A comprehensive physiological model of circulation enables automatic piloting of hemodynamics in patients with acute heart failure

A comprehensive physiological model of the whole circulation is mandatory to quantitatively diagnose pathophysiology and to guide an appropriate treatment. Such a model would enable automatic piloting of hemodynamics in patients with acute heart failure. By extending Guyton's model, so as to deal with heart failure predominantly affecting left heart and to quantify left atrial pressure, we constructed such a model consisting of a venous return (VR) surface and a cardiac output (CO) curve. VR surface, the integrated property of systemic and pulmonary vascular beds, relates VR and left and right atrial pressures (P(LA), P(RA)) linearly as VR = V/W - G(R)P(RA) - G(L)P(LA), given total blood volume (V). CO curve, the pumping ability of hearts, relates CO and either P(LA) or P(RA) approximately by logarithmic functions, respectively, as CO = S(L) [ln (P(LA)- B(L)) + C(L)] = S(R) [ln (P(RA) - B(R)) + C(R)]. The slopes (S(R) and S(L)) of CO curve mainly describes the pump performance. W, G(R), G(L), B(R), B(L), C(R) and C(L) are parameters. We validate the model with animal experiments. Parameters W, G's, B's and C's were relatively constant in 19 dogs. In other 8 dogs, with or without acute left heart failure, we determined V and S's from a single set of CO, P(RA), P(LA) and standard parameter values. We then predicted hemodynamics (CO, P(RA), P(LA)) for altered V from -8 to +8 ml/kg. We identified standard values of parameters as W (0.13 min), G(R) (19.6 ml/min/kg/mmHg), G(L) (3.5 ml/min/kg/mmHg). B(L) (2.1 mmHg), C(L) (1.9), B(R) (2.0 mmHg) and C(R) (0.80). Using these, we could accurately predict CO (y = 0.93x + 6.5, r2 = 0.96, Figure 2), P(RA) (y = 0.87x + 0.4, r2 = 0.91) and P(LA) (y = 0.90x + 0.48, r2 = 0.93). Our comprehensive physiological model of circulation is useful in accurately predicting hemodynamics from the measurement of a single set of CO, P(RA) and (P(LA) following blood volume changes. Therefore, this model enables continuous monitoring of blood volume and pump performance for automatic hemodynamic piloting.

High resolution optical mapping of cardiac action potentials in freely beating rabbit hearts

Optical mapping of action potentials (APs) has become important tools for the cardiac electrophysiology. However, cardiac contraction unavoidably produces motion artifacts (MA) in optical signal. We developed a method to suppress motion artifacts without arresting the hearts.

METHODS

Using a dual-wavelength optical mapping system, APs were recorded on the surface of an isolated rabbit heart. Transmembrane APs were simultaneously recorded using glass microelectrodes. We eliminated MA in a frontal plane by a motion tracking technique. Subsequently, a dual-wavelength ratiometric method was used to remove MA in a vertical direction to a frontal plane.

RESULTS

MA were effectively removed from optical signals. Action potential duration measured from optical signals corresponded with those measured from microelectrodes (r2=0.9677). Our method enables us to map action potentials in freely beating hearts.

Immunohistochemical detection of ubiquitin-positive intracytoplasmic eosinophilic inclusion bodies in diffuse alveolar damage

AIMS

To clarify the relationship between ubiquitin-positive pneumocytes and intracytoplasmic eosinophilic inclusion bodies (IB) in patients who died of diffuse alveolar damage (DAD).

METHODS AND RESULTS

Eighteen patients with DAD were studied, in whom hyaline membranes were present in one or more out of five sections from each lobe of the lungs and 15 patients with no DAD. Light microscopy revealed hyaline membrane in over 25% of lobes from 18 patients with DAD. The cytoplasm of pneumocytes from six of 18 cases of DAD contained IB. Immunohistochemically, all IBs were characteristically positive for both ubiquitin (Ub) and cytokeratin KL-1. Cytoplasmic granules were also Ub+ in four cases of DAD without IB. IB+ or Ub+ pneumocytes were undetectable in non-DAD patients. We evaluated DAD severity based on hyaline membrane formation; the mean score in DAD with IB (3.60; n = 6) was significantly higher than that in Ub- (2.92; n = 8). Ub+ pneumocytes were found with or without IB among those cases with high DAD scores.

CONCLUSIONS

These findings suggest that disordered proteolysis in the Ub-mediated proteasome system leads to the accumulation of abnormal ubiquitinated protein, which includes cytokeratin, in pneumocytes. This is the first report to suggest that Ub+ pneumocytes are associated with disease severity in patients with DAD.

A fatal case of ruptured giant coronary artery aneurysm

A 5-year-old Japanese boy died because of a ruptured left coronary artery aneurysm (CAA). He was diagnosed as having Kawasaki disease (KD) on the 5th day from onset, with all of the principal signs. On the 7th day of illness, bilateral CAAs were already found via echocardiography, and he was treated with intravenous (IV) gamma globulin and oral ASA. However, the fever persisted and the CAA progressed rapidly. Echocardiography on the 12th illness day showed a giant (18-mm) left anterior descending (LAD) artery aneurysm. Oral propranolol and nifedipine were administered, in conjunction with warfarin/aspirin anti-coagulation therapy. On the 13th day of illness, cardiac arrest developed abruptly, and, despite cardiopulmonary resuscitation (CPR), the patient remained unresponsive and died one hour later. The final pathological diagnosis was a ruptured LAD artery aneurysm and cardiac tamponade. Microscopic investigation of the ruptured vascular wall revealed marked neutrophilic infiltration, with fewer macrophages and lymphocytes. CAA ruptures are a very rare, but fatal, complication of KD. Based on a review of previous reports on CAA ruptures, we consider it useful to distinguish aneurysms which rapidly dilate and continue to expand beyond a diameter of 10 mm with ongoing vasculitis (these CAAs can be termed "super-giant") from the more common giant CAAs limited to a diameter of 8 or 9 mm, because a decision must be made as to whether to start intensive care or to intervene surgically, in order to ensure the survival of patients with such a potentially critical complication.

Filter cigarette smoking and lung cancer risk; a hospital-based case--control study in Japan

Recent changes in the histology of lung cancer, namely a relative increase of adenocarcinoma compared to squamous cell carcinoma, might be due to a temporal shift from nonfilter to filter cigarettes. To investigate the association between type of cigarette and lung cancer by histological type, we conducted a case–control study in Japan, comprising 356 histologically confirmed lung cancer cases and 162 controls of male current smokers, who provided complete smoking histories. Overall, logistic regression analysis after controlling for age and prefecture revealed decreased risk, as shown by adjusted odds ratios, for both squamous cell carcinoma and adenocarcinoma among lifelong filter-exclusive smokers as compared to nonfilter or mixed smokers. This decrease was greater for squamous cell carcinoma than for adenocarcinoma. Among men under 54 years, filter-exclusive smokers displayed increased risk of adenocarcinoma, but decreased risk of squamous cell carcinoma. The recent shift in histology from squamous cell carcinoma to adenocarcinoma, particularly among younger smokers, might be due to changes in cigarette type. However, among subjects aged 65 years or more, no differences in histological type appeared related to type of cigarette smoked, implying that other factors are associated with increases in adenocarcinoma among older Japanese population.

Bezold-Jarisch Reflex Blunts Arterial Baroreflex via the Shift of Neural Arc toward Lower Sympathetic Nerve Activity

Although the Bezold-Jarisch (BJ) reflex is potentially evoked during acute myocardial ischemia or infarction, its effects on the static characteristics of the arterial baroreflex remain to be analyzed in terms of an equilibrium diagram between the neural and peripheral arcs. The neural arc represents the static input-output relationship between baroreceptor pressure input and efferent sympathetic nerve activity (SNA), whereas the peripheral arc represents that between SNA and arterial pressure (AP). In 8 anesthetized rabbits, we increased carotid sinus pressure stepwise from 40 to 160 mmHg in increments of 20 mmHg at one-minute intervals while measuring renal SNA and AP under control conditions and during the activation of the BJ reflex by intravenous administration of phenylbiguanide (PBG, 100 microg.kg(-1).min(-1)). The neural arc approximated a sigmoid curve whereas the peripheral arc approximated a straight line. PBG decreased AP at the operating point from -91.3 +/- 2.4 to -71.7 +/- 3.1 mmHg (P < 0.01), and attenuated the total loop gain at the operating point from -1.31 +/- 0.44 to -0.51 +/- 0.14 (P < 0.05). The equilibrium diagram indicated that PBG caused a parallel shift of the neural arc toward lower SNA such that the maximum SNA was reduced to approximately 60% of control. PBG decreased neural and peripheral arc gains at the operating point to approximately 43% and 77%, respectively. In conclusion, the BJ reflex blunts arterial baroreflex via the shift of the neural arc toward lower SNA.

Attenuation of hypercapnic carbon dioxide chemosensitivity after postinfarction exercise training: possible contribution to the improvement in exercise hyperventilation

OBJECTIVE

To elucidate the responsible mechanisms of increased slope of minute ventilation relative to carbon dioxide production (VE/VCO(2)) during exercise after acute myocardial infarction without overt signs of heart failure, patients who had an acute myocardial infarction were examined after participating in a three month supervised exercise training programme.

DESIGN

Exercise testing, hypercapnic CO(2) chemosensitivity measurement (rebreathing method), and pulmonary function test were repeated at entry and after three months in 50 acute myocardial infarction patients with neither symptoms nor signs of heart failure who completed the training programme. Ten patients who performed initial inhospital training served as controls.

RESULTS

Age, peak oxygen uptake, left ventricular ejection fraction, CO(2) chemosensitivity, respiratory parameters (percentage of predicted normal vital capacity (%VC), forced expiratory volume in one second, and carbon monoxide transfer factor (%TLCO)) were all significantly correlated with VE/VCO(2) slope. Multivariate regression analysis showed that age (beta = 0.29, p = 0.01), %TLCO (beta = -0.27, p = 0.01), and CO(2) chemosensitivity (beta = 0.49, p < 0.001) were independent determinants of VE/VCO(2) slope. After three months, there was no significant change in these parameters in the control group. Peak oxygen uptake, %TLCO, and %VC and attenuation in CO(2) chemosensitivity increased significantly in the training group. The VE/VCO(2) slope decreased marginally (p = 0.11). The changes in VE/VCO(2) slope were correlated only with those in CO(2) chemosensitivity (r = 0.50, p < 0.001).

CONCLUSION

After acute myocardial infarction, exercise hyperventilation is seen in association with aging, enhanced hypercapnic CO(2) chemosensitivity, and reduced TLCO, even in the absence of overt heart failure. The correlation of VE/VCO(2) attenuation after training with the reduction in CO(2) chemosensitivity suggests that exercise training may reduce increased VE/VCO(2) slope, at least partially by reducing CO(2) chemosensitivity.

Changes in single unit activity in the lateral hypothalamic area of goats during feeding

In order to clarify the location of feeding centers in the ruminant brain, this study used a single-unit activity (SUA) recording electrode to investigate the existence of appetite-regulating neurons in the lateral hypothalamic area (LHA) in goats. Seven male Japanese Saanen goats were used in the experiment. The animals were fed twice daily, once in the morning (1000 to 1200) with 1.5 kg of roughly crushed alfalfa hay cubes, and once in the afternoon (1600 to 1800) with 200 g of commercial ground concentrate feed. The animals were allowed free access to drinking water. In this study, the animals were surgically operated on to position the recording electrode in the LHA. Recordings of SUA were carried out continuously over a 2.25-h period beginning 15 min prior to the commencement of morning feeding. The eating rates of crushed alfalfa hay cubes were highest 10 min after feeding commencement, but decreased sharply by the time 40 min had elapsed. The cumulative feed intake after the completion of the 2-h feeding period was 1164 +/- 38 g. The cumulative water intake upon the conclusion of the 2-h feeding period was 2422 +/- 107 mL. This study recorded 31 units, of which five showed a response to feeding and altered their firing rates. In response to a sharp increase in eating rates, all five units increased their firing rates to a level higher than that of prefeeding (P < 0.05). As the animals reached a level of satiety (eating rates declined to very low levels), firing of units I and II stopped completely, while the firing rates of units III, IV, and V decreased. Examination of a serial histological section confirmed that the five units in which changes in firing rates with feeding were observed were all located in the dorsolateral hypothalamic area close to the fornix. The LHA neurons recorded in this experiment characteristically showed neuronal activity increases at high levels of feeding, but decreases at low levels. The results suggest that there are cells located in the LHA of goats that are active in the physiological regulation of hay (dry forage) intake.

Serious bradyarrhythmia that was possibly induced by lopinavir-ritonavir in 2 patients with acquired immunodeficiency syndrome

We describe 2 patients with acquired immunodeficiency syndrome who had potentially fatal bradyarrhythmia that occurred shortly after commencement of antiretroviral therapy. Lopinavir-ritonavir was the only drug that both patients were using.

Gene-specific response of dynamic ventricular repolarization to sympathetic stimulation in LQT1, LQT2 and LQT3 forms of congenital long QT syndrome

AIMS

Differences in the sensitivity of the genotype of the congenital long QT syndrome to sympathetic stimulation have been suggested. This study compared the influence of sympathetic stimulation on continuous corrected QT (QTc) intervals between LQT1, LQT2 and LQT3 forms of the congenital long QT syndrome.

METHODS AND RESULTS

We recorded a 12-lead electrocardiogram continuously before and after bolus injection (0.1 microg x kg(-1)) of epinephrine followed by continuous infusion (0.1 microg x kg(-1) min(-1)) in 12 LQT1, 10 LQT2, 6 LQT3, and 13 control patients. The QT intervals and previous RR intervals of all beats were measured semi-automatically, and the QTc intervals of all beats were calculated by Bazett's method. The dynamic response of the RR interval to epinephrine was no different between the four groups. The QTc was prolonged remarkably (477+/-42 to 631+/- 59 ms; P<0.0005, % delta prolongation =+32%) as the RR was maximally decreased (at peak of epinephrine), and remained prolonged at steady state conditions of epinephrine (556+/-56 ms; P<0.0005 vs baseline, +17%) in LQT1 patients. Epinephrine also prolonged the QTc dramatically (502+/-23 to 620+/-39 ms; P<0.0005, +24%) at peak of epinephrine in LQT2 patients, but this shortened to baseline levels at steady state (531+/-25 ms; P=ns vs baseline, +6%). The QTc was much less prolonged at peak of epinephrine in LQT3 (478+/-44 to 532+/-41 ms; P<0.05, +11%) and controls (394+/-21 to 456+/-18 ms; P<0.0005, +16%) than in LQT1 and LQT2 patients, and shortened to the baseline levels (LQT3; 466+/-49 ms, -3%, controls; 397+/-16 ms, +1%; P=ns vs baseline) at steady state.

CONCLUSION

Our data suggest that the dynamic response of ventricular repolarization to sympathetic stimulation differs between LQT1, LQT2 and LQT3 syndromes, and may explain why the trigger of cardiac events differs between the genotypes.

Heart rate-independent vagal effect on end-systolic elastance of the canine left ventricle under various levels of sympathetic tone

BACKGROUND

Although we have shown that in rabbits the direct (heart rate [HR]-independent) vagal effect on left ventricular end-systolic elastance (E(es)) was negligible under minimal sympathetic tone, how underlying sympathetic tone modulates the inotropic response to vagal stimulation remains unknown.

METHODS AND RESULTS

We used an isolated canine heart preparation with functioning autonomic nerves. We examined the direct vagal inotropic effect by measuring E(es) under fixed-rate atrial pacing with or without concomitant sympathetic nerve stimulation. Right and left vagal stimulation at 20 Hz decreased HR by 27+/-3% and 14+/-2%, respectively, and decreased E(es) by 11+/-2% and 6+/-2%, respectively. When we fixed HR by atrial pacing, right and left vagal stimulation at 20 Hz did not decrease E(es) (0.01+/-0.3% and 0.3+/-0.4%; NS). Concomitant left sympathetic nerve stimulation at 4 Hz enhanced direct vagal negative inotropism to -19+/-3% and -34+/-5% for 20-Hz right and left vagal stimulation (interaction, P<0.01).

CONCLUSIONS

Direct vagal negative inotropism was unobservable with minimal sympathetic tone in dogs but was enhanced with concomitant sympathetic stimulation.

Effects of ketamine on in vivo cardiac sympathetic nerve endings

Using the dialysis technique, we examined the effect of ketamine on dialysate norepinephrine (NE) levels in the myocardial interstitial space in anesthetized cats. Dialysis probes were implanted in the left ventricular myocardium, and we measured the dialysate NE levels serving as an indicator of NE output at the cardiac sympathetic nerve endings. During local administration of ketamine (10 mM), we examined the time-course of the change in dialysate NE levels and the dialysate NE response to coronary occlusion. Dialysate NE levels significantly increased from 39+/-7 pg/ml at control to 133+/-22 pg/ml 30 min after beginning the ketamine administration. Addition of either omega-conotoxin GVIA (N-type calcium channel blocker) at 10 microg/kg intravenously or desipramine (neuronal NE transport blocker) at 100 microM did not inhibit the increment in dialysate NE evoked by ketamine. These findings suggest that the increase in dialysate NE evoked by ketamine is dependent neither on the activity of NE exocytosis nor on the neuronal NE transport. Left descending coronary artery occlusion evoked increments in dialysate NE. The addition of ketamine augmented the dialysate NE response to coronary occlusion. A ketamine-induced increment in dialysate NE might occur as a consequence of NE exocytosis independent or membrane NE transport insensitive efflux of NE.

In vivo assessment of acetylcholine-releasing function at cardiac vagal nerve terminals

We examined whether the ACh concentration measured by cardiac microdialysis provided information on left ventricular ACh levels under a variety of vagal stimulatory and modulatory conditions in anesthetized cats. Local administration of KCl (n = 5) and ouabain (n = 7) significantly increased the ACh concentration in the dialysate to 4.3 +/- 0.8 and 7.3 +/- 1.3 nmol/l, respectively, from the baseline value of 0.6 +/- 0.5 nmol/l. Intravenous administration of phenylbiguanide (n = 5) and phenylephrine (n = 6) significantly increased the ACh concentration to 5.4 +/- 0.9 and 6.0 +/- 1.5 nmol/l, respectively, suggesting that the Bezold-Jarisch and arterial baroreceptor reflexes affected myocardial ACh levels. Modulation of vagal nerve terminal function by local administration of tetrodotoxin (n = 6), hemicholinium-3 (n = 6), and vesamicol (n = 5) significantly suppressed the electrical stimulation-induced ACh release from 20.4 +/- 3.9 to 0.6 +/- 0.1, 7.2 +/- 1.9, and 2.7 +/- 0.6 nmol/l, respectively. Increasing the heart rate from 120 to 200 beats/min significantly reduced the myocardial ACh levels during electrical vagal stimulation, suggesting a heart rate-dependent washout of ACh. We conclude that ACh concentration measured by cardiac microdialysis provides information regarding ACh release and disposition under a variety of pathophysiological conditions in vivo.

Integrative sympathetic baroreflex regulation of arterial pressure

The baroreflex system is the most important negative feedback control system functioning physiologically to attenuate the effects of rapid perturbation in arterial pressure. However, the complexity of the system resulting from the closed-feedback loop, nonlinearity, and system memory makes detailed quantitative characterization of the baroreflex system difficult. To overcome such limitations, we proposed a framework to decompose the baroreflex loop into two major arcs, that is, the mechanoneural arc and neuromechanical arc. Steady state analysis indicated that such decomposition allowed us to analytically determine the operating point by equilibrating two respective function curves. Dynamic analysis suggested that the mechanoneural arc accelerated the slow mechanical response of the neuromechanical arc. The acceleration mechanism in the mechanoneural are optimized arterial pressure regulation in achieving both stability and quickness. Establishment of such an integrative framework allowed the development of an artificial feedback control system able to regulate sympathetic vasomotor tone.

Laterality in direct and indirect inotropic effects of sympathetic stimulation in isolated canine heart

Although sympathetic nerve stimulation is known to increase ventricular contractility, concomitant increases in heart rate (HR) make it difficult to separate its direct inotropic effect from indirect inotropic effect through a force-frequency mechanism. We stimulated the stellate ganglia in 8 isolated canine hearts with functional sympathetic nerves. Right sympathetic stimulation at 10 Hz increased ventricular end-systolic elastance (E(es)) by 95.7 +/- 7.5% (p < 0.001) and HR by 32.5 +/- 4.2% (p < 0.05). In contrast, left sympathetic stimulation at 10 Hz increased E(es) by 70.7 +/- 6.5% (p < 0.001) without significant changes in HR. Preventing the chronotropic response by fixed-rate pacing attenuated the E(es) response to right sympathetic stimulation at 5 Hz (52.0 +/- 5.1 vs. 22.8 +/- 2.8%, p < 0.001), but not to left sympathetic stimulation at 5 Hz (54.5 +/- 3.4 vs. 53.3 +/- 2.2%, NS). In the isolated canine heart, the right sympathetic nerve affected E(es) by both the direct inotropic effect and the indirect HR-dependent inotropic effect. In contrast, the left sympathetic nerve regulated E(es) primarily by its direct inotropic effect.

A new model-based method of reconstructing central aortic pressure from peripheral arterial pressure

We have shown in our previous study that the transfer function between central aortic pressure and tonometric radial arterial pressure can be modeled as a pure elastic lossless tube terminated with a modified Windkessel. We hypothesized, using the model-derived radial arterial flow, that central pressure could be reconstructed by adding the time-shifted forward and backward pressure components (Stergiopulos et al.: Am J Physiol 274: H1386---H1392, 1998). In eight patients (age 16--75), central micromanometric and radial arterial tonometric pressure were measured simultaneously. We imposed measured tonometric pressure to the terminal modified Windkessel to estimate radial arterial flow, with which tonometric pressure was separated into forward and backward components. These components were then appropriately time shifted, and summed to central pressure. We used average parameter values for the terminal impedance, but individualized the transmission delay. The poor correlation (r(2)) between tonometric and central pressure (0.264--0.765) was improved by both central pressure reconstruction methods (generalized transfer function: 0.887--0.974, model-based method: 0.849--0.979). The sensitivity analysis indicated that the key model parameter in reconstructing central pressure was the transmission delay. We conclude that our model-based method was capable of reconstructing central pressure as precisely as the generalized transfer function method, and also capable of individualizing the transfer function by changing the transmission delay.

Dynamic sympathetic control of atrioventricular conduction time and heart period

Although power spectra of R-R and P-R intervals in response to random respiration show similar frequency distributions, the way in which dynamic sympathetic regulation contributes to such similarity remains unknown. We estimated the transfer function from sympathetic stimulation to the atrioventricular interval (AV conduction time; T(AV)) with and without constant atrial pacing in seven anesthetized cats. The transfer function from sympathetic stimulation to T(AV), except for absolute gain values, approximated a low-pass filter similar to that from sympathetic stimulation to the A-A interval (heart period; T(AA)). The 90%-rise times did not differ between the T(AA) and T(AV) step responses (32.3 +/- 1.8 vs. 29.6 +/- 3.2 s). Constant pacing augmented the T(AV) step response (-0.58 +/- 0.10 vs. -0.86 +/- 0.12 ms/Hz, P < 0.05) without affecting the 90%-rise time. These findings suggest that the dynamic characteristics of sympathetic control are similar between T(AA) and T(AV) despite the different electrophysiological mechanisms determining T(AA) and T(AV). A numerical simulation indicated that if the dynamic characteristics of the sympathetic control do not match between T(AA) and T(AV), a critical condition for initiation of reentrant tachycardia would be encountered.

Differential dynamic baroreflex regulation of cardiac and renal sympathetic nerve activities

Although regional difference in sympathetic efferent nerve activity has been well investigated, whether this regional difference exists in the dynamic baroreflex regulation of sympathetic nerve activity remains uncertain. In anesthetized, vagotomized, and aortic-denervated rabbits, we isolated carotid sinuses and randomly perturbed intracarotid sinus pressure (CSP) while simultaneously recording cardiac (CSNA) and renal sympathetic nerve activities (RSNA). The neural arc transfer function from CSP to CSNA and that from CSP to RSNA revealed high-pass characteristics. The increasing slope of the transfer gain in the frequencies between 0.03 and 0.3 Hz was significantly greater for CSNA than for RSNA (2.96 +/- 0.72 vs. 1.64 +/- 0.73 dB/octave, P < 0.01, n = 9). The difference was hardly explained by the difference in static nonlinear characteristics of CSP-CSNA and CSP-RSNA relationships or by the difference in conduction velocities in the multifiber recording. These results indicate that the central processing in the brain stem differs between CSNA and RSNA. The neural arc of the baroreflex may exert differential effects on the heart and kidney in response to dynamic baroreflex activation.