The recent interest in minimal access surgery throughout

Transcription

1 Partial Versus Full Sternotomy for Aortic Valve Replacement Michael F. Szwerc, MD, Daniel H. Benckart, MD, Robert J. Wiechmann, MD, Edward B. Savage, MD, Gary W. Szydlowski, MD, George J. Magovern, Jr, MD, and James A. Magovern, MD Department of Cardiothoracic Surgery, Allegheny University Hospitals, Allegheny General Hospital, Pittsburgh, Pennsylvania Background. Recent trends suggest that smaller incisions reduce postoperative morbidity. This study tests the hypothesis that a partial upper sternotomy improves patient outcome for aortic valve replacement. Methods. A group of 50 patients who underwent aortic valve surgery through a partial upper sternotomy (group I) were compared to 50 patients who underwent aortic valve replacement through a median sternotomy during the same time period (group II). The mean age (60 2 versus 63 2 years; mean SEM) and preoperative ejection fractions (53 2 versus 54 2) were similar. Operations were performed with central cannulation, and antegrade/retrograde blood cardioplegia. Results. There was one death in each group. No differences were found in aortic occlusion time, mediastinal drainage, transfusion incidence, narcotic requirement, length of stay, or cost. The incidence of pleural and pericardial effusions was increased (18.4% versus 3.9%, p < 0.03), and the need for postoperative inotropic support was higher (38.7% versus 19.6%, p < 0.03) in the partial sternotomy group. Conclusions. Aortic valve replacement can be performed through a partial sternotomy with results comparable to full sternotomy. The partial sternotomy offers a cosmetic benefit, but does not significantly reduce postoperative pain, length of stay, or cost. (Ann Thorac Surg 1999;68: ) 1999 by The Society of Thoracic Surgeons The recent interest in minimal access surgery throughout all surgical disciplines is based upon the theory that smaller surgical incisions lead to less postoperative pain, shorter hospitalization and faster functional recovery. There is a cosmetic benefit to smaller incision, but safety and efficacy are more important factors, especially for heart operations. It has been shown that minimally invasive direct coronary artery bypass grafting (MID- CABG) surgery is safe, has good midterm results, and can lead to shorter length of stay when compared to conventional coronary artery bypass surgery [1, 2]. However, it is unclear whether the reduction in surgical morbidity stems from the smaller incision or from the absence of cardiopulmonary bypass. This interest in reduced incisions for heart surgery has been applied to valve surgery. There have been numerous reports [3 13] on smaller incisions for aortic valve replacement. These reports have described various surgical approaches to aortic valve replacement, including a partial upper sternotomy with central cannulation [3, 4] and transverse sternotomy with division of both internal mammary arteries [5]. In addition, various permutations of a limited sternotomy, including a J incision [6, 7], a right-sided partial sternotomy [8], and a reversed Z sternotomy [9] have been described. Other surgical options which do not use a sternotomy have also been Accepted for publication June 5, Address reprint requests to Dr James Magovern, Department of Cardiothoracic Surgery, Allegheny General Hospital, 320 E North Ave, Pittsburgh, PA 15212; jmagover@pgh.auhs.edu. described, including a right parasternal incision with femoral arterial and venous cannulation [10], and a limited right thoracotomy-pocket aortic valve replacement [11]. Many of these reports provide practical technical information and several have suggested that smaller incisions lessen surgical morbidity [3, 4, 7, 11, 12]. The presumption is that a smaller incision, by itself, leads to a shorter hospital stay, decreased postoperative pain, and earlier functional recovery, even though the duration of cardiopulmonary bypass may be longer when surgical access is reduced. We began performing minimally invasive aortic valve replacement in March 1997 through a partial upper sternotomy. During this same time period we also continued to perform aortic valve surgery through a traditional full sternotomy. The focus of this report is to validate the efficacy of this approach to aortic valve surgery, and compare these two groups of patients. It was our aim to determine if a smaller incision for aortic valve replacement, by itself, leads to a clinically significant reduction in surgical morbidity. Material and Methods This is a retrospective, nonrandomized review of a single institutional series of aortic valve replacements using either a partial or full sternotomy. From March 1997 to September 1998, 100 patients underwent isolated aortic valve replacement at Allegheny General Hospital. The mean age of the group was 61 years (range 22 85) and 1999 by The Society of Thoracic Surgeons /99/$20.00 Published by Elsevier Science Inc PII S (99)

2 2210 SZWERC ET AL Ann Thorac Surg PARTIAL STERNOTOMY FOR AVR 1999;68: Table 1. Preoperative Clinical Profile Variable Age (y) Male (%) Clinical risk score Aortic valve area (mm 2 ) Valve gradient (mm Hg) Ejection fraction (%) Aortic stenosis Aortic insufficiency AS/AI 10 9 Bicuspid valve 6 2 Endocarditis 1 0 AS aortic stenosis; AI aortic insufficiency 61% were male. Over half (56%) of the patients had aortic stenosis, 25% had aortic insufficiency, and 19% had a combination of aortic stenosis and insufficiency. Eight of the valves were bicuspid and 1 patient had evidence of endocarditis on preoperative echocardiogram. None of the patients had mitral stenosis and none had mitral regurgitation that was worse than 2 on preoperative echocardiogram. The mean clinical risk score (CRS) was 6, which places these patients in the low risk category. The CRS is a composite score, developed at Allegheny General Hospital, that is used to stratify patients into risk categories based on preoperative data [14]. Higher scores indicate higher risk, with a maximum score of 20. The CRS reflects urgency of operation, left ventricular function, comorbid disease, age, and laboratory studies. The mean ejection fraction was 53%. Patients having a partial sternotomy were defined as group I and those having a full sternotomy were defined as group II. Patients having aortic valve replacement with a cryopreserved homograft or stentless tissue valve were not included in this analysis, because all were done through a full sternotomy incision. The decision to perform a minimally invasive or a traditional operation was based on surgeon or patient preference. Several surgeons used a partial sternotomy as the incision of choice for all patients, and other surgeons used a full sternotomy for all patients. Relevant preoperative clinical data comparing the two groups is depicted in Table 1. All patients had 1-month follow-up following hospital discharge. Operative Technique The operative technique for the minimally invasive group is as follows. The patient is placed in the supine position. All patients were given Ancef (1 gm IV) (Smith- Kline Beecham, Pittsburgh, PA). Standard hemodynamic and monitoring lines are used. A shoulder roll is placed behind the scapula to elevate the sternum. External defibrillator pads are placed. A 6-cm midline skin incision is made beginning 3 cm from the sternal notch. Short superior and inferior subcutaneous tissue flaps are raised prior to exposing the sternum. The sternum is divided from the manubrium to the third or fourth interspace with a pneumatic saw. The sternal incision is then angled toward either the third or fourth interspace on the patient s right side, creating an incision in a J configuration. The internal mammary artery and vein are left undisturbed. A standard Finochietto retractor is used to separate the sternum and expose the mediastinum. Wide excision of the thymic remnant and the pericardial fat pat is performed to improve subsequent operative exposure. Only the upper portion of the pericardium is divided, exposing the ascending aorta and the right atrium. The pericardial edges are suspended from the skin edges to elevate the heart into the operative field. After full heparinization, central venous cannulation using a two-staged venous cannula is placed through the right atrial appendage and the ascending aorta is cannulated in the usual location. A retrograde cardioplegia cannula is placed just inferior to the right atrial appendage. Placement of this cannula is blind, but transesophageal echocardiography (TEE) can be used to confirm its position. The left ventricle is vented through a cannula placed in the right superior pulmonary vein. The exposure of the aortic valve through a partial sternotomy is excellent. Myocardial protection consists of mild systemic hypothermia (32 C), and both antegrade and retrograde cold blood cardioplegia. Topical myocardial protection is not used. Aortotomy and the remainder of the valve excision and replacement is performed in the same fashion as in full sternotomy. Transesophageal echocardiography is used in all cases to evaluate valve performance and to monitor intracardiac air when coming off cardiopulmonary bypass. Temporary pacing wires are placed on the heart. The mediastinum is drained with a 32F chest tube and a 19F Blake drain (Johnson and Johnson Medical, Inc, Arlington, TX). The pleural cavities are not routinely opened. The chest tube is removed on the first postoperative day and the Blake tube on the second postoperative day. The sternum is approximated with four stainless steel No. 5 wires. In patients who underwent full sternotomy, two chest tubes are placed in the mediastinum at the completion of each case and removed on the first postoperative day. Six wires are used to reapproximate the sternum. Statistical Analysis The two groups of patients were compared with Student s t test for unpaired data, 2 test, or Fisher s exact test where appropriate. A p value less than 0.05 was taken to indicate statistically significant trends. Results A total of 100 patients underwent aortic valve replacement. Minimally invasive aortic valve replacement was attempted in 50 patients and successful in 49 (49/50; 98%) (group I). One procedure was converted to a full sternotomy to improve exposure of the right atrium. This patient was subsequently analyzed with the full sternotomy group (group II). Fifty-one patients had aortic valve replacement through a full sternotomy. Operative data is listed in Table 2. The cross-clamp time for both procedures was identical (72.1 minutes). The cardiopulmonary

3 Ann Thorac Surg SZWERC ET AL 1999;68: PARTIAL STERNOTOMY FOR AVR 2211 Table 2. Operative Data Table 4. Morbidity Variable Event Cross-clamp time (min) CPB time (min) Mechanical valve Pericardial valve CPB cardiopulmonary bypass time. bypass times in both groups were similar ( versus minutes). Thirty-six patients had a mechanical valve placed through a partial sternotomy, whereas 35 patients had a mechanical valve placed through a full sternotomy. The remainder of the patients had a bioprosthetic valve replacement. All patients in both groups were successfully weaned from cardiopulmonary bypass without the need for mechanical circulatory support. The mean intubation time in the postoperative period for patients in group I was 15 hours, which was similar to patients in group II (16 hours) (Table 3). Intubation time includes the period of operation, so that the postoperative length of intubation was approximately 4 hours less than the total intubation time. Mean postoperative mediastinal drainage in group I was 466 ml, which was not statistically significant from group II (568 ml). Fifty-one percent of the patients in group I required a postoperative blood transfusion. The average number of units of packed red blood cells (PRBCs) transfused was 2.6. Neither the percent transfusion nor the number of units transfused was statistically significant between the groups. Most of the patients were transferred from the intensive care unit on the morning of the first postoperative day. The average length of ICU stay in both groups was identical (1.1 days). There was one postoperative death in each group. One patient in group I developed fungal endocarditis and multisystem organ dysfunction several weeks after aortic valve replacement. One patient in group II developed mediastinitis and multisystem organ failure several weeks after surgery. Two patients in group I required Table 3. Outcome Variable Intubation (h) Mediastinal drainage (ml) Transfusion PRBCs/transfusion Inotropic support (postop) a Narcotic requirements (morphine equivalents) ICU stay (days) LOS (days) a Significant difference between groups ( p value of 0.05). LOS length of stay; PRBCs packed red blood cells. Death 1 1 Complications Reintubation 2 1 Reexploration for bleeding 1 1 Atrial fibrillation Complete heart block 2 1 Pleural effusion 5 0 a Soft-tissue infection 4 0 a Mediastinitis 0 1 Pneumonia 1 0 GI bleed 1 0 Cerebrovascular accident 1 1 Respiratory insufficiency 2 1 Aspiration 1 0 Total % of total a Significant difference between groups ( p value of 0.05). reintubation. One patient in each group required reexploration for bleeding. The use of inotropes in the postoperative setting was higher in group I (19 versus 10, p 0.03). Atrial fibrillation developed in 31% of the patients who underwent full sternotomy, versus 20% in the partial sternotomy (p 0.31). Two patients in group I required a pacemaker for complete heart block versus one patient in group II. In 5 patients in group I, a significant pleural effusion developed that required drainage, whereas no significant pleural effusion developed in any patient in group II (p 0.05). In addition, in 4 patients in group I, a postoperative pericardial effusion developed that required either percutaneous (n 3) or open (n 1) drainage, compared to 2 patients in group II. No mediastinal wound infections developed in any patient in the partial sternotomy group. However, in 4 patients, a superficial wound infection developed that was successfully treated with oral antibiotics and local wound care in 2 patients and intravenous antibiotics in 2 other patients. In 1 patient in group II, mediastinitis developed following re-exploration for postoperative bleeding, requiring subsequent sternal debridement and muscle-flap coverage. The total incidence of complications in group I was higher than in group II (35 of 49, 71% versus 25 of 51, 49%) (Table 4). When comparing the two groups, the most notable difference in complications came from the development of a soft-tissue infection and a pericardial or pleural effusion in the partial sternotomy group. The cumulative total of narcotic medications administered throughout the hospitalization was not statistically significant between the two groups. (Table 3). This includes the total dose of narcotics administered in the immediate postoperative period and all of the subsequent narcotic use until the fifth postoperative day. The overall length of stay, excluding 1 outlier in each of the

4 2212 SZWERC ET AL Ann Thorac Surg PARTIAL STERNOTOMY FOR AVR 1999;68: groups who stayed greater then 14 days ( 5.7 versus 6.4 days), was similar. The principle costs for the two groups, which includes all direct and indirect costs, were not significantly different ($17,410 $7485 versus $16,382 $9674, group I versus group II). In group I, 6 patients required readmission to the hospital during the first 30 days after surgery. The cause for the readmission was a pericardial effusion in 2 patients, atrial fibrillation in 1 patient, urosepsis in 1 patient, and a wound infection in 2 patients. In group II, 3 patients required readmission. Two patients developed atrial fibrillation at home and 1 patient developed a pericardial effusion that required percutaneous drainage. Comment This study presents the operative results of 100 consecutive patients who underwent aortic valve replacement through either a partial upper sternotomy or a full sternotomy. Our data demonstrate that a partial upper sternotomy is a safe and effective way to perform aortic valve replacement. There was one perioperative death in the partial sternotomy group (2%). The remaining 48 patients were discharged to home and have experienced no mechanical problems related to their valve in the early follow-up period. It has been suggested that minimally invasive valve operations take longer to perform [4, 15], but we found no difference in cross-clamp time or total cardiopulmonary bypass times between the groups. Although a partial upper sternotomy is a safe procedure, we were unable to demonstrate any clinically significant improvement over traditional sternotomy in the early postoperative period. Valve replacement through the smaller incision did not decrease the amount of mediastinal drainage or the rate of postoperative transfusion. Additionally, the partial sternotomy did not shorten the intubation time, length of ICU stay, or overall hospital stay. There was also no evidence that the smaller incision resulted in less postoperative pain, because the postoperative narcotic usage was identical between the two groups. We found that partial sternotomy patients had a higher incidence of superficial wound infection (4 of 48 versus 0 of 51). Deep sternal or mediastinal infection did not develop in any of the patients. It is possible that tension placed on the soft tissue while working through the smaller wound may have contributed to this problem. We also found that the incidence of pleural and pericardial effusions in the partial sternotomy group was higher (9 versus 2; p 0.05). We suggest that this resulted from incomplete mediastinal drainage at the time of surgery. An additional difference between the groups was a higher usage of postoperative inotropes in the partial sternotomy patients, which may be related to the inability to use topical hypothermia with this incision. There have been many recent reports about alternate incisions for minimally invasive aortic valve replacement. The right parasternal approach popularized by Cosgrove and Sabik [10] has been abandoned by most surgeons because of patient discomfort, postoperative lung herniation, and the potential complications from femoral cannulation. Other approaches, which include the Z sternotomy or transverse sternotomy, are not ideal because they require division of both internal mammary arteries. Cohn and associates reported their experience with 41 minimally invasive valve operations [4]. This included 20 minimally invasive aortic valve replacements done with central cannulation through a partial sternotomy. These patients had a longer ischemic time (103 versus 82 minutes) when compared to a cohort of patients who had a full sternotomy, and to the patients in our study. Patients in both groups had similar pain scores within the first 10 days of operation; however, patients in the minimally invasive group appeared to return to normal activity sooner. In a report of 100 consecutive ministernotomies for aortic valve surgery, Dr Gundry reported a median length of stay of 2.6 days with no readmissions or postoperative wound complications [3]. Many of these patients were children and therefore it is difficult to extrapolate this experience to the adult population. In our series, 30% of the patients were older than 70 years. It has been suggested that minimally invasive aortic valve replacement is less expensive than valve replacement through of full sternotomy [4], but we did not confirm this observation. The principal costs for the two groups were nearly identical, which is not surprising because the duration of cardiopulmonary bypass and the length of intensive care unit stay and hospital length of stay were similar for the two groups. In addition, the relative morbidities of the two procedures were similar. It is possible that the partial sternotomy might result in earlier functional recovery and return to work, but many of these patients are in their retirement years. There are several important limitations in this study. First, this is a retrospective analysis of the results of two operations, not a prospective randomized study, which introduces the question of systematic bias. Second, several of the surgeons in our group chose to do all or most of their aortic valve replacements through a partial sternotomy while others chose to do them through a traditional sternotomy. Thus, there may be some selection bias. Third, although we have incorporated standardized protocols in our ICU, the duration of endotracheal intubation and the length of postoperative stay is ultimately the decision of the individual surgeon in conjunction with the cardiologist. Aortic valve replacement through a partial upper sternotomy is a safe procedure and it appears to be as efficacious as valve replacement through a traditional sternotomy. Although the smaller incision is cosmetically appealing, there is no evidence, at least from this study, that it reduces the morbidity associated with aortic valve replacement. We share the concern of others that the smaller incision limits exposure to the remaining heart, thus possibly making it difficult to deal with any intraoperative complications should they occur. However, conversion to a full sternotomy can be accomplished in an expeditious manner should this be necessary. We also caution that in patients with right ventricular dysfunction, topical cooling serves as an important adjunct to

5 Ann Thorac Surg SZWERC ET AL 1999;68; PARTIAL STERNOTOMY FOR AVR 2213 myocardial protection. The exposure in a partial sternotomy approach limits the amount of topical cooling that can be used. The partial sternotomy approach should not be considered in these patients. Some subsets of patients may benefit from a partial sternotomy approach for aortic valve replacement. Young patients could be one such group, because they generally have more interest in cosmesis. In addition, the use of a partial sternotomy might make reoperative surgery less difficult, because of less mediastinal tissue disturbance. This would also be a potential benefit for younger patients who face a higher chance of needing additional cardiac surgery. It would be helpful to evaluate the impact of partial sternotomy on functional recovery following heart surgery. The study by Cohn and colleagues is provocative [4], but certainly not conclusive. More study is needed on this topic before any firm conclusions can be drawn. References 1. Magovern JA, Benckart DH, Landreneau RJ, Sakert T, Magovern GJ Jr. Morbidity, cost, and six-month outcome of minimally invasive direct coronary artery bypass grafting. Ann Thorac Surg 1998;66: Calafiore AM, Di Giammarco G, Teodori G, et al. Left anterior descending coronary artery grafting via left anterior small thoracotomy without cardiopulmonary bypass. Ann Thorac Surg 1996;61: Gundry SR. Aortic valve surgery via limited incisions. In: Oz MC, Goldstein DJ, eds. Contemporary cardiology: minimally invasive cardiac surgery. Totowa, New Jersey: Humana Press, 1999: Cohn LH, Adams DH, Couper GS, et al. Minimally invasive cardiac valve surgery improves patient satisfaction while reducing costs of cardiac valve replacement and repair. Ann Surg 1997;226: De Amicis V, Ascione R, Iannelli G, Di Tommaso L, Monaco M, Spampinato N. Aortic valve replacement through a minimally invasive approach. Tex Heart Inst J 1997;24: Svensson LG, D Agostino RS. Minimal-access J or j sternotomy for valvular, aortic, and coronary operations or reoperations. Ann Thorac Surg 1997;64: Svensson LG, D Agostino RS. J incision minimal-access valve operations. Ann Thorac Surg 1998;66: Kasegawa H, Shimokawa T, Matsushita Y, Kamata H, Ida T, Kawase M. Right-sided partial sternotomy for minimally invasive valve operation: open door method. Ann Thorac Surg 1998;65: Nair RU, Sharpe DAC. Minimally invasive reversed Z sternotomy for aortic valve replacement. Ann Thorac Surg 1998; 65: Cosgrove DM III, Sabik JF. Minimally invasive approach for aortic valve operations. Ann Thorac Surg 1996;62: Pau KK, Yakub A, Awang Y. Minimally invasive aortic valve surgery: pocket AVR. J Thorac Cardiovasc Surg 1998; 115: Massetti M, Babatasi G, Lotti A, Bhoyroos S, LePage O, Khayat A. Less invasive cardiac operations through a median sternotomy: 100 consecutive cases. Ann Thorac Surg 1998;66: Tam RKW, Almeida AA. Minimally invasive aortic valve replacement via partial sternotomy. Ann Thorac Surg 1998; 65: Magovern JA, Sakert T, Magovern GJ Jr, et al. A model that predicts morbidity and mortality after coronary artery bypass graft surgery. J Am Coll Cardiol 1996; Cooley DA. Minimally invasive valve surgery versus the conventional approach. Ann Thorac Surg 1998;66: INVITED COMMENTARY Widespread interest in minimally invasive direct access cardiac valve surgery has led to numerous reports in the literature over the past few years. This unrandomized and retrospective review by Szwerc and associates has further clarified the utility of a limitedsternotomy approach to replacement of the aortic valve. It has been previously established that this approach is safe and reproducible, with potential benefit in terms of improved patient satisfaction and earlier return to normal activity [1]. This report again demonstrates that the aortic valve can be safely replaced with limited exposure. Surprisingly, it can be replaced in the same timely fashion as a full sternotomy. So the question is should a limited-sternotomy approach be the gold standard for aortic valve replacement? Additional data provided in this study suggest not. The authors have described three morbidities that occurred with greater frequency in the partial sternotomy group which deserve emphasis. They noted an increased need for post-operative inotropic therapy in the small incision group and noted that limited exposure of the heart may have compromised their myocardial protection because of their inability to topically cool, emphasizing concern regarding right ventricular dysfunction. Echocardiographic documentation of specific wall motion abnormalities would have strengthened this argument, but their finding is still important given the significance of a potential compromise in protection, especially in patients with aortic stenosis and myocardial hypertrophy. The increased incidence of pleural and pericardial effusions in the partial sternotomy group suggests that perioperative drainage of the thoracic cavity may be compromised with a smaller incision and special care should be taken when placing drains in these patients. Finally, their observation of superficial wound inflammation in several smaller incisions, presumably due to traction induced dermal ischemia, suggests that adding incision length may in fact decrease the morbidity of the procedure. The information provided in this manuscript should help surgeons select appropriate patients for minimally invasive aortic valve surgery and provide them with a warning regarding potential complications in these selected patients. We are continuing to refine our selection criteria for the hemi-sternotomy approach and are in agreement with the authors that younger patients concerned with cosmetics and facing a poten By The Society of Thoracic Surgeons /99/$20.00 Published by Elsevier Science Inc PII S (99)