Prognostic Value of Biventricular Function in Hypotensive Patients After Cardiac Surgery as Assessed by Transesophageal Echocardiography Constant

L.A. Reichert,

MD, Cees A. Visser, MD, Renee B.A. van den Brink, MD, Jacques J. Koolen,

Harry B. van Wezel, MD, Adriaan

C. Moulijn,

In patients after cardiac surgery, hypotension, defined as a mean arterial pressure less than 65 mmHg despite adequate filling pressures and positive inotropic medication, poses a problem. In addition, it is often difficult to determine whether these patients have suffered irreversible myocardial injury or if they are likely to recover. In this study, left and right ventricular function, as assessed by transesophageal echocardiography (TEE), was related to mortality both (1) quantitatively, using fractional area change (FAC), and (2) qualitatively, using a segmental wall motion analysis, which assigned a score to myocardial wal1 segments, in order to determine whether this technique can be used to predict survival. Mortality rate was very high in patients with biventricular and especially right ventricular failure (FAC < 35%). Left and

H

YPOTENSION, defined as a mean arterial pressure (MAP) < 65 mmHg, is common in patients following cardiac surgery. It can be caused by a variety of disorders,lT2 but if it persists despite optimization of loading conditions and administration of inotropic medication, it is usually due to severe circulatory failure, which is either secondary to an advanced state of preoperative myocardial dysfunction or to perioperative myocardial ischemia or necrosis.3 Hypotension under these circumstances is associated with a substantial mortality rate of 50% to 70%.3 From the clinical viewpoint it is difficult to determine which patients have suffered irreversible myocardial injury and which patients can be expected to recover. Nuclear angiocardiography and transthoracic echocardiography have been used to determine the degree of cardiac dysfunction in the ICU.4 The former, however, is not a technique readily available at the bedside.4 The latter is usually not practicable on patients after cardiac surgery, particularly not when they are mechanically ventilated, because of interference of the ultrasonic beam with the lungs and blood/air interfaces in the mediastinum. Transesophageal echocardiography (TEE) is not hampered by these acoustic window problems, is readily available at the bedside, and can provide high-quality images suitable for precise qualitative and quantitative analysis of cardiac function.5,6 The present study was undertaken to determine whether left (LV) and right (RV) ventricular function, as assessed by TEE in hypotensive patients following cardiac surgety, can be used to predict sutvival.

MD, Arend J. Dunning,

MD,

MD

right ventricular wall motion abnormality indices were significantly better in survivors compared to nonsurvivors, but no distinct cut-off value could be determined. A wal1 motion index derived from only 6 segments at the mid-papillary muscle leve1 was found to be as reliable as one based on 16 segments of the entire left ventricle. Thus, TEE provided information about the degree of left and right ventricular dysfunction by using a single cross-section at the papillary muscle level. It identified patients at high risk of death, ie, those with compromised right and biventricular function. Copyright o 1992 by W.B. Saunders Company KEY WORDS: transesophageal echocardiography, hypotension, ventricular function, cardiac surgery, ejection fraction

The coronary artery bypass (CABG) procedures included 10 reoperations and 2 emergency procedures following coronary angioplasty failures. Two aortic valve replacements were performed during the acute phase of endocarditis and 1 patient received a composite graft with reimplantation of the coronary arteries, having suffered an acute dissection of the ascending aorta. Two patients underwent mitral valve replacement because of severe mitral incompetente, which was caused by papillary muscle rupture. Of these 60 patients, those with echocardiographic signs of cardiac tamponade were excluded from further analysis (see below). Follow-up examinations were carried out on all patients during their stay in the ICU; mortality was considered to be the end-point. Oral informed consent was obtained from all patients prior to surgery. Two-dimensional TEE and color-flow Doppler examinations were performed with a commercially available 5 MHz transducer (Hewlett Packard, Palo Alto, CA) mounted on the distal end of an endoscope and interfaced with a phased array imaging system. The echoscope was inserted blindly, but laryngoscopic control was used when necessary. A complete 2D TEE and color-flow Doppler examination was performed, as described earlier,5,h with special emphasis on the examination of both the LV and RV wal1 segments. Repeat echocardiograms, sometimes performed during follow-up, were not taken into consideration. All studies were recorded on VHS videotape. Patients with echocardiographic signs of tamponade (defined as indentation or collapse of the left or right atrium andior RV due to pericardial effusion or blood clot formation): were excluded from the study, because ventricular function cannot be adequately assessed in this condition. Quantitative analysis of both LV and RV function was per-

METHODS In the intensive care unit (ICU), 60 consecutive patients with a MAP ~6.5 mmHg, despite positive inotropic medication (dopamine, 10 to 15 uglkgimin, and dobutamine, 15 to 25 uglkgimin) and mechanica1 support with an intra-aortic balloon pump (IABP) in 29 patients, were studied within 48 hours after cardiac surgery. These patients were selected on the basis of the above criteria from a group of 1,516 patients operated on between January 1,198s and December 31, 1989. Overall mortality for the total patient group was 4.9% (74 patients). There were 26 women and 34 men with a mean age of 66.7 2 9.4 years (range, 21 to 88 years) studied; surgical procedures of the 60 patients are summarized in Table 1.

Journal

of Cardiothoracic

and

VascularAnesthesia,

Vol 6, NO 4

From the Depattments of Cardiology, Cardiac Surgery. and Anesthesiology, Academie Medical Center, Amsterdam, The Netherlands, and the Interuniversity Cardiology Institute of the Netherlands. Utrecht, The Netherlands. Presented in part at the 63nd Scientifc Sessions of the American Heat? Association, Dallas, TX, 1990. Address reprint requests to C.L.A. Reichett, MD, Department of Cardiology-Kl, Medical Center Alkmaar, Wilhelminalaan 12, 1815 JD Alkmaar, The Netherlands. Copyright 0 1992 by W.B. Saunders Company 1053.077019210604-0007$3.0010

(August), 1992: pp 429-432

429

REICHERT ET AL

430

Table 1. Surgical Procedures of 60 Patients With Postoperative Hypotension in the Intensive Care Unit Number

of

Patients

CABG

30

CABG + AVR CABG + VSR repair AVR

6 10 6

MVR

3

AVR + MVR

3

Right atrial tumor resection

1

Chest trauma

1

Abbreviations:

AVR, aortic valve replacement;

al2 S

CABG, coronary ar-

tery bypass grafting; MVR, mitral vake replacement; VSR, ventricular

APEX

septal rupture.

formed using short-axis cross-sections at the leve1 of the papillary muscles. Al1 echocardiographic examinations were reviewed at normal speed, slow motion, and stop-frame format. The R wave of the electrocardiogram (ECG) was used to select end-diastolic stop-frames. End-systolic frames were defined as those that displayed the minima1 cross-sectional area of the ventricle. Endsystolic and end-diastolic endocardial boundaries of both ventricles were traced by an experienced observer using a commercially available contouring system (Datavue Micro-Sonics Inc, Indianapolis, IN). Papillary muscle outlines were excluded from the contour. From these outlines, end-diastolic area, end-systolic area, and fractional area change (FAC) were calculated. (FAC = [(diastolic area-systolic area)/diastolic area) x 100%). As beat-to-beat variability may exist, three consecutive beats were analyzed and averaged. Following the classification for ejection fraction, a FAC of

50%

I

> 50 %

9111

010

9110

2l5

114

010

319

010

2113

14125 (56 %)

114 (25 %)

11123 (46 %)

18/21@6%)

3/9

(33 %)

5/22 (23 %) 26152 (50 %) 1

Fig 2. Diagram showing the relation between different subgroups of left and right ventricular FACs and mortality. The numerator shows the number of nonsurvivors, the denominator the total number of patients in the subgroup. Abbreviations: FAC, fractional area change; LV, left ventricle; RV. right ventricle.

TEE IN POSTOPERATIVE

431

HYPOTENSION

patients) and 90% (9 patients), respectively. In the group in which both the LV and RV FAC was > 50% (13 patients), 2 patients (15%) died from hemorrhage. The mortality rate was affected by the degree of RV failure (x2 = 18.26, DF = 2, P < O.OOOl), but not by the degree of LV failure (x2 = 1.4, DF = 2, not significant). For the calculation of LV WMI, 520 segments were found to be suitable for grading (63%); 312 segments were not adequately visualized; of these 176 were apical segments (56%) 10 mid-papillary (3%), and 126 basal (41%). For RV WMI, al1 but 4 segments (98%) could be graded. Figure 3 shows the LV WMI in 52 patients, which was significantly lower in survivors compared to nonsurvivors (1.71 + 0.09 versus 2.06 ? 0.12; P = 0.019). The RV WMI was also significantly lower in survivors (1.5 + 0.09 versus 2.15 * 0.11; P = 0.0003) compared to nonsurvivors. For calculation of the WMIp (the 6 segments of the LV at the papillary muscles level), 302 of the 312 segments were adequate for grading (97%). Figure 3 also shows the WMIp, which was of the same order of magnitude as the WMI; survivors: 1.66 ? 0.08 versus 1.71 + 0.09 (not significant) and nonsurvivors: 2.06 ? 0.12versus 2.06 -I 0.12 (not significant). The WMIp in survivors was significantly lower compared to nonsurvivors (P = 0.008). Comparison of WMIp and WMI yielded a correlation coefficient of 0.94 (n = 52). WMIp and FAC, both obtained at the papillary muscles level, were compared and regression analysis showed a correlation coefficient of 0.70 (95% confidence interval - 0.81 to - 0.53) as could be expected, because both methods quantitate the same phenomenon. DISCUSSION

Hypotension after cardiac surgery is a serious complication, which requires immediate determination of its cause to allow therapeutic measures to be takenre Despite optimization of loading conditions, positive inotropic medication, and mechanica1 assist devices, the systemic arterial

Wal1 Motion Index

and mortality

3 ,

l

t

I

LVP

Fig 3. Diagram showing wal1 motion index of the RV, 16 segments of the LV, and the 6 segments of the LV at papillary muscle leve1 (LVp) for survivors and nonsurvivors. See text for WMI values.

pressure can stil1 be at a leve1 to compromise vita1 organ systems. Hypotension in this setting is usually caused by severe circulatory failure, due either to preexisting myocardia1 dysfunction andior perioperative myocardial infarction. From the clinical viewpoint it remains difficult to determine which patients have suffered irreversible myocardia1 injury and which patients can be expected to recover from the burden of cardiac surgery after adequate therapy has been instituted. In this respect, LV and RV function, as assessed by TEE, and its relation to mortality were studied. The mortality rate was 50%, which is in keeping with previous reports.1,3J-12 Considering FAC as a systolic function parameter, the majority of nonsurvivors were assigned to the patient subcategory of either severe biventricular or isolated RV failure (mortality rates of 82% and 90%, respectively). The better prognosis for patients with isolated LV dysfunction (mortahty rate in this series was 36%) can be explained by the use of the IABP, thus decreasing the workload of the LV.” Several reports have described the beneficial effects of assist devices or balloon pumping in the pulmonaty artery, t2-ld in patients with severe RV failure. Application of these assist modalities could gain time to allow the heart to recover from the burden of cardiac surgery, which causes myocardial edema, ischemia, and loss of compliance. It can take as long as 8 to 11 days for the LV to recover, whereas the RV usually recovers in 3 to 5 days in the absente of pulmonary hypertension.rl-rh Besides off-line calculation of the FAC as a systolic function parameter, a reasonable estimate of LV and RV function can be derived from visual segmental wal1 motion analysis, which can readily be done at the bedside.‘: In this study, the ló-segment model for LV function assessment was compared with a single short-axis cross-section at the mid-papillary muscle level, including only 6 segments, but representing the perfusion areas of 3 major coronary artery branches. WMIp, derived from these 6 segments, was found to be comparable to the WMI, calculated on the basis of the ló-segment model. It also provided useful information about ventricular dysfunction with respect to survival. Both WMI and WMIp of the LV were significantly lower in survivors as compared to nonsurvivors. Furthermore, visualization of al1 16 segments is time-consuming and not always feasible using TEE. Thus, visual assessment of regional wal1 motion at the leve1 of the papillary muscles appears to provide a reliable estimate of global LV dysfunction. Although a good estimate of LV and RV function can be derived from the echocardiographic examination, it is not possible to determine a distinct cut-off value to distinguish between survivors and nonsurvivors. This is not surprising, however, because severe failure of one ventricle can cause death, even in the presence of good function of the other ventricle. Although the results of this study are promising, there are several limitations. Myocardial wal1 motion does not depend on the contractility of the heart muscle alone, but also on preload and afterload. In this study, ventricular function was probably overestimated because ah patients

432

REICHERT ET AL

received inotropic medication and many patients were supported with an IABP, reducing afterload. Furthermore, the possible degree of valvular insufficiency, which also influences wal1 motion, was not taken into consideration. The FAC was considered as a global systolic function parameter, similar to ejection fraction as determined by nuclear angiocardiography, despite the apparent differente of geometrie proportions of the LV and RV. Echocardiography is a tomographic 2D technique, possibly making comparison of FAC with ejection fraction unreliable, because the latter is a 3D estimate of LV function. However, recent reports suggest a good correlation between these methods.r7Jx Although a good correlation was found between WMIp and FAC, WMI provides more information about ventricular function. However, using single-plane transducers, adequate grading cannot be done in a substantial number of myocardial wal1 segments (37% in this study); biplane transducers would improve visualization.

The present study confirms that TEE performed on hypotensive patients after cardiac surgery, at the bedside in the ICU, readily provides high-quality images that allow qualitative and quantitative assessment of LV and RV function. An advanced degree of ventricular impairment, as assessed by either the FAC, the WMI based on al1 the myocardial wal1 segments, or the WMIp derived from only 6 segments of a single cross-section at the mid-papillary muscle leve], was associated with an increased mortality rate; patients with signs of severe RV or biventricular dysfunction were at very high risk of death. Future prospective studies need to determine whether this knowledge about the degree of LV or RV impairment can be used to predict the final outcome, and to select additional therapeutic measures to improve survival. ACKNOWLEDGMENT

We thank Dr F.J.L. van Capelle for his statistical advice and M. de Bruin for secretarial assistance.

REFERENCES

1. Dietzman RH, Ersek RA, Lillehei CW, et al: Low-output syndrome: Recognition and treatment. J Thorac Cardiovasc Surg 57:139-150,1969 2. Lappas DG, Powell WMJ, Daggett WM: Cardiac dysfunction in the perioperative period: Pathophysiology, diagnosis, and treatment. Anesthesiology 47:117-137,1977 3. Pennington DG, McBride LR, Kanter KR, et al: Effect of perioperative myocardial infarction on survival of post-cardiotomy patients supported with ventricular assist devices. Circulation 78 (Suppl3):111-110-115, 1988 4. Parker MM, Cunnion RE, Parrillo JE: Echocardiography and nuclear cardiac imaging in the critical care unit. JAMA 254:29352939,1985 5. Visser CA, Koolen JJ, van Wezel HB, Dunning AJ: Transesophageal echocardiography: technique and clinical applications. J Cardiothorac Anesth 2:74-91,1988 6. Seward JB, Khandheria BK, Oh JK, et al: Transesophageal echocardiography: Technique, anatomie correlations, implementation, and clinical applications. Mayo Clin Proc 63:649-680, 1988 7. Feigenhaum H: Echocardiography. Philadelphia, PA, Lea & Febiger; 1986 (ed 4): 558-565 8. Parssonnet V, Dean D, Bernstein AD: A method of uniform stratihcation of risk for evaluating the results of surgety in acquired adult heart disease. Circulation 79 (Suppl l):I-3-1-12, 1989 9. Schiller NB, Shah PM, Crawford M, et al: Recommendations for quantification of the left ventricle by two-dimensional echocardiography. J Am Sec Echo 2:358-368,1989 10. Gussenhoven EJ, Taams MA, Roelandt JRTC, et al: Trans-

esophageal two-dimensional echocardiography: Its role in solving clinical problems. J Am Col1 Cardiol8:975-999, 1986 ll. Chan KL: Transesophageal echocardiography for assessing cause of hypotension after cardiac surgery. Am J Cardiol 62:11421143,1988 12. Pennington DG, Meijavy JP, Swartz MT, et al: The importante of biventricular failure in patients with postoperative cardiogenie shock. Ann Thorac Surg 39:16-26,1985 13. McEnany MT, Kay HR, Buckley MJ, et al: Clinical experience with intra-aortic balloon pump support in 728 patients. Circulation 58:124-132,1978 14. Moran JM, Opravil M, Gorman AJ, et al: Pulmonary artery balloon counterpulsation for right ventricular failure: 11 Clinical experience. Ann Thorac Surg 38:254-259,1984 15. Dreyfus G, Guillemain R, Jebara VA, et al: Assistente ventriculaire droite par counterpulsion intrapulmonaire. Arch Mal Coeur 82:81-86,1989 16. Pierce WS, Parr GVS, Myers JL, et al: Ventricular assist pumping in patients with cardiogenic shock after cardiac operations. N Eng1 J Med 305:1606-1610, 1981 17. Amice AF, Lichtenberg GS, Reisner SA, et al: Superiority of visual versus computerized echocardiographic estimation of radionuclide left ventricular ejection fraction. Am Heart J. 118:12591265,1989 18. Urbanorvicz JH, Shaaban MJ, Cohen NH, et al: Comparison of transesophageal echocardiographic and scintigraphic estimates of left ventricular end-diastolic volume index and ejection fraction in patients following coronaty artery bypass grafting. Anesthesiology 72:607-612, 1990