Gilbert E. Levinson

Evidence for Cardiomyopathy in Familial Diabetes Mellitus

Recent epidemiologic studies have suggested that cardiac disease in common in diabetics and may often have a noncoronary basis. To examine the status of the left ventricle, 17 adult-onset diabetics of familial type without hypertension or obesity underwent hemodynamic study and were compared to 9 controls of similar age. Of the 17, 12 subjects had no significant occlusive lesions by coronary angiography. From this group eight without heart failure had a modest, but significant, elevation of left ventricular end-diastolic pressure. End-diastolic and stroke volumes were reduced, but ejection fraction and mean rate of fiber shortening were within normal limits. The left ventricular end-diastolic pressure/volume ratio was significantly higher than controls. Afterload increments effected a significant increase of filling pressure compared to normals without a stroke volume response, consistent with a preclinical cardiomyopathy. Four patients with prior heart failure had similar but more extensive abnormalities. None had local dyskinesia by angiography, and lactate production was not observed during pacing-induced tachycardia. Left ventricular biopsy in two patients without ventricular decompensation showed interstitial collagen deposition with relatively normal muscle cells. These findings suggest a myopathic process without ischemia. Postmortem studies were performed in 11 uncomplicated diabetics. Nine were without significant obstructive disease of the proximal coronary arteries, and the majority succumbed with cardiac failure. On left ventricular sections, none had evident luminal narrowing of the intramural vessels. All nine exhibited periodic acid-Schiff-positive material in the interstitium. Collagen accumulation was present in perivascular loci, between myofibers, or as replacement fibrosis. Multiple samples of left ventricle and septum revealed enhanced triglyceride and cholesterol concentrations, as compared to controls. Thus, a diffuse extravascular abnormality may be a basis for cardiomyopathic features in diabetes.

Assessing Clinical Skills of Residents with Standardized Patients

Current techniques do not provide a reproducible, reliable, or valid basis for assessing clinical skills. The need for large-scale direct observation and standardized assessment procedures has precluded development of better techniques. A project using standardized patients presenting with common clinical problems evaluated the skills of 336 internal medicine residents at 14 New England residency programs in 1289 standardized patient and resident encounters. Results indicated that reproducible assessment of the clinical skills could be achieved in approximately 1 day of testing time using standardized patients. Resident performance improved with years of training, and senior residents and those from programs with stronger reputations performed better and were more homogeneous in ability. Low correlations between standardized-patient-based measures of clinical skills and other evaluation techniques suggested that standardized patients provided unique information. Reactions of residents and faculty to standardized-patient-based evaluations were favorable.

Ventricular function in noncardiacs with alcoholic fatty liver: role of ethanol in the production of cardiomyopathy

Since many patients with cardiomyopathy have a history of chronic ethanolism often associated with malnutrition, we have evaluated left ventricular (LV) function in alcoholics with fatty liver, who had no clinical evidence of cardiac or nutritional disease. During an afterload test of LV function the pressor response to angiotensin evoked a threefold rise of enddiastolic pressure in the alcoholic group which was substantially greater than the 4 mm Hg rise in control subjects. The stroke volume and stroke work response in the noncardiac alcoholic was significantly less than in controls. Diminished LV function was corroborated in the noncardiac alcoholic at rest, using a contractility index. To evaluate the dose-response relationship of ethanol in the production of cardiac malfunction, two groups of noncardiac alcoholic subjects were studied acutely at low and moderate dose levels. After 6 oz, ventricular function, myocardial blood flow, and metabolism were not significantly affected. After 12 oz, there was a progressive rise of end-diastolic pressure and decrease of stroke output at a mean blood alcohol level of 150 mg/100 ml, reverting toward control by 4 hr. The coronary effluent transiently evidenced leakage of cell constituents, despite an increase of coronary blood flow, suggesting a direct but reversible cardiac injury. Myocardial extraction of triglyceride was enhanced, whereas FFA uptake was reduced. A possible role of myocardial triglyceride accumulation in heart muscle was considered in pathogenesis. Chronic ingestion of 16 oz of Scotch daily by an alcoholic subject while on a normal diet produced, after 12 wk, a progressive increase of heart rate and size, circulation time, and venous pressure, and a ventricular diastolic gallop. Normal values were restored within 7 wk after interrupting alcohol. These several studies suggest that the cumulative effects of repeated ingestion of ethanol in intoxicating doses can produce diminished LV function before clinical evidence of cardiac abnormality, or heart disease not necessarily related to malnutrition.

Hemodynamic effects of staged hematocrit reduction in patients with stable cor pulmonale and severely elevated hematocrit levels

Patients with cor pulmonale and high hematocrit levels are often subjected to phlebotomy in the belief that the adverse effects of high viscosity may outweigh the benefit of increased oxygen carrying capacity. To evaluate this, 12 patients with stable cor pulmonale and hematocrit values greater than 55 per cent were studied before and after a series of venesections. Right heart and aortic pressures, cardiac output and blood gases were measured at three mean hematocrit levels, 61 per cent (stage I), 50 per cent (stage II) and 44 per cent (stage III), with blood volume unchanged. From stages I to II, there were significant decreases in both man pulmonary artery pressure and total pulmonary resistance. Oxygen transport fell but not oxygen consumption. Right ventricular end-diastolic pressure and cardiac output did not change. Right ventricular work either fell or was maintained by increased output. Frank-Starling performance (supine exercise) improved. No significant changes occurred with further reduction in hematocrit to normal levels (stage III). The findings of this study support the concept of overcompensating erythrocytosis in cor pulmonale, and the effects of moderate hematocrit reduction should not be overlooked in these severely ill patients.

Left ventricular function during the early and late stages of scar formation following experimental myocardial infarction

Abstract To evaluate changes in left ventricular (LV) function after myocardial infarction (MI), anesthetized closed-chest dogs were studied following survival from acute nonmassive MI induced by thrombus formed on a catheter electrode in either branch of the left main coronary artery. Seven dogs were studied 3 to 4 weeks after MI (early scar formation) and 12 dogs 6 to 8 weeks after MI (late scar formation). Ten normal dogs with heart rates and aortic pressures similar to the MI groups served as controls. LV function was evaluated at rest and during afterloading with angiotensin. In both the early and late stages of scar formation, the infarcted ventricles had normal end-diastolic volumes and normal isovolumetric velocity-force length relationships but elevated resting end-diastolic pressures. LV ejection-rate functions (stroke power, mean rate of ejection, and mean rate of circumferential fiber shortening) were reduced after MI, the differences becoming statistically significant by 6 to 8 weeks. There were no significant differences found between the two infarct groups. With augmented afterloading, LV function after MI appeared depressed when stroke work was plotted against LV end-diastolic pressure but not when plotted against end-diastolic circumference. Thus, although its length-tension (Frank-Starling) characteristics and isovolumetric velocity-force-length (contractility) relationships were not altered, the healing infarcted ventricle exhibited reduced velocity of shortening during ejection and increased LV end-diastolic pressure which, in the absence of evidence of cardiac decompensation, probably reflects a reduced compliance.

An index of the contractile state of the myocardium in man

There is a profound need, on both clinical and physiologic grounds, for a measure of the contractile state of the intact ventricle. Such a measure can be obtained by evaluating the force-velocity relationship with a correction for myocardial fiber length. The force-velocity relation can be expressed as the ratio of maximum rate of pressure rise to maximum isovolumetric pressure, a quantity which was described by Hill as the maximum rate of proportional rise of pressure and which is similar to the velocity constant of a chemical reaction. Division of this ratio by an estimate of ventricular circumference corrects for variations due to differences in initial fiber length. This index was evaluated in 11 normal subjects and 46 patients with cardiac disease during left heart catheterization. Maximum rate of pressure rise was obtained by electronic differentiation of the ventricular pressure pulse, and ventricular circumference, assuming a spherical ventricle, was calculated from volumes measured by indicator washout. The contractility index of normal subjects did not differ from that of patients with mitral stenosis, atrial septal defect, or chronic pulmonary disease (patients without left ventricular overloading). In contrast, in patients with left ventricular failure, the indices were more than two standard deviations below the mean value for normal subjects. Such a reduction was not noted in patients with pressure or volume overloading of the left ventricle before the onset of myocardial failure. During exercise, the index rose uniformly in patients without left ventricular disease, responded variably in compensated patients with volume or pressure overloading, and was virtually unchanged in patients with left ventricular decompensation. The administration of isoproterenol or digitalis resulted in increased contractility regardless of the patients status. It is concluded that the use of this index in physiologic studies of the ventricle and in diagnostic and therapeutic decisions is justified.

Studies of Cardiopulmonary Blood Volume Measurement of Left Ventricular Volume by Dye Dilution

The measurement of ventricular end-diastolic volume from washout of an indicator requiring blood sampling was studied in a heart model and in dogs and applied to the left ventricle in 34 human subjects. The model, under the ideal conditions of constant ejection fraction, uniform cycle length, and complete or nearly complete mixing, demonstrated that dye dilution accurately measures chamber volumes if the distorting effects of catheter sampling are obviated by clearance of the sampling system at least once per two cardiac cycles. The studies in dogs and in human beings demonstrated that the required sampling conditions are achievable.In 11 normal human subjects, left ventricular end-diastolic volume ranged from 72 to 99 with a mean of 82 ± 12 ml/m,2 end-systolic volume, from 22 to 60 with a mean of 37 ± 11 ml/m,2 and ejection fraction, from 0.39 to 0.71 with a mean of 0.55 ± 0.08. These measurements are in substantial agreement with results by radiocardiography and thermal dilution and are systematically, but only slightly, higher than results by quantitative angiocardiography.In six considerably older patients with chronic pulmonary disease, cardiac output was normal for the age group but ejection from a ventricle of normal size was reduced. In 12 subjects with pure mitral stenosis and in five with atrial septal defect, end-diastolic volume and ejection fraction were significantly reduced.End-diastolic volume correlated with stroke volume in all groups and in the series as a whole, but a correlation between end-diastolic volume and cardiac output was demonstrated only in the patients with mitral stenosis or atrial septal defect. This suggests that chronic restriction to ventricular filling results in decreased end-diastolic volume. The absence of a significant correlation between end-diastolic volume and end-diastolic pressure indicates that variation in ventricular compliance precludes reliance on end-diastolic pressure as a valid index of end-diastolic fiber length.

Studies of Cardiopulmonary Blood Volume Measurement of Total Cardiopulmonary Blood Volume in Normal Human Subjects at Rest and during Exercise

The volume of blood in the heart and lungs can be measured, by the Stewart-Hamilton principle, as the product of cardiac output and the mean transit time from right atrium to the aortic root. Although previous investigators have estimated a variety of central blood volumes, measurements of the true cardiopulmonary blood volume in man have not previously been reported.Eighty-one measurements of the total cardiopulmonary blood volume were obtained in 15 normal human subjects. At rest, total cardiopulmonary blood volume ranged from 301 to 546 ml/m2, with a mean of 422 ml/m2, and it represented 15% of estimated total blood volume. Cardiopulmonary blood volume was significantly larger in the male subjects than in the female. Reproducibility of measurements was good: the mean discrepancy between successive replications was 25 ml/m2 and the mean coefficient of variation 3.7%. There was no correlation between cardiac output and cardiopulmonary blood volume but a significant correlation (r=0.79, P<0.0001) was evident between cardiopulmonary blood volume and stroke volume.With elevation of the legs to the pedals of a bicycle ergometer, a small but statisticallysignificant increase occurred in cardiopulmonary blood volume, but no significant changes occurred in cardiac output, heart rate, or stroke volume. With exercise, no further significant change in cardiopulmonary blood volume occurred, despite significant increases in output, rate, and stroke volume.Analysis of cardiac output measurements, both at rest and during exercise, indicates that aortic root sampling is characterized by an appreciably higher reproducibility than that reported for peripheral arterial sampling.

Measurement of Aortic Regurgitation by Upstream Sampling with Continuous Infusion of Indicator

A direct and theoretically valid method for the measurement of aortic regurgitation involves the recording of indicator concentrations from the left ventricle and a downstream site during aortic root injection. However, this method has yielded erratic results when applied to man in our laboratory when using the sudden injection technique. Moreover, others have found a systematic overestimate of backflow when the sudden injection technique was compared with flowmeter measurements in dogs with experimental aortic regurgitation. If defects in the method are due primarily to a critical dependence of indicator distribution on the timing of injection or to beat-to-beat variations in the forward or regurgitant stroke volumes, these difficulties could conceivably be overcome by substituting continuous infusion of indicator for sudden single injections. Therefore, the upstream sampling method, using continuous infusion of indicator, was evaluated in 18 patients with aortic regurgitation during retrograde aortic and transseptal left ventricular catheterization. The continuous infusion technique was compared withthe technique of sudden injection in 10 patients and with aortic valvulography in 14 patients.Measurements of forward flow obtained with continuous infusions into the aortic root were not significantly different from measurements obtained with sudden injections into the pulmonary artery. Recordings of indicator concentrations from the left ventricle, during continuous infusions into the aortic root, demonstrated readily evident equilibrium plateaus. The resultant measurements of regurgitant flow were highly reproducible and not impaired by nonsimultaneity of upstream and downstream sampling. The percentage error of estimate at 95% confidence limits was 22% of the measurement for regurgitant flow, 13% for total flow, and 9% for the regurgitant fraction of total flow. The corresponding errors of estimate for the sudden injection technique were four times larger.Regurgitant flow by the continuous infusion method ranged from 0.8 to 30.0 L/min, total flow from 3.0 to 36.0 L/min, and the regurgitant fraction of total flow from 12 to 86%. Ranking of patients by the magnitudes of regurgitant and total flow did not correspond to ranking by angiographic criteria of severity. However, an excellent correlation prevailed between angiographic grade and the regurgitant fraction of total flow, demonstrating that this variable is the most meaningful expression of severity. The correlation (0.997) between the angiographic grade and the regurgitant fraction measured by the continuous infusion technique was clearly superior to that obtained with the sudden injection technique (0.894). Mild regurgitation was equivalent to a regurgitant fraction of <25%, moderate regurgitation to a fraction of 25 to 50%, moderately severe regurgitation to a fraction of 50 to 75% and severe regurgitation to a fraction of >75%.It is concluded that the upstream sampling method during continuous infusion of indicator, because of its sensitivity, reliability, applicability to multiple measurements, and validity in the presence of mitral regurgitation, is the most useful method for quantifying aortic regurgitation in man.

The state of the left ventricular myocardium in mitral stenosis

Summary Subnormal stroke volume (SV) and ejection fraction (EF) in mitral stenosis (MS) have been attributed to abnormality in the left ventricular (LV) myocardium despite studies in this laboratory showing normal LV myocardial contractility and oxidative metabolism. To clarify this problem, LV function was studied in 10 patients with pure MS, seven with cardiomyopathy (MYO) who had normal LV preload (end-diastolic volume) and afterload (mean systolic tension), six with atrial septal defect (ASD), and in 10 normal subjects (N). All three patient groups were characterized by subnormal SV, EF, and contractile telement velocity (Vce) at peak dP/dt despite normal preload and afterload. However, the Frank-Levinson contractility index, which normalizes Vce for diastolic fiber length, was depressed only in MYO (0.73±0.07 vs 1.39±0.07 in the N, P