Assessment of cardiac risk before peripheral vascular surgery: a comparison of myocardial perfusion imaging and long axis echocardiography at rest.

OBJECTIVE: To compare resting long axis echocardiography with adenosine thallium-201 emission tomography in detecting myocardial ischaemic abnormalities and surgical related risk in patients before peripheral vascular surgery. DESIGN: A prospective and blinded pre-operative examination of resting left ventricular minor and long axes and myocardial perfusion during adenosine vasodilation using thallium-201 emission tomography. SETTING: A tertiary referral centre for cardiac and vascular disease equipped with invasive, non-invasive and surgical facilities. SUBJECTS: 65 patients (40 male) with significant peripheral vascular disease, mean age 63+/-10 (S.D.) years, and 21 normal subjects of similar age. RESULTS: Thallium-201 myocardial perfusion tomography was abnormal in 50/65 patients; 27 had fixed, 23 reversible abnormalities (19 of whom had both). Long axis was considered abnormal if one or more of two systolic long axis disturbances, reduced extent of total excursion <1 cm at any of the three (left, septal and posterior left ventricular) sites or prolonged shortening >1 mm after A2, and two diastolic abnormalities, delayed onset of lengthening >80 ms after A2 or reduced peak lengthening velocity <4.5 cm/s, was present. Long axis score (maximum 12) was based on the presence or absence of these four disturbances at each of the three sites. Myocardial perfusion imaging with thallium-201 classified the patients into three different groups according to their liability to low, moderate or high surgical risk (summed stress perfusion score of 36). Thirteen of 50 patients were identified as subjects at high surgical risk, with a perfusion score of 22/36 and below. Twelve of these demonstrated significantly greater impairment of systolic and diastolic long axis function, compared to those at low surgical risk, with a total long axis echo score of 6/12 or more. Seventeen of 18 patients identified as being at low surgical risk, with a perfusion score of 32/36 and above, had total long axis score of less than 6/12. The remaining 19 moderate risk patients had a wide range of long axis scores. In the 65 patients studied there were two post-operative deaths, one post-discharge death due to cerebrovascular accident, and one due to renal failure. CONCLUSION: The combination of both systolic and diastolic long axis disturbances in patients with peripheral vascular disease can be used to predict the thallium assessment of surgical risk. Long axis echocardiography may thus have value as a screening test before non-cardiac surgery as well as providing a means of monitoring myocardial perfusion.

Perioperative assessment of aortic homograft, Toronto stentless valve, and stented valve in the aortic position.

We investigated aortic valve hemodynamic performance and perioperative left ventricular function in 50 patients (mean [+/- SD] age, 64 +/- 9 years; 34 men, 16 women) undergoing elective aortic valve replacement, using an aortic homograft (n = 20), a Toronto stentless porcine valve (n = 20), or a stented bioprosthesis (n = 10), by transesophageal echocardiography combined with high-fidelity cavity pressure recordings and thermodilution cardiac output measurements. Thirty-nine patients had aortic stenosis; 11 had predominant regurgitation. Thirteen patients with concomitant coronary artery stenosis underwent grafting. Left ventricular mass index in all patients was 280 +/- 110 g/m2. The transvalvular pressure drop and energy consumption were significantly higher with stented than stentless valves (5 with aortic homograft and 11 with Toronto valve, with matched age and valve size; 20 +/- 12 versus 3 +/- 9 mm Hg; 21% +/- 13% versus 8% +/- 8%, both p < 0.01). However, there was no difference in these variables between the Toronto valve and the aortic homograft (3 +/- 12 versus 2 +/- 10 mm Hg; 5% +/- 14% versus 2% +/- 12%, both p > 0.05), although the Toronto valves (normalized to body surface area) were larger than the aortic homografts (14.4 +/- 1.9 versus 12.6 +/- 1.8 mm/m2, p < 0.01). There was no significant difference in left ventricular stroke volume index or stroke work index in the systemic circulation, either between stentless and stented valves or between aortic homografts and Toronto valves, although the cross-clamp time required to insert a stentless valve was 20 minutes longer than that for a stented valve.(ABSTRACT TRUNCATED AT 250 WORDS)

Assessment of the peak tricuspid regurgitant velocity from the dynamics of retrograde flow.

We describe a simple, non-invasive and practical method to determine the peak velocity of tricuspid regurgitant flow (and hence derive systolic pulmonary artery pressure) from examination of the dynamics of retrograde tricuspid flow on Doppler. Based on a previously described relationship between right ventricular systolic pressure and the time interval between pulmonary valve closure and tricuspid valve opening, our technique does not require the peak tricuspid regurgitant velocity to be recorded; nor, as in previous studies does it rely upon recording the jugular venous pulse, right ventricular apexcardiogram or invasive pressure measurements. We have studied 65 patients with right ventricular disease (53 with pulmonary hypertension), and 24 with dilated cardiomyopathy, with M-mode, two-dimensional echocardiography, Doppler, and phonocardiography. The peak tricuspid regurgitant velocity could be predicted from the interval between pulmonary closure and the end of the tricuspid regurgitant signal on Doppler in patients with pulmonary hypertension and those with right ventricular disease with normal pulmonary artery pressure, but not in patients with dilated cardiomyopathy. In patients with pulmonary hypertension or right ventricular dilatation, this may be a useful alternative method in estimating pulmonary artery pressure from Doppler, in cases where it is not possible to record the peak tricuspid regurgitant velocity.

Cardiac function after surgery for subaortic stenosis: non-invasive assessment of left ventricular performance.

OBJECTIVE: To examine left ventricular function after surgical resection of subaortic stenosis during childhood. DESIGN: Left ventricular performance was measured non-invasively in all patients who responded to an invitation for formal assessment. SETTING: Outpatient study, tertiary referral centre. PATIENTS: Twenty three (12 male and 11 female) patients (age range 3 to 31 years) of 43 consecutive patients with fixed subaortic stenosis undergoing surgical resection between 1975 and 1989 reattended for formal assessment 16 months to 15 years (median 4 years 4 months) after operation. MAIN OUTCOME MEASURES: Left ventricular dimension, left ventricular wall thickness, left ventricular Doppler inflow velocities, and left ventricular diastolic pressure (measured from apexcardiograms). Results were compared with those in controls individually matched for age and sex. RESULTS: All patients were symptom free. Left ventricular cavity dimensions were normal, as was the mean fractional shortening. Posterior wall thickness tended to be greater in the patients and there was a significant increase in septal thickness. Normalised peak rate of posterior wall thinning was significantly lower in the patients and the isovolumic relaxation time was significantly shorter. Doppler inflow velocity measurements showed that early diastolic mitral flow acceleration time was normal but deceleration time was significantly shorter in the patients. The ratio of mitral flow in early diastole (E) to E plus mitral flow in late diastole (A) was significantly higher in the patients and in two patients there was complete absence of A wave flow despite large A waves on the apexcardiogram. CONCLUSIONS: Systolic function was well preserved in patients after operation for subaortic stenosis. A restrictive pattern of left ventricular filling was common, however, and presumably reflected a response to the chronic pressure load and to surgery in the paediatric heart.

Diastolic disease in left ventricular hypertrophy: comparison of M mode and Doppler echocardiography for the assessment of rapid ventricular filling.

OBJECTIVE: To investigate possible discrepancies between M mode and Doppler echocardiography in assessing early diastolic filling. DESIGN: Forty seven patients with left ventricular hypertrophy due to aortic stenosis and 26 healthy controls with a similar age range were studied by M mode, Doppler, apexcardiography, and phonocardiography. The patients also underwent cardiac catheterisation. M mode echograms were digitised by a computer. Early diastolic filling in both groups as assessed by the two techniques was compared. SETTING: A tertiary cardiac referral centre with facilities for non-invasive and invasive investigations. SUBJECTS: Patients referred for assessment of aortic stenosis who had left ventricular hypertrophy. MAIN OUTCOME MEASURES: Filling velocities on Doppler and rates of wall thinning and dimension increase on M mode. RESULTS: Digitised M mode indices of diastolic filling (peak wall thinning rate 6.4 (3.0) v 10.0 (3.0) cm/s and peak rate of dimension increase 9.3 (3.3) v 16 (4.5) cm/s) in the patients and controls were consistently different. In contrast, the Doppler A/E ratio and peak E wave velocity were not; they varied widely among patients with left ventricular hypertrophy. In part, this variability was because the Doppler A/E ratio, but not the digitised M mode indices, was very sensitive to the abnormalities of isovolumic relaxation frequently present in left ventricular hypertrophy. The Doppler A/E ratio varied similarly with age in both normal and hypertrophied hearts; in the patients with ventricular hypertrophy the peak rate of dimension increase depended on age only, whereas the thinning rate was independent of age in both the patients and controls. Neither the A/E ratio nor the M mode indices could be related to the left ventricular end diastolic pressure or the peak aortic pressure difference. CONCLUSIONS: When Doppler and M mode techniques are used to assess rapid filling in patients with left ventricular hypertrophy the M mode indices are more consistently abnormal. The two methods measure different aspects of left ventricular diastolic function and should be regarded as complementary rather than interchangeable.

An echocardiographic assessment of atrial mechanical behaviour.

Relations between movement of the atrioventricular ring and changes in left atrial and ventricular dimensions were studied by echocardiography and compared with apexcardiography and Doppler mitral flow velocity traces in 20 healthy controls and in patients with left ventricular hypertrophy (n = 28) or dilatation (n = 16). During left ventricular systole the atrioventricular ring, a structure common to ventricle and atrium, moved towards the ventricular apex, thus increasing left atrial volume. This action matched pulmonary venous return because it was in phase with the transverse left atrial dimension measured from aortic root to posterior left atrial wall. During early diastole, the mitral ring moved rapidly towards the atrium as transmitral flow accelerated. This requires a force directed from ventricle to atrium, likely to be the result of elastic recoil arising from compression of the ventricular myocardium or stretching of the atrial myocardium during ventricular systole. Two additional mechanisms of ventricular filling with atrial systole were recognised: (a) an increase in ventricular volume as the atrioventricular ring moved upwards and (b) transverse left ventricular expansion by pressure driven transmitral flow. The former is undetectable by Doppler from the apex; it accounted for 10% of ventricular filling in the healthy controls, but for significantly less in those with ventricular dilatation. In left ventricular hypertrophy, left ventricular filling was maintained by both mechanisms compensating for the reduced increase in volume early in diastole. Interactions between the atrium and ventricle are functionally important during ventricular systole, early diastole, and in atrial systole. They are not included in the traditional separation of atrial function into reservoir, conduit, and pump functions.

Assessment of left ventricular structure and function by cross sectional echocardiography.

Cross sectional echocardiography can be used to give semiquantitative estimates of ventricular volumes and ejection fraction which are very valuable in detecting the severe abnormalities seen clinically. The technique has also been widely used to study regional abnormalities of wall motion. However, with a frame rate of 30 s-1 and lateral resolution of 3-4 mm critical analysis of disturbances of timing is not possible, so the method should be used in conjunction with M-mode and Doppler techniques. In spite of physical limitations, based on the underlying mechanisms of image generation, its advantages of real time application and noninvasive nature, its ability to demonstrate myocardial thickness as well as echo intensity, and its relative cheapness have made cross sectional echocardiography a major tool in documenting abnormalities of left ventricular function occurring in disease.

Coronary arterial surgery in the elderly: its effect in the relief of angina.

Financial and other constraints, such as operative risk, may prevent older patients being considered for coronary arterial bypass grafting. Grafting was performed in 315 elderly patients (244 males, 71 females, age 65-79, mean 69 years) between 1981-1986. All patients had limiting angina, 38% had rest pain, 90% were housebound and 80% had triple-vessel disease. Impairment of left ventricular function was absent in 46%, mild in 20%, moderate in 23% and severe in 10% of patients. Grafts (saphenous vein or internal mammary artery) were inserted into 3 vessels (52%), 4 vessels (42%), 5 vessels (6%), 6 vessels (0.5%). Death during surgery occurred in 1.6% and a further 3.5% of patients died later during the same admission (70% of deaths were among the 33% with preoperative moderate or severe left ventricular impairment). Surgical complications included myocardial infarction (8%), cerebrovascular accident (1%), transient cerebral vascular ischaemia (5%), chest infection (10%) and wound infection (4%). Median stay on the intensive care unit was 1 day and median total hospital stay 12 days. 299 patients therefore survived to leave hospital and follow-up data are available for 217 (72%) of these. 96% were subjectively improved by surgery, 88% being free of angina on no antianginal drugs a median of 72 weeks (range 8-307) and a further 8% not limited by angina on medical therapy a median of 85 weeks (range 9-302) after surgery. We conclude that coronary arterial surgery is an effective treatment for angina in the elderly. This will have consequences for future resource allocation if the elderly are not to be denied effective therapy because of financial rather than clinical restraints.

Assessment of the severity of mitral regurgitation from the dynamics of retrograde flow.

Sixty four consecutive patients with isolated mitral regurgitation referred for Doppler echocardiography were divided into three groups: group 1, 20 patients with severe mitral regurgitation that required operation; group 2, 22 patients with severe left ventricular dysfunction and secondary mitral regurgitation; and group 3, 22 patients with mild to moderate mitral regurgitation that did not require valve operation. M mode and continuous wave Doppler traces with a simultaneous electrocardiogram and phonocardiogram were analysed to identify time intervals that could be used to distinguish patients who needed valve operation from those who did not. An interval of less than 55 ms between the aortic component of the second heart sound (A2) and the cessation of mitral retrograde flow was a powerful predictor that the patient required operation (sensitivity 100% and specificity 86%). The mean (SD) value of this variable in group 1 (40(15) ms) was significantly lower than in group 2 (90(35)ms) and group 3 (75(20)ms). Mean isovolumic relaxation time was less than normal in group 1 but did not differ significantly between groups. Deceleration of regurgitant velocity at end ejection was greater in group 1. The pressure drop from the left ventricle to the left atrium at A2 of less than 50% of the peak gradient also identified patients who needed valve operation (sensitivity 75% and specificity 68%). These findings may help to identify patients who require operation. They suggest that there are significant differences in the dynamics of flow velocities in patients with mitral regurgitation, possibly related to the relative resistances to retrograde and anterograde and anterograde flow.

Computer assisted echocardiographic assessment of left ventricular function before and after anatomical correction of transposition of the great arteries.

Left ventricular function before and after anatomical correction of transposition of the great arteries was assessed by computer assisted analysis of 78 echocardiographs from 27 patients obtained one year before to five years after operation. Sixteen patients had simple transposition, and 11 had complex transposition with additional large ventricular septal defect. Immediately after correction mean shortening fraction fell from 46(9)% to 33(8)%. There was a corresponding drop in normalised peak shortening rate from 5.4(3.7) to 3.3(1.1) s-1 and normal septal motion was usually absent. Systolic shortening fraction increased with time after correction and left ventricular end diastolic diameter increased appropriately for age. The preoperative rate of free wall thickening was significantly higher in simple (5.6(2.8) s-1) and complex transposition (4.5(1.8) s-1) than in controls (2.9(0.8) s-1). After operation these values remained high in both the short and long term. Thus, computer assisted analysis of left ventricular dimensions and their rates of change before and after anatomical correction showed only slight postoperative changes which tended to become normal with time. Septal motion was commonly absent after operation. This was associated with an increase in the rate of posterior wall thickening that suggested normal ventricular function associated with an altered contraction pattern. Computer assisted echocardiographic analysis may be helpful in the long term assessment of ventricular function after operation for various heart abnormalities.

Assessment of hypertrophic cardiomyopathy in potential aircrew.

On account of the risks of: (1) sudden death; (2) significant arrhythmias; (3) reduction in effort tolerance due to severe cardiac hypertrophy; (4) syncope; (5) systemic embolism and (6) side effects of drugs used to treat the condition, no patient with established hypertrophic cardiomyopathy should be licensed to fly. This includes those who may have undergone surgery. Those with minor degrees of hypertrophy on 2-D echo may be considered for restricted licensing subject to their having a normal exercise test and no significant arrhythmia on 24 h ambulatory electrocardiogram. If licensed, six-monthly review by a cardiologist should be required.

Assessment of mitral Björk-Shiley prosthetic dysfunction using digitised M mode echocardiography.

Digitised M mode echocardiograms were analysed in 22 patients with possible Björk-Shiley mitral prosthetic dysfunction. Patients with paraprosthetic mitral regurgitation had a significantly greater shortening fraction, an increased peak rate of dimension change during systole, and an increased peak velocity of circumferential fibre shortening than those with poor left ventricular function. Patients with a clotted prosthesis had lower values for shortening fraction and peak rate of dimension change during systole than patients with paraprosthetic regurgitation. In this latter group, the peak rate of dimension change during diastole and peak lengthening rate were greater than in either those patients with poor left ventricular function or those with a clotted prosthesis. In addition, the peak lengthening rate was greater in those with a clotted prosthesis than in those with poor left ventricular function. Thus M mode echocardiography is a useful method of assessing mitral prosthetic dysfunction and allows patients with paraprosthetic regurgitation to be distinguished from those with either poor left ventricular function or a clotted prosthesis.

Non-invasive assessment of changes in left ventricular relaxation by combined phono-, echo-, and mechanocardiography.

In order to investigate interrelations between early diastolic events, simultaneous apex-, phono-,f and echocardiograms were recorded in 10 normal and 55 patients with ischaemic heart disease. In normal subjects isovolumic relaxation time measured as A2 to mitral valve opening was 72 +/- 9 ms, the interval A2-0 point was 116 +/0 15 ms, and the amplitude of the f wave of the apexcardiogram relative to the total amplitude was 11 +/- 2%. In patients with a normal left ventricular end-diastolic pressure (less than 15 mmHg), isovolumic relaxation time was prolonged to 99 +/- 18 ms, A2-0 point to 162 +/- 5 ms, and the relative height of the f wave was increased to 21 +/- 4%. Isovolumic relaxation time and A2-0 point interval both showed significant inverse correlation with left ventricular diastolic pressure. Glyceryl trinitrate administration or isometric stress both caused well recognised changes in arterial pressure and cavity size. Neither had any additional effect during early diastole in normal subjects. In the patients, however, glyceryl trinitrate prolonged isovolumic relaxation time, delayed the 0 point of the apexcardiogram with respect to A2, increased the interval 0 point - f wave, and reduced the relative amplitude of the f wave. Isometric stress had the opposite effects. These changes extend beyond the end of ventricular relaxation and so are more readily explained on a mechanical basis, possibly as oscillations whose periodicity and degree of damping were increased by glyceryl trinitrate and decreased by isometric stress.

Echocardiographic assessment of left ventricular filling and septal and posterior wall dynamics in idiopathic hypertrophic subaortic stenosis.

In order to study left ventricular function in idiopathic hypertrophic subaortic stenosis (IHSS), left ventricular echograms were analyzed by computer and compared with results in normal subjects. Systolic function was consistently normal or above normal even in the presence of severe diastolic abnormalities. Wide variation in diastolic function in IHSS allowed separation of patients into three groups on the basis of the left ventricular peak filling rate. Because of the severe septal hypertrophy and hypokinesia, peak left ventricular filling rate is predominantly determined by the rate of free wall thinning. Patients in group 1 had rapid left ventricular filling rates, those in group 2 had normal filling rates, and those in group 3 had slow filling rates. With reduction in left ventricular peak filling rate caused by impaired free wall thinning, there was progressive increase in 1) duration of the rapid filling phase, 2) delay of mitral valve opening, 3) asynchrony between septum and posterior wall, 4) incidence of angina, and 5) incidence of atrial fibrillation.

Echocardiographic assessment of left ventricular filling after mitral valve surgery.

In order to investigate the functional effects of mitral valve surgery, echocardiograms showing left ventricular dimension were recorded and digitised in 14 normal subjects and 129 patients after mitral valve surgery. Measurements were made of peak rate of increase of dimension (dD/dt) and duration of rapid filling, studies on left ventriculograms in 36 patients having shown close correlation between these values and changes in cavity volume. In 14 patients with mitral stenosis, peak dD/dt was reduced to 7-2 +/ 1-5 cm/s, and filling period prolonged to 330 +/- 65 ms, compared with normal (16-0 +/- 3-2 cm/s, and 160 +/- 50 ms, respectively), and after mitral valvotomy, these values improved significantly (10-4 +/- 2-7 cm/s and 245 +/- 55 ms). Characteristic abnormalities were found in 67 patients with mitral prostheses. Values for the Björk-Shiley (10-5 +/- 4-2 cm/s and 180 +/- 80 ms) and Hancock (10-3 +/- 3-7 cm/s, 245 +/- 80 ms) values were similar, and both superior to the Starr-Edwards (7-4 +/- 3-0 cm/s, 295 +/- 105 ms). Results after mitral valve repair in 30 cases were not significantly different from normal (14-4 +/- 5-0 cm/s, 170 +/- 50 ms). Values outside the 95 per cent confidence limits for the valve in question allowed diagnosis of value malfunction in 18 cases. The method is value in comparing different operative procedures and in following up patients after mitral valve surgery.

Non-invasive assessment of left ventricular function after correction of severe aortic regurgitation.

Twenty patients were studied with simultaneous left ventricular cavity echocardiograms and apex cardiograms during the first two weeks after correction of severe aortic regurgitation. Endocardial echoes and apex cardiograms were digitized, so that left ventricular dimensions, their rates of change, and echo dimension-apex cardiogram relations could be studied. After aortic valve replacement, there was an early reduction in end-diastolic dimension, within 2 days, from 7-0 +/- 0-8 cm to 5-7 +/- 1-0 cm (P less than 0-001), while peak normalized shortening rate (peak Vcf) dropped from 1-9 +/- 0-6 to 1-4 +/- 0-6 S-1 (P less than 0-01), and remained unchanged for the remainder of the study. Immediately after operation, striking abnormalities of isovolumic contraction and, to a lesser extent, of early relaxation, could be seen, which regressed over 4 to 7 days, except in 2 patients who developed a low output state. These changes in left ventricular dimension, Vcf, and isovolumic contraction could not have been described by an single "measure" of left ventricular function.

Echocardiographic assessment of abnormal left ventricular relaxation in man.

In 64 patients requiring cardiac catheterization for chest pain, echocardiograms showing anterior mitral leaflet and left ventricular cavity simultaneously were recorded. These were digitized and their first derivatives computed in order to study time relations between mitral valve and left ventricular wall movement in early distole. In 10 patients with normal left ventricular angiograms and coronary arteriograms, mitral valve opening began 1-1 +/- 9-3 ms (mean +/- SD) before the onset of outward wall movement, and reached peak opening velocity 2-0 +/- 13 ms after maximum rate of change of dimension. Virtually identical time relations were seen in 15 patients with normal left ventricular angiograms but with obstructive coronary artery disease (3-6 +/- 9-3 ms and 0-7 +/- 7-3 ms, respectively). These close relations were lost in patients with segmental abnormalities of contraction on left ventricular angiogram. In 19 such patients with normal septal motion, outward wall movement began 53 +/- 31 ms before the onset of anterior movement of the mitral valve leaflet, and this isovolumic wall movement accounted for 31 per cent of the total diastolic excursion. In 9 patients with reversed septal movement, these abnormalities were greater, 92 +/- 39 ms and 33 per cent, respectively, while in 11 patients with diffuse left ventricular involvement they were small, 5-5 +/- 13 ms and 3 per cent. Frame-by-frame digitization of cineangiograms was used to confirm these findings which appear to reflect an abnormal change in left ventricular cavity shape during isovolumic relaxation.

Assessment of left ventricular systolic function in man from simultaneous echocardiographic and pressure measurements.

Simultaneous measurements of left ventricular dimension and wall thickness by echocardiography, and of pressure by micromanometer, were made at cardiac catheterization in 30 patients with heart disease, in order to study mechanisms of impairment of left ventricular systolic function. Echocardiograms and pressure traces were digitized so that continuous measurements of left ventricular wall thickness and dimensions with their rates of change could be obtained. Ejection fraction was estimated from digitized cineangiograms. In all patients, except those with severe mitral regurgitation, there was close correlation (r=0.92) between peak left ventricular dP/dt and peak rate of reduction of dimension. Myocardial power values, calculated as the product of circumferential shortening rates and wall stress were plotted throughout the cardiac cycle, and peak values in patients with normal left ventricular function were in the range 30 to 60 mW cm-3 myocardium. Pressure-dimension loops were constructed, which reflected the relation between the function of a localized region of cavity studied by echocardiography and that of the ventricle as a whole in the pressure wave form. Incoordinate contraction was associated with distortion of the loop and a reduction in its area to less than 75% that of the maximum for the cycle in question (cycle efficiency). In patients with left ventricular disease, ejection fraction, peak power, and cycle efficiency were all reduced, either singly or in combination. There was no consistent pattern, however, suggesting that clinical left ventricular disease may be the resultant of a number of different types of disturbance. These include structural abnormalities, reduction in peak rates of myocardial shortening or power development, and incoordinate contraction. The present investigation suggests ways in which these may be separated and studied in individual patients.

Continuous assessment of left ventricular shape in man.

Frame by frame analysis of left ventriculograms has been performed in 10 normal subjects and 40 patients with heart disease. Left ventricular shape index was derived as 4 pi (cavity area)/(perimeter)2, which has a maximum value of 1 when the outline is circular. In normal subjects systole was always associated with progressive reduction in shape index, indicating that the cavity projection had become less circular. This change was smaller in patients with low ejection fraction and also when inferior or anterior hypokinesia was present, even though ejection fraction was normal. During early diastole shape index rose rapidly due to an increase in minor diameter occurring throughout the period of rapid filling. In some cases this preceded any change in long axis, which was due to upward movement of the aortic root as well as outward movement of the apex. These results have functional implications, suggesting in particular that wall movement during filling may be non-uniform and that assumptions about cavity shape used in the derivation of wall properties from estimates of ventricular volume may require modification.