Among the many challenges presented to the cardiovascular. Impact of Retrograde Cerebral Perfusion on Ascending Aortic and Arch Aneurysm Repair

Transcription

1 Impact of Retrograde Cerebral Perfusion on Ascending Aortic and Arch Aneurysm Repair Hazim J. Safi, MD, George V. Letsou, MD, Dimitrios C. Iliopoulos, MD, Mahesh H. Subramaniam, MS, Charles C. Miller III, PhD, Heitham Hassoun, MS, Panayiotis J. Asimacopoulos, MD, and John C. Baldwin, MD Department of Surgery, Baylor College of Medicine, The Methodist Hospital, Houston, Texas Purpose. The effect of retrograde cerebral perfusion on the incidence of stroke and death among patients undergoing repair of aneurysms of the ascending aorta and transverse arch was determined. Material and Methods. Between January 1991 and March 1995, 161 patients were operated on for aneurysms of the ascending aorta and transverse arch. Thirty-three of the patients (20%) had an aneurysm of the ascending aorta only and 128 (80%) had aneurysms of both the ascending aorta and the transverse arch. All the patients underwent cardiopulmonary bypass, profound hypothermia, and circulatory arrest, and 120 (74%) also underwent retrograde cerebral perfusion. Median pump time was 143 minutes (range, 21 to 461 minutes). Median circulatory arrest time was 42 minutes (range, 8 to 111 minutes), and median myocardial ischemic time was 71 minutes (range, 14 to 306 minutes). Results. The overall 30-day mortality rate was 6% (9 patients) and the incidence of stroke was 4% (7 patients). The use of retrograde cerebral perfusion demonstrated a protective effect against stroke (3 of 120 patients, or 3%) compared with no retrograde cerebral perfusion (4 of 41 patients, or 9%; odds ratio, 0.24; confidence interval, 0.06 to 0.99; p < 0.049). This was most significant in patients more than 70 years of age; none of the 36 elderly patients who received retrograde cerebral perfusion had a stroke, compared with 3 of the 13 (23%) who did not (p < 0.003). Only pump time was associated with an increased risk of stroke (odds ratio, 1.01; 95% confidence interval, 1.00 to 1.02; p < 0.005). Pump time also was associated with increased mortality (odds ratio, 1.01; 95% confidence interval, 1.00 to 1.02; p < 0.008). Conclusion. Retrograde cerebral perfusion decreased the incidence of stroke in patients undergoing repair of aneurysms of the ascending aorta and transverse arch. (Ann Thorac Surg 1997;63:1601 7) 1997 by The Society of Thoracic Surgeons Among the many challenges presented to the cardiovascular surgeon during the repair of ascending aorta and transverse arch aneurysms is the provision of both heart and brain protection. Cardiopulmonary bypass, profound hypothermia, and circulatory arrest were a great development toward this purpose in the mid- 1970s [1 3]. Profound hypothermia provided protection of the brain and other organs. Cardiopulmonary bypass and circulatory arrest cleared the cluttered operative field of shunts and clamps, and made graft replacement of the ascending aorta and transverse arch easier. A safe circulatory arrest period, however, has been difficult to quantify. Circulatory arrest for less than 30 minutes appears to be safe, but arrest for more than 45 minutes has been found to increase the incidence of stroke. Arrest for more than 65 minutes has been found to increase the incidence of death [4]. Mills and Ochsner [5] in 1980 first used retrograde cerebral perfusion (RCP) through the superior vena cava in a patient to treat a major air embolus and prevent Presented at the Forty-third Annual Meeting of the Southern Thoracic Surgical Association, Cancun, Mexico, Nov 7 9, Address reprint requests to Dr Safi, 6550 Fannin, Suite 1603, Houston, TX damage to the brain. By 1986, Ueda and colleagues [6] had developed the clinical application of RCP as it is currently used for brain protection during extended circulatory arrest and profound hypothermia. Their innovative design used continuous rather than intermittent perfusion and simplified the perfusion circuit of earlier designs by Lemole and associates [7]. Continuing studies indicate that RCP is a valuable defense against brain damage during ascending aorta and transverse arch repair, with the potential to extend the duration of safe circulatory arrest [8, 9]. During oxygenation and cooling, the brain is protected against reperfusion injury and provided with metabolic support and removal of toxic metabolites [10 12]. In 1993, my colleagues and I reported our preliminary clinical experience with RCP [13]. The results were promising for future application, and in 1995, the excellent results of our animal study prompted us to use RCP in all our patients during ascending aorta and transverse arch This article has been selected for the open discussion forum on the STS Web site: by The Society of Thoracic Surgeons /97/$17.00 Published by Elsevier Science Inc PII S (97)

2 1602 SAFI ET AL Ann Thorac Surg RETROGRADE CEREBRAL PERFUSION 1997;63: Fig 1. (A) Cooling period. Blood drains from the cannulated superior vena cava and inferior vena cava to the reservoir and is pumped into the femoral artery. (AO aorta.) (B) Hypothermic circulatory arrest and retrograde cerebral perfusion. The pump is turned off. Blood flows from the reservoir to the superior vena cava in retrograde fashion to the brain; pressure and flow are monitored. (C) Warming period. A sidearm arterial line is cut and placed into the transverse arch graft, which is perfused in antegrade fashion to the brachiocephalic arteries. Fig 2. (A) Artist s illustration (left) and computed tomographic scans (A, B, C, D) of aortic dissection, occluded left coronary artery, and massive aortic aneurysm. (B) Finished repair in the same patient by staged repair (elephant-trunk technique), including left coronary artery bypass.

3 Ann Thorac Surg SAFI ET AL 1997;63: RETROGRADE CEREBRAL PERFUSION 1603 aneurysm repair [14]. Here, we report our latest progress and analysis of data regarding ascending aorta and transverse arch aneurysm repair, as well as the relation of stroke and death to the absence or presence of RCP. Material and Methods Between January 1991 and March 1995, 161 patients underwent repair of ascending aorta and transverse arch aneurysms. Seventy-two of the patients (45%) were women and 89 (55%) were men. The median age was 62 years (range, 21 to 83 years). Hypertension was present in 89 patients (55%). Twenty-six patients (16%) had chronic obstructive pulmonary disease, 11 (7%) had cerebrovascular disease, and 35 (22%) had heart disease. Twentyfour patients (15%) underwent aortic valve and ascending aorta replacement; 40 (25%) underwent aortic valve, ascending aorta, and transverse arch replacement; 9 (6%) underwent ascending aorta replacement; and 88 (55%) underwent ascending aorta and transverse arch replacement. Aortic aneurysm replacement for 47 patients (29%) was the first stage of the elephant-trunk technique (followed later by descending or thoracoabdominal aortic aneurysm repair). Thirty-nine patients (24%) had undergone previous surgery of the ascending aorta or transverse arch. Twenty patients (12%) had undergone previous replacement of a descending, thoracoabdominal, or abdominal aortic aneurysm. All patients underwent cardiopulmonary bypass, profound hypothermia, and circulatory arrest. After February 1994, all patients underwent RCP, for a total of 120 of 161 patients (74%). Intraoperative Data The median times for the operative procedures were as follows: pump time, 54 minutes (range, 14 to 306 minutes); cooling time, 24 minutes (range, 4 to 63 minutes); warm-up time, 68 minutes (range, 37 to 189 minutes). The median circulatory arrest time was 42 minutes (range, 8 to 111 minutes). The median myocardial ischemic time was 71 minutes (range, 14 to 306 minutes), and the median RCP flow time was 30 minutes (range, 8 to 85 minutes). Operative Technique Fig 3. Probability of stroke by pump time. Retrograde cerebral perfusion (RCP) blunts the rise in the probability of stroke seen with increasing pump times. The operative technique used has been described and illustrated in step-by-step fashion previously [13]. All patients underwent graft replacement of the ascending aorta, usually the aortic arch, and in 40% of cases, also the aortic valve. All the operations were accompanied by cardiopulmonary bypass, profound hypothermia, circulatory arrest, and, in most patients (74%), RCP. The chest was entered through a median sternotomy. After an incision was made in the right groin, we cannulated the right femoral artery and vein for cardiopulmonary bypass. The cooling phase is illustrated in Figure 1A. Cardiopulmonary bypass followed systemic heparinization and patient cooling until the electroencephalogram was isoelectric. Circulation then was arrested and cardiopulmonary bypass was terminated. The heart was protected with the retrograde input of cold blood cardioplegia through the coronary sinus, keeping myocardial temperature less than 15 C. Retrograde cerebral perfusion was begun, perfusing oxygenated blood at 10 C through the superior vena cava to the internal jugular vein and then to the brain while the patient was in a head-down position (see Fig 1B). The superior vena cava pressure was maintained at 25 mm Hg. Flow rate fell to less than, but was not permitted to exceed, 500 ml/min. If the safeguard nasopharyngeal temperature fell to less than 12 C, cooling was halted. The aneurysm was opened and excised with the aortic valve resuspended or replaced. When the aortic arch was involved, the preserved island of the brachiocephalic arteries was attached to the graft. Retrograde cerebral perfusion was stopped and cardiopulmonary bypass was restarted while the patient was still in a head-down position. For rewarming, an arterial sidearm was inserted into the transverse aortic graft and flow was delivered in antegrade fashion (see Fig 1C). We always aspirated the left ventricle to remove any potentially trapped air and then completed the proximal anastomosis. Protamine sulfate was administered to reverse the anticoagulated state. Patients were weaned from cardiopulmonary bypass and the chest was closed in the usual fashion. Figure 2 provides an example of the sort of case for which we typically use RCP. Statistical Methods Univariate statistical tests for dichotomous variables were computed by contingency table methods. Ninetyfive percent confidence intervals are test-based. Univariate tests for continuous variables were computed two ways by 2 analysis across the data divided into quartiles, and by continuous univariate logistic regression analysis.

4 1604 SAFI ET AL Ann Thorac Surg RETROGRADE CEREBRAL PERFUSION 1997;63: Table 1. Patient Characteristics: Association With Stroke Variable Patients (%) 95% Confidence Strokes (%) Odds Ratio a Interval b p Value c All patients 161 (100) 7 (4.4) Age (y) (24.8) 1 (2.5) (24.8) 2 (5) (0.36) (23) 2 (5.4) (27.4) 2 (4.6) Female 72 (44.7) 4 (5.6) Male 89 (55.3) 3 (3.4) 1 Hypertensive 89 (55.3) 5 (5.6) Normotensive 72 (44.7) 2 (2.8) 1 COPD 26 (16.2) 1 (3.9) No COPD 135 (83.8) 6 (4.4) 1 CVD 11 (6.8) 1 (9.1) No CVD 150 (93.2) 6 (4) 1 Heart disease 35 (21.7) 1 (2.9) No heart disease 126 (78.3) 6 (4.8) 1 RCP 120 (74.5) 3 (2.5) No RCP 41 (25.5) 4 (9.8) 1 Pump time (min) (22.4) 2 (5.6) (26.1) 0 (0) (0.005) (26.1) 1 (2.4) (25.5) 4 (9.8) Cerebral circulatory arrest time (min) Missing (24) 1 (2.6) (26) 3 (7.3) (0.55) (22.2) 1 (2.9) (27.8) 2 (4.6) Total clamp time (min) Missing (22.5) 2 (5.6) (28.1) 1 (2.2) (0.85) (24.4) 2 (5.1) (25) 2 (5) a Odds ratios derived from tables with zero cells were computed using the 0.5 per cell addition correction and logic estimation; b This reflects the units against which its companion odds ratio is computed. Confidence intervals are test-based; c Probability of type I statistical error (common p value). Values without parentheses are Pearson 2 probabilities. Probability values in parentheses are univariate logistic regression likelihood ratio p values. COPD chronic obstructive pulmonary disorder; CVD cerebrovascular disease; RCP retrograde cerebral perfusion. Although continuous data are presented in quartiles in the tables, the odds ratios are against the continuous variable. Multivariate analysis was conducted by Mantel- Haenszel stratification and by multiple logistic regression for dichotomous and continuous variables, respectively. Odds ratios for continuous variables are scaled to a one-unit change in the variable. Logistic regression odds were converted to the probabilities shown in the figure using the transformation OR/(1 OR), where OR is the multivariate odds ratio computed from the exponent of the logistic regression equation. All tests were two-tailed. The null hypothesis was rejected at p Results Stroke was defined as a focal deficit whose correlation with a defect on cerebral imaging such as computed tomographic scanning or magnetic resonance imaging was determined by an independent neurologist. The overall incidence of stroke was 4% (7 of 161 patients). For all patients, RCP demonstrated a protective effect against stroke (2 of 120 patients, or 3%) versus no RCP (4 of 41 patients, or 9%; odds ratio, 0.24; 95% confidence interval, 0.06 to 0.99; p 0.049). This became more significant in patients more than 70 years of age; none of the 36 elderly patients who received RCP had a stroke, compared with 3 of the 13 (23%) who did not receive RCP (p 0.003). The overall 30-day mortality rate was 6% (9 of 161 patients). The 30-day mortality rates for patients who did and did not receive RCP were 6% (7 of 120 patients) and 5% (2 of 41 patients), respectively (odds ratio, 1.21; 95% confidence interval, 0.24 to 6.06; p 0.81). Although patients more than 70 years of age who received RCP had

5 Ann Thorac Surg SAFI ET AL 1997;63: RETROGRADE CEREBRAL PERFUSION 1605 Table 2. Patient Characteristics: Association With 30-Day Mortality Rate Variable Patients (%) 95% Confidence Deaths (%) Odds Ratio a Interval b p Value c All patients 161 (100) 9 (5.6) Age (y) (24.8) 5 (12.5) (24.8) 1 (2.5) (0.13) (23) 1 (2.7) (27.4) 2 (4.6) Female 72 (44.7) 4 (5.6) Male 89 (55.3) 5 (5.6) 1 Hypertensive 89 (55.3) 4 (4.5) Normotensive 72 (44.7) 5 (6.9) 1 COPD 26 (16.2) 1 (3.9) No COPD 135 (83.8) 8 (5.9) 1 CVD 11 (6.8) 0 (0) 0.65 a No CVD 150 (93.2) 9 (6) 1 Heart disease 35 (21.7) 0 (0) 0.17* No heart disease 126 (78.3) 9 (7.1) 1 RCP 120 (74.5) 7 (5.8) No RCP 41 (25.5) 2 (4.9) 1 Pump time (min) (22.4) 3 (8.3) (26.1) 0 (0) (0.0009) (26.1) 0 (0) (25.5) 6 (14.6) Cerebral circulatory arrest time (min) Missing (24) 1 (2.6) (26) 3 (7.3) (0.24) (22.2) 1 (2.9) (27.8) 4 (9.1) Total clamp time (min) Missing (22.5) 1 (2.8) (28.1) 2 (4.4) (0.004) (24.4) 1 (2.6) (25) 5 (12.5) Stroke 7 (4.4) 1 (14.3) No stroke 154 (95.7) 5 (5.2) 1 a Odds ratios derived from tables with zero cells were computed using the 0.5 per cell addition correction and logic estimation; b This reflects the units against which its companion odds ratio is computed. Confidence intervals are test-based; c Probability of type I statistical error (common p value). Values without parentheses are Pearson 2 probabilities. Probability values in parentheses are univariate logistic regression likelihood ratio p values. COPD chronic obstructive pulmonary disorder; CVD cerebrovascular disease; RCP retrograde cerebral perfusion. relatively fewer early deaths (1 of 36, or 2%) than those who did not receive RCP (1 of 13, or 8%), these numbers were not statistically significant (odds ratio, 0.34; 95% confidence interval, 0.02 to 5.92; p 0.45). Overall, only pump time (odds ratio, 1.01; 95% confidence interval, 1.00 to 1.02; p 0.005) was significantly associated with an increased risk of stroke (odds ratio, 1.01, 95% confidence interval, 1.00 to 1.02; p 0.008). The increased probability of stroke associated with increased pump time with or without RCP is illustrated in Figure 3. Table 1 outlines the patient characteristics that were analyzed for association with the incidence of stroke and Table 2 outlines the characteristics that were associated with early mortality. Table 3 shows the multiple logistic regression analysis of stroke odds versus pump time and RCP use. The odds ratios for stroke and death in regard to pump time are continuous and reflect a 1.01-fold increase in odds per minute. Comment Cardiopulmonary bypass, profound hypothermia, and circulatory arrest have been the mainstay of heart and brain protection during major cardiovascular procedures since the early 1970s. Although ascending aorta and transverse arch aneurysm repair have been greatly enhanced by these adjuncts, the vexing problem of an entirely safe period of circulatory arrest and ischemia to

6 1606 SAFI ET AL Ann Thorac Surg RETROGRADE CEREBRAL PERFUSION 1997;63: Table 3. Stroke Rate by Cerebral Ischemic Time Between RCP and No-RCP Groups Cerebral Circulatory 95% Confidence Arrest Time (min) No. of Patients (%) No. of Strokes (%) Odds Ratio a Interval b p Value c With RCP Missing (19.5) 0 (0) (28.8) 2 (5.9) (0.54) (24.6) 1 (3.5) (27.1) 0 (0) Without RCP Missing (37.5) 1 (6.7) (17.5) 1 (14.3) (0.81) (15) 0 (0) (30) 2 (16.7) a Odds ratios derived from tables with zero cells were computed using the 0.5 per cell addition correction and logic estimation; b This reflects the units against which its companion odds ratio is computed. Confidence intervals are test-based; c Probability of type I statistical error (common p value). Values without parentheses are Pearson 2 probabilities. Probability values in parentheses are univariate logistic regression likelihood ratio p values. RCP retrograde cerebral perfusion. the brain remains. Our preliminary clinical experience with RCP in 1993 was very positive, and after February 1994, all our patients, unless prevented by anatomic limitations, underwent RCP [13]. In this study of 161 patients undergoing ascending aorta and transverse arch aneurysm repair, RCP reduced the incidence of stroke, which is consistent with the findings of our earlier animal experiment and clinical study. The beneficial effects of RCP were especially significant in patients more than 70 years of age. Age has been explored as a risk factor in many studies involving ascending aorta and transverse arch aneurysm repair procedures that used cardiopulmonary bypass, moderate hypothermia, cold cardioplegia or cardiopulmonary bypass, profound hypothermia, and circulatory arrest. For patients more than 70 years of age, the early mortality rate has varied from 14% to 33% [4, 15, 16]. In these studies, age was declared a significant factor for patient outcome. Of interest is the fact that in these studies, except for 50 of the 656 patients (8%) in the study by Svensson and colleagues [4], there was no perfusion of head vessels. In our current series, age was not a significant factor. We found in previous studies that RCP is differentially more protective in older patients than in younger patients [17]. We hypothesized that because stroke is a multifactorial occurrence that combines host factors with external sources of risk (ie, ischemic time), and because younger patients have less atherosclerosis and better brain perfusion than older patients, they can better tolerate ischemic insult. Younger patients, therefore, received less benefit from RCP than older patients for any given ischemic time. Recently, we also examined the effect of RCP in the staged extensive aneurysm repair known as the elephant-trunk technique [18]. We found that no patients undergoing the elephant-trunk technique with RCP had a stroke, compared with 20% of those without RCP. In conclusion, RCP had a dramatic effect in reducing the incidence of stroke in patients undergoing ascending aorta and transverse arch aneurysm repair. This was most noticeable for patients more than 70 years of age, who ordinarily are at higher risk. Also observed was a benefit to patients undergoing staged repair of extensive aneurysms. We will continue to use RCP with cardiopulmonary bypass, circulatory arrest, and profound hypothermia in all our patients undergoing ascending aorta and transverse arch repair. In the future, a preoperative psychometric analysis to compare the postoperative cognitive problems of these patients should be addressed. We also strongly believe that a comparison of the effectiveness of antegrade perfusion of the brachiocephalic arteries versus RCP should be settled by a randomized study. We extend special thanks to our editor, Amy Wirtz Newland, and grateful acknowledgment to Graeme Hammond, MD, Yale Medical School, for their editorial assistance. References 1. Griepp RB, Stinson EB, Hollingsworth JF, Buehler D. Prosthetic replacement of the aortic arch. J Thorac Cardiovasc Surg 1975;70: Crawford ES, Saleh SA. Transverse aortic arch aneurysm: improved results of treatment employing new modifications of aortic reconstruction and hypothermic cerebral circulatory arrest. Ann Surg 1981;194: Cooley DA, Livesay JJ. Technique of open distal anastomosis for ascending and transverse arch resection. Bull Tex Heart Inst 1981;8: Svensson LG, Crawford ES, Hess KR, et al. Deep hypothermia with circulatory arrest. Determinants of stroke and early mortality in 656 patients. J Thorac Cardiovasc Surg 1993;106: Mills NL, Ochsner JL. Massive air embolism during cardiopulmonary bypass. Causes, prevention, and management. J Thorac Cardiovasc Surg 1980;80:

7 Ann Thorac Surg SAFI ET AL 1997;63: RETROGRADE CEREBRAL PERFUSION Ueda Y, Miki S, Kusuhara K, et al. Surgical treatment of aneurysm or dissection involving the ascending aorta and aortic arch, utilizing circulatory arrest and retrograde cerebral perfusion. J Cardiovasc Surg 1990;31: Lemole GM, Strong MD, Spagna PM, et al. Improved results for dissecting aneurysms. Intraluminal sutureless prosthesis. J Thorac Cardiovasc Surg 1982;83: Pagano D, Carey J, Patel R, et al. Retrograde cerebral perfusion: clinical experience in emergency and elective aortic operations. Ann Thorac Surg 1995;59: Usui A, Tohsio A, Murase M. Early clinical results of retrograde cerebral perfusion for aortic arch operations in Japan. Ann Thorac Surg 1996;62: Yamashita C, Nakamura H, Nishikawa Y, et al. Retrograde cerebral perfusion with circulatory arrest in aortic arch aneurysms. Ann Thorac Surg 1992;54: Ueda Y, Miki S, Okita Y, et al. Protective effect of continuous retrograde cerebral perfusion on the brain during deep hypothermic systemic circulatory arrest. J Card Surg 1994;9: Kimura T, Muraoka R, Chiba Y, et al. Effect of intermittent deep hypothermic circulatory arrest on brain metabolism. J Thorac Cardiovasc Surg 1994;108: Safi HJ, Brien HW, Winter JN, et al. Brain protection via cerebral retrograde perfusion during aortic arch aneurysm repair. Ann Thorac Surg 1993;56: Safi HJ, Iliopoulos DC, Gopinath SP, et al. Retrograde cerebral perfusion during profound hypothermia and circulatory arrest in pigs. Ann Thorac Surg 1995;59: Lytle BW, Mahfood SS, Cosgrove DM, et al. Replacement of the ascending aorta, early and late results. J Thorac Cardiovasc Surg 1990;99: Crawford ES, Svensson LG, Coselli JS, et al. Surgical treatment of aneurysm and/or dissection of the ascending aorta, transverse aortic arch, and ascending aorta and transverse aortic arch. Factors influencing survival in 717 patients. J Thorac Cardiovasc Surg 1989;98: Safi HJ, Miller CC III, Iliopoulos DC. Thoracic aortic aneurysms in the elderly. In: Recent advances in cardiovascular disease in the elderly. Churchill Communications of Japan (in press). 18. Safi HJ, Miller CC III, Iliopoulos DC, et al. Staged repair of extensive aortic aneurysm: improved neurologic outcome. Ann Surg (in press). DISCUSSION DR GEORGE W. JOHNSON (Denison, TX): This was a nice paper, Dr Safi, and I thank you very much for the opportunity to look over your manuscript ahead of time. There are two areas that I have questions about. First, I saw that 20% of your patients had only ascending aortic aneurysms and total circulatory arrest was used with RCP. Is it necessary to do this in those patients who have aneurysms without arch involvement? If so, how is it advantageous in these patients? Second, in patients who have anticipated total circulatory arrest times of less than 30 minutes (one patient had only 8 minutes of arrest time), is it advantageous to do RCP? Can it be deleterious to do RCP in these patients? DR SAFI: Thank you, these are two good questions. I prefer to use circulatory arrest and RCP in cases of ascending aortic aneurysm for several reasons. One reason is that with these adjuncts and no clamps, we are sure to be able to remove all the aneurysm. No residual bits will remain and there is less chance of recurrent aneurysm in that small segment just proximal to the innominate artery. If the aortic arch is dilated, we have ample time to replace it. Another reason is that I have found that the distal anastomosis is more secure when we do not apply clamps. Also, although the procedure is generally faster when using these adjuncts, one never knows when graft replacement will exceed 30 minutes, which would send us into the danger zone for circulatory arrest alone. Retrograde cerebral perfusion for circulatory arrest of less than 30 minutes is not deleterious. So, in other words, we prefer to use these adjuncts as a safeguard in ascending aortic aneurysm repair. I agree with you that for operations lasting less than 30 minutes, the patient would likely do well without RCP, but we like to guard against the surprises that often await us, especially aortic dissection. DR CHRISTOPHER J. KNOTT-CRAIG (Oklahoma City, OK): I enjoyed the paper very much and compliment you on your excellent results. In using a similar technique, my colleagues and I have placed a retrograde cardioplegic coronary sinus cannula in the superior vena cava, because it has its own pressure-monitoring line, and when the balloon inflates, it is very easy to make sure all the blood is going up the superior vena cava. How do you monitor the pressure in the superior vena cava, and have you used this at all? DR SAFI: I have not tried this kind of pressure-monitoring line, but it sounds like a good one. Personally, to keep the method as simple as possible, I have preferred to keep the monitor on the arterial line to the superior vena cava. All cannulas are placed at the start of the operation with Y connections that are clamped and unclamped to change the direction of blood flow from the cooling to the RCP mode.