Starting minimally invasive valve surgery using endoclamp technology: safety and results of a starting surgeon Herbert De Praeterea,b, Peter Verbrugghea,b, Filip Regaa,b, Bart Meurisa,b and Paul Herijgersa,b,* a b
Department of Cardiac Surgery, University Hospitals Leuven, Leuven, Belgium Department of Cardiovascular Sciences, KU Leuven, Leuven, Belgium
* Corresponding author. Department of Cardiac Surgery, U.Z. Leuven, Herestraat 49, 3000 Leuven, Belgium. Tel: +32-16-344260; fax: +32-16-344616; e-mail: [email protected]
(P. Herijgers). Received 22 May 2014; received in revised form 17 October 2014; accepted 27 October 2014
Abstract OBJECTIVES: To critically review the learning curve, safety issues and outcome of a single surgeon while starting up minimally invasive mitral valve surgery (MIMVS). METHODS: We performed a descriptive, retrospective study of 138 patients with minimally invasive mitral valve surgery between March 2004 and December 2010. The learning curve was assessed using a logarithmic curve-ﬁt regression analysis of the cardiopulmonary bypass parameters and deﬁned as the end of the steepest part. Complexity was assessed by the number of different techniques performed on the mitral valve and the number of concomitant procedures. Follow-up was obtained for embolic events, endocarditis, bleeding, reintervention, echocardiographic data and NYHA class. RESULTS: The learning curve was found in the last 30 cases. There was a signiﬁcant reduction in aortic cross-clamp time before and after the end of the learning curve [Patients 1–30: 120.77 (±28.28); Patients 31–138: 97.57 (±5.66); P 480 s], venous [QuickDraw® Venous Cannula Kit (21 or 23 Fr), Edwards Lifesciences Corp., Irvine, CA, USA] and arterial cannulas [EndoReturn™ Arterial Cannula Kit (21 and 23 Fr), Edwards Lifesciences Corp.] are installed. Then, the endoaortic balloon (EndoClamp® Intra-aortic Occlusion Device, Edwards Lifesciences Corp.) is positioned using TEE guidance and since 18 August 2010 also under radiographic guidance in a hybrid operation theatre. CPB is then started with moderate cooling of the patient. Right lung is deﬂated and the pericardium is transected at least 2 cm anterior of the phrenic nerve. The EndoClamp is inﬂated and myocardial protection is started. We used adenosine (0.25 mg/kg) in shot followed by single-shot modiﬁed NIH crystalloid solution (20 ml/kg). The left atrium is entered through the Waterston groove and a left atrium retractor is placed parasternally in the third or fourth IS. After evaluation of the valve, the proper surgical technique is chosen to treat the mitral valve. De-airing of the heart is performed using the aortic root sucker and the left ventricular vent. A temporary ventricular epicardial pacemaker is placed. After echocardiographic evaluation of the repair and rewarming of the patient, CPB is discontinued and all wounds are closed.
Table 1: Experience of the studied surgeon: number of performed mitral valve operations between 2000 and 2012 2000 Standard MVP isolated MIMVS MVP isolated Standard MVP concomitant MIMVS MVP concomitant Standard MVR isolated MIMVS MVR isolated Standard MVR concomitant MIMVS MVR concomitant Total
37 2 22
24 5 22
20 9 18
23 4 27
17 3 27
19 3 28
24 15 11 3 11 6 29 1
17 11 20 2 9 7 31
17 18 23 3 20 2 17 3
19 32 16 5 7 3 13 1
11 32 22 1 10 2 23
6 35 35 2 5 5 18 2
MVP: mitral valve plasty; MVR: mitral valve replacement; MIMVS: minimally invasive mitral valve surgery.
Operative time, hospitalization and intensive care unit stay CPB parameters (perfusion time, aortic cross-clamp time and reperfusion time) were chosen as the most important predictors of improving surgical skills and were studied throughout the whole series of consecutive patients. Using the operation time and the consecutive order of the case, we evaluated the effect on hospitalization and intensive care unit (ICU) stay.
Statistical analysis The learning curve was assessed using a logarithmic curve-ﬁt regression analysis of the aortic clamp time. To compare both groups, we used a two-tailed Student’s t-test for continuous data. Every variable was assessed for normality. For categorical data, the Fisher’s exact test was used. Pearson’s product–moment correlation coefﬁcient was used to measure linear relationships. Mean ± standard deviation were calculated for appropriate variables. Excel 2010 (Microsoft Corp., Redmond, WA, USA) and Graphpad Prism 5 (GraphPad Software, Inc., San Diego, CA, USA) were used as data management and statistical software, and P-values