Noninvasive evaluation of ventricular function in cystic fibrosis

September 1979 TheJournalofPEDIATRICS

379

Noninvasive evaluation of ventricular function in cystic fibrosis The cardiac function of 21 patients with cystic fibrosis was studied using radionuclides and M-mode echocardiography. The patients (mean age 13.2 years, range 4 to 27) had a wide range of clinical and pulmonary function abnormalities (mean Shwachman-Kulczycki score 62.1). Decreased right ventricular ejection fraction was found in 13 of 18 patients (72%); an additional four patients had abnormal septal motion on ECHO. Left ventricular ejection fraction was abnormal at rest in four patients (19~ an additional four patients had borderline low L VEF. The L V pre-ejection period to ejection time ratio increased significantly (i.e, worsening L V performance) with deterioration of S-K score, chest radiograph score, and forced expiratory volume in 1 second Three of eight patients with normal L V E F at rest had an abnormal response to supine bicycle exercise: L VEF fell in two patients and was unchanged in one. Thus evidence of L V dysfunction was observed in seven of 21 (33%) of the patients," four at rest and in three only during exercise.

Bradley E. Chipps, M.D.,* Philip O. Alderson, M.D., Jean-Michel A. Roland, M.D., Shirley Yang, M.D., Andries van Aswegen, Ph.D., C a r l o s R. M a r t i n e z , M . D . , and Beryl J. R o s e n s t e i n , M . D . , B a l t i m o r e , M d .

COR PULMONALE has been demonstrated in at least 70% of children dying from cystic fibrosis? The diagnosis of cor pulmonale by physical examination may be difficult in patients with CF owing to their pulmonary disease. 2 Signs usually ascribed to heart failure of the right side, such as tachycardia, tachypnea, hepatomegaly, and cyanosis, may be manifestations of the underlying lung disease. Noninvasive evaluation of cardiac function in patients with CF using echocardiography, electrocardiography, vectorcardiography, pulmonary function studies, and clinical assessment have been reported, l, ~-5 Right ventricular systolic time intervals determined by ECHO correlate well with pulmonary artery pressures measured at cardiac catheterization and with RV wall thickness measured at postmortem examination in patients with From the Departments of Pediatrics and Radiology, The Johns Hopkins Medical Institutions. Supported in part by United States Public Health Service Grant GM-10548, the Hospital for Comsumptives of Maryland (Eudowood), and the Cystic Fibrosis Foundation. *Reprint address':Blaloek 1139, Johns Hopkins Hospital, Baltimore, MD 21205.

0022-3476/79/090379+06500.60/0 9 1979 The C. V. Mosby Co.

CF.1, 3, 4 However, owing to pulmonary hyperinflation, the frequency of satisfactory RV ECHO studies in patients with CF ranges from only 50 to 75%. 1' ~' 5 Abbreviations used CF: cystic fibrosis S-K: Schwachman-Kulczycki ECG: electrocardiogram ECHO: eehocardiography EF: ejection fraction LVEF: left ventricular ejection fraction RVEF: fight ventricular ejection fraction PEP/LVET: left ventricular pre-ejection period to ejection time ratio LV: left ventricle RV: right ventricle COPD: chronic obstructive pulmonary disease PEP: pre-ejection period FEVI: forced expiratory volume in one second FVC: forced vital capacity FRC: functional residual capacity The frequency of left ventricular dysfunction in patients with CF is unknown. LV dysfunction occurs in adults with chronic obstructive pulmonary disease, but other factors which contribute to LV dysfunction, such as

VoL 95, No. 3, pp. 379-384

380

Chipps et aL

coronary artery disease, hypertension, and valvular disease, are often present. '~-1~Patients with CF are younger and unlikely to have atherosclerotic or hypertensive cardiac involvement. Thus, any LV function abnormalities are more likely to be related to their primary disease. Pathologic studies have documented myocardial fibrosis in a small percentage of patients with CF, TM 14 but the functional significance of this finding is unknown. Radionuclide studies of the heart can provide quantitative indices of both left and right ventricular function noninvasively while exposing patients t o minimal amounts of radiation. 1~-18Accordingly, the present study was designed to evaluate the frequency of RV and LV dysfunction in patients with CF as determined by radionuclide studies and other noninvasive techniques such as ECHO, ECG, pulmonary function tests, arterial blood gas values, and chest radiographs. MATERIALS

AND METHODS

Twenty-one patients with CF were studied-15 males and six females. Their ages ranged from 4 to 27 years with a mean of 13.2 years. Informed consent was obtained from patients or parents prior to the study. The diagnosis of CF was based on characteristic clinical features and a positive quantitative pilocarpine iontophoresis sweat test (C1 concentration > 60 mEq/1). Patient 2 had a history of compensated biventricular failure and was taking digitalis. All others wer e clinically stable at the time of evaluation. Clinical assessment and scoring was based on the Shwachman-Kulczycki score? 9 Chest radiographs were evaluated according to a recently reported system. `-'~ Pulmonary function studies of forced expiration were obtained using a 9-liter Collins water seal spirometer. The best of three forced expirations was chosen for analysis. Lung volumes were measured with the patient in the seated position in a variable pressure body plethysmograph (Warner E. Collins), Blood gas values were obtained from a radial or brachial artery puncture while the patient was breathing room air, and were analyzed by Radiometer BMS MK2 (Copenhagen). Radionuclide heart studies were performed in two stages. The RVEF was determined during first transit using a multicrystal scintillation camera (Baird-Atomic)? 8 The range for RVEF using this technique in normal adults (n = 10) is 40 to 55%, which is similar to the range (35 to 67%) found by Gentzler et al ~ using contrast techniques. The patient was then moved beneath an Anger camera interfaced with a laboratory computer (Informatek Simis-3) where LVEF and ventricular function indices were determined. The LVEF obtained by this technique in normal adults (n = 10) ranges from 40 to 70%? ~ Camera images synchronized with the ECG were

The Journal of Pediatrics September 1979

obtained at 16 intervals during the cardiac cycle and displayed as a motion picture to evaluate LV wall motion. LV studies during supine bicycle exercise were attempted in the last 12 children studied. Four of these children were unable to cooperate because their legs were too short to reach the bicycle pedals, but the other eight children successfully completed the exercise. A constant load bicycle ergometer was used, and the E C G was monitored during the procedure. After a two-minute warm-up period pedaling at 25 watts, the workload was increased to 75 watts and exercise was continued for five minutes. LV function data were acquired from the left anterior oblique 40-degree view during this last exercise. M-mode ECHO was performed using a 2.25 MHz, 5 cm focused transducer. Patients were in the supine position when examined. Four areas were used for analysis: (1) the aorta, traversing the RV outflow tract, aortic root, and left atrium; (2) the mitral valve area, including the RV, intraventricular septum, and mitral valve at the atrioventricular valve level; (3) the ventricular area, viewing the RV, intraventricular system and LV; and (4) the tricuspid area, recording the tricuspid valve within the RV inflow tract. An attempt to record a pulmonary valve ECHO was made in each case. Measures of LV function including EF and systolic time indices (PEP,LVET and PEP/LVET) were ascertained from several cardiac cycles and averaged to minimize respiratory Variation. The results were compared to previously reported normal values, ~3 which are identical to normal values in our laboratory. A strip chart ECG was recorded simultaneously using a Honeywell 1856 recorder. A standard twelve-lead ECG was obtained along with a scalar vectorcardiogram on the same day. RESULTS The S-K scores of the patients ranged from 28 to 90 with a mean of 62, and the chest radiograph scores ranged from 8 to 22 wit~'a mean of 15.4. Only Patient 2 had evidence of ventricular dysfunction by history and physical examination. He had been digitalized and was clinically stable at the time of the study. Pulmonary function results were expressed as a percentage of predicted values for age, height, and sex. 24 The mean forced expiratory volume in one second was 46.0% of predicted (range 13 to 83), and the mean forced vital capacity was 63% of predicted (range 23 to 94). Airway obstruction as measured by FEV1/FVC averaged 65%. The mean functional residual capacity was 141% of predicted (range 102 to 217). The mean arterial oxygen tension was 69.5 mm Hg (range 45 to 95) and the mean arterial carbon dioxide tension was 34.7 mm Hg (range 25 to 47).

Volume 95 Number 3

Ventricular f i m c t i o n in cystic fibrosis

38 1

Table, Cardiac studies Echoc~trdiology

Radionuc~de wchniqub Patient No.

R VEF %

Normal

48 • 4

1

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 Mean

42 45 35 26 25 28 17 32 34 50 24 37 37 31 47 41 27 31 34

L VEF % (res0

L VEF % &tres~

56 • 8 43 32 56 56 45 55 47 59 36 48 43 37 48 43 62 54 46 49 54 65 38 48

68 • 11

S'eptal motion

P E P / L VET

L VEF %

PEP/ L VET

60 • 5 67 38 54 65

0.35 • 0.04 0.29 0.54 0.29

NI NI Abn Abn N1

66 58 61 60 60 54 54 56 48

0.33 0.26 0.193 0.31 0.26 0.19 0.20 0.40 0:33

N1 Abn N1 Abn Abn Abn N1 Abn NI

48 53 59 54 72

0.38 + 0.08 0.57 0.73 0~33 0.29 0.39 0.33 0.19 0.25 0.24 0.30 0.18 0.41 0.53 0.36 0.39 0.33 0.35 0.22 0.32 0.30 0.35

0.30 0.40 0.30 0.34 0.25 0.20 0.30

NI NI NI Abn N1 NI

54

54 47 62 53 65 80 57

44 52

50

Abbreviations used: RVEF Right ventricular ejection fraction; LVEF = left ventricular ejection fraction; PEP = pre-ejection period; N1 ~ normal; Abn = abnormal; LVET= left ventricular ejection time. Electrocardiogram showed RV hypertrophy in Patients t, 5, a n d 6 and LV hypertr0ptiy in Patient 2. Vectorcardiogram s confirmed these findings and were normal in aH other patients. Overall, 17 of the 21 patients (81%) had evidence o f RV dysfunction. The R V E F obtained from the first transit radionuclide studies was abnormal ( < 4 0 % ) in 13 of the 18 patients studied (range 17 to 37%) (Table). R V systolic time intervals could not be obtained in 15 of the 21 patients (;/1%)because oi" inadequate E C H O of the RV and pulmonary valve. A b n o r m a l septal m o t i o n was documented.in eight of the 19 patients (Table). F o u r of the eight patients with abnormal septal motion had normal first transit RVEF. Left ventricular ejection fraction as measured by the ECG-synchronized blood pool technique was abnormal in four patients, who had a mean value of 35.8% (range 3 2 to 38%; normal L V E F = 40 to 72%) (Table). 17 Each of these patients h a d a diffusely tiypokinetic 9 as seen in their blood pool image. An additional four patients had borderline abnormal L V E F (mean 43.5%, range 43 to 45%). E C H O showed definitely abnormal L V E F in three childrenl with a mean value of 44.3% (range 38 to 48%) and a borderline low L V E F in one child (53%). In three

patients there was lack of agreement between the two techniques. The radionuclide study classified Patients 9, 12, and 21 as abnormal but the L V E F by E C H O was normal in Patient 21, borderline low in Patient 12, and Patient 9 had abnormal septal motion which may lead to over-estimation of LVEF. The patients with borderline or low L V E F by radionuclide angiocardiography had S-K scoreS, chest radiograph scores, pulmonary function values, arterial blood gas values, and R V E F that were similar to those in patients with normal LVEF. There was no correlation between L V E F and R V E F (r = - 0 . 2 5 , P > 0.2). The P E P / L V E T as determined by E C H O was signifiCantly increased (indicating worsening LV performance) as the S-K score, chest radiograph score, and FEV~ decreased (Fig. 1). The P E P / L V E T did not significantly correlate with any of the other pulmonary function measurements. Similar results were found with the P E P / L V E T data obtained f r o m t h e equilibrium radionuclide studies, and these measurements of P E P / L V E T correlated significantly with those obtained by E C H O (r = 0.70, P < 0.01). As expected, the S-K score correlated well with pulmonary function measurements. TM As S-K score decreased,

Chipps et al.

382

The Journal of Pediatrics September 1979'

o.8 0.8

9

0.7

r = -0.74 sl =-0.02 p < .001

r = -0,70

9

-6.oo5

sl =

0,6

9

0.7-

P < ,OOI

r = -0.50 sl ~ - 0 . 0 0 2

0.7

p < .05 0.6

0.60.5 - ~

w

9 0,5 I-" LIJ

0.50,4-

9

Ld

9 e

0,4 0,4-

9

L# n 0.3-

9

o_

*

|

0.3

0.3-

0.2-

OA-

0

(9

, 0

,

,

20

, 40

i

,

60

S-K SCORE

,

,

80

,

i

I00

|

0.2-

0,2

0.[-

0.1

0-

4

~

~2

,~

! 2'o

CHEST RADIOGRAPH SCORE

~

,~

~2'0

4o

'

6'0

8'o

'

~oo

FEVI

Fig. 1. A, LV performance (PEP/LVET) by ECHO on ordinate and clinical score (S-K) on abscissa. B, PEP/LVET by ECHO on the ordinate an d chest radiograph score on abscissa. C, PEP/LVET by ECHO on the ordinate and FEV, as a percent of predicted on abscissa.

the F E V 1 (r = 0.80, P < 0.001), FVC 9 (r = 0.79, P < 0.00i), and Po~ (r = 0.55, P < 0.05.) als0 decreased. None of the other functions studied, except PEP/LVET, correlated with the S-K score. In the eight Children who were exercised, the mean heart rate increased from 82 beats per minute (range 61 to 115) to 129 beats per minute (range 107 to 157). The LVEF decreased during exercise in Patients 10 and 16 and w a s unchanged in Patient 19 (Fig. 2). Each of these 'children had normal radionuclide a n d ECHO LVEF at rest. In each the heart rate increased with exercise (mean increase = 39 beats per minute). DISCUSSION At present, there are no .practical therapeutic means, other than control of the underlying pulmonary disease, to prevent progression of cor pulmonale in patients with CF. Intravascular voltime depletion, administration of oxygen, and intravenous lise of tolazoline hydrocholoride have been shown, either a l o n e or in combination, to decrease the right heart filling pressure and pulmonary artery pressure? The long-term effects of these therapies have not been studied. In the present study, first transit radi0nuclide angiocardiography provided a noninvasive means for deriving RVE F and m a y be useful for following the effects of therapy. The inability to determine RV function by ECHO was probably caused by hyperinflation of the chest. The frequency of LV dysfunction in COPD in children

and adults is unknown. 7 Williams et al 8 found no evidence of LV dysfunction in patients with COPD even when RV failure was present. Frank et al 9 studied a group oi~adult patients with COPD and concluded that LV function was preserved in chronic cor pulmonale unless the myocardium was involved with a second process, Steele et al TM studied a group of adult patients with COPD using a radionuclide technique and concluded that LVEF was normal unless coronary artery disease was present. Baum et a111reported increased LV wall thickness and increased LV end-diastolic volume in a small group of patients with COPD and no evidence of other forms of heart disease. A pathologic study by Michelson I2 showed a high frequency of LVH in patients dying of COPD. Hooper and Whitcomb 25studied L~Zunction in a group of 28 adult patients with C O P D using systolic time intervals, significant prolongation of the systolic time interval ratio PEP/LVET was noted with increasing airway obstruction. There is also evidence for LV dysfunction in CF. Pathologic studies have shown increases in LV free wall weight 29 and areas of myocardial fibrosis. 1~Zinman et a127 evaluated LV function by ECHO before and after digitalization in nine children with CF and no evidence of cor pulmonale; LV posterior wall motion and Systolic time intervals were normal but paradoxical intraventricular septal excursions were present in all patients. After digitalizati0n there was an increase in the velocity of posterior ventricular wall motion and reversal of paradoxical septal motion. Rosenthal et al, 3 in an ECHO study of

Volume 95 Number 3

94 patients with CF, showed that increasing LV dysfunction tended to parallel increasing RV involvement. As the RV wall thickened, the septum and posterior wall of the LV began to thicken. Prolongation of systolic time intervals was noted. Ejection fraction, stroke volume, and other cardiac indices remained normal. The patients in our series had moderately advanced lung disease based on S-K and chest radiograph scores. The pulmonary }'unction measurements were consistent with moderate airway obstruction and correlated well with the S-K and chest radiograph scores. RV function by first transit nuclear angiocardiography was abnormal in 72% of the 18 patients studied. Four patients (19%) had an abnormally low LVEF at rest. In addition, three of eight patients who had normal LVEF at rest failed to increase their LVEF during exerciseiThus seven of 21 patients (33%) with CF had evidence of abnormal LV function as measured by ejection fraction at rest or during exercise. There are several possible explanations for LV dysfunction in CF. Myocardial fibrosis, reported in some patients with C F , TM 14 may result in a noncompliant ventricle which is unable to respond to the stress of exercise. The markedly negative swings in pleural pressure seen in COPD may serve to functionally afterload the LV by increasing the transmural pressure gradient across the LV wall? ~-3~ Theoretically RV dysfunction may alter LV function by the mechanism of ventricular interdependence. 2s-3~However. in the present study there was no correlation between RV and LV ejection fractions. These data are similar to those Of Berger et al? 1 who found no significant correlation between RVEF and LVEF in adults with COPD. In our study a good correlation was found between the ratio of systolic time intervals PEP/LVET and S-K score. chest radiograph score, and FEV]. Since systolic time intervals correlate well with changes in myocardial performance? ~ our study suggests that LV involvement occurs in CF and correlates with the clinical state of the patient. Stress testing identified several patients with abnormal LV function who would not have been identified by any of the other cardiac, radiographic, or pulmonary function studies. Radionuclide LV function studies during exercise have been used to evaluate adult patients with suspected coronary.artery disease? 3 and have improved the detection o f LV functional abnormalities. Normal patients increase EF by 7% or more during exercise, but patients with coronary artery disease and other types of LV dysfunction usually show decreased EF during exercise? 3 These supine exercise tests can also yield valuable information in children, and. as the present study shows. patients with CF will cooperate with supine bicycle

Ventricular funclion in cystic fibrosis

383

80-

70-

o~

60(19) (17l

la_ ILl

(tt)

50-

C~4) (t6/

40-

(iO/

50

REST

EXERCISE

Fig. 2. LVEF from radionuclide angiography at rest and after exercise9 Patient numbers corresponding to the Table are in parentheses.

exercise beneath a gamma camera. The fact that three of the seven children with LV dysfunction were identified only during stress suggests that the evaluation of LV function in a patient with CF should include exercise testing. The results of this study may also have therapeutic implications. In CF. progressive dyspnea with advancing pulmonary involvement may be related to lung involvement. but could also be on the basis of LV failure. No combination of noncardiac studies could identify the patients in this study who had LV dysfunction. Radionuclide studies and ECHO provide the opportunity to determine quantitatively both -right and left ventricular function, and they may be useful in evaluating patients with suspected LV dysfunction. Digoxin has not been shown to be helpful in the treatment of cor pulmonale. but may be a useful therapeutic adjunct in selected patients with CF whose course is complicated bY LV dysfunction. The authors thank Dr. H. N. Wagner. Jr.. for his advice. Drs. R. C. Talamo and J. L. R-obothamreviewed the manuscript and provided helpful comments. Terry Langbaum. Cathy Hensley, and Paula Pyzik provided invaluable technical assistance. REFERENCES

1. Liebman J. Lucas RV. Moss A. and Rosenthal A: Cor pulmonale and related cardiovascular effects of cystic fibrosis, in Mangos J. and Talamo R. editors: Cystic fibrosis: Projectons into the future. New York. 1976. Symposia Specialists, pp 41-82. 2. SiassiB. Moss AJ. and Dooley RR: Clinical recognition of cor pulmonale in cystic fibrosis..t PEDIATR78:794. 1971.

384

3.

4. 5.

6.

7. 8.

9.

10.

11.

12. 13.

14.

15.

16.

17.

18. 19.

Chipps et al.

Rosenthal A, Tucker CR, Williams RG, et al: Eehocardiographic assessment of cor pulmonale in cystic fibrosis, Pediatr Clin N Am 23:327, 1976. Gewitz M, Eshaghpour E, Holsclaw DS, et al: Echocardiography in cystic fibrosis, .Am J Dis Child 131:275, 1977. Fleming DG, Hirschfeld SS, Stern RC, et al: Echocardiographic evaluation of right ventricular function in patients with cystic fibrosis, Pediatr Res 12:560, 1978 (abstr). Marshall RC, Berger HJ, Costin JC, et al: Assessment of cardiac performance with quantitative radionuclide angiocardiography: Sequential left Ventricular ejection fraction, normalized left ventricular ejection rate and regional wall motion, Circulation 56:820, 1977. Kachel RG: Left ventricular function in chronic obstructive pulmonary disease, Chest 74:286, 1978. Williams JF, Childress RH, Boyd DL, et al: Left ventricular function in patients with chronic obstructive pulmonary disease, J Clin Invest 47:1143, 1968. Frank MJ, Weisse AB, Moschos CB, et al: Left ventricular function, metabolism, and blood flow in chronic cot pulmonale, Circulation 47:798, 1973. Steele P, Ellis JH, Van Dyke D, et al: Left ventricular ejection fracfioff in severe chronic airways disease, Am J Med 59:21, 1975. Baum GL, Schwartz A, Llmas R, et al: Left ventricular function in chronic obstructive lung disease N Engl J Med 285:361, 1971. Michelson N: Bilateral ventricular hypertrophy due to chronic pulmonary diseases, Dis Chest 38:425, 1960. Barnes GL, Gwynne JF, and Watt JM: Myocardial fibrosis in cYstic fibrosis of the pancreas; A u s t P e d i a t r J 6:81, 1970. Oppenheimer EH, and Esterly JR: Myocardial lesions in patients with cystic fibrosis of the -pancreas, John s Hopkins Med 133:]252, 1973. Strauss HW, Zaret BL, Hurley PJ, et al: A scintiphotographic method for measuring left ventrieular ejection fraction in man withot~t cardiac catheterization, Am J Cardiol 28:575 1971. Burow RD, Strauss HW, Singleton R, et al: Analysis of left ventricular function from multiple gate d acquisition c~irdiac blood pool imaging: Comparison to contrast angiography, Circulation 56:1024, 1977. Qureshi S, Wagner HN, Alderson PO, et al: Evaluation of left ventricular function in normal persons and patients with heart disease, J Nucl Med 19:135, 1978. Pitt B, and Strauss HW: Evaluation of ventricular function by radioisotope technics, N Engl J Med 2~6:1097, 1977. Shwachman H, and Kulczycki LL: Long-term study of one

The Journal of Pediatrics September 1979

20.

21.

22.

23.

24. 25.

26.

27.

28.

29.

30.

31.

32. 33.

hundred five patients with cystic fibrosis, Am J Dis Child 96:6, 1958. Brasfield D, Hicks G, Soong S, et al: The chest x-ray in cystic fibrosis a new scoring system, Pediatrics 63:24, 1979. Gentzler R, Briselli M, and Gault J: Angiographic estimation of right ventricular volume in man, Circulation 50:324, 1974. Douglass KH, Wagner HN Alderson PO, et al: Characteristics of left ventricular time-activity curves in normal volunteers, J Nucl Med 19:670, 1978 (abstr). Cantor A, Wanderman KL Kar01evitch T, et al: Systolic time intervals in children: Normal standards in clinical use, Circulation 58:1123, 1978. Polgar G, and Promadhut V: Pulmonary function testing in children, Philadelphia, 1971, WB Saunders Co. Hooper RG, and Whitcomb ME: Systolic time intervals in chronic obstructive pulmonary disease, Circulation 50:1205, 1974. Bowden DH, Fischer VW, and Wyatt JP: Cor pulmonale in cystic fibrosis: A morphometrie analysis, Am J Med 38:226, 1965. Zinman R. Kavey RE. Thomas D. et al: Left ventricular dysfunction and the use of digoxin assessed by echocardiography in patients with cystic fibrosis. Abstract of 17th Meeting of CF Center Directors. San Francisco. April 1977. Summer WR. Permutt S. Sagawa K. et al: Effects of spontaneous respiration on left ventricular function. Circ Res (in press~, Robotham JL. Lixfeld W. Holland L. et al: Effect of respiration on cardiac performance. J Appl Physiol 44:703. 1978. Robotham JL. and Mitzner W: A model of the effects of respiration on left ventricular performance. J Appl Physiol 46:411. 1979. Berger HJ. Matthay RA. Loke J. et al: Assessment of cardiac performance with quantitative radionuclide anglocardiography: Right ventricular ejection fraction with reference to findings in chronic obstructive pulmonary disease. Am J Cardiol 41:897. 1978, Weissler AM. Garrard CL: Systolic time intervals in cardiac disease (Ilk Mod Concepts Cardiovasc Dis 40:5. 1971. Borer JS. Bacharach SL. Green MV. et al: Real-time radionuclide clneangiogr~)hy in the non-invasive evaluation of gIobal and regional left ventricular function at rest and during exercise in patients with coronary arte(y disease. N Engl J Med 296:839. 1977.