Effect of aminoguanidine on lung fluid filtration after endotoxin in awake sheep.

It has been suggested that enhanced generation of nitric oxide by inducible nitric oxide synthase (iNOS) may contribute to acute lung injury. We hypothesized that aminoguanidine (AG), a proposed selective inhibitor of iNOS, would alter pulmonary hemodynamics, fluid filtration, and gas exchange after endotoxin in chronically instrumented awake sheep. Eighteen sheep were randomly assigned to receive either AG (10 mg/kg + 1 mg/kg/h), or NaCl 0.9% intravenously for 4 h, beginning 2 h after injection of Escherichia coli endotoxin (1 microgram/kg). After endotoxin, pulmonary artery pressure (Ppa), capillary pressure (Pc), and vascular resistance index (PVRI) rose concomitantly with six-fold increments in lung lymph flow (Q L) and protein clearance (CL). Extravascular lung water (EVLW) doubled, as assessed with the thermal dye dilution technique; Pa(O(2)) decreased, AaPO(2) and venous admixture (Q S/Q T) increased. After AG, Q L and CL increased further by approximately 30%, whereas EVLW remained unchanged, despite an additional increase in Pc. Ppa, PVRI, and systemic vascular resistance index rose, whereas cardiac index and pulmonary blood volume index declined. In addition, Pa(O(2)) rose, and AaPO(2) and Q S/Q T decreased. We conclude that in endotoxemic sheep, AG improves gas exchange and increases Q L and CL, whereas EVLW remains unchanged in spite of enhanced Pc. Apparently, increased lymphatic drainage prevents EVLW from rising after AG.

Coronary endothelium-derived vasodilation during cooling and rewarming of the in situ heart.

The integrity of coronary vascular endothelial vasodilator function during core cooling and rewarming was investigated in a pentobarbital-anesthetized open-chest dog model. Vasodilator response was assessed as the change from baseline blood flow by injecting the endothelial-dependent vasodilator acetylcholine (ACh) (1.0 microg) or the endothelial-independent vasodilator nitroglycerin (NTG) (50 microg) into the left anterior descending (LAD) coronary artery. Change in blood flow was measured using a transit time ultrasonic volume flowmeter technique. During cooling and rewarming LAD blood flow was significantly decreased. After rewarming, aortic pressure was artificially elevated to reach control. This procedure restored heart work (LV-RPP, left ventricular rate pressure product) and coronary perfusion pressure, but LAD blood flow remained lowered. Ability to dilate the vascular bed supplied by LAD, after injections of ACh or NTG, was present both during cooling and rewarming. At 25 degrees C coronary blood flow (LAD) increased from 3 +/- 1 to 9 +/- 1 mL x min(-1) in response to both ACh and NTG. Posthypothermic blood flow increased from 7 +/- 1 to 19 +/- 2 and 20 +/- 3 mL x min(-1) in response to ACh and NTG, respectively. Measured as the percent change from baseline LAD blood flow, the response was not significantly different from the one obtained in prehypothermic hearts. In conclusion, coronary vasodilator function, both endothelium dependent and endothelium independent, is present but not maintained at the same level during cooling to 25 degrees C and rewarming. In spite of the deterioration of cardiac function, no selective defect in the endothelium-dependent response was detected, either during hypothermia or after rewarming.

Errors in the intensive care unit (ICU). Experiences with an anonymous registration.

BACKGROUND: In order to obtain information about the occurrence and severity of errors in an ICU, this investigation was conducted in a combined ICU and postoperative ward at a Norwegian University Hospital. METHODS: An anonymous registration was conducted in order to reveal as many as possible of all errors in the unit. A separate registration form was used, recording the type of error, date and time, sex and age of the patient, patient condition (unstable/stable) and where the error occurred (on the ward or during transport). The registration started in October 1995, and reports until November 1996 are included (13 months). Consequences of the errors were graded using a 6-point scale (0=no consequences and 5=fatal). Two experienced intensivists and two experienced ICU nurses independently evaluated the errors using a visual analogue scale (VAS) with 10 as the worst imaginable error. All four were blinded to consequences of the error. RESULTS: A total of 87 errors was reported: 36 (41.3%) were medication errors, 17 (19.5%) related to intravenous infusions, 15 (17.2%) were due to technical equipment failure, and the rest (19 errors, 21.8%) miscellaneous. No consequences could be detected in 55 cases (63%) (grade 0). Six errors were graded as 1, and 22 (25%) as grade 2 (therapeutic intervention necessary, no damage recorded). Five errors had more serious consequences, and one was fatal. The scoring of errors varied a great deal. Mean VAS score was 4.2 (SD 1.7). The sum of VAS score (max. 40) on each error followed a normal distribution, and 12 errors had a score >25. CONCLUSION: Errors happen frequently in the ICU. Probably, our data do not represent the true incidence of errors in the period, which we believe was higher. Many errors are graded as serious or severe, but still have limited consequences for the patient.

Left ventricular dysfunction following rewarming from experimental hypothermia.

This study was aimed at elucidating whether ventricular hypothermia-induced dysfunction persisting after rewarming the unsupported in situ dog heart could be characterized as a systolic, diastolic, or combined disturbance. Core temperature of 8 mongrel dogs was gradually lowered to 25 degreesC and returned to 37 degreesC over a period of 328 min. Systolic function was described by maximum rate of increase in left ventricular (LV) pressure (dP/dtmax), relative segment shortening (SS%), stroke volume (SV), and the load-independent contractility index, preload recruitable stroke work (PRSW). Diastolic function was described by the isovolumic relaxation constant (tau) and the LV wall stiffness constant (Kp). Compared with prehypothermic control, a significant decrease in LV functional variables was measured at 25 degreesC: dP/dtmax 2,180 +/- 158 vs. 760 +/- 78 mmHg/s, SS% 20.1 +/- 1.2 vs. 13.3 +/- 1.0%, SV 11.7 +/- 0.7 vs. 8.5 +/- 0.7 ml, PRSW 90.5 +/- 7.7 vs. 29.1 +/- 5.9 J/m. 10(-2), Kp 0.78 +/- 0.10 vs. 0.28 +/- 0.03 mm-1, and tau 78.5 +/- 3.7 vs. 25.8 +/- 1.6 ms. After rewarming, the significant depression of LV systolic variables observed at 25 degreesC persisted: dP/dtmax 1,241 +/- 108 mmHg/s, SS% 10.2 +/- 0.8 J, SV 7.3 +/- 0.4 ml, and PRSW 52.1 +/- 3.6 m. 10(-2), whereas the diastolic values of Kp and tau returned to control. Thus hypothermia induced a significant depression of both systolic and diastolic LV variables. After rewarming, diastolic LV function was restored, in contrast to the persistently depressed LV systolic function. These observations indicate that cooling induces more long-lasting effects on the excitation-contraction coupling and the actin-myosin interaction than on sarcoplasmic reticulum Ca2+ trapping dysfunction or interstitial fluid content, making posthypothermic LV dysfunction a systolic perturbation.

Improved oxygenation using the prone position in patients with ARDS.

BACKGROUND: The prone position is known to increase oxygen uptake in patients with Adult Respiratory Distress Syndrome (ARDS). METHODS: In this clinical study from 1995-96, 14 ARDS patients with severe respiratory failure were treated for at least 1 h in the prone position. Responders, defined as having more than 10% increase in PaO2/FiO2 ratio from baseline after 1 h, were treated at least 6 h in the prone position. RESULTS: 11 patients responded during the first period of the prone position (primary responders). Two of the 3 non-responders were turned prone a second time with increase in the PaO2/FiO2 ratio (secondary responders). Mean PaO2/FiO2 ratio (mean +/- SEM) in the supine position was 11.7 +/- 0.8 kPa, increasing to 16.6 +/- 1.8 kPa and 18.0 +/- 1.4 kPa after 1 and 6 h respectively (P = 0.009). Mean time spent in the prone position was 69 h (range 3-256 h), and mean ventilatory time was 17 d (3-52 d). The mortality in this subgroup of our patients with ARDS was 42%, compared to 58% in 19 patients not turned prone in the same period. CONCLUSION: The prone position together with PEEP appears to improve ventilation-perfusion matching. The prone position is simple, effective and readily available and could be used early in most patients with ARDS.

Glucose and fatty acid oxidation by the in situ dog heart during experimental cooling and rewarming.

BACKGROUND: Reduced myocardial function after hypothermia may be metabolic in origin, but the relationship between myocardial metabolism and the various components of hypothermia-mediated dysfunction has not been thoroughly investigated. METHODS: In the present study we measured myocardial uptake and oxidation of glucose and oleate in mongrel dogs undergoing cooling to 25 degrees C followed by rewarming to 37 degrees C, using radiolabeled substrates. RESULTS: Segment work index declined from 39.3 +/- 5.1 to 15.1 +/- 2.4 mm Hg in response to cooling from 37 degrees to 25 degrees C and did not recover completely on rewarming (27.2 +/- 4.2 mm Hg, p < 0.05). Oleate uptake declined from 3,251 +/- 619 to 1,043 +/- 356 nmol.min-1.100 g-1 (p < 0.05) when the dogs were cooled from 37 degrees to 25 degrees C. Simultaneously, oxidation rate fell from 1,089 +/- 158 to 354 +/- 83 nmol.min-1.100 g-1 (p < 0.05). On rewarming, oleate uptake was restored to prehypothermic values, whereas its rate of oxidation remained depressed (480 +/- 129 nmol.min-1.100 g-1; p < 0.05). Uptake and oxidation of glucose also declined significantly during cooling. However, both uptake and oxidation of glucose recovered fully on rewarming. CONCLUSIONS: The results of the present study demonstrate a reduced capacity to oxidize fatty acids by the myocardium during rewarming after hypothermia.

Increase in leukotrienes in the coronary circulation by cooling: a study in the anesthetized dog.

Pentobarbital-anesthetized dogs were instrumented for hemodynamic measurements and cooled by heat exchange tubes. Through a catheter in the coronary sinus, blood samples were obtained and plasma leukotrienes measured using a high-pressure liquid chromatography technique. Hemodynamic function was significantly reduced during cooling, and during subsequent rewarming hemodynamic function was only partly restored. Leukotriene C4 and B4 were significantly increased at a body core temperature of 31 degrees C and 25 degrees C during cooling but not during rewarming (28 degrees C) and after rewarming (37 degrees C). This indicates that during decreasing body temperature elevated leukotriene levels may increase vascular permeability, inflammation, and vasoactivity and counteract temperature-dependent decrease in these functions. Disturbed microcirculation may thus diminish cardiac recovery during rewarming.

Extracorporeal membrane oxygenation (ECMO) as lung or heart assist.

Extracorporeal membrane oxygenation (ECMO) may serve as extracorporeal lung assist (ECLA) in patients with acute respiratory failure (ARF) or as extracorporeal heart assist (ECHA) in patients with low output syndrome (LOS) after open heart surgery. From 1988 to 1992 seven patients underwent ECMO in our hospital; four suffered from ARF and three from LOS. Various bypass techniques were employed. Two ARF patients, aged 58 and 18 years, had veno-venous bypass; in the latter, ECMO was reinstituted as a veno-arterial bypass one week after weaning. In a three-year-old boy, the ECMO outflow tubing was primarily connected to the pulmonary artery, and shortly afterwards relocated to the common carotid artery. In a 31-year-old man with ARF, and three LOS patients, a 56-year-old woman, and two men aged 68 and 70 years, ECMO was veno-arterial with direct access to the ascending aorta. A heparin-coated system was used, and all but one patient, who was treated with warfarin, received a daily low dose of heparin, which was withdrawn after from one to nine days. Six patients were weaned off ECMO after 4.5 to 21 days. Three ARF patients recovered completely; the child died. In one LOS patient, ECMO was withdrawn due to a poor general condition. Two others were weaned off ECMO and the intra-aortic balloon pump, and the inotropic support was significantly reduced, but both died of multiple system organ failure. Although no firm conclusions can be drawn from these few case reports, the heparin-coated system used as ECLA appears promising, whereas ECHA seems to imply a poor prognosis in patients who are not candidates for cardiac transplantation.

Experimental hypothermia: effects of core cooling and rewarming on hemodynamics, coronary blood flow, and myocardial metabolism in dogs.

Conflicting results have been reported as to the extent that cardiovascular function can be reestablished after rewarming from hypothermia. We measured hemodynamic function, myocardial metabolism and tissue water content in dogs core-cooled to 25 degrees C and later rewarmed. At 25 degrees C left ventricular (LV) systolic pressure (LVSP) was 54% +/- 4%, maximum rate of LV pressure rise (LV dP/dtmax) 44% +/- 5%, aortic pressure (AOP) 50% +/- 6%, heart rate (HR) 40% +/- 0%, cardiac output (CO) 37% +/- 5%, myocardial blood flow (MBF) 34% +/- 5%, and myocardial oxygen consumption (MVO2) 8% +/- 1%, compared to precooling. Stroke volume (SV) and LV end-diastolic pressure (LVEDP) were unchanged. As normothermia (37 degrees C) was reestablished, the depression of cardiac function and myocardial metabolism remained the same as that at 25 degrees C: LVSP 71% +/- 6%, LV dP/dtmax 73% +/- 7%, SV 60% +/- 9%, AOP 70% +/- 6%, CO 57% +/- 9%, MBF 53% +/- 8%, and MVO2 44% +/- 8% HR, in contrast, recovered to precooling values. The arterial concentrations of glucose and free fatty acids (FFA) did not change significantly during the experimental period, whereas an increase in lactate of nonmyocardial origin appeared after rewarming. Increased myocardial contents of creatine phosphate and water were found during both hypothermia and rewarming. The present study demonstrates a persistent depression of cardiac function after hypothermia and rewarming in spite of adequate energy stores. Thus, a direct influence on myocardial contractile function by the cooling and rewarming process is suggested.

Left ventricular diastolic function during short-term positive end-expiratory pressure ventilation in dogs.

Simultaneous recordings of left ventricular (LV) pressure and volume (sonomicrometry) were made in acutely instrumented dogs anaesthetized with pentobarbital during intermittent positive pressure ventilation with zero and positive end-expiratory pressure at 10 and 20 cmH2O (PEEP10 and PEEP20). Pericardial pressure was measured continuously in order to obtain transmural LV pressure. PEEP reduced LV end-diastolic volume and transmural pressure significantly. This was accompanied by significantly reduced stroke volume. LV peak diastolic filling rate, calculated as dV/dtmax, was significantly reduced when PEEP was applied, independent of LV volume alterations. LV diastolic compliance, assessed by the slope of LV pressure-volume relationship during LV filling, decreased significantly with increasing PEEP levels. A positive correlation was observed between reductions in peak diastolic filling rate and reductions in end-diastolic volume. The reduced peak diastolic filling rate, on the other hand, was closely correlated to reduced LV diastolic compliance. Isovolumetric relaxation rate (T) increased slightly at the highest PEEP level. This could, however, not be related to a reduced LV diastolic filling rate. The close association between reduced LV diastolic compliance and reduced diastolic filling rate may indicate that a tamponade-like effect is involved in the reduction of LV preload observed during PEEP ventilation.

Hemodynamic and metabolic effects of hypothermia and rewarming.

There is a lack of detailed knowledge of the pathophysiologic mechanisms initiated during and after rewarming. To study cardiac function after rewarming from hypothermia sodium pentobarbital anesthetized open chest-dogs were cooled to 25 degrees C and rewarmed. Myocardial blood flow was measured at different temperatures, and blood samples were drawn from the aorta and the coronary sinus for metabolic measurements. Mean aortic blood pressure (AOP) and aortic blood flow were recorded. Compared to precooling, AOP and heart rate were both significantly reduced during hypothermia. During rewarming stroke volume (SV) decreased significantly. At the end of rewarming AOP and SV were significantly lower than before cooling and myocardial blood flow, as well as oxygen and lactate uptake were only 50% of precooling levels. The present study demonstrated that hypothermia and rewarming depress cardiovascular function. Changes in peripheral vascular function, myocardial metabolism and contractility, may lead to the observed reduction in recovery upon rewarming.

Mechanism of ECG changes and arrhythmogenic properties of low osmolality contrast media during coronary arteriography in dog.

STUDY OBJECTIVE: The aim of the study was to investigate the mechanisms behind ECG changes and ventricular arrhythmias during coronary arteriography. DESIGN: Transmembrane action potentials were recorded from isolated heart muscle preparations superfused with contrast media. Conductivity of plasma diluted with contrast media was determined in vitro. Epicardial ECG recordings were made during coronary arteriography. SUBJECTS: Epicardial ECG recordings were made in eight mongrel dogs of either sex, weight 14-22 kg. Atrial appendages were excised from the same dogs and used for heart muscle preparation studies. MEASUREMENTS AND MAIN RESULTS: Iohexol and ioxaglate affected the action potentials similarly: resting potential, amplitude, rate of depolarisation, and action potential duration increased; effective refractory period decreased. Both contrast media reduced plasma conductivity. During coronary arteriography both media increased R wave amplitude, depressed ST segment, and prolonged QT time on epicardial ECG. CONCLUSIONS: ST segment deviation on ECG reflects hyperpolarization and increased amplitude of action potential rather than depolarisation consistent with ischaemia. Increased rate of depolarisation and amplitude of action potential together with reduced conductivity after contrast media may explain increased amplitude of QRS voltage on ECG. Observed regional changes in depolarisation, repolarisation, and refractoriness may be important in the genesis of ventricular arrhythmias.

Low osmolality contrast media and metabolic changes in the myocardium during free and reduced coronary flow in the dog.

The authors assessed whether intracoronary injection of low osmolality contrast media induces metabolic and electrocardiographic changes consistent with myocardial ischemia in anesthetized dogs. Ioxaglate and iohexol were injected into the left main coronary artery (eight dogs) and into a carotid-coronary artery shunt (eight dogs), during free coronary flow and during 50% flow reduction. Blood samples were obtained simultaneously from a femoral artery and from a small cardiac vein draining the contrast perfused area. Contrast media had no immediate or late effects on lactate balance during free or reduced flow. Early depression of the ST segment in epicardial ECG did not reflect ischemia. The authors conclude that the two low-osmolality contrast media, iohexol and ioxaglate, did not induce ischemic changes in the myocardium.

Effect of intracoronary contrast medium on left ventricular diastolic pressure-volume relationships.

The purpose of this study was to investigate the effect of intracoronary injection of the contrast medium sodium-meglumine diatrizoate (CM) on left ventricular (LV) diastolic pressure-volume relationships. Seven closed-chest dogs were instrumented with pressure catheters in the left ventricle and aorta, a balloon transducer to measure pericardial pressure, and an aortic flow meter to determine stroke volume. We estimated LV volume from two diameters by sonomicrometry. Six milliliters of CM was injected into the left main coronary artery. Transmural LV end-diastolic pressure increased from 3.3 +/- 1.1 to 7.2 +/- 0.9 mm Hg and LV end-diastolic volume index from 40.8 +/- 6.8 to 44.7 +/- 7.4 ml. There was only a minor increase in pericardial pressure. Stroke volume decreased by 31 +/- 7%. There was no change in the intracavitary or transmural diastolic pressure-volume relationship, indicating unchanged LV "compliance." Increased LV filling pressure by CM reflected reduced systolic function.

Effects of contrast media on myocardial function during acute left ventricular failure in the dog.

We studied the hemodynamic side effects of intracoronary injection of contrast media during acute ischemic heart failure by using anesthetized dogs. Induction of failure was performed by microembolization of the area supplied by the left main coronary artery. Six ml of iohexol (Omnipaque) increased contractility during the normal state, while this contrast medium induced no alterations in any of the recorded hemodynamic variables during left ventricular failure. Ioxaglate (Hexabrix) was also well tolerated during the normal state, while sodium-meglumine diatrizoate (Renografin) markedly decreased systolic variables. However, in the failing heart both ioxaglate and diatrizoate resulted in greater reduction in all systolic variables than in the normal heart. We conclude that both ionic contrast media may be harmful in acute ischemic heart failure. Non-ionic iohexol appears safer in this condition.

Cardiac effects of secretin; an approach to its mechanisms of action as shown by beta-adrenergic blockade and measurement of left ventricular dimensions in dogs.

In previous studies, the peptide secretin has demonstrated the ability to increase cardiac output and peripheral organ flow. In this investigation the mechanisms of the myocardial effects of secretin were studied. The secretin effects on cardiac output, stroke volume and systemic resistance were unaltered after propranolol, whereas the effect on LVdP/dt was reduced and the heart rate effect negligible. Systemic pressure, myocardial blood flow and myocardial oxygen consumption were unchanged by secretin infusion both before and after beta-receptor blockade. Secretin caused reduction of all three end-systolic left ventricular diameters. Thus, an inotropic effect by secretin was confirmed. Beta-receptor blockade reduced the inotropic effect and almost abolished the chronotropic effect, whereas the vasodilating effect was unaltered. The main haemodynamic effect of secretin was caused by activation of receptors other than the beta-adrenergic receptors.

Plasma concentrations of atrial natriuretic factor during positive end-expiratory pressure ventilation in dogs with normal or impaired left ventricular function.

The influence of positive end-expiratory pressure (PEEP) ventilation on plasma concentrations of atrial natriuretic factor (ANF) was studied in dogs anesthetized with sodium pentobarbital during normal cardiac function and during acutely impaired left ventricular function. Left ventricular impairment was induced by injecting repeated doses of polystyrene microspheres with a diameter of 50 microns into the main left coronary artery, causing a severe depression of left ventricular performance. This was accompanied by doubling of ANF concentrations measured in blood sampled from aorta. Application of PEEP (10 cmH2O (0.98 kPa] reduced plasma ANF in dogs both with normal and impaired left ventricular function. The decrease was significantly greater during left ventricular impairment compared to control, 31 and 19%, respectively. A positive correlation was observed between plasma ANF and transmural left ventricular end-diastolic pressure when all data were pooled, but not between ANF and transmural right atrial pressure. This implies that transmural left ventricular end-diastolic and hence transmural left atrial pressure probably is the principal determinant of acute ANF release in this model. Reduced plasma ANF in response to PEEP even during acute left ventricular impairment when ANF release was augmented, was probably due to diminished atrial distension during PEEP ventilation.

Myocardial blood flow and oxygen consumption during positive end-expiratory pressure ventilation at different levels of cardiac inotropy and frequency.

Effects of PEEP on cardiac function, myocardial blood flow (MBF) and myocardial oxygen consumption (mVO2) were studied in eight mongrel dogs anesthetized with pentobarbital. Myocardial oxygen demand was increased by isoproterenol infusion or atrial pacing, or decreased by beta-receptor blockade. PEEP was set to 15 cm H2O in all groups. The greatest reduction in cardiac output due to PEEP was seen during isoproterenol infusion (44%), and the smallest during beta-receptor blockade (18%). This is attributed to increased sensitivity to the reduced left ventricular (LV) preload induced by PEEP, when cardiac inotropy is augmented by isoproterenol, compared to normal and reduced cardiac inotropy. PEEP decreased MBF similarly and significantly in all groups. However, myocardial oxygen extraction did not increase, and reduction in MBF caused by PEEP was closely related to concomitant reduction in mVO2. A significant correlation was also observed between reductions in LV work and reduction in mVO2 when PEEP was applied in all groups. We conclude that the reduced MBF observed with use of PEEP was probably due to reduced myocardial oxygen demand.

Cardiovascular effects of positive end-expiratory pressure during acute left ventricular failure in dogs.

Haemodynamic and metabolic effects of ventilation with positive end-expiratory pressure (PEEP) were studied in closed-chest dogs anaesthetized with sodium pentobarbital during normal cardiac function and during acute left ventricular (LV) failure. LV failure was induced by embolizing the left coronary bed with 50 micron plastic microspheres causing a marked depression of LV function. End-expiratory pressure was set to 0, 5, 10 and 15 cm H2O both before and after coronary embolization. During normal cardiac function PEEP above 5 cm H2O depressed cardiac output (CO) significantly. However, following coronary embolization after which LV function was seriously impaired, CO was maintained as PEEP was applied. This is attributed to reduced sensitivity to the LV pre-load reductions induced by PEEP during LV failure. PEEP reduced MBF both during normal and impaired LV function. This did not result in ischaemic myocardial metabolism assessed by lactate extraction either in normal hearts or following coronary embolization. The reduced MBF was, however, associated with reduced MVO2 both during normal cardiac function and during LV failure. It is suggested that the reduced MBF is mainly due to reduced myocardial oxygen demand probably caused by reduced LV wall tension during PEEP ventilation.

Plasma concentrations of atrial natriuretic factor in acute left ventricular failure in the dog.

Acute ischemic left ventricular failure was induced in anesthetized dogs by repeated coronary embolization with 50 microns microspheres. Plasma concentrations of atrial natriuretic factor (ANF) were measured by radioimmunoassay of arterial and venous samples before and after failure induction. Heart failure was accompanied by a doubling of arterial ANF concentration, whereas there were only insignificant changes on the venous side. The increase in arterial ANF correlated significantly to the increase in left ventricular end-diastolic pressure, but not to the increase in right atrial pressure. Measurements of pericardial pressure indicated that a pericardial constraint acted to reduce atrial distension and thereby cardiac ANF release during failure.