Routine minimally invasive aortic valve procedures

Cardiovascular Surgery, Vol. 8, No. 6, pp. 484–490, 2000  2000 The International Society for Cardiovascular Surgery. Published by Elsevier Science Ltd All rights reserved. Printed in Great Britain 0967–2109/00 $20.00

PII: S0967-2109(00)00067-3

www.elsevier.com/locate/cardiosur

Routine minimally invasive aortic valve procedures Jae-Won Lee, Sang-Kwon Lee, Suk-Joong Choo, Hyun Song and Meung-Gun Song Department of Thoracic and Cardiovascular Surgery, Asan Medical Center and University of Ulsan, 388-1 Pungnap-Dong, Songpa-Gu, Seoul 138-736, South Korea Background: Due to the lack of objective evidence supporting the advantages and early technical difficulties, minimally invasive aortic valve procedures were performed on a highly selective rather than routine basis. Methods: From September 1997 to February 1999, one surgeon routinely used upper or transverse minimally invasive sternotomy to perform 46 consecutive cases of aortic valve procedures (M), whereas two other surgeons performed 40 aortic valve procedures through a conventional sternotomy (C). Results: More time consuming and technically demanding surgeries were done in M. There was one death in each group. Aortic clamp time was longer in M (93 ± 40 vs 59 ± 24 min, P = 0.001). There were no differences in operating time, pump time, intubation duration, bleeding and intensive care unit stay. The advantages of minimally invasive aortic valve operation included better postoperative ejection fraction (58 ± 17 vs 51 ± 10%, P = 0.04), decreased pain score (3 ± 2 vs 5 ± 2, P = 0.004), less transfusion (19 vs 55%, P = 0.02), shorter duration of chest tube drainage, and cosmetically more acceptable surgical wound (6.8 ± 2.2 vs 5.2 ± 2.0, P = 0.018). From our series, we could not find any negative effects of minimal access surgery. Conclusions: Our study demonstrated that aortic valve surgeries could be performed routinely by the minimally invasive approach with a high degree of effectiveness and safety.  2000 The International Society for Cardiovascular Surgery. Published by Elsevier Science Ltd. All rights reserved Keywords: upper sternotomy, transverse sternotomy, aortic valve replacement, minimally invasive surgery, minimally invasive valve surgery, aortic valve surgery

The best approach for most cardiac operations has been by the median sternotomy with a 20 cm skin incision, since the beginning. It allows wide exposure of the heart and the origin of the great vessels, but the patients sometimes pay off for this quality of exposure with postoperative morbidity [1]. Since the first reports of minimally invasive aortic valve procedures (M), there has been increasing interest in the development of small incision valve surgery [2, 3]. The damaging effects of extracorporeal circulation Correspondence to: Jae-Won Lee. Tel: + 82-2-224-3584; fax: + 822224-6966; E-mail: [email protected]

484

were very clearly understood by most cardiac surgeons and cardiologists, but the benefits of a small incision in the immediate postoperative period were more difficult to verify scientifically [4, 5]. Moreover, the surgical problems of operating through a small incision has the potential to be critical, especially in the early learning period [6, 7]. Due to such shortcomings, these approaches in valve surgery have failed to become widely accepted and established for routine practice [8–11]. To assess the value of these relatively new surgical approaches, we analyzed the results of 46 consecutive cases of minimally invasive aortic valve procedures and compared the results with a group of 40 CARDIOVASCULAR SURGERY

OCTOBER 2000 VOL 8 NO 6

Routine minimally invasive aortic valve procedures: Jae-Won Lee et al.

patients that underwent aortic valve surgery through conventional median sternotomy (C). Our goals were to determine (1) whether any aortic valve disease can be treated routinely by the minimally invasive approach and (2) to see if there were any differences in the intraoperative and postoperative outcomes between the two groups.

Methods Between September 1997 and February 1999, a total of 46 consecutive patients with aortic valve disease received minimally invasive aortic valve procedures (Group M). Surgery was performed by one surgeon in all of these patients whereas two other surgeons performed aortic valve replacements by full sternotomy during the same period (Group C). In group M, transverse sternotomy was performed through the 3rd or 4th intercostal space in nine patients who were older than 60 yr of age with normal coronary angiograms while upper sternotomy was performed in the remaining 37 patients. Patients with aortic dissection and concomitant procedures such as mitral valve procedures and coronary bypass grafting were excluded from the study. One patient requiring double valve replacement, and two patients requiring patch closure of subarterial VSD with AR, and another with sinus of Valsalva rupture were included in group M. The amount of postoperative bleeding was evaluated on a 24-h basis twice. To evaluate the neurologic status, each patient had one or more neurologic examinations beginning on the 3rd postoperative hour and daily thereafter until stable. The decision to transfer to the general ward was based on the availability of beds in the ward. All patients had routine echocardiography 7–10 days, 6 months and one year postoperatively. Discharge was based on the patient’s family consent and on adequacy of anticoagulant regulation but, not on the patient’s subjective sense of well being. We did not adopt any fast track policy. The amount and number of blood transfusions were calculated after termination of cardiopulmonary bypass in order to determine the effect of bleeding and transfusion by the approach adopted. Pain level was evaluated 2 weeks after discharge with a visual analog scale (1–10)[12]. The patients were also asked to subjectively evaluate the cosmetic effect of the incision (using a numerical visual scale of 1–10). Operative technique Anesthesia was achieved with a combination of fentanyl–isoflurane–nocurone. Clinical monitoring routinely included ECG, arterial pressure, central venous pressure, pulse oxymetry, and capnography, but TEE and pulmonary arterial catheterization were done only in selected cases. We routinely applied CARDIOVASCULAR SURGERY

OCTOBER 2000 VOL 8 NO 6

external defibrillator patches in preparation of possible difficulties with weaning that may require defibrillation. Upper sternotomy A midline skin incision of about 3 inches was made from the 2nd rib down to the level of the 3rd or 4th rib and deepened to the sternum. The sternum was split from the suprasternal notch to the 3rd intercostal space level. At the 3rd intercostal level, the sternum was then split transversely with a narrow oscillating saw. If an ascending aortic aneurysm was present, the sternum was transected one space lower at the 4th intercostal space level. No attempt was made to identify the internal thoracic artery. On the right half, the sternum was split completely, but on the left half, the sternum was only partially split allowing greenstick fracture to occur with later retraction. After opening the pericardium, the edges of the pericardium were sutured tightly to the external tables of both sternal edges. A small pediatric sternal retractor was then applied and the pericardial edges were spread together with the sternum. The sternum was spread to a width of about 5 cm. This important maneuver resulted in raising the cardiac structures from the pericardial cradle thereby bringing them closer to the surgeon’s view. Cannulations of the aortic arch and right atrial appendage were routinely done. If possible, preoperative autologous blood donation was done just prior to institution of cardiopulmonary bypass employing hemo-dilution with perfusion at a tepid temperature of 35°C. Only one pint was collected. The aorta was cross clamped with full venous drainage. After aortic cross clamping, the heart was emptied with further drainage of venous blood, and the ascending aorta was transected 1cm above the right coronary ostium. The distal aorta was retracted with retention stitch toward the right side of the wound, and the three commissures of the aortic valve were retracted radially to bring the leaflets comfortably into the operating field. Blood cardioplegic solution (1000– 1500 cc), which was cold and high in potassium, was initially infused directly into both coronary ostia with Polystan self-inflating cannula for rapid induction of diastolic arrest. The ostial cannulae were fixed to the aortic wall with fine sutures. Afterwards, 500 cc of warm blood cardioplegic solution was infused every 20 min or whenever needed. After the main operation, low potassium warm blood cardioplegic solution was infused during aortic closure for at least 3minutes. Immediately after this last dose of cardioplegia warm blood was reperfused through the same cardioplegic catheters, still in place until the aortic closure was nearly complete. The heart usually started beating immediately. After regaining a vigorous heartbeat and adequate deairing through the aortotomy, a vent cannula was placed on the ascend485

Routine minimally invasive aortic valve procedures: Jae-Won Lee et al.

ing aorta for further deairing. An obligatory 10% of the arterial input was returned through this vent. With vent suctioning and removal of blood and air from the still cross-clamped ascending aorta, the aortic cross clamp was removed without any difficulties. During the early learning curve period, we had some difficulties with reperfusion ventricular fibrillation — reportedly a very troublesome problem with minimal incisional surgery [13, 14]. The causes of this problem are thought to be due to the massive hypertrophy in these hearts, poor myocardial protection, and distention of the left ventricle before beating. In such cases we usually opened the pulmonary artery and fully vented the left ventricle. Secondary cardioplegia was instituted and magnesium and lidocaine were infused intravenously. In the end if fibrillation persisted external defibrillation was performed, but these procedures were usually sufficient to induce sinus rhythm.

Table 1 Preoperative clinical data*

Patients M/F Age EF(%)

M

C

46 31/15 50.1 ± 13.9 (16–73) 56 ± 15 (32–81)

40 26/14 48.9 ± 15.1 (13–76) 56 ± 12 (25–72)

Valve gradient (mm Hg) 80 ± 42 PPG† MPG‡ 49 ± 27

77 ± 33 49 ± 22

NYHA functional class II 25 III 17 IV 4

21 15 4

12.8 ± 2.7 38.3 ± 7.9

Hb(mg%) Hct(%)

12.7 ± 3.3 38.2 ± 9.9

Transverse sternotomy A transverse 3 inch skin incision was made over the 3rd or 4th intercostal space, and deepened to the sternum. After dissection and careful ligation of both internal thoracic arteries and veins, the sternum was transected. The rib spreader was then inserted between the two separated ends of the sternum. The natural course of the ribs resulted in the elevation of the upper sternal plate and depression of the lower sternum by this maneuver, providing a comfortable field of vision even though the level of incision was significantly lower than those typically reported in the literature [15]. The pericardium was opened longitudinally. The cannulation and operative techniques were not different from the methods employed in the upper sternotomy. Full sternotomy After routine full sternotomy, in which the skin incision from the level of the second rib to the tip of the xiphoid process, and cannulation of the aorta and right auricle are included, the route and method of cardioplegic delivery were decided by the operating surgeon. Both surgeons in this group used intermittent antegrade cardioplegia with a soft hand held coronary ostial cannula. There were two surgeons in this group. One surgeon used cold intermittent antegrade blood cardioplegia while the other surgeon used warm intermittent antegrade cardioplegia. Valve fixation In the minimal group, mechanical valve prostheses were inserted with multiple interrupted multifilament stitches (as many as the size of the prostheses in mm), and bioprostheses were inserted with 12 pledgetted mattress sutures in which the pledgets 486

*No statistically significant difference between the two groups were found; † PPG, peak pressure gradient; ‡ MPG, mean pressure gradient

were placed in the ventricular side. In the conventional group, all the prostheses were fixed with 12 pledgetted everting mattress sutures. Data analysis Data are presented as mean ± standard deviation with range. Statistical analysis was performed using the Student’s unpaired t-test. For comparison of qualitative variables, the χ2 test was applied. A Pvalue of 0.05 or less was considered significant.

Results There were no differences in patient demographics (Table 1). Patients with active endocarditis were included in group M (Table 2). The procedures perTable 2 Preoperative diagnosis Pathophysiology Aortic stenosis Aortic regurgitation Mixed type

(N = 46) 12 22 12

(N = 40) 10 24 16

Etiology Rheumatic Bicuspid Endocarditis Degenerative Annuloaortic ectasia Takayasu’s disease Valsalva Sinus Rupture Congenital VSD, AR Quadricuspid Subaortic stenosis

(N = 46) 20 8 8 5 2 1 1 1 1

(N = 40) 19 5 1 8 2 3

CARDIOVASCULAR SURGERY

1

OCTOBER 2000 VOL 8 NO 6

Routine minimally invasive aortic valve procedures: Jae-Won Lee et al.

formed in group M included AVR with the St Jude Medical valves (N = 29), Carpentier Edward perimount pericardial valves (N = 8), homograft root replacements (N = 4) and insertion of a SJM valved conduit (N = 1) as free standing root replacement device, one aortic valve repair, and three David procedures. The six cases of annulus enlarging procedures were done at the surgeon’s discretion in group M (Table 3). One case of AVR in which aortic valve repair was done 4 years previously through a conventional incision was included in the M group. In the C group, there were 39 cases of aortic valve replacements and one aortic valve repair. Associated procedures and type of incisional approach are listed in Table 3. Operative mortality One patient died in each group. In group C, an elderly female whose aortic valve was replaced with a 19-mm tissue valve died intraoperatively of cardiogenic shock, probably from poor myocardial protection. In group M, one woman died of septic shock which was present preoperatively that later developed into acute respiratory distress syndrome 5 days postoperatively despite an uneventful initial recovery. Operative times (Table 4) Operating time and total bypass time showed no difference. The mean aortic clamp time was prolonged in group M (93 ± 40 vs 59 ± 24 min, P = 0.001). Comparing the mean aortic clamp time, 79 ± Table 3 Surgical procedures M

C

St Jude Medical valve St Jude Medical valved conduit Carpentier-Edward tissue valve Homograft root replacement David procedure Valve repair

29 1 8 4 3 1

33

1

Associated procedures Annular enlargement VSD patch closure MVP MVR LVOT myectomy

4 2 1 1 1

1

6

33 min, of isolated aortic valve replacements alone in group M with group C still showed significantly longer clamp times in group M (P = 0.005). This difference may be attributed to the difference in suture technique as AVR in group M was performed with multiple interrupted sutures where as in group C, AVR was performed with 12 pledgetted everting mattress sutures. Postoperative course Echocardiogram performed within 7–10 days postoperatively showed a significantly higher ejection fraction in group M as compared to group C, 59 + 17 vs 59 + 10 despite the absence of a significant difference in the maximal rise in the MB fraction of CK enzyme (25.2 + 19.9 vs 22.4 + 22.1) between the two groups. Intubation duration was same. Predonation of 1 pint of autolgous blood was performed in all of the patints in group M, and in every case this blood was reinfused upon termination of bypass. Further autologous transfusion was performed in the form of reinfusing shed mediastinal blood if this amount exceeded 200 cc in the fist 5 h postoperatively. Any significant bleeding beyond this point were not recycled. The change in the mean hemoglobin levels in group M and group C were, 12.8 ± 2.7 to 11.8 ± 3.7 and 12.7 ± 3.3 to 11.5 ± 3.0 respectively (Tables 1 and 4) During the first postoperative 24 h, there were no differences in the amount of tube drainage, but nine (19%) patients in M and 22 (55%) in group C required transfusion (P = 0.02). The duration of tube drainage was also shorter in group M. The amount of transfusion per patient was 1.3 ± 1.3 vs 1.1 ± 2.3 units. By definition, infusion of autologous blood was not considered transfusion of blood but only homologous blood. One patient from each group required reoperation for bleeding control. Because we made no efforts to adopt a ‘fast tract’ policy, the mean length of stay in the intensive care unit and total hospital stay were the same for both groups. There was no case of switch over to median sternotomy from minimal incision. There was no sternal wound complication in M, but there was one wound infection in group C. Another patient in group C required reoperation for paravalvular leak. Two patients in group M and one in group C had pericardial effusion more than 10 mm on echocardiogram, but none required any invasive treatment. Postoperative pain and cosmetic evaluation

Incision Upper sternotomy Transverse sternotomy

37 9

Redo case

1(AVP)

CARDIOVASCULAR SURGERY

1(VSD)

OCTOBER 2000 VOL 8 NO 6

Evaluation of pain by visual analog scale allowed numerical calculation of the degree of pain. Average pain scores were significantly less in M (3 ± 2 vs 5 ± 2, P = 0.004). Degree of esthetic satisfaction by visual scale was also significantly greater in M (6.8 ± 2.2 vs 5.2 ± 2.0, P = 0.018). 487

Routine minimally invasive aortic valve procedures: Jae-Won Lee et al. Table 4 Postoperative data

Duration skin to skin (min) Total bypass time (min) Aortic clamp time (min) CK-MB (IU) Post operative EF (%) Duration of intubation (postoperative, h) Hb (mg%) Hct (%) Duration of tube drainage (h)

Postoperative bleeding (cc) for 24 h after POD 24 h

Patients transfused (%) Mean blood requirement Patients needing analgesia after POD3 Pain (visual scale 1–10) Esthetic satisfaction (1–10)

M (n = 46)

C(n = 40)

P value

259 ± 56(165–430) 119 ± 55(59–298) 93 ± 40(48–219) 25.2 ± 19.9(6.1–102.4) 59 ± 17(19–80) 12 ± 8(3–48)

244 ± 60(140–400) 93 ± 40(43–190) 59 ± 24(34–140) 22.4 ± 22.1(5.1–87.2) 51 ± 10(26–79) 13 ± 5(4–29)

ns ns 0.001 ns 0.04 ns

11.8 ± 3.7 33.0 ± 5.3 31 ± 14(14–82)

11.5 ± 3.0 34.1 ± 9.0 67 ± 35(24–179)

ns ns ⬍ 0.001

295 ± 388(80–1990) 179 ± 300(0–1640)

324 ± 237(90–600) 251 ± 388(0–870)

ns ns

19% 1.1 ± 2.3 units 14(30%)

55% 1.3 ± 1.3 units 16(40%)

0.02 ns ns

3±2 6±2

5±2 5±2

0.004 0.018

Discussion As modern cardiac surgery improved the prognosis of most types of cardiac operations, application of less invasive approaches and enhancement of esthetic results have become important issues [1, 16]. However, the definition of what is ‘minimally invasive’ remains unsettled and is often focused on the size and location of the incision [1]. Some of the approaches involve routine resection of ribs or cartilages, opening of pleural spaces, and peripheral cannulation for extracorporeal circulation. Each option should be considered as a trade-off between the exposure provided, the ease and speed of the operation, the easy and rapid conversion to standard sternotomy whenever necessary, and postoperative recovery. There are still different opinions regarding which incision is best for aortic valve procedures. The optimal incision must be centered on the part of the heart in which the operation takes place. The aortic valve and ascending aorta are located immediately behind the upper sternum, and the line of vision to the aortic valve is from the upper to downward direction from the right shoulder, with both incision types [9]. The only benefit of splitting the lower sternum is to see the ventricle under direct vision in the event of massage or defibrillation. That is why we chose the upper sternotomy as a routine incision in aortic valve surgery. The only drawback to this incision is that an upper anterior chest wound is unavoidable, becoming visible over clothing. With transverse ster488

notomy at the 3rd or 4th intercostal space level, we can secure the same operative field of vision, without an upper anterior chest scar. Unfortunately both internal thoracic arteries have to be cut with this incision. So, the indication for this incision is confined to elderly patients with normal coronary angiographic findings. Although there have been many reports of minimally invasive cardiac surgery, the main advantages of these procedures are said to be the avoidance of cardiopulmonary bypass and subsequent complications of extracorporeal circulation, such as cerebrovascular accidents and coagulopathy [4]. In this respect, minimally invasive valve surgery may be considered more invasive than minimally invasive coronary surgery. However, the promise of minimally invasive heart surgery continues to fascinate patients, particularly when it is heard over the mass media. The advantages of these small incisions should include reduction in surgical trauma and pain, bleeding and transfusion, shortened hospital stay, and a faster mobilization combined with cost reduction. In the current study, we found that there was less pain and bleeding with minimal incisional surgery. The difference in the bleeding between the two groups however, in itself would not have been sufficient to account for the difference in the transfusion requirement which was found to be significantly less in group M compared to group C. This discrepancy may be attributable to the different level of tolerance to bleeding by each surgeon before makCARDIOVASCULAR SURGERY

OCTOBER 2000 VOL 8 NO 6

Routine minimally invasive aortic valve procedures: Jae-Won Lee et al.

ing a decision to transfuse. Furthermore, in group M, autologous transfusion was more actively employed. As a result all the patients in group M predonated one pint of blood just prior to institution of cardiopulmonary bypass and this blood was reinfused upon termination of bypass. In the ICU, shed mediastinal blood within the first five hours in excess of 200 cc were reinfused to the patient. Any bleeding beyond this point were not reinfused. These measures were not adopted by the surgeons in group C. Therefore, when transfusion was indicated, homologous blood was given as needed. Through these adjunctive measures, it was possible to reduce the amount of bleeding in group M. However, the point is despite the more difficult and complex nature of the cases being performed in group M this did not lead to a greater incidence of significant bleeding compared to group C, but rather a slightly less amount or nearly the same amount of bleeding. Therefore, as far as bleeding was a concern we did not find minimally invasive procedure to be a separate risk. Less importantly or perhaps just as important, the operation scar was cosmetically more appealing. There were no differences in postoperative complications (Table 5, albeit the more technically demanding surgical procedures performed in group M. It is possible that the risk of surgery associated with limited access surgery is exaggerated to the extent that it is simply taken for granted that these operations, which actually involve low risk, are just too dangerous [5]. Considerable skepticism thus remains regarding the value of minimally invasive valve surgery, as they are currently considered to be specialized procedures that should be set aside for very highly selected group of patients [5, 11, 17]. We wanted to see whether aortic valve procedures could be done through minimally invasive surgery on a routine basis. In our 18 months’ experience of routinely performing 46 non randomized consecutive minimally invasive aortic valve procedures, we found no increase in mortality, lower morbidity, lower complications, less need for transfusion, shorter duration of tube drainage, no significant lengthening of

Table 5 Complications*

Pericardial effusion ( > 1 cm) Reoperation for bleeding Wound infection Reoperation for paravalvular leak

Mortality Cardiogenic shock ARDS (preop sepsis)

CARDIOVASCULAR SURGERY

M

C

2 1

1 1 1 1

1 1

OCTOBER 2000 VOL 8 NO 6

the pump and operation times. However, the aortic clamp time was definitely longer in group M. This increase in aortic clamp time was clearly present in group M compared to group C even in the patients that underwent simple AVR. However, the resultant prolongation of global ischemic time of about 20 min had no deleterious effect on patient safety. The postoperative ejection fraction was significantly higher in M, there was no case of low cardiac output and the rise in postoperative CKMB was unremarkable even though the operations in group M were more difficult and time consuming. Olinger [10] proposed three issues to address for minimally invasive valve surgery. (1) Security of valve surgery, (2) Complications of perfusion and myocardial protection, and (3) Incisional morbidity. These issues were properly addressed in our series. (1) There was no case of reoperation, giving support to the quality of valve surgery. (2) There was no case of cardiac dysfunction or cerebrovascular damage. (3) There was no increase in bleeding, patient satisfaction in terms of both pain sensation and esthetic appeal were greater, and the need for transfusion in group M was decreased. Although the duration of tube drainage was significantly shorter in group M (31 ± 14 vs 67 ± 35 h), the difference in the amount of bleeding between the two groups failed to reach statistical significance. This was due to the profuse postoperative bleeding in a few of the more complex surgical cases in group M. Furthermore, all of the complex aortic valve procedures were included in group M. Therefore, in consideration of these facts, lesser bleeding can safely be expected to occur in the minimally invasive group. The limitations of this study lie in that the minimally invasive aortic surgeries were performed by only one surgeon rather than by all three surgeons. Therefore, it may be argued that the results reflect the comparison of the surgical results of the different individual surgeons rather than those of the different incisional approaches. Another limitation is that there was no randomization of patients in the group receiving minimally invasive surgery. Despite these purported limitations, they do not negate nor disprove the hypothesis that any valve procedure confined to the aortic valve can be performed through the minimally invasive approach — with safety. By performing these procedures consecutively, it was shown that they could be performed on a routine basis. Out of moral and ethical considerations, all patients assigned to the surgeon in group M were treated without randomization to either group C or group M. In conclusion, our study demonstrated the practicability and reliability of performing this new method on a routine basis, which will surely become an established method for aortic valve procedures. As an indication, we started performing the Ross procedure successfully through the 489

Routine minimally invasive aortic valve procedures: Jae-Won Lee et al.

minimally incisional approach after the current study period.

References 1. Massetti, M., Babatasi, G., Lotti, A., Bhoyroo, S. et al., Less invasive cardiac operations through a median sternotomy: 100 consecutive cases. Annals of Thoracic Surgery, 1998, 66, 1050– 1054. 2. Cosgrove III, D.M. and Sabik, J.F., Minimally invasive approach for aortic valve operations. Annals of Thoracic Surgery, 1996, 62, 596–597. 3. Konertz, W., Waldemberger, F., Schmutzler, M., Ritter, J. and Liu, J., Minimal access valve surgery through superior partial sternotomy: a preliminary study. Journal of Heart Valve Disease, 1996, 5, 638–640. 4. Westaby, S. and Benetti, F.J., Less invasive coronary surgery: consensus from Oxford meeting. Annals of Thoracic Surgery, 1996, 62, 924–931. 5. Aris, A., Camara, M.L., Montiel, J., Delgado, L.J. et al., Ministernotomy versus median sternotomy for aortic valve replacement: a prospective, randomized study. Annals of Thoracic Surgery, 1999, 67, 1643–1647. 6. Walther, T., Falk, V., Metz, S., Diegeler, A. et al., Pain and quality of life after minimally invasive versus conventional cardiac surgery. Annals of Thoracic Surgery, 1999, 67, 1643–1647. 7. Riess, F.C., Lower, C. and Bleese, N., Prevention of potential complications after minimal access aortic valve replacement. Annals of Thoracic Surgery, 1998, 66, 1866.

490

8. Lytle, B.W., Minimally invasive cardiac surgery. Journal of Thoracic and Cardiovascular Surgery, 1996, 111, 554–555. 9. Shennib, H. and Mack, M.J., Facts and myths of minimally invasive cardiac surgery: current trends in thoracic surgery IV. Annals of Thoracic Surgery, 1998, 66, 995–1120. 10. Olinger, G.N., Informed advice. Annals of Thoracic Surgery, 1999, 67, 1545–1546. 11. Cooley, D.A., Minimally invasive valve surgery versus the conventional approach. Annals of Thoracic Surgery, 1998, 66, 1101–1105. 12. Sriwatanakul, K., Kelvie, R. and Lasagna, L., Studies with different types of visual analogue scales for measurement of pain. Clinical Pharmacology Therapeutics, 1983, 34, 234–239. 13. Holman, W.L., Spruell, R.D., Vicente, W.V.A. and Pacifico, A.D., Electrophysiologic mechanisms for postcardioplegia reperfusion ventricular fibrillation. Circulation, 1994, 90(Pt II), 293– 298. 14. Holman, W.L., Vicente, W.V.A., Spruell, R.D., Digerness, S.B. and Pacifico, A.D., Effect of postcardioplegic reperfusion rhythm on myocardial blood flow. Annals of Thoracic Surgery, 1994, 58, 351–358. 15. Cosgrove III, D.M., Sabik, J.F. and Navia, J., Minimally invasive valve operations. Annals of Thoracic Surgery, 1998, 65, 1535– 1539. 16. Gundry, S.R., Shattuck, O.H., Razzouk, A.J., Del Rio, M.J. et al., Facile minimally invasive cardiac surgery via ministernotomy. Annals of Thoracic Surgery, 1998, 65, 1100–1104. 17. Verrier, E.D., Editorial (pro) re minimally invasive port-access mitral valve surgery. Journal of Thoracic and Cardiovascular Surgery, 1998, 115, 565–566. Paper accepted 9 June 2000

CARDIOVASCULAR SURGERY

OCTOBER 2000 VOL 8 NO 6