Ann Thorac Surg 2003;75:1422-1428
© 2003 The Society of Thoracic Surgeons
Original article: cardiovascular
Supraclavicular control of patent internal thoracic artery graft flow during aortic valve replacement
Erkan Kuralay, MDa*,
Faruk Cingöz, MDa,
Celalettin Günay, MDa,
Bilgehan Sava
Öz, MDa,
Nezihi Küçükarslan, MDa,
Vedat Yildirim, MDb,
S.Yavuz Sanisoglu, MDc,
Ertu
rul Özal, MDa,
Ufuk Demirkiliç, MDa,a,a,a,
Mehmet Arslan, MDa,
Harun Tatar, MDa
a Departments of Cardiovascular Surgery, Gülhane Military Medical Academy, Etl
k, Ankara, Turkey
b Anesthesiology, Gülhane Military Medical Academy, Etlk, Ankara, Turkey
c Biostatistics, Gülhane Military Medical Academy, Etlk, Ankara, Turkey
Accepted for publication December 12, 2002.
* Address reprint requests to Dr Kuralay, Yazanlar sokak No = 31, 11, Asagi Ayranci, Ankara, Turkey 06540
e-mail: ekural{at}gata.edu.tr
 |
Abstract
|
|---|
BACKGROUND: The mortality and morbidity of aortic valve replacement (AVR) after prior coronary artery bypass surgery (CABG) with patent left internal thoracic artery (LITA) is significant. The risk of LITA injury and inadequate myocardial preservation during the cross-clamp period may cause myocardial pump failure.
METHODS: A total of 43 patients with a patent LITA graft underwent AVR. The patients were divided into the two groups. Group 1 included 19 patients who underwent AVR with deep hypothermia (20°C) without LITA clamping. Group 2 included 24 patients in whom LITA flow was controlled through supraclavicular occlusion and AVR performed with moderate hypothermia (28°C).
RESULTS: Average cardiopulmonary bypass time (CPB) time was 118.79 ± 20.36 minutes in group 1 and 102.67 ± 9.66 minutes in group 2 (p = 0.006). Average cross-clamp time was 53.79 ± 7.26 minutes in group 1 and 49.63 ± 6.7 minutes in group 2 (p = 0.022). Inotropic support was required in 12 patients in group 1 and 4 patients in group 2 (p = 0.002). Average intensive care unit stay was 4.68 ± 2.24 days in group 1 and 2.29 ± 0.46 days in group 2 (p < 0.001). Average hospital stay was 11.84 ± 2.91 days in group 1 and 8.04 ± 2.38 days in group 2 (p < 0.001). Mortality due to myocardial failure developed in 4 patients in group 1 but in none of the patients in group 2 (p = 0.02).
CONCLUSIONS: Proximal control of LITA flow by extrathoracic supraclavicular occlusion reduces the incidence of myocardial failure due to nonhomogenous cardioplegia delivery to the anterior wall of the heart, resulting in improved myocardial protection and the elimination of the need for deep hypothermia.
|
Introduction
|
|---|
Aortic valve replacement (AVR) after coronary artery bypass grafting (CABG) can be extremely difficult technically when the left internal thoracic artery (LITA) is patent. Repeat sternotomy, the handling of the patent vein or LITA grafts, and the need to relocate sites for aortic perfusion, aortic cross-clamping, cardioplegia administration, and aortotomy because of previously placed bypass grafts all add technical complexity to the procedure and potentially increase the perioperative risk [1–6].
The patent LITA can result in inadequate myocardial protection of the dependent myocardium (anteroseptal region) due to cardioplegia washout. Various approaches to myocardial protection designed to circumvent this problem have been proposed. These include total circulatory arrest, performing the operation without occluding LITA flow in deep hypothermia [7], and intrathoracic occlusion of the patent LITA. In our study, the patent LITA flow was controlled using a soft bulldog clamp applied from a supraclavicular approach to the proximal LITA in one group of patients group (group 2). AVR was then performed utilizing moderate (28°C) hypothermia. The control patients (group 1) had AVR performed with patent LITA and deep hypothermia. Pertinent perioperative morbidity and mortality in these groups are compared.
|
Material and methods
|
|---|
Patients
A total 43 consecutive post-CABG patients with a patent pedicle LITA graft underwent elective AVR at the Gülhane Military Medical Academy Cardiovascular Surgery Department between June 1991 and May 2002. Patients were divided into two Groups. Group 1 included 19 patients who underwent AVR with moderate-to-deep hypothermia (20°C), aortic cross-clamping, retrograde cardioplegia, and an open uncontrolled LITA graft. Group 2 included 24 patients who underwent AVR with moderate hypothermia (28°C), aortic cross-clamping, retrograde cardioplegia, and proximal control of LITA graft flow. The LITA was exposed by a supraclavicular approach just before sternotomy and the proximal LITA occluded with a soft bulldog clamp just before aortic cross-clamping. Group 1 patients were operated upon between June 1991 and March 1997, and group 2 patients since March 1997.
The interval between CABG and AVR was an average of 8.32 ± 4.09 years (range, 1.3 and 17 years) in group 1 and an average of 10.37 ± 4.5 years (range, 3.1 and 21 years) (p = 0.210) in group 2. The average patient age was 65.89 ± 6.0 (range, 55 to 81 years) in group 1 and 66.21 ± 6.28 (range, 49 to 84 years) in group 2 (p = 0.980). There were 4 patients greater than 70 years of age in group 1 and 6 patients greater than 70 years of age in group 2. There were 3 patients less than 60 years of age in each group.
Average aortic valve systolic gradient was 64.95 ± 9.82 mm Hg in group 1 and 62.29 ± 8.51 mm Hg in group 2 (p = 0.698). The predominant pathology was aortic stenosis in both groups. Severe aortic regurgitation was found in only 1 patient in group 1 (who underwent prior bioprosthesis implantation) and in 2 patients in group 2. Ejection fraction (EF) was 42.68% ± 7.8% in group 1 patients and 44% ± 6.8% in group 2 patients (p = 0.538). Patient demographics are summarized in Table 1.
Surgical procedures
Group 1
A full sternotomy incision was utilized in all patients. Median sternotomy and mediastinal dissection was accomplished before full systemic heparinization. Ascending aortic and right atrial cannulation were utilized in 14 patients. Femoral artery cannulation was used for cardiopulmonary bypass (CPB) in 5 patients before median sternotomy. Left internal thoracic artery graft clips were noted to cross the midline on chest roentgenogram in 3 of these patients, and the LITA graft was adjacent to the posterior table of the sternum in the other 2. Moderate to deep hypothermia (20°C) was utilized. Aortic valve replacement was accomplished with 12 St. Jude bioprosthesis and seven Sorin mechanical aortic valve prostheses. Cardiopulmonary bypass flow was reduced when excessive blood flow was detected through the coronary ostium. Blood pressure was controlled by metaraminol in this period. Right coronary artery bypass was done in 5 patients. Obtuse marginal bypass grafts were done in 3 patients, with minimal dissection to avoid LITA injury. The left anterior descending artery (LAD) was bypassed in 2 patients who had a LITA injury on sternal entry. The heart and LITA were dissected using the technique described by Elami and coworkers [8]. Proximal anastomoses were done under a single cross-clamp. De-airing was performed with a long needle inserted into the left ventricle through the right ventricle and interventricular septum when the left side of heart was not freed. Transesophageal echocardiographic (TEE) control of de-airing was used in all patients.
Group 2
Extrathoracic proximal control of LITA flow was used in all patients just before sternal reentry. The patient was placed in supine position with the shoulders elevated. The head was turned to the right side and an 8- to 10-cm skin incision was made. The left subclavian artery, the vertebral artery, and the LITA were exposed as described by Haimovici [9]. The subclavian artery was controlled both proximal and distal to the LITA and vertebral artery. The bag under the left shoulder was removed and sternal reentry accomplished (Fig 1)
View larger version (28K):
[in this window]
[in a new window]
|
Fig 1. (A) Supraclavicular skin incision. Patient’s head turned to right with the shoulders elevated with pad. The skin incision is made 1 cm above the clavicle and extends from the sternoclavicular joint to the lateral portion of the supraclavicular region for about 8 to 10 cm. (B) Extrathoracic proximal control of LITA. Dissection of the deep cervical layer was completed. Anterior scalene muscle was divided. Both LITA and vertebral artery were identified. LITA was clamped just before aortic cross-clamping. (LITA = left internal thoracic artery.).
|
|
Ascending aorta and right atrial cannulation were used in 14 patients. Femoral artery cannulation and institution of cardiopulmonary bypass (CPB) were used in 8 patients. LITA graft clips were noted to cross the sternal midline on chest roentgenogram in 3 of these patients, and the LITA graft was adjacent to the posterior sternum in 5 of these patients. Right atrial venous cannulation was used in all patients after mediastinal dissection was completed. Arterial cannulation was done on the left subclavian artery in 1 patient with severe peripheral vascular disease using a Dacron tube graft. Moderate hypothermia (28°C) was used in all patients. Just before applying the aortic cross-clamp, the proximal LITA was occluded with a soft bulldog clamp. AVR was performed with 17 St. Jude bioprostheses and seven Sorin mechanical aortic valve prostheses. Right coronary artery bypass was necessary in 6 patients, and obtuse marginal grafting was done in 4 patients. The LAD was bypassed in 1 patient. Proximal anastomoses were performed under a single cross-clamp. De-airing was done as in group1.
Myocardial preservation
A balloon-tipped retrograde cannula was inserted through the right atrium into the coronary sinus in all patients. Both induction cardioplegia (St. Thomas II solution) and repeated doses of cold blood and terminal "hot shot" blood cardioplegia were delivered retrograde in both groups. Selective right coronary ostia or coronary grafts were perfused directly to protect right ventricular performance. Eleven patients (5 patients in group 1and 6 patients in group 2) had either severe right coronary native disease or graft occlusion (3 patients) precluding direct perfusion.
Statistical analyses
Statistical analyses were performed by using SPSS software version 10.0 (SPSS for Windows; SPSS Inc, Chicago, IL) statistical package. Means ± standard deviation notation were used for continuous variables. Two-group comparisons were made by using Mann-Whitney U test. The effects of the variables on mortality were investigated by using the univariate and multivariate logistic regression analyses. Spearman rank correlation procedure was used for the correlations between variables and hospital mortality.
|
Results
|
|---|
Hospital mortality occured in 5 patients in group 1 and 1 patient in group 2. Myocardial failure was the principal cause of mortality in 4 patients in group 1, but mortality due to myocardial failure did not occur in group 2 patients (p = 0.02). Multisystem organ failure (MSOF) developed in 1 patient in both groups (p = 0.867). Average CPB time was 118.79 ± 20.36 minutes in group 1 and 102.67 ± 9.64 minutes in group 2 (p = 0.006). Average cross-clamp time was 53.79 ± 7.26 minutes in group 1 and 49.63 ± 6.70 in group 2 (p = 0.022). The operative data for both patient groups are summarized in Table 2.
The LITA was injured in 3 patients during sternal reentry. Saphenous vein bypass to the LAD was performed in these 3 patients. The right ventricle was injured in 3 patients. Small tears in the right ventricle were repaired using pletgetted sutures in these 3 patients. Right atrial lacerations occured in 2 patients, and these were repaired easily using Prolene sutures.
Three patients in group 1 and 1 patient in group 2 (p = 0.198) sustained perioperative myocardial infarction. This was defined as a new Q wave in the electrocardiogram, creatinine kinase-MB (CK-MB) equal or greater than 50 IU/L with a CK-MB/CK ratio greater than 5%, or a new wall motion abnormality on echocardiography. Renal failure requiring dialysis developed in 1 patient. This patient also required long-term mechanical ventilation due to respiratory insufficiency necessitating a tracheostomy. The patient died due to MSOF on postoperative day 27. Respiratory insufficiency in group 2 developed in 1 patient who required tracheostomy. Sepsis also developed in this patient. The patient died due to MSOF on postoperative day 35. Inotropy was required in 12 patients in group 1 and 4 patients in group 2 during weaning from CPB (p = 0.002).
Myocardial pump failure, defined as postoperative need for high doses of inotropic agents or the use of an intraaortic balloon pump, developed in 9 patients in group 1 and 3 patients in group 2 (p = 0.012). Three patients in group 1 and 1 patient in group 2 required delayed sternal closure to improve early postoperative hemodynamics. One patient required delayed sternal closure to accommodate diffuse postoperative bleeding. Two patients sustained transient ischemic attacks (TIA) and 3 patients sustained a cerebrovascular accident (CVA) in group 1. All of these patients showed steady improvement on discharge. Neurologic deficit did not develop in group 2 patients (p = 0.108 for TIA, p = 0.046 for CVA). Postoperative atrial fibrillation developed in 9 patients in group 1 and 13 patients in group 2 (p = 0.479). Average intensive care unit (ICU) stay was 4.68 ± 2.24 days in group 1 and 2.29 ± 0.46 days in group 2 (p < 0.001). Average hospital stay was 11.84 ± 2.91 days in group 1 and 8.04 ± 2.38 days in group 2 (p < 0.001). Postoperative variables are summarized in Table 3.
Univariate, multivariate, and Spearman’s correlation test results are shown in Table 4.
View this table:
[in this window]
[in a new window]
|
Table 4. Effects of Variables on Mortality Investigated by Univariate and Multivariate Logistic Regression Analyses
|
|
Predictors of hospital mortality
Data for both groups were combined, and 21 variables were subjected to statistical analysis as a predictor of hospital mortality. Cox-Snell R2 value of multivariate analyses was 0.750. As shown in Table 4, univariate analysis identified extrathoracic control of LITA flow, age, diabetes mellitus, hypertension, chronic obstructive pulmonary disease, ejection fraction, elapsed time between CABG and AVR, systolic aortic gradient, CPB time, cross-clamp period, inotropy requirement, myocardial pump failure, ICU stay, hospital stay, cardioplegia from right coronary ostium, and bypass to the right coronary artery as predictors of hospital mortality. Multivariate analysis confirmed all above variables as predictors of hospital mortality. There was a statistically significant correlation in Spearman rank correlation analysis between hospital mortality and extrathoracic control of patent LITA, cross-clamp period, inotropy requirement, myocardial pump failure, dialysis, ICU and hospital stay, and cardioplegia from the right coronary ostium.
|
Comment
|
|---|
AVR in those patients who previously underwent CABG with patent LITA poses a difficult challenge for the cardiovascular surgeon. These operations are particularly challenging, due to the risk of injury of patent saphenous vein grafts and, in particular, the LITA during the dissection of heart or during sternal reentry. Prevalence of LITA graft injury was cited as 9% by Odell and associates [6], 5.3% by Gillinov nd associates [10], and 5% by Byrne and associates [11] during reoperation. This has been associated with a mortality rate of up to 50% [4, 8, 10, 12, 13]. Gillinov [10] reported an 8.6% mortality and 40% perioperative myocardial infarction rate in patients with LITA injury despite LAD-supplemented saphenous vein graft. The incidence of LITA injury was 10.5% for group 1 and 4.2% for group 2. Bypass to LAD after LITA injury did not statistically affect mortality in our univariate (p = 0.564) and multivariate (p = 664) analysis.
Some studies indicate that the presence of a patent LITA to the LAD artery decreases operative mortality at coronary reoperation. The proposed mechanisms for this salutary effect include preserved anterior wall function and the absence of atherosclerotic embolization from the LITA graft [13–16]. However, a patent LITA graft also creates a specific technical challenge at reoperation, which is the delivery of cardioplegia to the LAD territory. A patent LITA may cause discrepancy in myocardial protection between the anterior myocardium supplied by the LITA, where cardioplegia is gradually washed out, and the remainder of the myocardium. The effectiveness of retrograde cardioplegia when the patent LITA is unclamped has also remained rather obscure, especially in the hypertrophied left ventricle.
There are three strategies for AVR in patients with a patent LITA. The commonly known technique involves clamping both the aorta and the LITA graft. This has the advantage of shorter CPB times and permits the use of moderate rather than deep hypothermia. The identification of the intrathoracic LITA is difficult, if not impossible, in most reoperations. Another strategy involves deep hypothermia and circulatory arrest without both aortic cross-clamping and LITA clamping. This procedure is frequently used in the presence of a porcelain aorta. Recently, the most frequently used strategy is the open LITA technique, which is favored by Byrne and associates [11, 17]. This strategy has the advantage of permitting minimally invasive incisions for isolated AVR and avoiding dissection of the LITA pedicle. A feature of this strategy is the temperature discrepancy of the myocardium at risk. While the myocardium is protected by cardioplegia (
4°C) cooler than CPB temperature (20°C), the LAD territory may approach CPB temperature due to cardioplegia washout by the patent LITA. Cardioplegia washout reduces myocardial protection; therefore, deep hypothermia is mandatory. Electrical activity frequently resumed in our group 1 patients, and the administration of additional retrograde cardioplegia was required to achieve cardiac quiescence. If blood returning from the left main coronary ostia (through the open LITA) obscures the operative field during the AVR, CPB flows can be temporarily decreased for placement of sutures. Reduced blood pressure may, however, cause neurologic complications in elderly patients despite deep hypothermia. In group 2 patients, deep hypothermia was not used and the mean arterial pressure was kept near normal, eliminating the occurrence of neurologic deficits. Deep hypothermia was used in group 1 patients, TIA developed in 2 patients (p = 0.108) and CVA (p = 0.046) occurred in 3 patients. Reduced pump flow during CPB circulation may have played a role in the development of neurologic complications in these patients.
Extrathoracic control of LITA flow has significant effect on hospital mortality both by univariate (p < 0.001) and multivariate (p < 0.001) analysis. Extrathoracic control of LITA flow significantly improved hospital mortality (p = 0.02) and resulted in decreased myocardial pump failure in group 2 patients. Uniformity of myocardial preservation may be achieved and pump failure is reduced with our approach. The main limitation of this study is the small number of patients in both groups. A larger patient population is necessary to be able to reach more definitive conclusions.
|
Acknowledgments
|
|---|
The authors thank Dr James Diehl for language editing this manuscript.
|
References
|
|---|
- Grover F.L., Hammermeister K.E., Burchfiel C. Initial report of the Veterans Administration Preoperative Risk Assessment Study for Cardiac Surgery. Ann Thorac Surg 1990;50:12-28.[Abstract]
- Sindhi R., Belisle J., Cleveland R.J., Diehl J.T. Patch aortotomy for aortic valve replacement after previous coronary artery bypass grafting. Ann Thorac Surg 1991;51:676-677.[Abstract]
- Antunes M.J. Coronary artery bypass and mild aortic stenosis: to replace or not to replace? An alternative to repair. J Heart Valve Dis 1994;3:235.[Medline]
- Collins J.J., Jr, Aranki S.F. Management of mild aortic stenosis during coronary artery bypass graft surgery. J Cardiac Surg 1994;9(Suppl 2):145-147.[Medline]
- Safi H.J., Ilipoulos D.C., Duff D.S. Repeat replacement of aortic bioprosthesis. Ann Thorac Surg 1995;59:1217-1219.[Abstract/Free Full Text]
- Odell J.A., Mullany C.J., Schaff H.V., Orszulak T.A., Daly R.C., Morris J.J. Aortic valve replacement after previous coronary artery bypass grafting. Ann Thorac Surg 1996;62:1424-1430.[Abstract/Free Full Text]
- Mills N.L., Everson C.T., Hockmuth D.R. Technical consideration for myocardial protection during the course of coronary artery bypass reoperation: the impact of functioning saphenous vein and internal mammary artery grafts. J Card Surg 1991;6:34-40.[Medline]
- Elami A, Laks H, Merin G. Technique for reoperative median sternotomy in the presence of a patent left internal mammary artery graft. J Card Surg 1994;19:123–7
- Haimovici H. The upper extremity. In: Haimovici H, ed. Vascular surgery: Principals and Techniques, 4th ed. Blackwell Science, Inc: Cambridge, MA, 1996:373–84
- Gillinov A.M., Casselman F.P., Lytle B.W., et al. Injury to a patent left internal thoracic artery grafts at coronary reoperation. Ann Thorac Surg 1999;67:382-386.[Abstract/Free Full Text]
- Byrne J.G., Karavas A.N., Filsoufi F., et al. Aortic valve surgery after previous coronary artery bypass grafting with functioning internal mammary artery grafts. Ann Thorac Surg 2002;73:779-784.[Abstract/Free Full Text]
- Ivert T.S., Ekestrom S., Peterffy A., Welti R. Coronary artery reoperations: early and late results in 101 patients. Scand J Thorac Cardiovasc Surg 1988;22:111-118.[Medline]
- Lytle B.W., McElroy D., McCarty P. Influence of arterial coronary bypass grafts on the mortality in coronary reoperations. J Thorac Cardiovasc Surg 1994;107:675-683.[Abstract/Free Full Text]
- Coltharp W.H., Decker M.D., Lea J.W., et al. Internal mammary artery graft at reoperation: risks, benefits, and methods of preservation. Ann Thorac Surg 1991;52:225-229.[Abstract]
- Velebit V., Christenson J.T., Maurice J., Simonet F., Schmuziger M. A patent internal mammary artery graft decreases the risk of reoperative coronary artery bypass surgery. Texas Heart Inst J 1994;21:125-129.[Medline]
- Christenson J.T., Velebit V., Maurice J., Simonet F., Schmuziger M. Risks, benefits, and results of reoperative coronary artery surgery with internal mammary grafts. Cardiovasc Surg 1995;3:163-169.[Medline]
- Byrne J.G., Aranki S.F., Couper G.S., Adams D.H., Allred E.N., Cohn L.H. Reoperative aortic valve replacement: partial upper hemisternotomy versus conventional full sternotomy. J Cardiovasc Surg 1999;118:991-997.
Top
Abstract
Introduction
