Asian Cardiovasc Thorac Ann 2007;15:507-510
© 2007 Asia Publishing EXchange Ltd
Learning Curve of Arch-First Technique Analyzed by Cumulative Sum
Min-Ho Song, MD,
Yoshiyuki Tokuda, MD,
Masaya Hirai, MD1,
Yuichi Ueda, MD2
Department of Cardiovascular Surgery, Gifu Prefectural Tajimi Hospital, Tajimi
1 Department of Cardiac Surgery, Aichi Medical College, Nagakute
2 Department of Cardiothoracic Surgery, Nagoya University Graduate School of Medicine Nagoya, Japan
For reprint information contact: Min-Ho Song, MD, Tel: 81 57 222 5311, Fax: 81 57 224 1325, Email: p41558{at}govt.pref.gifu.jp, 5-161 Maehata-cho, Tajimi, Gifu 507-8522, Japan.
 |
ABSTRACT
|
|---|
This study was undertaken to verify efficacy of the arch-first technique in the light of its learning curve. From April 2002 to September 2005, 10 consecutive elective cases of total arch replacement were retrospectively examined. The learning curve of the arch-first technique was constructed using cumulative sum analysis. There were no operative deaths. The mean deep hypothermic circulatory arrest time was 28.4 ± 13.7 min, the lower body ischemic time was 91.3 ± 35.1 min, aortic cross clamp time was 133.2 ± 18.1 min, cardiopulmonary bypass time was 198.8 ± 21.5 min, and operation time was 383 ± 24 min. The durations of deep hypothermic circulatory arrest, bypass, and operation were under the 90% lower alarm line in all 10 cases. The lower body ischemic time and cardiac arrest time were between the 80% upper and lower alert lines. Cumulative sum analysis of total arch replacement using the arch-first technique showed satisfactory rates of improvement in reconstruction of the 3 arch vessels, cardiopulmonary bypass time, and overall mortality.
|
INTRODUCTION
|
|---|
According to the annual report by the Japanese Association for Thoracic Surgery, total arch replacement still carries one of the highest mortality rates of all cardiac procedures: 8.4% (89 deaths in 1,055 patients undergoing total arch replacement in 2003) for non-dissecting and unruptured thoracic aortic aneurysm, and 36% (60 deaths in 166 patients) for ruptured aortic arch aneurysm.1 This gives us the clinical impetus to make every effort to improve patient survival after arch replacement. The arch-first technique was introduced by Rokkas and Kouchoukos2 in 1999 to minimize cerebral ischemic time and the technical difficulty of the intricate distal anastomosis. The rationale of this technique is to perform the brachiocephalic vessel connections first, and then to use selective antegrade cerebral perfusion through the completed graft while the distal anastomosis is being constructed. Since Kouchoukos and Masetti3 reported superior clinical results to those with conventional selective brain perfusion and retrograde brain perfusion, Sasaki and colleagues4 adopted a modified arch-first technique and reported excellent results with significantly fewer cerebral complications. This consists of reconstruction of the 3 neck vessels with a branched graft under deep hypothermic circulatory arrest, elephant trunk insertion into the distal aortic arch which is sewn together with the distal aortic arch, graft-to-graft anastomosis, and ascending aortic anastomosis. To provide an objective assessment of the efficacy of the arch-first technique performed by a surgeon other than the originator (Yuichi Ueda), we examined the learning curve using cumulative sum (CUSUM) analysis.5
|
PATIENTS AND METHODS
|
|---|
The performance of the arch-first technique by a cardiothoracic surgeon certified by the Japanese Association of Thoracic Surgery was retrospectively analyzed, and CUSUM was applied to construct learning curves according to operative variables. From April 2002 to October 2005, he performed 10 elective arch-first operations for unruptured aortic arch aneurysms > 6 cm in maximal diameter. There were 7 men and 3 women, with a mean age of 70.8 years. There were 9 atherosclerotic true aneurysms and one chronic dissecting aneurysm. All aneurysms were located at the aortic arch and distal aortic arch.
CUSUM is a statistical method for analyzing the changes in mortality and morbidity rates in different practice years. In CUSUM analysis, the operative variables and results of Yuichi Ueda were assumed to represent success.4 The following operative variables were used to construct CUSUM learning curves: mortality (defined as 30-day mortality plus all deaths within the same hospitalization), morbidity (according to the Society of Thoracic Surgeons’ reporting database), operation time, cardiopulmonary bypass (CPB) time, aortic cross clamp time (i.e., myocardial ischemic time), ischemic time of the lower body (down from the left subclavian artery), and deep hypothermic circulatory arrest time (i.e., cerebral ischemic time).
CUSUM was defined as Sn = 
(Xi–X0), where Xi = 0 for success, Xi = 1 for failure, and X0 is the reference or was set strictly at 0.07 (7%) target value. In this study, X0 in analyses of mortality and morbidity because 30-day mortality and morbidity was 8.4% according to the Japanese Association of Thoracic Surgery database, and it was set at 0.02 (2%) in analyses of operative variables because 2% is the average failure rate in operative skill according to the Japanese Association of Thoracic Surgery database. Success in terms of operation time, CPB time, aortic cross clamp time, lower body ischemic time, and deep hypothermic circulatory arrest time was defined as a time that did not exceed the originator’s averages plus 1 standard deviation. CUSUM curves were constructed manually for overall mortality and operative variables. Furthermore, an 80% alert line and a 95% alarm boundary line were constructed for the CUSUM curve of each investigation, according to published formulae.6
The detailed operative procedure of the arch-first technique has already been described.4 The patient was placed in the supine position and the aortic arch was exposed via a median sternotomy. Cardiopulmonary bypass was applied with bicaval drainage, and the ascending aorta was perfused after careful inspection by direct echocardiography. Core cooling was initiated, with left ventricular venting via the left upper pulmonary vein. Perfusion was discontinued once the esophageal temperature fell below 20°C. Circulatory arrest was obtained. Cold blood cardioplegia was administered antegradely. The arch vessels were transected and reconstructed with a 4-branch graft, and antegrade cerebral perfusion was resumed through the graft. The distal anastomosis was performed with another section of graft, using the elephant trunk technique. The distal graft was drawn back, and distal visceral perfusion commenced with a Foley catheter inserted into the elephant trunk. The distal graft and the arch graft were anastomosed, and rewarming was started. Proximal anastomosis between the arch graft and ascending aorta was carried out, and the total aortic arch reconstruction was completed.
|
RESULTS
|
|---|
There were no operative deaths. Figure 1
shows the CUSUM analysis of mortality during the learning curve. The curve remained below the 90% lower alarm line from beginning to end. CUSUM analyses of the learning curves with respect to operative variables are shown in Figure 2. The mean operation time was 383 ± 24 min and the curve remained below the 90% lower alarm line throughout. The mean CPB time was 198.8 ± 21.5 min and the CUSUM curve remained below the 90% lower alarm line. The mean aortic cross clamp time was 133.2 ± 18.1 min. With the exception of the first case, the curve remained below the 80% upper alert line; from the 4th case, it showed constant improvement. The mean lower body ischemic time was 91.3 ± 35.1 min; the curve tended to approach the 80% upper alert line but remained under the line. The mean circulatory arrest time was 28.4 ± 13.7 min and the curve remained below the 90% lower alarm line. Table 1 shows the postoperative variables and complications. There were statistically significant differences between the results of the originator and those of the surgeon under study regarding the rates of stroke, renal failure, and re-operation for bleeding. Other variables showed no significant differences. Other digestive organs complications were not noted.
View larger version (10K):
[in this window]
[in a new window]
|
Figure 2. Learning curve regarding operation time, bypass time, cross clamp time, lower body ischemic time, and circulatory arrest time analyzed by CUSUM.
|
|
|
DISCUSSION
|
|---|
The CUSUM technique was introduced by Page5 in 1954 to detect changes in the frequencies of failures by repeated application of a sequential probability ratio test. The CUSUM technique recognizes the importance of time as a hidden variable in clinical studies, and avoids the statistical problems associated with repeated significance testing. Since the CUSUM method detects decreases in performance during the patient care process rather than after an arbitrary unit of time, it might be helpful in alerting the surgeon to suboptimal results in advance of standard chi-squared and analysis of variance methods. CUSUM analysis was first used to monitor surgical performance in 1994 by de Leval and colleagues7 at the Great Ormond Street Hospital, UK. In the coronary surgery field, Novick and Stitt6 from the University of Western Ontario, Canada, reported their learning curve with the use of CUSUM in 1999 and found that it provided a more sensitive indicator of a cluster of surgical failures than standard statistical methods. We reported the first study in Japan of a coronary surgery learning curve with the use of CUSUM in 2005.8
In total aortic arch replacement surgery, several technical issues remain to be solved, such as how to prevent fatal bleeding from the distal anastomosis site, and the optimal method of brain protection. We consider the arch-first technique to be a good solution to these difficulties because the necessary skill can be acquired quickly, as shown by the overall safe performance in this study. We can infer that it was not difficult to master reconstruction of the 3 neck branches from the very first case, and also to perform the proximal anastomosis because CUSUM showed a safe steady learning curve below the 90% lower alarm line. It is difficult to determine how many cases are required for optimal mastery of the technique because this preliminary experience in 10 patients did not result in any mortality. The knack of the arch-first technique lies in whether or not we can place good sutures into the friable and the diseased atherosclerotic aortic wall. As the atheromatous quality of the distal aortic wall varies from patient to patient, experience of more than 10 cases may be needed to ensure that the distal anastomosis is always completed safely and swiftly. The technical difficulty of the distal anastomosis is reflected in lower body ischemic time, and remains a challenging issue. Several minor adaptations were applied: a modified elephant trunk technique was used with a replicated graft, four 2/0 polyvinyl retracting sutures were employed to make the distal anastomosis site shallow, and a tight running suture was used to achieve perfect anastomosis. Graft perfusion with a Foley catheter has the further advantage of confirming complete hemostasis of the distal anastomosis. We were fortunate not to encounter uncontrollable bleeding from the distal anastomosis.
We are aware of the main limitation of this study: the small number of cases. However, aortic arch aneurysm does not occur often. Even in a university-based hospital, the originator treated just 50 cases in 6 years.4 Rokkas and colleagues2 reported results of just 6 patients using the arch-first technique. Some might have theoretical concerns about the use of CUSUM methods to examine continuous outcomes rather than categorical and dichotomous outcomes of procedures. However, we dealt with success or failure; this is a categorical parameter of a continuous variable.9 Thus it was concluded that CUSUM analysis of total arch replacement using the arch-first technique showed satisfactory rates of improvement in reconstruction of the 3 arch vessels, CPB time, and overall mortality. However, because the quality of the distal aortic aneurysmal wall varied among patients, it might require more experience to perform the optimal distal anastomoses.
|
REFERENCES
|
|---|
- Kazui T, Wada H, Fujita H; Japanese Association for Thoracic Surgery Committee of Science. Thoracic and cardiovascular surgery in Japan during 2003: annual report by The Japanese Association for Thoracic Surgery. Jpn J Thorac Cardiovasc Surg 2005;53:517–36.[Medline]
- Rokkas CK, Kouchoukos NT. Single-stage extensive replacement of the thoracic aorta: the arch-first technique. J Thorac Cardiovasc Surg 1999;117:99–105.[Abstract/Free Full Text]
- Kouchoukos NT, Masetti P. Total aortic arch replacement with a branched graft and limited circulatory arrest of the brain. J Thorac Cardiovasc Surg 2004;128:233–7.[Abstract/Free Full Text]
- Sasaki M, Usui A, Yoshikawa M, Akita T, Ueda Y. Arch-first technique performed under hypothermic circulatory arrest with retrograde cerebral perfusion improves neurological outcomes for total arch replacement. Eur J Cardiothorac Surg 2005;27:821–5.[Abstract/Free Full Text]
- Page ES. Continuous inspection schemes. Biometrika 1954;41:100–15.[Free Full Text]
- Novick RJ, Stitt LW. The learning curve of an academic cardiac surgeon: use of the CUSUM method. J Card Surg 1999;14:312–22.[Medline]
- de Leval MR, Francois K, Bull C, Brawn W, Spiegelhalter D. Analysis of a cluster of surgical failures. Application to a series of neonatal arterial switch operations. J Thorac Cardiovasc Surg 1994;107:914–24.[Abstract/Free Full Text]
- Song MH, Tajima K, Watanabe T, Ito T. Learning curve of coronary surgery by a cardiac surgeon in Japan with the use of cumulative sum analysis. Jpn J Thorac Cardiovasc Surg 2005;53:551–6.[Medline]
- Grunkemeier GL, Wu YX, Furnary AP. Cumulative sum techniques for assessing surgical results. Ann Thorac Surg 2003;76:663–7.[Free Full Text]
TOP
ABSTRACT
INTRODUCTION
