Fractal or biologically variable delivery of cardioplegic solution prevents diastolic dysfunction after cardiopulmonary bypass.

OBJECTIVE: To determine whether myocardial protection is improved by restoring physiologic variability to the cardioplegia pressure signal during cardiopulmonary bypass, we compared cardiac function in pigs in the first hour after either conventional cold-blood cardioplegia (group CC) or computer-controlled biologically variable pulsatile cardioplegia (group BVC). METHODS: Invasive monitors and sonomicrometry crystals were placed, and cardiopulmonary bypass was initiated. The aorta was crossclamped, and cold blood cardioplegic solution was infused intermittently through the aortic root with either conventional cardioplegia (n = 8) or biologically variable pulsatile cardioplegia (n = 8; mean pressure, 75 mm Hg for 85 minutes). The crossclamp was released, cardiac function was restored, and separation from cardiopulmonary bypass was completed. With stable temperature and arterial blood gases, hemodynamics and systolic and diastolic indices were compared at 15, 30, and 60 minutes after cardiopulmonary bypass. RESULTS: Diastolic stiffness doubled from 0.027 +/- 0.016 mm Hg/mm (mean +/- SD) at baseline to 0.055 +/- 0.036 mm Hg/mm (P =.003) at 1 hour after bypass in group CC, associated with increased left ventricular end-diastolic pressure from 9 +/- 2 to 11 +/- 2 mm Hg (P =.001), mean pulmonary artery pressure from 14 +/- 2 to 20 +/- 3 mm Hg (P =.003), and serum lactate levels from 2.0 +/- 0.5 to 5.6 +/- 2.3 mmol/L (P =.008). Systolic function was not affected. In group BVC diastolic stiffness, left ventricular end-diastolic pressure, and pulmonary artery pressure values were not different from control values at any time after bypass, and serum lactate levels were significantly less than with conventional cold blood cardioplegia. Peak pressure variability with biologically variable pulsatile cardioplegia fit a power-law equation (exponent = -3.0; R(2) = 0.97), indicating fractal behavior. CONCLUSION: Diastolic cardiac function is better preserved after cardiopulmonary bypass with biologically variable pulsatile cardioplegia and fractal perfusion. This may be attributed to enhanced microcirculatory perfusion with improved myocardial protection. A model supporting these results is presented.

Biologically variable ventilation increases arterial oxygenation over that seen with positive end-expiratory pressure alone in a porcine model of acute respiratory distress syndrome.

OBJECTIVES: We compared biologically variable ventilation (BVV) (as previously described) (1) with conventional control mode ventilation (CV) in a model of acute respiratory distress syndrome (ARDS) both at 10 cm H2O positive end-expiratory pressure. DESIGN: Randomized, controlled, prospective study. SETTING: University research laboratory. SUBJECTS: Farm-raised 3- to 4-month-old swine. INTERVENTIONS: Oleic acid (OA) was infused at 0.2 mL/kg/hr with FIO2 = 0.5 and 5 cm H2O positive end-expiratory pressure until PaO2 was < or =60 mm Hg; then all animals were placed on an additional 5 cm H2O positive end-expiratory pressure for the next 4 hrs. Animals were assigned randomly to continue CV (n = 9) or to have CV computer controlled to deliver BVV (variable respiratory rate and tidal volume; n = 8). Hemodynamic, expired gas, airway pressure, and volume data were obtained at baseline (before OA), immediately after OA, and then at 60-min intervals for 4 hrs. MEASUREMENTS AND MAIN RESULTS: At 4 hrs after OA injury, significantly higher PaO2 (213+/-17 vs. 123+/-47 mm Hg; mean+/-SD), lower shunt fraction (6%+/-1% vs. 18%+/-14%), and lower PaCO2 (50+/-8 vs. 65+/-11 mm Hg) were seen with BVV than with CV. Respiratory system compliance was greater by experiment completion with BVV (0.37+/-0.05 vs. 0.31+/-0.08 mL/cm H2O/kg). The improvements in oxygenation, CO2 elimination, and respiratory mechanics occurred without a significant increase in either mean airway pressure (14.3+/-0.9 vs. 14.9+/-1.1 cm H2O) or mean peak airway pressure (39.3+/-3.5 vs. 44.5+/-7.2 cm H2O) with BVV. The oxygen index increased five-fold with OA injury and decreased to significantly lower levels over time with BVV. CONCLUSIONS: In this model of ARDS, BVV with 10 cm H2O positive end-expiratory pressure improved arterial oxygenation over and above that seen with CV with positive end-expiratory pressure alone. Proposed mechanisms for BVV efficacy are discussed.

Biologically variable or naturally noisy mechanical ventilation recruits atelectatic lung.

Biologically variable mechanical ventilation (Vbv)-using a computer-controller to mimic the normal variability in spontaneous breathing-improves gas exchange in a model of severe lung injury (Lefevre, G. R., S. E. Kowalski, L. G. Girling, D. B. Thiessen, W. A. C. Mutch. Am. J. Respir. Crit. Care Med. 1996;154:1567-1572). Improved oxygenation with Vbv, in the face of alveolar collapse, is thought to be due to net volume recruitment secondary to the variability or increased noise in the peak inspiratory airway pressures (Ppaw). Biologically variable noise can be modeled as an inverse power law frequency distribution (y approximately 1/f(a)) (West, B. J., M. Shlesinger. Am. Sci. 1990;78:40-45). In a porcine model of atelectasis-right lung collapse with one-lung ventilation-we studied if Vbv (n = 7) better reinflates the collapsed lung compared with conventional monotonously regular control mode ventilation (Vc; n = 7) over a 5-h period. We also investigated the influence of sigh breaths with Vc (Vs; n = 8) with this model. Reinflation of the collapsed lung was significantly enhanced with Vbv-greater Pa(O(2)) (502 +/- 40 mm Hg with Vbv versus 381 +/- 40 mm Hg with Vc at 5 h; and 309 +/- 79 mm Hg with Vs; mean +/- SD), lower Pa(CO(2)) (35 +/- 4 mm Hg versus 48 +/- 8 mm Hg and 50 +/- 8 mm Hg), lower shunt fraction (9.7 +/- 2.7% versus 14.6 +/- 2.0% and 22.9 +/- 6.0%), and higher respiratory system compliance (Crs) (1.15 +/- 0.15 ml/cm H(2)O/kg versus 0.79 +/- 0.19 ml/cm H(2)O/kg and 0.77 +/- 0.13 ml/cm H(2)O/kg)-at lower mean Ppaw (15.7 +/- 1.4 cm H(2)O versus 18.8 +/- 2.3 cm H(2)O and 18.9 +/- 2.8 cm H(2)O). Vbv resulted in an 11% increase in measured tidal volume (VT(m)) over that seen with Vc by 5 h (14.7 +/- 1.2 ml/kg versus 13. 2 ml/kg). The respiratory rate variability programmed for Vbv demonstrated an inverse power law frequency distribution ( y approximately 1/f(a)) with a = 1.6 +/- 0.3. These findings provide strong support for the theoretical model of noisy end-inspiratory pressure better recruiting atelectatic lung. Our results suggest that using natural biologically variable noise has enhanced the performance of a mechanical ventilator in control mode.

Biologically variable ventilation prevents deterioration of gas exchange during prolonged anaesthesia.

We have studied the time course of changes in gas exchange and respiratory mechanics using two different modes of ventilation during 7 h of isoflurane anaesthesia in pigs. One group received conventional control mode ventilation (CV). The other group received biologically variable ventilation (BVV) which simulates the breath-to-breath variation in ventilatory frequency (f) that characterizes normal spontaneous ventilation. After baseline measurements with CV, animals were allocated randomly to either CV or BVV (FIO2 1.0 with 1.5% end-tidal isoflurane). With BVV, there were 376 changes in f and tidal volume (VT) over 25.1 min. Ventilation was continued over the next 7 h and blood gases and respiratory mechanics were measured every 60 min. The modulation file used to control the ventilator for BVV used an inverse power law frequency distribution (I/fa with a = 2.3 +/- 0.3). After 7 h, at a similar delivered minute ventilation, significantly greater PaO2 (mean 72.3 (SD 4.0) vs 63.5 (6.5) kPa) and respiratory system compliance (1.08 (0.08) vs 0.92 (0.16) ml cm H2O-1 kg-1) and lower PaCO2 (6.5 (0.7) vs 8.7 (1.5) kPa) and shunt fraction (7.2 (2.7)% vs 12.3 (6.2)%) were seen with BVV, with no significant difference in peak airway pressure (16.3 (1.2) vs 15.3 (3.7) cm H2O). A deterioration in gas exchange and respiratory mechanics was seen with conventional control mode ventilation but not with BVV in this experimental model of prolonged anaesthesia.

Biologically variable pulsation improves jugular venous oxygen saturation during rewarming.

BACKGROUND: Conventional pulsatile (CP) roller pump cardiopulmonary bypass (CPB) was compared to computer controlled biologically variable pulsatile (BVP) bypass designed to return beat-to-beat variability in rate and pressure with superimposed respiratory rhythms. Jugular venous O2 saturation (SjvO2) below 50% during rewarming from hypothermia was compared for the two bypass techniques. A SjvO2 less than 50% during rewarming is correlated with cognitive dysfunction in humans. METHODS: Pigs were placed on CPB for 3 hours using a membrane oxygenator with alpha-stat acid base management and arterial filtration. After apulsatile normothermic CPB was initiated, animals were randomized to CP (n = 8) or BVP (roller pump speed adjusted by an average of 2.9 voltage output modulations/second; n = 8), then cooled to a nasopharyngeal temperature of 28 degrees C. During rewarming to stable normothermia, SjvO2 was measured at 5 minute intervals. The mean and cumulative area for SjvO2 less than 50% was determined. RESULTS: No between group difference in temperature existed during hypothermic CPB or during rewarming. Mean arterial pressure, arterial partial pressure O2, and arterial partial pressure CO2 did not differ between groups. The hemoglobin concentration was within 0.4 g/dL between groups at all time periods. The range of systolic pressure was greater with BVP (41 +/- 18 mm Hg) than with CP (12 +/- 4 mm Hg). A greater mean and cumulative area under the curve for SjvO2 less than 50% was seen with CP (82 +/- 96 versus 3.6% +/- 7.3% x min, p = 0.004; and 983 +/- 1158 versus 42% +/- 87% x min; p = 0.004, Wilcoxon 2-sample test). CONCLUSIONS: Computer-controlled BVP resulted in significantly greater SjvO2 during rewarming from hypothermic CPB. Both mean and cumulative area under the curve for SjvO2 less than 50% exceeded a ratio of 20 to 1 for CP versus BVP. Cerebral oxygenation is better preserved during rewarming from moderate hypothermia with bypass that returns biological variability to the flow pattern.

Computer-controlled cardiopulmonary bypass increases jugular venous oxygen saturation during rewarming.

BACKGROUND: Conventional roller pump apulsatile cardiopulmonary bypass (CPB) was compared with computer-controlled pulsatile bypass, which was designed to recreate biological variability (return of beat-to-beat variability in rate and pressure with superimposed respiratory rhythms). The degree of jugular venous oxygen saturation (SjvO2) less than 50% during rewarming from hypothermic CPB was compared for the two bypass techniques. An SjvO2 less than 50% during rewarming from hypothermic CPB is correlated with cognitive dysfunction in humans. METHODS: Pigs were placed on CPB for 3 hours using a membrane oxygenator with alpha-stat acid-base management and arterial filtration. After baseline measurements and normothermic CPB, the animals were randomized to apulsatile CPB (n = 12) or computer-controlled pulsatile CPB (roller pump speed adjusted by an average of 2.9 voltage output modulations/s; n = 12). The animals were then cooled to a nasopharyngeal temperature of 28 degrees C. During rewarming to stable normothermic temperatures, SjvO2 was measured at 5-minute intervals. The mean and cumulative areas for an SjvO2 less than 50% were determined for all animals. RESULTS: No between-group differences in temperature were noted during hypothermic CPB or during rewarming. The rate of rewarming was not different between groups. Mean arterial pressure, partial pressure of oxygen in arterial blood, and partial pressure of carbon dioxide in arterial blood also did not differ between groups. The hemoglobin concentration was within 0.4 g/dL between groups at all time periods. Mean pulse pressure was 10.0 +/- 4.8 mm Hg in the apulsatile CPB group and 20.7 +/- 5.2 mm Hg in the pulsatile CPB group (p = 0.0002; unpaired t test). Markedly greater mean and cumulative areas under the curve for SjvO2 less than 50% were seen with apulsatile CPB (164 +/- 209 versus 1.9 +/- 3.6% x min, p = 0.021; and 1,796 +/- 2,263 versus 23 +/- 45% x min, p = 0.020, respectively). CONCLUSIONS: Computer-controlled pulsatile CPB was associated with significantly greater SjvO2 during rewarming from hypothermic CPB. Both the mean and cumulative areas under the curve for SjvO2 less than 50% exceeded a ratio of 75:1 for apulsatile versus computer-controlled pulsatile CPB. These experiments suggest that cerebral oxygenation was better preserved during rewarming from moderate hypothermia with computer-controlled pulsatile CPB, which returned biologic variability to the flow pattern.

Cerebral hypoxia during cardiopulmonary bypass: a magnetic resonance imaging study.

BACKGROUND: Neurocognitive deficits after open heart operations have been correlated to jugular venous oxygen desaturation on rewarming from hypothermic cardiopulmonary bypass (CPB). Using a porcine model, we looked for evidence of cerebral hypoxia by magnetic resonance imaging during CPB. Brain oxygenation was assessed by T2*-weighted imaging, based on the blood oxygenation level-dependent effect (decreased T2*-weighted signal intensity with increased tissue concentrations of deoxyhemoglobin). METHODS: Pigs were placed on normothermic CPB, then cooled to 28 degrees C for 2 hours of hypothermic CPB, then rewarmed to baseline temperature. T2*-weighted, imaging was undertaken before CPB, during normothermic CPB, at 30-minute intervals during hypothermic CPB, after rewarming, and then 15 minutes after death. Imaging was with a Bruker 7.0 Tesla, 40-cm bore magnetic resonance scanner with actively shielded gradient coils. Regions of interest from the magnetic resonance images were analyzed to identify parenchymal hypoxia and correlated with jugular venous oxygen saturation. Post-hoc fuzzy clustering analysis was used to examine spatially distributed regions of interest whose pixels followed similar time courses. Attention was paid to pixels showing decreased T2* signal intensity over time. RESULTS: T2* signal intensity decreased with rewarming and in five of seven experiments correlated with the decrease in jugular venous oxygen saturation. T2* imaging with fuzzy clustering analysis revealed two diffusely distributed pixel groups during CPB. One large group of pixels (50% +/- 13% of total pixel count) showed increased T2* signal intensity (well-oxygenated tissue) during hypothermia, with decreased intensity on rewarming. Changes in a second group of pixels (34% +/- 8% of total pixel count) showed a progressive decrease in T2* signal intensity, independent of temperature, suggestive of increased brain hypoxia during CPB. CONCLUSIONS: Decreased T2* signal intensity in a diffuse spatial distribution indicates that a large proportion of cerebral parenchyma is hypoxic (evidenced by an increased proportion of tissue deoxyhemoglobin) during CPB in this porcine model. Neuronal damage secondary to parenchymal hypoxia may explain the postoperative neuropsychological dysfunction after cardiac operations.

Improved arterial oxygenation after oleic acid lung injury in the pig using a computer-controlled mechanical ventilator.

We compared computer-controlled mechanical ventilation programmed for biologic variability of respiratory rate (RR) and tidal volume (VT) with conventional intermittent positive-pressure ventilation (IPPV) in an oleic acid (OA) lung injury model. Seventeen pigs were ventilated with an Ohio 7000 anesthesia ventilator. Minute ventilation (VE) was adjusted to maintain PaCO2 at 30 to 35 mm Hg at baseline and was not altered further. OA was infused at 0.2 ml/kg/h until PaO2 decreased to < 125 mm Hg (F(I)O2 = 0.5). Animals were randomly assigned to continue with conventional IPPV (control group; n = 8) or had IPPV computer-controlled (computer group; n = 9). Hemodynamic, respiratory gas, airway pressure, and volume data were obtained at baseline (before OA infusion), at Time 30 (after infusion), and at 30-min intervals for 240 min after OA. At experiment completion, the lungs were removed to determine the wet:dry weight ratios. The control group had RR fixed at 20 breaths/min. The computer group had a RR of 20 +/- 2.3 breaths/min (range, 15 to 27 breaths/min), comprising 369 different RR values with reciprocal changes in VT over 1,089 s before the program looped to repeat itself. There was no difference between groups in the volume of OA infused. By 120 min after lung injury, animals in the computer group had significantly greater PaO2, associated with a lower Qs/QT. Mean airway pressures and mean peak airway pressures were not different in the two groups. By 180 min, respiratory system compliance (Crs) was significantly lower in the control group. The wet:dry lung weight ratios were greater in the control group. Thus, in a porcine model of OA lung injury, computer-controlled mechanical ventilation, which is programmed for biologic variability, resulted in improved blood oxygenation without increasing mean airway pressures when compared with conventional IPPV.

Neurological outcome in a porcine model of descending thoracic aortic surgery. Left atrial-femoral artery bypass versus clamp/repair.

BACKGROUND AND PURPOSE: In a porcine model of thoracic aortic cross-clamping (AoXC), we compared the incidence and severity of paraplegia with two surgical techniques: left atrial-femoral artery (LA-FA) bypass (BP group; n = 9) and clamp/repair (CR group; n = 8). The descending thoracic aorta was clamped near its origin and distal to the third intercostal artery for 30 minutes. The intervening three intercostal arteries were ligated and divided. METHODS: All animals received methohexital anesthesia and were hyperventilated to a Paco2 of 28 to 32 mm Hg. Animals in the CR group received mannitol, and after AoXC, proximal hypertension was controlled with phlebotomy. In the BP group, proximal hypertension was controlled with LA-FA bypass using a centrifugal pump (Biomedicus 520C). Proximal mean arterial pressure, distal mean arterial pressure, central venous pressure, and cerebrospinal fluid pressure were measured; radioactive microspheres were injected at baseline, at AoXC + 5 minutes, at AoXC + 20 minutes, at AoXC off + 5 minutes, and after resuscitation. Neurological function was assessed at 24 hours. The animals were killed, and the spinal cord was removed to determine spinal cord blood flow. Histological cross sections of the lumbar spinal cord were stained with cresyl violet/acid fuchsin and then examined with light microscopy to determine the ratio of altered to total spinal cord neurons. RESULTS: Fifteen animals survived (one death in each group) and were assessed neurologically at 24 hours after AoXC. Despite better distal perfusion and lumbar spinal cord blood flow in the BP group, during AoXC, and at AoXC off + 5 minutes, there was no significant difference in the severity of spinal cord ischemic injury between groups as assessed neurologically by Tarlov score (P = .90, Mann-Whitney U test). As well, the ratio of altered to total lumbar spinal cord neurons did not differ between groups (P = .24). CONCLUSIONS: In this chronic porcine model, distal circulatory support with LA-FA bypass afforded better distal perfusion and improved lumbar spinal cord blood flow but did not influence the severity of spinal cord ischemic injury when compared with a clamp/repair technique.

Neuroanesthesia adjunct therapy (mannitol and hyperventilation) is as effective as cerebrospinal fluid drainage for prevention of paraplegia after descending thoracic aortic cross-clamping in the dog.

We compared cerebrospinal fluid (CSF) drainage (Group D; n = 8) to neuroanesthesia adjunct therapy (hyperventilation and mannitol administration; Group N; n = 8) for the prevention of paraplegia using a canine model of descending thoracic aortic cross-clamping (AXC; 2.5 mm distal to the left subclavian artery for 30 min). We expected no difference in neurologic outcome between groups. After surgical preparation and a 30-min stabilization period, dogs in Group D had CSF drained prior to application of the AXC. During the period of AXC, CSF was allowed to drain freely in an attempt to have cerebrospinal fluid pressure (CSFP) no greater than central venous pressure (CVP). Dogs in Group N were hyperventilated (PaCO2 28-32 mm Hg) and received 2 g/kg of mannitol prior to AXC and then 1 g.kg-1.hr-1 during clamping. Systemic hemodynamics, CSFP, and arterial blood gases were measured at 1) baseline, 2) 2 min after AXC, 3) 20 min after AXC, 4) 5 min after AXC release, and 5) 30 min after resuscitation. With release of the AXC, PaCO2 was not controlled in Group D; in Group N the minute ventilation was further increased to maintain PaCO2 constant. At precisely 24 h after AXC, the animals were assessed for incidence and severity of paraplegia, using the Tarlov score, by an observer unaware of the experimental protocol. The animals were then killed, and the entire spinal cord was removed for histologic assessment. Multiple sections of the lumbar spinal cord were processed and stained with hematoxylin and eosin, then examined by light microscopy for nonviable neurons in the anterior spinal cord.(ABSTRACT TRUNCATED AT 250 WORDS)

Use of neuroanesthesia adjuncts (hyperventilation and mannitol administration) improves neurological outcome after thoracic aortic cross-clamping in dogs.

BACKGROUND AND PURPOSE: Using a canine model of thoracic aortic cross-clamping, we compared the incidence and severity of paraplegia with and without standard neuroanesthesia adjuncts (mannitol administration and deliberate hyperventilation). Better outcome was predicted for animals treated with mannitol and hyperventilation. METHODS: Nineteen dogs (mean +/- SD weight, 21 +/- 3 kg) were anesthetized with methohexital to an isoelectric electroencephalogram. Animals were randomized to group C (control; n = 9) or group M (mannitol administration and deliberate hyperventilation; n = 10). In group C, animals were maintained normocapnic (PaCO2, 38 to 42 mm Hg). In group M, animals were hyperventilated to a PaCO2 of 28 to 32 mm Hg and received mannitol 2 g.kg-1 during surgical preparation, then 1 g.kg-1.h-1 by continuous infusion. The thoracic aorta was cross-clamped for 30 minutes. Systemic hemodynamics, cerebrospinal fluid pressure, and arterial blood gases were measured at (1) baseline, (2) 2 minutes after cross-clamp, (3) 20 minutes after cross-clamp, (4) 5 minutes after cross-clamp release, and (5) 30 minutes after resuscitation. No attempt was made to control the hemodynamic consequences of cross-clamping in either group. With release of the cross-clamp, PaCO2 was not controlled in group C; in group M the minute ventilation was further increased to maintain PaCO2 constant. At precisely 24 hours after cross-clamp the animals were assessed for incidence and severity of paraplegia, using the Tarlov score, by an observer unaware of the experimental protocol. The animals were killed, and the entire spinal cord was removed for histological assessment. Multiple sections of the lumbar spinal cord were processed and stained with hematoxylin and eosin. RESULTS: With application of the cross-clamp, cerebrospinal fluid pressure and central venous pressure increased significantly in both groups. However, in group M the maximal mean cerebrospinal fluid pressure never exceeded baseline values in group C. With cross-clamp release, spinal cord perfusion pressure (distal mean aortic pressure minus mean cerebrospinal fluid pressure) was significantly greater in group M (86 +/- 23 vs 65 +/- 17 mm Hg; P = .0017 between groups). Acid-base balance was better maintained in group M. The incidence and severity of paraplegia were significantly lower in group M (P = .043; Mann-Whitney rank-sums test, two-tailed). In this group 10 of 10 animals could walk and 4 of 10 had complete recovery. In group C 4 of 9 animals were paraplegic. There was a strong negative correlation between the Tarlov score and the ratio of dead to total anterior spinal cord neurons in the lumbar region as assessed by light microscopy (P = .0004; Spearman's rank test). CONCLUSIONS: We conclude that a protocol using standard neuroanesthesia adjuncts (mannitol administration and deliberate hyperventilation) is associated with improved neurological outcome after thoracic aortic cross-clamping of 30 minutes' duration in dogs anesthetized with methohexital.

Effects of altered left atrial pressure on pulmonary vascular pressure-flow relationships.

We studied the effects of changes in pulmonary capillary wedge pressure (PCWP) on the slope (incremental resistance) and the extrapolated pressure intercept (PI) of the mean pulmonary artery pressure (PAP)-cardiac output (CO) relationship. Multipoint plots of PAP against CO were obtained in intact anesthetized dogs. Group 1 consisted of six dogs entirely in West zone 3 and group 2 of four dogs with mixed West zone 2-3. The four conditions studied were the following: 1) fixed low PCWP, 2) fixed high PCWP, 3) variable PCWP, and 4) time-control repeat of condition 1. The PI significantly exceeded PCWP at fixed low PCWP (group 1, 9.3 vs. 11.1 mmHg, group 2, 6.6 vs. 3.9 mmHg). PI became identical to PCWP only at fixed high PCWP in group 1 (19 +/- 2.0 vs. 19 +/- 1.1 mmHg). Thus PCWP reflects the effective vascular outflow pressure when PCWP is fixed and high. For both groups of dogs in condition 3, when PCWP was varied with CO, the slope of the resulting PAP-CO plot was significantly greater than when PCWP was constant. Also in 9 of 10 dogs, PI was less than PCWP when PCWP was varied. These findings demonstrate that when changes in PCWP are allowed to occur during the generation of a pulmonary artery pressure-flow plot, the resulting slope and intercept, as defined by a Starling resistor model, do not accurately represent the incremental resistance and outflow pressure of the pulmonary vasculature.

Pulmonary vascular pressure-flow relationship in canine oleic acid pulmonary edema.

We tested the hypothesis that the increased impedance to flow in canine oleic acid (OA) lung injury is predominantly due to an increase in effective downstream pressure (EDP), obtained by extrapolating to zero flow the linear portion of the pulmonary artery pressure (PAP)/flow (Q) relationship. PAP-Q coordinates were obtained in eight anesthetized, O2-ventilated dogs by varying Q through systemic arteriovenous fistulae. PAP-Q lines were obtained before and approximately 5 h after injection of OA. A second group of six dogs served as a time control (TC) group. There was a linear relationship between PAP and Q in both experimental and control groups (mean r value 0.948). The presence of pulmonary edema in the OA group caused the EDP to almost double, from 7 to 12 mmHg (P less than 0.01). In contrast, EDP remained constant in TCs. Incremental vascular conductance (IVC), slope of the PAP/Q line, decreased (P less than 0.05) a similar amount in both groups. The above findings are consistent with the modeling of the pulmonary circulation according to a Starling resistor in that large amounts of edema changed EDP but not incremental conductance.

Hemodynamic management in clinical acute hypoxemic respiratory failure. Dopamine vs dobutamine.

We investigated short-term hemodynamic effects of dopamine and dobutamine in eight patients with acute hypoxemic respiratory failure. We tested the hypothesis that for a similar increase in cardiac output, left ventricular filling pressure (pulmonary capillary wedge pressure [PCWP]) would increase with dopamine and decrease with dobutamine. Dopamine increased cardiac output (p less than 0.05), stroke volume (p less than 0.05), and PCWP (p less than 0.01). Cardiac output increased almost 20 percent when PCWP increased 50 percent with dopamine. In contrast, despite a mean 30 percent increase in cardiac output with dobutamine (p less than 0.01), PCWP decreased. In six of these patients, left ventricular end-diastolic volumes and end-systolic volumes were measured using scintigraphic techniques. In all patients, end-diastolic volume increased with dopamine (p less than 0.05); and in four of six, end-systolic volume increased. In contrast, with dobutamine, in five of six patients, end-diastolic volume decreased; and in all six patients, end-systolic volume decreased. There was a small increase in intrapulmonary shunt with both drugs. We conclude that if an inotropic agent is required to increase cardiac output in patients with acute hypoxemic respiratory failure, dobutamine is probably preferred over dopamine.

Pulmonary vascular effects of hydralazine in a canine preparation of pulmonary thromboembolism.

Pulmonary arterial pressure (PAP)-flow coordinates were obtained in 14 anesthetized dogs before and after pulmonary hypertension was induced with autologous blood clots. Cardiac output (CO) was altered by systemic arteriovenous fistulas. The PAP-CO coordinates were always rectilinear. Before emboli, the mean vascular closing or outflow pressure (the pressure intercept of the PAP-CO line) was 8.8 +/- 2.1 (SD) mm Hg. Emboli increased PAP (15.1 +/- 1.6 to 36.5 +/- 3.5 mm Hg; p less than .001) and decreased CO (3.8 +/- 0.6 to 2.4 +/- 0.8 liters X min-1; p less than .001). Incremental resistance (the slope of the PAP-CO line) only increased slightly. On the other hand, the marked increase in PAP was predominantly due to an increase in effective outflow pressure (from 8.8 +/- 2.1 to 28.6 +/- 3.6; p less than .001). Hydralazine was administered in a dose sufficient to double CO. This did not affect PAP and caused an inconsistent and small decrease in incremental resistance. However, a consistently significant decrease in effective outflow pressure, averaging 23%, was observed. In this canine preparation of pulmonary hypertension the predominant effect of hydralazine appears to be a decrease in the mean vascular closing or outflow pressure.

Treatment of canine permeability pulmonary edema: short-term effects of dobutamine, furosemide, and hydralazine.

The effects of treatment of oleic acid pulmonary edema with dobutamine, furosemide, and hydralazine on cardiopulmonary function in 24 dogs were investigated. Pulmonary capillary wedge pressure (PCWP) was adjusted to approximately 7 mm Hg; 45 min after oleic acid (0.08 ml/kg), dogs were randomly divided into a control group, in which PCWP was maintained at approximately 7 mm Hg, and into treatment groups as described above. Mean time-averaged PCWP was 2.3 mm Hg in dogs treated with dobutamine, 4.1 mm Hg with furosemide, and 4.4 mm Hg with hydralazine. Four hours of treatment with dobutamine and furosemide significantly (p less than .01) reduced accumulation of lung water compared with the control and hydralazine groups. Qs/Qt was lower (p less than .05) with dobutamine and furosemide compared with the other groups. In dogs given hydralazine, cardiac output (CO) and systemic vascular resistance (SVR) remained constant over the 4 hr treatment interval. In contrast, in all other groups, SVR increased and CO decreased (both p less than .05). The short-term pulmonary effects of the above drugs are probably explained by differences in PCWP and/or by regional pulmonary vascular effects.

Acute cardiopulmonary effects of nitroglycerin in canine oleic acid pulmonary edema.

In a canine model of acute respiratory failure, the authors investigated acute cardiopulmonary effects of nitroglycerin (TNG) and compared the results with those obtained after phlebotomy. Oleic acid increased intrapulmonary shunt (Qs/Qt) from 7.4 to 31% (P less than 0.001) and decreased (P less than 0.01) cardiac output (CO). In the presence of assumed low-pressure pulmonary edema, TNG was infused to decrease mean blood pressure (BP) by 40%; this was associated with a 26% decrease (P less than 0.05) in CO. Qs/Qt increased from 31 to 42% (P less than 0.01). There was a slight increase (P less than 0.01) in pulmonary vascular resistance (PVR) with TNG, and mean pulmonary artery pressure (PAP) decreased (P less than 0.05). In contrast, when CO was decreased by a similar amount with phlebotomy, mean Qs/Qt did not significantly change. There were similar changes in PVR and PAP and mixed venous O2 tension with TNG and phlebotomy. Accordingly, current results rule out increased flow, increased PVO2, and mechanical alterations in pulmonary vascular pressures as contributory to the increase in Qs/Qt with TNG. Alternatively, the increase in Qs/Qt with TNG may be explained by a direct pharmacologic decrease in pulmonary hypoxic vasoconstriction and/or by nonspecific pharmacologic effects.

Effects of vasodilators on canine cardiopulmonary function when a decrease in cardiac output complicates an increase in right ventricular afterload.

In canine oleic acid pulmonary edema, we investigated acute cardiopulmonary effects of nitroprusside (NP) before (NP1), and after (NP2) pulmonary vascular resistance (PVR) was increased via glass bead embolization. In the setting of increased PVR and reduced cardiac output (CO), acute cardiopulmonary effects of NP and hydralazine were compared. Oleic acid increased (p less than 0.05) pulmonary shunt (Qs/Qt) from 15 to 24%, but did not alter PVR. Cardiac output decreased (p less than 0.01) 31% with oleic acid from 4.2 to 2.9 1 X min-1 and systemic vascular resistance (SVR) increased (p less than 0.01). When PVR was normal, NP reduced (p less than 0.05) blood pressure (BP) from 148 to 123 mmHg, decreased SVR 31%, and increased (p less than 0.05) CO and Qs/Qt. Glass bead embolization increased (p less than 0.001) PVR from 2.2 to 20 mgHg X 1-1 X min and reduced (p less than 0.01) CO 23%, from 2.6 to 2 L/min. The Qs/Qt did not increase with embolization. In contrast to effects of NP1, when RV afterload was increased, CO fell (p less than 0.05) with NP2 from 2 to 1.6 1 X min-1. Alternatively, hydralazine improved cardiopulmonary function. In the setting of increased RV afterload, SVR and PVR decreased (p less than 0.01) 48 and 29%, respectively, with hydralazine. Corresponding to the decrease in resistance, CO increased (p less than 0.001) 84% with hydralazine, from 1.9 to 3.5 1 X min-1. Also, BP and Qs/Qt remained constant and arterial O2 tension increased (p less than 0.05) with hydralazine, from 113 to 152 mmHg.(ABSTRACT TRUNCATED AT 250 WORDS)

Effects of reduced resistive afterload on left ventricular pressure-volume relationship.

In seven anesthetized, beta-blocked dogs, we investigated the effects of a reduction in systemic vascular resistance (SVR) on left ventricular (LV) systolic mechanics. LV pressure and volumes (scintigraphic techniques) were measured in base-line condition, after opening one and then two arteriovenous fistulas (AVF). Volume was infused to maintain LV end-systolic pressure (LVESP). Despite a constant ESP, the mean end-systolic volume (LVESV) fell from 42 to 31 ml (P less than 0.025) when the SVR fell from 81 to 48 units (P less than 0.0025), and the LVESV fell further to 24 ml (P less than 0.0025) when the SVR was decreased to 30 units (P less than 0.025). In six similarly prepared dogs, aortic flow was measured, and when resistive afterload decreased, instantaneous flow increased. Since end-diastolic volume was not significantly changed when resistive afterload decreased, instantaneous LV volume decreased despite constant systolic LV pressure. In two of these dogs, LV pressure-volume (PV) trajectories were drawn for the ejection period. When SVR decreased there was a marked leftward shift of the PV trajectory as the end of ejection was approached. It is concluded that at a given contractile state and ventricular pressure, alterations in resistive load directly affect rate and extent of ventricular shortening.

Direct effects of nitroprusside do not alter gas exchange in canine oleic acid edema.

The authors investigated why intrapulmonary shunt (QS/QT) increases with sodium nitroprusside (SNP) in canine oleic acid pulmonary edema. To determine the effects of flow alone on QS/QT, a peripheral arteriovenous fistula with a variable resistor was employed to increase cardiac output (Q) 26 and 52% above base line in a stepwise fashion (P less than 0.01). To examine the direct effects of SNP, distinct from changes in flow, the drug was given to produce matched increments in Q in each dog (P less than 0.01). To control for time, base-line measurements were obtained before and after each intervention, the sequence of which was alternated. At each increment in Q, SNP and the arteriovenous fistula increased QS/QT a similar amount. The mixed venous O2 tension (P-vO2) followed Q similarly in each group. Pulmonary vascular resistance (PVR) fell more (P less than 0.01) with SNP than with the arteriovenous fistula at identical Q and P-vO2. The authors conclude that, in this model, a direct pharmacological effect of SNP does not contribute to the deterioration in QS/QT. In fact, SNP exerts a pulmonary vasoactive effect that does not adversely affect gas exchange.