Efficacy and Cost-Effectiveness of Preoperative IABP in Patients with Ejection Fraction of 0.25 or Less

Transcription

1 Efficacy and Cost-Effectiveness of Preoperative IABP in Patients with Ejection Fraction of 0.25 or Less Charles A. Dietl, MD, Marie D. Berkheimer, RN, Edward L. Woods, MD, Christian L. Gilbert, MD, William F. Pharr, MD, and Charles H. Benoit, MD Department of Cardiovascular and Thoracic Surgery, Geisinger Medical Center, Danville, Pennsylvania Background. The purposes of this study are to determine whether patients with severe left ventricular dysfunction benefit from prophylactic insertion of an intraaortic balloon pump and to evaluate its costeffectiveness. Methods. Between January 1991 and December 1995, 163 consecutive patients with a left ventricular ejection fraction of 0.25 or less underwent isolated coronary artery bypass grafting. An intraaortic balloon pump was inserted before operation in 37 patients (group A). The remaining 126 patients underwent operation without preoperative insertion of the device (group B). Preoperatively, 91.9% (34/37) of group A patients and 54.8% (69/126) of group B patients were in New York Heart Association functional class III or IV (p < 0.001). Results. The 30-day mortality rate was 2.7% (1/37) and 11.9% (15/126) for groups A and B, respectively (p < 0.005). All deaths occurred in patients in functional class III or IV. In group B, 28 patients (22.2%) required an intraaortic balloon pump after cardiotomy for low cardiac output, 42.9% (12/28) of whom died. Median postoperative hospital stay was 9.9 days and 12.0 days, and mean hospital charges were $50,627 and $54,818 for survivors in groups A and B, respectively (p = not significant). Conclusions. Our experience suggests that patients with severe left ventricular dysfunction undergoing coronary artery bypass grafting may benefit from preoperative intraaortic balloon pump insertion, especially patients in functional class III or IV. This approach improved survival significantly, reduced hospital stay, and was more cost-effective. (Ann Thorac Surg 1996;62:401-9) S evere left ventricular (LV) dysfunction is an incremental risk factor for patients undergoing coronary artery bypass grafting (CABG), especially if the LV ejection fraction (LVEF) is 0.25 or less [1-5]. Despite the major advances in monitoring devices, anesthesia, techniques of myocardial preservation, and perioperative pharmacologic support, the operative mortality rate remains elevated, ranging from 4.8% to 14.3% according to several reports published during this decade [1-5]. Christakis and colleagues [3] suggested that to improve these results, strategies should be focused on patient selection, improvement of myocardial protection, and more aggressive use of the intraaortic balloon pump (IABP) before operation to reduce myocardial ischemia. Diastolic counterpulsation is the only means currently available to increase aortic diastolic pressure while reducing myocardial oxygen demand in patients with critical coronary artery stenosis [6]. Gunstensen and associates [6] recommended insertion of an IABP under local anesthesia prior to induction of general anesthesia in patients with acute coronary insufficiency, critical left main coronary artery stenosis, or Presented at the Thirty-second Annual Meeting of The Society of Thoracic Surgeons, Orlando, FL, Jan 29-31, Address reprint requests to Dr Dietl, Department of Cardiovascular and Thoracic Surgery, Geisinger Medical Center, Danville, PA severe LV dysfunction. These authors believed that prophylactic use of balloon counterpulsation in such patients is cost-effective because in their experience, the postoperative stay in the intensive care unit was shorter, and the prevalence of respiratory and renal complications was also reduced. The purpose of the present study is to determine whether patients with severe LV dysfunction benefit from prophylactic insertion of an IABP and to evaluate its cost-effectiveness. A multivariate analysis of associated risk factors was done to predict which subgroups of patients are more likely to benefit from prophylactic use of balloon counterpulsation. Material and Methods A retrospective analysis was done on 163 consecutive patients with an LVEF equal to or less than 0.25 who underwent isolated CABG between January 1991 and December 1995 at Geisinger Medical Center, Danville, PA. Patients undergoing concomitant valve repair or replacement or resection of an LV aneurysm and patients in cardiogenic shock prior to operation with postinfarction ventricular septal defect or rupture of the papillary muscles of the mitral valve were excluded. Cardiac catheterization and coronary angiography were performed in all patients. The preoperative LVEF 1996 by The Society of Thoracic Surgeons /96/$15.00 Published by Elsevier Science Inc PII S (96)

2 402 DIETL ET AL Ann Thorac Surg COST-EFFECTIVENESS OF PREOPERATIVE [ABP 1996;62:401-9 Table 1. Prevalence of Associated Risk Factors '~ Risk Factor Group A Group B p Value Age -> 70 y 8/37 (21.6) 29/126 (23.0) NS Female sex 7/37 (18.9) 31/126 (24.6) NS Diabetes mellitus 16/37 (43.2) 60/126 (47.6) NS Previous CABG 10/37 (27.0) 12/126 (9.5) Congestive heart failure (70.3) 72/126 (57.1) NS LVEF < /37 (48.6) 53/126 (42.1) NS NYHA class III 16/37 (43.2) 48/126 (38.1) NS NYHA class IV 18/37 (48.6) 21/126 (16.7) <0.001 NYHA classes III& W 34/37 (91.9) 69/126 (54.8) <0.001 CCS classes II & 1V 33/37(89.2) 91/126 (72.2) NS Elective operation 7/37 (18.9) 44/126 (34.9) 0.05 Urgent operation 20/37 (54.1) 79/126 (62.7) NS Emergency operation 10/37 (27.0) 3/126 (2.4) <0.001 Recent MI (0-7 days) 16/37 (43.2) 36/126 (28.6) Unstable angina 29/37 (78.4) 82/126 (65.1) NS 1V nitroglycerin 19/37 (51.4) 31/126 (24.6) <0.01 Left main stenosis 14/37 (37.8) 17/126 (13.5) <0.01 LVEDP > 20 mm Hg 27/35 (77.1) 74/118 (62.7) NS Mitral regurgitation 14/37 (37.8) 42/126 (33.3) NS Numbers in parentheses are percentages. CABG = coronary, artery bypass grafting; CCS = Canadian Cardiovascular Society; 1V = intravenous; LVEDP = left ventricular end-diastolic pressure; LVEF - left ventricular ejection fraction; MI = myocardial infarction; NS not significant; NYHA = New York Heart Association. was measured by the area-length method from biplane ventricular angiography. Left main coronary artery stenosis was defined as a lesion obstructing 50% or more of the lumen. Preoperative functional class of congestive heart failure was rated according to the criteria of the New York Heart Association (NYHA) [7], and angina pectoris was graded using the classification of the Canadian Cardiovascular Society [8]. An IABP was placed percutaneously before operation in 37 patients (group A) either in the catheterization laboratory (n - 10) or in the operating room (n = 27) before induction of general anesthesia. The remaining 126 patients underwent CABG without preoperative insertion of an IABP (group B). Mean age was 61.7 years (range, 39 to 83 years) and 62.7 years (range, 36 to 79 years) for groups A and B, respectively (p = not significant). Mean LVEF was 0.22 (range, 0.14 to 0.25) and 0.23 (range, 0.10 to 0.25) (p = not significant), and mean LV end-diastolic pressure (LVEDP) was 24.2 mm Hg (range, 12 to 40 mm Hg) and 22.5 mm Hg (range, 10 to 42 mm Hg) (p = not significant) for groups A and B, respectively. Group A had a significantly higher prevalence of associated risk factors, including previous CABG, emergency operation, NYHA class IV, recent myocardial infarction (0 to 7 days before operation), and left main stenosis (Table 1). Indications for preoperative use of an IABP included refractory unstable angina despite intravenous adminis- tration of heparin sodium and nitroglycerin and, in several patients, the association of poor LV function and severe left main stenosis, previous CABG, recent myocardial infarction, or congestive heart failure. Several of these risk factors frequently coexisted in the same patient. Surgical Technique Ten patients in group A required counterpulsation in the catheterization laboratory either before or after the study because of refractory unstable angina and were taken directly to the operating room for emergency CABG. The other 27 patients in group A underwent IABP insertion in the operating room using local anesthesia before induction of general anesthesia. In all these patients, a 9.5F Percor balloon (Stat-DL dual-lumen intraaortic balloon catheter with 40-mL balloon, Datascope System; Datascope Corp, Fairfield, NJ) was inserted percutaneously using the provided 12F sheath through the common femoral artery. The correct position of the balloon in the descending thoracic aorta was verified by fluoroscopy in the catheterization laboratory, or by transesophageal echocardiography when the IABP was placed in the operating room. All patients received prophylactic antibiotics. After IABP insertion, all patients were given 5,000 units of heparin and a continuous infusion of heparin (1,000 U/h) until they were fully heparinized and cannulated. Mild systemic hypothermia (30 to 32 C) was used during cardiopulmonary bypass. To maintain a pulsatile flow, counterpulsation was continued during cardiopulmonary bypass in several patients using the internal trigger of the console to activate the balloon, which was partially inflated with 20 ml of helium. Antegrade cold blood cardioplegia was used for myocardial protection in most patients in both groups. A combination of antegrade and retrograde cold blood cardioplegia was used in several patients, especially those having reoperations. Complete revascularization was accomplished in most instances. The average number of grafts per patient was 3.52 (range, 1 to 6) and 3.43 (range, 1 to 7) for groups A and B, respectively (p = not significant). A left internal mammary artery graft was used in 32 patients (86.5%) in group A and 116 (92.1%) in group B (p - not significant). A postcardiotomy IABP was used in 28 patients (22.2%) in group B, including 23 who required the device for weaning from cardiopulmonary bypass and 5 who required postoperative device insertion for low cardiac output. The percutaneous common femoral artery approach was used in 22 patients and a surgical cutdown was used in 6 to expose the common femoral artery. However, the IABP was inserted through the aortic arch in 2 of these patients because attempted insertion through the femoral vessels failed. Statistical Analysis Mean age and mean LVEF for the two groups were tested using unpaired Student's t test. Comparison of the prevalence of risk factors and of the characteristics of patients

3 Ann Thorac Surg DIETL ET AL ;62:401-9 COST-EFFECTIVENESS OF PREOPERATIVE IABP in group B who required postcardiotomy IABP use was carried out using the X 2 test. Comparison of the 30-day mortality for groups A and B was analyzed using a multiple logistic regression analysis. The postoperative hospital stay was analyzed by means of the Mann- Whitney test, and an analysis of covariance was used to compare the data for hospital charges in those instances when significant differences were found in the mean postoperative hospital stay. Calculated p values lower than 0.05 were considered significant. Results All 37 patients in group A were weaned from cardiopulmonary bypass and were supported postoperatively with the IABP for 2 to 48 hours (mean duration, 16.3 hours). One patient with an LVEF of 0.20, previous CABG, a recent anterior myocardial infarction, and "poor targets" died of low cardiac output. Thus, the overall 30-day mortality rate was 2.7% for group A. In group B, 98 patients (77.8%) were weaned from cardiopulmonary bypass without IABP support and did not require such support postoperatively (subgroup B1). The other 28 patients (22.2%) (subgroup B2) required postcardiotomy IABP placement for low cardiac output. For these patients, postoperative IABP support ranged from 1 hour to 96 hours (mean duration, 42.3 hours). Incremental risk factors for postcardiotomy IABP insertion included previous CABG, NYHA functional class III or IV, nonelective operation, recent myocardial infarction (0 to 7 days before operation), unstable angina, and left main stenosis (p < 0.05). Advanced age, female sex, diabetes, LVEDP higher than 20 mm Hg, and coexisting mitral regurgitation did not increase the need of postcardiotomy IABP support in these patients (Table 2). The overall 30-day mortality rate for patients in group Table 2. Incremental Risk Factors for Postcardiotomy Intraaortic Balloon Pump Use (Group B) ~ Risk Factor Present Absent p Value Age > 70 y 6/29 (20.7) 22/97(22.7) NS Female sex 9/31 (29.0) 19/95 (20.0) NS Diabetes mellitus 14/60 (23.3) 14/66 (21.2) NS Previous CABG 8/12 (66.7) 20/114 (17.5) <0.001 Congestive heart failure 18/72 (25.0) 10/54 (18.5) NS LVEF < /53 (26.4) 14/73 (19.2) NS NYHA class III or IV 25/69 (36.2) 3/57 (5.3) ~0.001 CCS class III or IV 22/91 (24.2) 6/35 (17.1) NS Nonelective operation 23/82 (28.0) 5/44 (11.4) <0.05 Recent MI (0-7 days) 13/36 (36.1) 15/90 (16.7) Unstable angina 23/82 (28.0) 5/44 (11.4) <0.05 1V nitroglycerin 12/31 (38.7) 16/95 (16.8) Left main stenosis 9/17 (52.9) 19/109 (17.4) LVEDP > 20 mm Hg 18/74 (24.3) 10/52 (19.2) NS Mitral regurgitation 11/42 (26.2) 17/84 (20.2) NS Abbreviations are the same as in Table 1. Table 3. Observed 30-Day Mortality a Patient Characteristic Group A Group B p Value Age -~ 70 y 0/8 (0.0) 2/29 (6.9) NS Age ~ 70 y 1/29 (3.4) 13/97 (13.4) Female sex 0/7 (0.0) 6/31 (19.4) NS Male sex 1/30 (3.3) 9/95 (9.5) Diabetes mellitus 0/16 (0.0) 10/60 (16.7) NS Nondiabetes 1/21 (4.8) 5/66 (7.6) NS Previous CABG 1/10 (10.0) 6/12 (50.0) NS First CABG 0/27 (0.0) 9/114 (7.9) NS NYHA class III or IV 1/34 (2.9) 15/69 (21.7) NYHA class I or II 0/3 (0.0) 0/57 (0.0) NS CCS class III or IV 1/33 (3.0) 14/91 (15.4) CCS class I or II 0/4 (0.0) 1/35 (2.9) NS Elective operation 0/7 (0.0) 1/44 (2.3) NS Nonelective operation 1/30 (3.3) 14/82 (17.1) Recent MI (0-7 days) 1/16 (6.3) 8/36 (22.2) NS No recent MI 0/21 (0.0) 7/90 (7.8) NS Left main stenosis 1/14 (7.1) 7/17 (41.2) NS No left main stenosis 0/23 (0.0) 8/109 (7.3) NS LVEDP ~ 20 mm Hg 1/27 (3.7) 11/74 (14.9) NS LVEDP -~ 20 mm Hg 0/8 (0.0) 4/44 (9.1) Mitral regurgitation 0/14 (0.0) 5/42 (11.9) NS No mitral regurgitation 1/23 (4.3) 10/84 (11.9) NS Total for group 1/37 (2.7) 15/126 (11.9) CABG - coronary artery bypass grafting; CCS Canadian Cardiovascular Society; LVEDP left ventricular end-diastolic pressure; M! - myocardial infarction; NS - not significant; NYHA - New York Heart Association. B was 11.9% (15/126 patients). Thus, the overall 30-day mortality rate was significantly lower in group A than in group B (2.7% versus 11.9%; p < 0.005). Using a multiple logistic regression model that included all the risk factors for which significant differences were found between groups A and B, the 30-day mortality rate was significantly lower (p < 0.05) in group A, especially in patients with the following characteristics: age of less than 70 years, male sex, NYHA functional class III or IV, Canadian Cardiovascular Society functional class III or IV with unstable angina or an LVEDP of 20 mm Hg or less, and nonelective operation (Table 3). However, the 30-day mortality rate was 3.1% (3/98) and 42.9% (12128) for patients in subgroups B1 and B2, respectively (p = 0.001). Significant differences (p < 0.05) were found for the following characteristics: age of less than 70 years, male sex, diabetes mellitus, previous CABG, NYHA class Ill or IV, nonelective operation, unstable angina, left main coronary stenosis, LVEDP higher than 20 mm Hg, and coexisting mitral regurgitation (Table 4). Vascular complications were observed in 3 patients (8.1%) in group A, of whom only 1 had leg ischemia, which resolved after the balloon was removed. The other 2 patients required vascular reconstruction in the operating room; 1 had profuse hemorrhage at the femoral insertion site after IABP removal and I had development of a pseudoaneurysm.

4 404 DIETL ET AL Ann Thorac Surg COST-EFFECTIVENESS OF PREOPERATIVE IABP 1996;62:401-9 Table 4. Observed 30-Day Mortality in Group B a Patient Characteristic Subgroup B1 Subgroup B2 p Value Age -~ 70 y 1/23 (4.3) 1/6 (16.7) NS Age < 70 y 2/75 (2.7) 11/22 (50.0) 0.02 Female sex 2/22 (9.1) 4/9 (44.4) NS Male sex 1/76 (1.3) 8/19 (42.1) Diabetes mellitus 2/46 (4.3) 8/14 (57.1) Nondiabetes 1/52 (1.9) 4/14 (28.6) NS Previous CABG 0/4 (0.0) 6/8 (75.0) First CABG 3/94 (3.2) 6/20 (30.0) NS NYHA class III or IV 3/44 (6.8) 12/25 (48.0) 0.01 NYHA class I or II 0/54 (0.0) 0/3 (0.0) NS CCS class III or IV 3/69 (4.3) 11/22 (50.0) CCS class I or II 0/29 (0.0) 1/6 (16.7) NS Elective operation 0/39 (0.0) 1/5 (20.0) NS Nonelective operation 3/59 (5.1) 11/23 (47.8) Recent MI (0-7 days) 3/23 (13.0) 5/13 (38.5) NS No recent MI 0/75 (0.0) 7/15 (46.7) Left main stenosis 1/8 (12.5) 6/9 (66.7) No left main stenosis 2/90 (2.2) 6/9 (31.6) NS LVEDP > 20 mm Hg 2/56 (3.6) 9/18 (50.0) LVEDP -< 20 mm Hg 1/34 (2.9) 3/10 (30.0) NS Mitral regurgitation 0/31 (0.0) 5/11 (45.5) No mitral regurgitation 3/67 (4.5) 7/17 (41.2) Total for group 3/98 (3.1) 12/28 (42.9) Abbreviations are the same as in Table 3. Vascular complications occurred in 14.3% (4/28) of the patients in group B who required postcardiotomy counterpulsation (subgroup B2). One patient underwent thrombectomy and embolectomy of the involved extremity, and 2 patients required thrombectomy of the femoral artery and fasciotomy. In the other patient, the extremity improved after IABP removal; no operation was necessary. There were no amputations, no aortic dissections, no cases of gangrenous bowel or paraplegia, and no deaths related to use of the IABP in either group A or B. The prevalence of other postoperative complications was significantly higher in group B. For example, 2.7% (1/37) and 10.3% (13/126) (p < 0.005) of the patients in groups A and B, respectively, sustained a perioperative myocardial infarction, and 3.2% (4/126) of the patients in group B had a cerebrovascular accident compared with none in group A. In group B, 2 patients required hemodialysis, and 1 had a wound infection at the IABP insertion site. Median postoperative hospital stay was 9.9 days (range, 5 to 40 days) and 12.0 days (range, 5 to 78 days) for groups A and B, respectively. However, an analysis of variance showed no significant difference in the mean length of postoperative stay between the two groups because of the wide variability. The mean duration of the postoperative hospital stay was 10.9 days (range, 5 to 78 days) for patients in subgroup B1 compared with 18.9 days (range, 6 to 56 days) for patients in subgroup B2 (postcardiotomy IABP use) who survived to be dis- charged from the hospital (p < 0.05). The following patient characteristics significantly increased the mean duration of postoperative hospital stay in subgroup B2: diabetes mellitus, NYHA class III or IV, and LVEDP higher than 20 mm Hg (Table 5). Mean hospital charges were $50,627 (range, $26,099 to $125,430) and $54,818 (range, $21,551 to $239,072) for survivors in groups A and B, respectively. Although the mean hospital charges were lower for group A patients, the analysis of covariance showed that there was no significant difference between the two groups because of the wide variability. Mean hospital charges were $48,583 (range, $21,551 to $239,072) and $79,238 (range, $35,822 to $149,720) for patients in subgroups B1 and B2, respectively (p ~ 0.001). Compared with subgroup B1, hospital charges were significantly higher (p ~ 0.05) for subgroup B2 patients with the following characteristics: age lower than 70 years, male sex, diabetes mellitus, previous CABG, NYHA class II or IV, nonelective operation, recent myocardial infarction (0 to 7 days prior to CABG), unstable angina, left main stenosis, and coexisting mitral regurgitation (Table 6). Compared with group A, mean hospital charges were less (but not significantly so) in subgroup B1 patients who were 70 years old or less and in patients undergoing an elective operation who were in NYHA class I or II, had an LVEDP lower than 20 mm Hg, and had no associated mitral regurgitation. Comment Review of the recent literature indicates that the 30-day mortality rate in patients undergoing coronary revascularization who have an EF of 0.25 or less ranges from 4.8% to 14.3% [1-5]. Wechsler and Junod [9] suggested that advanced age, poor EF, presence of mitral regurgitation, and left main coronary artery disease are incremental risk factors for operative death in patients undergoing myocardial revascularization who have chronic congestive heart failure. Creswell and associates [10] have shown that revascularization within 14 days of an acute myocardial infarction is also associated with an increased risk; however, in their experience, operative mortality was significantly reduced in patients who received an IABP preoperatively for postinfarction angina. In our experience, the 30-day mortality rate was also significantly reduced (p K 0.005) in a group of 37 patients in whom an IABP was inserted prophylactically before operation compared with a group of 126 patients who underwent CABG without preoperative device insertion (2.7% versus 11.9%). Using multiple logistic regression analysis, we identified the following characteristics to be indicative of increased risk for early death after myocardial revascularization in patients with severely depressed ventricular function (LVEF -< 0.25): previous CABG, NYHA class III or IV, left main stenosis, nonelective operation, and LVEDP greater than 20 mm Hg (see Table 4). Prophylactic placement of an IABP before anesthetic induction may be beneficial in such patients.

5 Ann Thorac Surg DIETL ET AL ;62:401-9 COST-EFFECTIVENESS OF PREOPERATIVE IABP Table 5. Mean Postoperative Hospital Stay '~ Patient Characteristics Group A Subgroup B1 Subgroup B2 p Value Age -~ 70 y 74/8 (9.3) 366/22 (16.6) 130/5 (26.0) NS Age < 70 y 282/28 (10.1) 668/73 (9.2) 172/11 (15.6) NS Female sex 102/7 (14.6) 195/20 (9.8) 170/5 (34.0) NS Male sex 254/29 (8.8) 839/75 (11.2) 132/11 (12.0) NS Diabetes mellitus 179/16 (11.2) 525/44 (11.9) 169/6 (28.2) Nondiabetes 177/20 (8.9) 509/51 (10.0) 133/10 (13.3) NS Previous CABG 74/9 (8.2) 46/4 (11.5) 62/2 (31.0) NS First CABG 282/27 (10.4) 988/91 (10.9) 240/14 (17.1) NS NYHA class III or IV 331/33 (10.0) 469/41 (11.4) 269/13 (20.7) NYHA class I or II 25/3 (8.3) 565/54 (10.5) 33[3 (11.0) NS CCS class III or 1V 321/32 (10.0) 766/66 (11.6) 216/11 (19.6) NS CCS class I or II 35/4 {8.8) 268/29 (9.2) 86/5 (17.2) NS Elective operation 51/7 (7.3) 369/39 (9.5) 73[4 (18.2) NS Nonelective operation 305/29 (10.5) 665/56 (11.9) 229/12 (19.1) NS Recent MI (0-7 days) 173/15 (11.5) 224/20 (11.2) 167/8 (20.9) NS No recent MI 183/21 (8.7) 810/75 (10.8) 135/8 (16.9) NS Left main stenosis 150/13 (11.5) 85/8 (10.6) 43/2 (21.5) NS No left main stenosis 206/23 (9.0) 949/87 (10.9) 259/14 (18.5) NS LVEDP > 20 mm Hg 261/26 (10.0) 646/54 (12.0) 206/9 (22.9) LVEDP -< 20 mm Hg 80 8 (10.0) 388/41 (9.5) 96/7 (13.7) NS Mitral regurgitation 141/14 (10.1) 415/31 (13.4) 116/6 (19.3) NS No mitral regurgitation 215[22 (9.8) 619/64 (9.7) 186/10 (18.6) NS Total for group (9.9) 1,034/95 (10.9) 302/16 (18.9) ~0.05 Abbreviations are the same as in Table 3. Counterpulsation with an IABP was first used in 1967 by Kantrowitz and associates [11] for circulatory support of patients in cardiogenic shock. The IABP augments the aortic diastolic pressure and consequently increases coronary blood flow; the IABP concurrently reduces afterload and myocardial oxygen demand. For these reasons, counterpulsation is particularly useful for providing temporary support of the ischemic myocardium before revascularization procedures. Isom and co-workers [12] observed that maximum release of the myocardial-specific isoenzyme of creatine kinase occurred in 92% of patients during the prebypass period after induction of general anesthesia. For this reason, Gunstensen and associates [6] recommended insertion of the IABP under local anesthesia prior to induction of general anesthesia in patients with acute coronary insufficiency, critical left main coronary artery stenosis, or severe LV dysfunction. In their series, prophylactic preoperative counterpulsation in patients undergoing CABG was associated with a 7.5% mortality rate compared with a 23.1% mortality rate when the IABP was used after cardiotomy. Recent studies [13-16] show that the early postoperative mortality rate in patients requiring postcardiotomy IABP assistance after CABG ranges from 37.5% to 48.4%. In our present series, use of an IABP after cardiotomy (subgroup B2) was associated with a 42.9% mortality rate. Golding and associates [15] showed that delayed use of the IABP was associated with an elevated mortality, whereas the prevalence of perioperative myocardial infarction and in-hospital mortality were significantly less in patients in whom IABP support was initiated preoperatively. Creswell [13], Bolooki [17], and their colleagues also demonstrated that earlier use of the IABP is associated with improved survival. Indications for prophylactic IABP assistance before revascularization procedures include acute coronary insufficiency with refractory unstable angina, severe left main coronary occlusion or its equivalent, and severe LV dysfunction [6, 17]. In this study, our criteria for elective preoperative IABP insertion included the combination of severe LV dysfunction and severe myocardial ischemia, especially in patients with left main stenosis or a recent myocardial infarction or in patients undergoing reoperation. According to Akins [18], the IABP is the most predictable method of controlling myocardial ischemia, and preoperative counterpulsation allows virtually all patients to have revascularization in a nonischemic state. Intraaortic balloon counterpulsation allows stabilization of the patient's hemodynamic status and permits safer cardiac catheterization and induction of general anesthesia [6]. Another potential advantage of preoperative counterpulsation is that an IABP already in place can be used to provide pulsatile flow during cardiopulmonary bypass, which may result in better organ function than the nonpulsatile flow of conventional bypass [19]. However, great caution must be exercised when using

6 406 DIETL ET AL Ann Thorac Surg COST-EFFECTIVENESS OF PREOPERATIVE IABP 1996;62:401-9 Table 6. Mean Hospital Charges" P Patient Characteristics Group A Subgroup B1 Subgroup B2 Value b Age -> 70 y $ 53,180 $ 57,218 $ 90,276 NS Age ~ 70 y $ 49,989 $ 45,944 $ 76, Female sex $ 58,111 $ 50,940 $ 99,447 NS Male sex $ 48,756 $ 47,862 $ 67, Diabetes mellitus $ 50,044 $ 49,045 $ 95, Nondiabetes $ 51,118 $ 48,140 $ 59,793 NS Previous CABG $ 52,134 $ 56,168 $ 95,733 <0.001 First CABG $ 50,106 $ 48,246 $ 72, NYHA class III or IV $ 50,868 $ 52,441 $ 82, NYHA class I or II $ 48,056 $ 45,187 $ 53,138 NS CCS class III or IV $ 51,073 $ 50,482 $ 82, CCS class I or II $ 45,875 $ 43,871 $ 71,430 NS Elective operation $ 46,312 $ 44,030 $ 62,488 NS Nonelective operation $ 51,706 $ 51,408 $ 84,822 <0.001 Recent MI (0-7 days) $ 54,369 $ 54,547 $ 80, No recent MI $ 47,821 $ 46,651 $ 78, Left main stenosis $ 51,526 $ 49,747 $ 69, No left main stenosis $ 50,096 $ 48,474 $ 83, LVEDP > 20 mm Hg $ 50,895 $ 52,397 $ 80,844 NS LVEDP <- 20 mm Hg $ 49,724 $ 43,889 $ 76, Mitral regurgitation $ 51,341 $ 54,337 $ 83, No mitral regurgitation $ 49,557 $ 45,751 $ 76,913 ~0.001 Mean for group $ 50,627 $ 48,583 $ 79,238 ~0.001 a Numbers in parentheses are mean hospital charges, b Comparison between subgroups B1 and B2. Abbreviations are the same as in Table 3. the IABP because vascular complications associated with use of this device are not uncommon. We observed vascular complications in 3 (8.1%) of the 37 patients in group A (preoperative IABP placement) and in 4 (14.3%) of the 28 patients in group B having postcardiotomy IABP insertion. However, no patient in our series required amputation. With the femoral approach, the prevalence of IABPrelated vascular complications ranges from 6.8% to 32.6% [20-25]. Kantrowitz and associates [23], however, suggested that most IABP complications are minor and resolve after balloon removal, are usually related to the vascular status of the patient, and, with the exception of bacteremia, are independent of the duration of counterpulsation. Deaths directly attributable to the IABP are very uncommon [13, 17, 20, 21, 25]. Several groups [21-26] have suggested that the percutaneous approach is associated with an increased rate of major vascular complications, especially in patients with peripheral vascular disease, in women, and in diabetic patients. To reduce the risk of ischemic complications, Phillips and associates [27] recommend a sheathless insertion of the percutaneous balloon in patients with small or diseased arteries. Alternatively, several authors [21, 24] suggest insertion of the IABP through the ascending aorta whenever the wire-guided balloon cannot be easily advanced from the femoral approach. Other major complications associated with the IABP include aortic dissection [21, 23, 25], paraplegia [21, 24], bacteremia [21-26], mesenteric infarction [22, 24, 28], balloon rupture [21, 22], and balloon entrapment [29]. We did not observe any of these complications in our series. Finally, with emphasis on cost containment in health care today, a very important aspect to consider is whether a procedure is cost-effective or not. Bolooki and colleagues [17] observed that the average hospital stay was 12 days for patients who had elective IABP insertion before operation compared with 18 days for patients requiring IABP support after CABG. They also noted an increased rate of renal and pulmonary complications in the latter group. Gunstensen and associates [6] had a similar experience and concluded that prophylactic IABP use is cost-effective in patients with severely depressed ventricular function, left main stenosis, or refractory unstable angina. We compared the average duration of postoperative hospitalization and the mean hospital charges for all surviving patients. We observed that the average hospital stay was 2.1 days shorter for patients with elective preoperative IABP placement (group A) compared with those without preoperative counterpulsation (group B) (mean, 9.9 days versus 12.0 days). Also, the mean hospital charges for patients with an IABP preoperatively were lower than those for patients without preoperative IABP insertion ($50,627 versus $54,818). Because of the wide variability, these differences between groups A and B were not significant. Nevertheless, we observed that prophylactic use of the

7 Ann Thorac Surg DIETL ET AL ;62:401-9 COST-EFFECTWENESS OF PREOPERATWE IABP Table 7. Trends in Four Categories Over 5 Years a Category IABP use No. of patients (n = 163) Total with IABP No. with preop IABP 1 (4.5) 6 (13.0) 4 (10.3) 14 (40.0) 12 (57.1) No. with postop IABP 3 (13.6) 12 (26.1) 10 (25.6) 3 (8.6) 0 (0.0) Overall 30-day mortality IABP mortality 2/4 (50.0) 4/18 (22.2) 4/14 (28.6) 2/17 (11.8) 1/12 (8.3) Overall mortality day mortality (%) Postop hospital stay Mean postop hospital stay (days) % of patients hospitalized -> days postop Mean hospital charges $48,322 $52,536 $64,107 $50,533 $41,316 IABP = intraaortic balloon pump. IABP did not prolong hospital stay and did not increase costs. Table 5, for example, shows that the mean postoperative hospital stay was slightly shorter (but not significantly so) for patients in group A compared with patients in subgroup B1 (no counterpulsation). However, mean hospital charges were higher (but not significantly higher) for patients in group A compared with patients in subgroup B1 (see Table 6). A possible explanation is that the hospital charges included heart catheterization and CABG (diagnostic-related group 106) done on an urgent or emergency basis during the same hospital admission in 81.1% (30/37) of the patients in group A compared with only 60.2% (59/98) of the patients in subgroup B1. Our study also demonstrated that the mean postoperative hospital stay and the mean hospital charges were significantly higher for patients in subgroup B2 (p < 0.05) compared with patients in group A and subgroup B1 (see Tables 5, 6). Thus, it is also important to identify the patients with severely impaired ventricular function who may require counterpulsation after CABG (see Table 2). Statistical analysis showed that the following patients were more likely to need an IABP after CABG (in group B): patients in NYHA class III or IV (odds were 5.8 times higher), patients undergoing reoperation (odds were 4.5 times higher), patients with a recent myocardial infarction (0 to 7 days before operation), and patients undergoing a nonelective operation with unstable angina, intravenous nitroglycerin, or left main stenosis (odds were between 2 and 3 times higher) (see Table 2). Table 7 shows the trend toward more frequent use of the IABP preoperatively during 1994 and 1995, which was associated with a lower overall mortality, a shorter hospital stay, and reduced hospital charges (especially during 1995). In summary, use of the IABP before operation effectively controls myocardial ischemia in patients with poor ventricular function. It is the most predictable method of myocardial protection during induction of general anes- thesia and during the prebypass period. Preoperative identification of patients who are more likely to require postcardiotomy counterpulsation is very important because these are the patients who have a significantly increased mortality rate, a significantly longer hospital stay, and significantly higher hospital charges. Although our results do not show any particular advantage for use of the IABP preoperatively in patients with an LVEF of 0.25 or less who are undergoing elective CABG and are in class I or II of the NYHA and Canadian Cardiovascular Society classifications, our experience suggests that most patients with severe LV dysfunction undergoing coronary bypass operations may benefit from prophylactic IABP insertion, especially patients undergoing reoperations, patients in NYHA class III or IV, patients with a recent myocardial infarction, or patients with left main stenosis. Elective preoperative IABP insertion significantly improved survival, reduced hospital stay, and was the more cost-effective approach for these patients. We are very grateful to Carmen Acufia, PhD (Department of Mathematics, Bucknell University, Lewisburg, PA), for the statistical analysis; to Christy W. Brouse and Dave Evans (Information Analysis, Geisinger Clinic System Services) for providing the inpatient hospital charges; and to the following perfusionists: Myra H. Klayman, Jeffrey T. Minnich, James D. Murdock, Benjamin T. Poon, Richard F. Slusser, and David L. Updegrove. References 1. Lansman SL, Cohen M, Galla JD, et al. Coronary bypass with ejection fraction of 0.20 or less using centigrade cardioplegia: long-term follow-up. Ann Thorac Surg 1993;56: Goor DA, Golan M, Bar-El Y, et al. Synergism between infarct-borne left ventricular dysfunction and cardiomegaly in increasing the risk of coronary bypass surgery. J Thorac Cardiovasc Surg 1992;104:983-9.

8 408 DIETL ET AL Ann Thorac Surg COST-EFFECTIVENESS OF PREOPERATIVE IABP 1996;62: Christakis GT, Weisel RD, Fremes SE, et al. Coronary artery bypass grafting in patients with poor ventricular function. J Thorac Cardiovasc Surg 1992;103: Milano CA, White WD, Smith LR, et al. Coronary artery bypass in patients with severely depressed ventricular function. Ann Thorac Surg 1993;56: Christenson JT, Maurice J, Simonet F, et al. Effect of low left ventricular ejection fractions on the outcome of primary coronary by-pass grafting in end-stage coronary artery disease. J Cardiovasc Surg (Torino) 1995;36: Gunstensen J, Goldman BS, Scully HE, Huckell VF, Adelman AG. Evolving indications for preoperative intraaortic balloon pump assistance. Ann Thorac Surg 1976;22: Seizer A, Cohn K. Functional classification of cardiac disease: a critique. Am J Cardiol 1972;30: Campeau L. Grading of angina pectoris [Letter]. Circulation 1976;54: Wechsler AS, Junod FL. Coronary bypass grafting in patients with chronic congestive heart failure. Circulation 1989;79 (Suppl 1): Creswell LL, Moulton MJ, Cox JL, Rosenbloom M. Revascularization after acute myocardial infarction. Ann Thorac Surg 1995;60: Kantrowitz A, Tjonneland S, Freed PS, Phillips SJ, Butner AN, Sherman JL Jr. Initial clinical experience with intraaortic balloon pumping in cardiogenic shock. JAMA 1968;203: Isom OW, Spencer FC, Feigenbaum H, Cunningham J, Roe C. Prebypass myocardial damage in patients undergoing coronary revascularization: an unrecognized vulnerable period [Abstract]. Circulation 1975;51(Suppl 2): Creswell LL, Rosenbloom M, Cox JL, et al. Intraaortic balloon counterpulsation: patterns of usage and outcome in cardiac surgery patients. Ann Thorac Surg 1992;54: Di Lello F, Mullen DC, Flemma RJ, Anderson AJ, Kieinman LH, Werner PH. Results of intraaortic balloon pumping after cardiac surgery: experience with the Percor balloon catheter. Ann Thorae Surg 1988;46: Golding LAR, Loop FD, Peter M, Cosgrove DM, Taylor PC, Phillips DF. Late survival following use of intraaortic balloon pump in revascularization operations. Ann Thorac Surg 1980;30: Baldwin RT, Slogoff S, Noon GP, et al. A model to predict survival at time of postcardiotomy intraaortic balloon pump insertion. Ann Thorac Surg 1993;55: Botooki H, Williams W, Thurer RJ, et al. Clinical and hemodynamic criteria for use of the intra-aortic balloon pump in patients requiring cardiac surgery. J Thorac Cardiovasc Surg 1976;72: Akins CW. Discussion of Creswell LL, Rosenbloom M, Cox JL, et al. Intraaortic balloon counterpulsation: patterns of usage and outcome in cardiac surgery patients. Ann Thorac Surg 1992;54: Pappas G, Winter SD, Kopriva CJ, et al. Improvement of myocardial and other vital organ functions and metabolism with a simple method of pulsatile flow (IABP) during clinical cardiopulmonary bypass. Surgery 1975;77: Leinbach RC, Goldstein J, Gold HK, Moses JW, Collins MB, Subramanian V. Percutaneous wire-guided balloon pumping. Am J Cardiol 1982;49: Pennington DG, Swartz M, Codd JE, Merjavy JP, Kaiser GC. Intraaortic balloon pumping in cardiac surgical patients: a nine-year experience. Ann Thorac Surg 1983;36: Gottlieb SO, Brinker JA, Borkon M, et al. Identification of patients at high risk for complications of intraaortic balloon counterpulsation: a multivariate risk factor analysis. Am J Cardiol 1984;53: Kantrowitz A, Wasfie T, Freed PS, Rubenfire M, Wajszczuk W, Schork MA. Intraaortic balloon pumping 1967 through 1982: analysis of complications in 733 patients. Am J Cardiol 1986;57: Macoviak J, Stephenson LW, Edmunds LH Jr, Harken A, MacVaugh HIII. The intraaortic balloon pump: an analysis of five years' experience. Ann Thorac Surg 1980;29: Shahian DM, Neptune WB, Ellis FH Jr, Maggs PR. Intraaortic balloon pump morbidity: a comparative analysis of risk factors between percutaneous and surgical techniques. Ann Thorac Surg 1983;36: Wasfie T, Freed PS, Rubenfire M, et al. Risks associated with intraaortic balloon pumping in patients with and without diabetes mellitus. Am J Cardiol 1988;61: Phillips SJ, Tannenbaum M, Zeff RH, Iannone LA, Ghali M, Kongtahworn C. Sheathless insertion of the percutaneous intraaortic balloon pump: an alternate method. Ann Thorac Surg 1992;53: Jarmolowski CR, Poirier RL. Small bowel infarction complicating intra-aortic balloon counterpulsation via the ascending aorta. J Thorac Cardiovasc Surg 1980;79: Aru GM, King JT Jr, Hovaguimian H, Floten HS, Ahmad A, Starr A. The entrapped balloon: report of a possibly serious complication. J Thorac Cardiovasc Surg 1986;91: DISCUSSION DR SHLOMO GABBAY (Newark, NJ): I congratulate Dr Dietl and his co-workers for this excellent investigation. This study seems to offer convincing proof that we should be less reluctant to use the intraaortic balloon. I would like to point out a misconception; most surgeons think that if the patient leaves the operating room with an intraaortic balloon, this represents a "failure"; they see it as a sign that maybe the operation did not go well. If we want to save lives, I believe this is the wrong attitude. The results of this investigation show that if we think about inserting the balloon, it must be done immediately. I am sure many of you have noticed that when you use the balloon prophylactically because the patient is considered a high risk, that patient gets through the procedure better than the so-called low-risk patient. This anecdotal experience has been proved by this study. I have long wanted to conduct a similar study to prove this point and am delighted that it was done by your group. I suggest that inserting the intraaortic balloon is a "virtue" and not a sign of surgical mishap. Use of the intraaortic balloon might save lives and, according to this study, might save some money, too. DR NOEL L. MILLS (New Orleans, LA): In any of your patients did you use thyroid hormone before you put them on the intraaortic balloon pump at the time of operation? DR DIETL: We have used triiodothyronine in a few patients, usually before weaning from cardiopulmonary bypass. I cannot say whether it was beneficial or not because of the small number of patients in whom it was used.

9 Ann Thorac Surg DIETL ET AL ;62:401-9 COST-EFFECTIVENESS OF PREOPERATIVE IABP DR MILLS: How many of you use triiodothyronine when you are in trouble before you put the patient on the balloon pump? It looks as though none of you do. Now, how many of you use triiodothyronine before you take off your gloves and go out to tell the family you are sorry, but you could not save the patient? There appear to be several of you. I have one more point. I believe there were no complications insofar as the balloon pump was concerned; there was no isehemia, no fasciotomies, and no femoral femoral bypasses? DR DIETL: In group A, 8.1% (3 patients) had vascular complications. One patient had an ischemic leg without pulses. The condition of the leg improved after the balloon was removed shortly after the operation. Another patient had severe hemorrhage from the femoral artery after balloon removal, and 1 patient had development of a pseudoaneurysm. Both required surgical intervention. In group B, 14.3% of the patients (4/28) in whom the balloon was used had ischemic complications in the lower extremity. Two of them required thrombectomy of the femoral artery and a fasciotomy; 1 patient underwent thrombectomy and embolectomy, and another patient was simply observed after balloon removal. There were no amputations in either group. DR ANTHONY L. MOULTON (Providence, RI): I have a suggestion for what may be an intermediate course in some of these patients who are potentially high risk. One of the things my colleagues and I do for patients with low ejection fractions is place a femoral artery line at the start of the procedure. We all know how difficult it is to predict the degre~ of reversibility of ischemic ventricular dysfunction. When completely revascularized, a number of these patients come off bypass reasonably well. Although you predicted that they would need a balloon pump postoperatively, they do extremely well without one. One of the advantages of having a femoral artery line already in place is that in a patient who is not coming off bypass well and has no palpable pulses, you can simply slide up a guidewire and put in the intraaortic balloon pump without any fuss. It certainly speeds up that insertion. If the arterial line is already in place and you do not need a balloon pump, you have not exposed the patient to the risk of the balloon itself. DR ALFONSO CHISCANO (San Antonio, TX): Thank you so much, Dr Dietl, for your beautiful paper. 1 stress the importance of the timing of the intraaortic balloon insertions. If the patient has had a recent myocardial infarction and has an ejection fraction of 0.15 to 0.20, and if the cardiologist gives you the luxury, ask him or her to insert the balloon the day before operation. I am aiming at 12 hours. In my experience, patients seem to do better than if they have the balloon inserted just preoperatively or postoperatively. DR DIETL: Thank you, Dr Chiscano, for your comments. The timing of prophylactic insertion of the balloon pump depends on the clinical presentation. If the patient has chest pain during or shortly after cardiac catheterization, the cardiologists usually insert the balloon, and we take the patient to the operating room the same day. However, if the patient has no pain during the study, we prefer to insert the balloon ourselves in the operating room, just before induction of general anesthesia, using local anesthesia in the groin. We start counterpulsation before induction, and we verify the position of the balloon with transesophageal echocardiography, after the patient is intubated.