Isaac Starr

Twenty-Year Studies with the Ballistocardiograph The Relation between the Amplitude of the First Record of "Healthy" Adults and Eventual Mortality and Morbidity from Heart Disease

A group of 211 healthy persons, gathered together from 23 to 17 years ago to provide normal standards for the ballistocardiograms, has been followed to the present time. This study is concerned with the group as a whole, or with subgroups. The study of individuals will appear in later communications. With a few minor exceptions, the group, at the time of the initial test, had ballistocardiograms normal in form. The death rate of the group during the period of the study was much lower than was to be anticipated from standard life expectancy tables for the state of Pennsylvania. Although the ballistocardiograms of this group were normal in form, there was great variation in the amplitude of the records. Since these records were calibrated in terms of force, this can be interpreted as due to differences in the force of the hearts contraction. Those whose hearts contracted with little force at the initial test later suffered from death and cardiac disability, chiefly coronary heart disease, in far greater numbers than those whose hearts contracted strongly. The interpretation of this striking finding is bound up with that of another; in our data, as in that of others, there is strong correlation between the ages of the subjects and the amplitudes of their ballistocardiograms. The heart tends to weaken and beat with less coordination as it grows older. Exact expressions can be given to the normal rate of decline of cardiac function with the years by the slope of the regression between age and ballistocardiogram amplitude found in the group of those who remained healthy for 17 years after the test. The normal rate of cardiac decline, as age advances, agrees closely with that calculated for several bodily functions. The cardiac decline seems a little more rapid than that of these other bodily functions, but the difference is small and its significance is doubtful. In our data, among those who later developed undoubted heart disease, there was no significant tendency for the older people to have smaller ballistocardiograms than the younger. From the chronologic age of each subject who later developed undoubted heart disease, and the age-amplitude regression of those who remained healthy, one can calculate for each cardiac subject, the amplitude expected if he had been in perfect health. On the average this expected value far exceeds the values found. So most of these hearts were performing abnormally for their age although manitest clinical evidence of heart disease had not yet appeared. When the effect of age is eliminated by pairing the results directly, the relation between the amplitudes of the initial ballistocardiograms and the later development of heart disease remains significant for p = 0.05. Indeed, in such pairs with similar ages, those who later developed heart disease had initial ballistocardiograms that averaged 25 per cent smaller than those of their mates who remained healthy for the next 17 years. This is additional evidence that the hearts of those who were to develop myocardial disease were, at the initial test, like the hearts of much older normal people; that is, the physiologic age of their hearts far exceeded the chronologic age.

Blind study on the action of digitoxin on elderly women

Abstract In 11 elderly women, all in normal rhythm, and some with manifest heart disease, digitoxin and control periods, each of about a months duration, were alternated. In 6 patients who had been taking digitoxin daily for many years, the drug was withdrawn. A few acute experiments were made. Both the patients and the observers were “blind”; the latter attempted to identify digitoxin action both by the ordinary clinical observations, and by records of the ULF force BCG and the carotid pulse derivative. The results were completely negative. Neither subjects nor observers could distinguish between the digitoxin and control periods. Statistical analysis disclosed no significant differences. Surprised by the negative results, the authors asked the Food and Drug Administration to test the digitoxin used; it met all official chemical and biological requirements, although it had failed to stimulate the frog heart in a type of test now discarded. In the great majority of our subjects, their hearts, though not stimulated by digitalis, were markedly stimulated by mild exercise. This demonstrates clearly: (1) that our failure to demonstrate digitoxin action in the majority cannot be attributed to incapacity of elderly hearts to respond to stimulation; (2) that the techniques used were fully capable of demonstrating cardiac stimulation when it occurred. That digitalis glycosides benefit some types of diseased hearts more than others has long been known and all types were not represented in our study. Our negative results do not indicate that such drugs should never be tried in elderly persons in normal rhythm; our studies detected no harm from digitoxin administration. Nevertheless, Mackenzies skeptical view is strongly upheld by our negative results, and our findings contrast sharply with the therapeutic expectations of many doctors.

Standardization of the Ballistocardiogram by Simulation of the Heart's Function at Necropsy; With a Clinical Method for the Estimation of Cardiac Strength and Normal Standards for It

The interpretation of ballistocardiograms has been attacked by a method entirely new; a physiologic experiment performed at necropsy. While the subject lies on the ballistocardiograph a normal diastolic pressure is created and the hearts function is simulated by injecting fluid into the aorta and pulmonary artery, the amount injected at each instant being recorded. The resulting ballistocardiograms can be directly compared with many aspects of cardiac function. The amplitude of the ballistocardiogram measures the maximum force exerted by the heart in moving the blood and preliminary normal standards for this estimate of cardiac strength have been set up.

A clinical study of the first derivative of the brachial pulse. Normal standards and abnormalities encountered in heart disease

Abstract Records of the first derivative of the brachial pulse have been secured in over 175 subjects, 68 of whom were healthy persons, and the rest of whom were ambulatory hospital patients. Such records emphasize certain features of the conventional pulse, such as angles, notches, and slurs, and suppress other features, such as the slow movements of the base line. From data secured in the healthy subjects, normal standards have been defined for contour, for amplitude of the chief deflection, and for the durations of the main waves on the base line. As age advances, the main wave of the pulse derivative diminishes in amplitude and becomes broader on the base line, even though health commensurate with ones age is retained. Conspicuous abnormalities of the contour of the pulse derivative were identified by inspection in 43 hospital patients, in all of whom conventional hospital studies gave strong evidence of cardiac abnormality. Four persons who believed themselves to be healthy showed similar abnormalities of the pulse derivative contour. The amplitude of the main wave of the pulse derivative indicates the maximum rate at which blood pressure rises on the advancing pulse wave front. Amplitudes greater than normal were found in 25 patients. No one was encountered who had an amplitude less than normal; the reason for this is discussed. Abnormalities of wave duration were identified in a small number of patients—the exact number depended on whether this duration was expressed as time or as per cent of the cardiac cycle. Changes in the amplitude and duration of the main wave of the pulse derivative permit one to identify aspects of the cardiac performance which change as age advances. These indicate the “normal” weakening of the myocardium as one grows older. Abnormalities of contour of the pulse derivative indicate an incoordination of the cardiac contraction.

The relation of the ballistocardiogram to cardiac function

Abstract From the foregoing account one can see that determined efforts have been made to identify sources of error in the force ballistocardiogram as a record of the resultant of the forces generated by the heart. In few other clinical methods has there been such an extensive hunt. Certainly nothing of much moment has turned up, although we found reason to believe that differences in vessel elasticity and in size of subject would have some effect on the relation between the record and the forces which originate it. An effect of the position of the heart and of arteriosclerosis would manifest itself only if the deviation from the normal was extreme. While various other uncertainties certainly exist, it seems unlikely that they are important. Clinicians are often forced to use methods which are not perfect and much trouble can be avoided by proper care in the preparation and use of normal standards. In the case of the ballistocardiogram the scatter of the normal population will be due not only to differences in the forcefulness of the hearts of healthy persons, but also to all the other factors which might have some effect on the proper recording of such forces, such as: differences in body size, body habitus, blood pressure, arteriosclerosis and the usual technical errors. But if we identify as abnormal only those outside the statistical limits of normality of such a healthy group, the odds are strongly in favor of a valid identification of abnormality of the cardiac forces. Using the data from the cadaver experiments as a test, one finds we can identify the magnitude of the acceleration of the blood with an accuracy about equal to that of the common clinical method of taking blood pressure. It is true that the Riva-Rocci method has greater errors inherent in it than most doctors realize; the point I wish to make is that despite such errors, much important clinical information has been secured by means of it, and its introduction into clinical medicine was responsible for a great advance. I am hopeful, therefore, that the time has come when doctors will no longer be content to regard heart diseases from the purely anatomic viewpoint, and when a statement of the strength and coordination of the hearts contraction will be considered an essential part of the record of every cardiac case. The ballistocardiogram provides a method of readily securing information on this point without causing the least discomfort to the patient.

The Effect of Aging and of the Development of Disease on the Ballistocardiogram A Study of Eighty Subjects, Originally Healthy, Followed from Ten to Fourteen Years

Second ballistocardiograms were taken on 80 subjects who were part of the series from which the original normal standards were secured in 1937, 1938 and 1939. Sixty-five of these subjects had retained good health during the long interval; by a comparison between their old and new records, the effect of normal aging on the ballistocardiogram has been determined. Fifteen subjects lost their health during the interval and two of them have died since the second test. The evolution of their ballistocardiograms has been compared with the development of their disabilities.

Studies Made by Simulating Systole at Necropsy VIII. Significance of the Pulse Pressure

In a previous publication we defined the relation between cardiac work and blood pressure measurements by means of multiple regression equations. In this communication we seek a simple method of estimating cardiac work from observations any doctor could make in the clinic.

Clinical Tests of the Simple Method of Estimating Cardiac Stroke Volume from Blood Pressure and Age

The results secured by the simple method of estimating cardiac output, described in the previous paper, have been applied to data already in the literature. Hundreds of comparisons are possible between estimates of cardiac output made by our blood pressure method from values given by the authors, and the results of estimates of cardiac output by some other well known method employed at the same time. The correlations between the results secured by our method and those obtained by Fick, acetylene, ethyl iodide, and nitrous oxide methods are often surprisingly good. There is an unexplained difference in level between the Fick and blood pressure estimates, but not between the blood pressure estimates and those of the three older methods.

Studies Made by Simulating Systole at Necropsy XII. Estimation of the Initial Cardiac Forces from the Ballistocardiogram

In a preceding communication1 I have suggested that we are now in a position to look more closely at the higher dynamic aspects of cardiac function in our patients. From an analogy with common experience in automobiles, it is reasonable to expect that the first sign of myoeardial weakness will manifest itself in diminished ability to accelerate the blood, that is, in diminished cardiac forces. In this study we aim to establish a clinical method of detecting abnormalities of the initial cardiac forces by means of the ballistocardiogram.

Studies Made by Simulating Systole at Necropsy XI. On the Higher Dynamic Functions of the Heart, and their Reflections in the Pulse Wave

The background of basic conceptions underlying this paper is familiar to everyone by experience. The powerful automobile, when the light turns from red to green, can make a jack-rabbit start which leaves the less powerful cars far behind. If the engine of the big car is weakened, as by a cylinder gone dead, it can still do 50 miles per hour on the parkway and get you to your destination, but the jack-rabbit start is no longer possible. The first sign of weakness manifests itself more conspicuously in differences in acceleration than in differences in speed or in the distance that can be traveled; indeed the effect on acceleration may be the only manifestation of weakness. There is every reason to expect that this analogy would hold for the cardiac performance also and that the first sign of a weakening myocardium might well manifest itself by a change in acceleration of the ejected blood. This paper is chiefly concerned with the development of a simple method, available to clinicians, which would aid in the detection of changes in the acceleration of the aortic blood. This opportunity has also been used to answer certain criticisms of our previously published work, some, we believe, due to misunderstanding of the mathematical background of statistical procedures, and others due to a frank difference of opinion concerning the accuracy of the Fick method used to test the results secured by one of our formulas published previously.