the diameter of all venules decreased approximately 10% at 10 mmHg venous pressure, and the slope of the pressure-diameter relationship increased approximately 50% above control. Blood volume expansion led to an approximately 10% increase in venule diameter at 10 mmHg and a 25% decrease in slope. Denervation of the vessels causes concomitant vasodilation, which was greater than the vasodilation caused by blood volume expansion. Hemorrhage after denervation caused no significant changes in the relationship when compared with denervated control. Nitroprusside caused an even greater vasodilation when compared with the pressure-diameter relationship after denervation. The results suggest that the slope and 10-mmHg intercept of the pressure-diameter relationship for the largest through smallest intestinal venules and, therefore, their vascular compliance and capacitance characteristics are primarily determined by sympathetic activity.

Does volume catheter parallel conductance vary during a cardiac cycle?

Absolute left ventricular volume measurement by the conductance (volume) catheter requires subtraction of the conductance contribution from structures extrinsic to the cavity blood pool. Previously, this parallel conductance volume (Vp) has been assumed constant throughout the cardiac cycle, and the technique described for its estimation in situ yields a single value. We present a new method for parallel conductance determination that yields multiple estimates during systole, enabling an assessment of Vp variability [Vp(t)]. For isolated blood-perfused ejecting canine left ventricles with empty (vented) right ventricles, Vp(t) displayed virtually no variation throughout systole. For in situ hearts, despite the presence of other cardiac chambers, Vp(t) also displayed little variation, with no statistically significant deviation from its mean value throughout systole. Volume signal simulations found the new technique to be less sensitive to signal noise and thus more robust than the one previously published. The isolated and in situ heart data indicate that for the left ventricle, the parallel conductance is relatively constant throughout normal ejection.

Effects of vasopressin on pulmonary and systemic vascular mechanics.

The direct effects of vasopressin on the resistance and capacitance properties of the pulmonary and systemic vasculature were studied in nine aneural dogs on systemic and pulmonary bypass. The systemic and pulmonary pressure-flow, the systemic and pulmonary arterial pressure-volume, and the systemic and pulmonary venous pressure-volume relationships were determined for five levels of infused vasopressin. Vasopressin levels of approximately 10, 30, 150, 300, and 500 pg/ml were achieved by intravenous infusions. Samples of venous blood were drawn before and after each set of pressure-flow and pressure-volume relationships for the determination of vasopressin level by radioimmunoassay. A linear relationship was found between vasopressin level and systemic vascular resistance. Systemic vascular resistance increased 0.072 +/- 0.011 mmHg.kg.min.ml-1 for a change in vasopressin level of 100 pg/ml. Vasopressin did not affect pulmonary vascular resistance or any vascular compliance. High doses of infused arginine vasopressin were necessary to elicit substantial vasoconstriction.