A simple method to determine anastomotic quality of coronary artery bypass grafting in the operating room

Cardiovascular Surgery, Vol. 9, No. 5, pp. 499–503, 2001  2001 The International Society for Cardiovascular Surgery. Published by Elsevier Science Ltd All rights reserved 0967-2109/01 $20.00

PII: S0967-2109(01)00043-6

www.elsevier.com/locate/cardiosur

A simple method to determine anastomotic quality of coronary artery bypass grafting in the operating room Yoshiyuki Takami and Hiroshi Ina Kasugai Municipal Hospital, Division of Cardiovascular Surgery, 1-1-1 Takagi-cho, Kasugai, Aichi 486-8510, Japan Anastomotic quality of coronary artery bypass grafting is directly associated with peri-operative and long-term clinical results. In this study, we investigated a cut-off value for intra-operative flow measurement. This value could be of use to a surgeon in determining the anastomotic quality of grafts. Intra-operative transit-time flow variables (mean flow, pulsatility index, % efficiency, fast Fourier transformation (FFT) of the flow curve) and the 2-week post-operative angiographic findings were examined in 66 coronary artery bypass grafts, including 33 internal thoracic arteries. There were significant differences between patent and non patent grafts in all of the intra-operative flow parameters. Only the FFT ratio, the ratio of powers of the fundamental frequency and its first harmonic, could be utilized as a cut-off value to distinguish patent from non patent grafts. All stenotic or occluded grafts showed an intra-operative FFT ratio of ⬍1.0, while all patent grafts yielded a ratio of >1.0. Based upon these results, we concluded that power spectral analysis of flow measurement might be useful for intra-operative differential diagnosis of the anastomotic quality in coronary artery bypass grafting.  2001 The International Society for Cardiovascular Surgery. Published by Elsevier Science Ltd. All rights reserved Keywords: coronary artery bypass grafting, graft flow, transit-time flowmeter, fast Fourier transformation

Introduction Coronary artery bypass grafting (CABG) has contributed to treatment of patients with ischemic heart disease, especially with left main coronary artery disease, three-vessel disease, proximal left anterior descending artery disease, and/or depressed left ventricular function, as demonstrated by increased survival and reduced ischemic complications [1]. Both peri-operative and long-term results of myocardial revascularization depend upon the patency and anastomotic quality of grafts [2]. Therefore, it is critical for surgeons to evaluate the quality of CABG anasCorrespondence to: Yoshiyuki Takami, MD. Tel.: +81-568-57-0057; Fax:+81-568-57-0067; e-mail: [email protected]

CARDIOVASCULAR SURGERY

OCTOBER 2001 VOL 9 NO 5

tomoses in the operating room. When the graft is not patent well, the surgeon should regraft it. Currently, off-pump CABG necessitating complex techniques are indicated for more patients [3–5], and surgeons increasingly realize the importance to assess the anastomotic quality of the graft in the operating room. Although there have been several reports demonstrating the usefulness of intraoperative assessment of the graft flow by electromagnetic, ultrasonic Doppler, or transit-time method [6–10], no cut-off values have been advocated to distinguish a poor graft anastomosis from an excellent one in CABG surgery. In the present study, we attempted to establish a cut-off value to evaluate the anastomotic quality in the operating room by analyzing intra-operative flow measurement and post-operative angiograms of anastomosed grafts. 499

Intra-operative flow analysis to evaluate CABG anastomotic: Yoshiyuki Takami and Hiroshi Ina

Patients and methods The present study included 32 consecutive patients (25 males and 7 females; mean age, 66.8±6.7 yr) who underwent CABG either with cardiopulmonary bypass (n = 25) or without (n = 7). Combined procedures included mitral valve replacement (n = 4), aortic valve replacement (n = 1),abdominal aneurysmectomy (n = 2), and femoro-femoral arterial bypass (n = 1). The patients received 70 grafts including 33 right or left internal thoracic arteries, 23 radial arteries, 9 saphenous veins, and 5 right gastroepiploic arteries. Graft flow tracing was obtained intraoperatively using a transit-time flowmeter (BF 2000; Medi-Stim AS, Oslo, Norway). A flow probe of 3 or 4 mm was placed around the graft when the hemodynamic condition became stable after cardiopulmonary bypass was weaned in a standard CABG, or when an anastomotic procedure was completed in a off-pump case. Based upon the flow profile obtained, the following variables were calculated: mean, maximal, and minimal graft flow (ml/min); pulsatility index (=[maximal flow⫺minimal flow]/mean flow]; percent efficiency (=volume of backward flow/volume of forward flow); and fast Fourier transformation (FFT) of the flow curve. FFT analysis is based upon the principal that all periodic waveforms can be broken down into a series of pure sine waves or harmonics [11, pp. 167–203]. Harmonics exist at frequencies that are multiplies of the frequency of the original waveform (‘the fundamental frequency’) and are described in terms of an amplitude and phase. The pulsatile waveforms of graft flow in CABG can be considered to be periodic waveforms possessing a fundamental frequency (i.e., the heart rate of the patient). As a parameter representing gradual decrease in power of the harmonics of the fundamental frequency, a FFT ratio ( = F0/H1, where F0 is a power of the fundamental frequency and H1 is a power of the first harmonic.) was calculated in the present study. Every patient underwent a post-operative cardiac catheterization 14±5 days after CABG with a standard technique through the femoral or brachial route. A dose of 2 mg isosorbide dinitrate was injected selectively in each bypass graft. All grafts were examined from at least three different views. The anastomotic sites were evaluated independently and graded as ‘stenosis or occluded’(>25% stenosis) and ‘patent well’ (⬉25% stenosis) by the cardiologists at our hospital. All data were expressed as the means±the standard deviations. Data between ‘stenotic or occluded’ and ‘patent well’ grafts were compared using the MannWhitney’s test. A P-value of less than 0.01 was considered to be statistically significant. 500

Figure 1 Data from a 68-yr-old male patient who underwent in situ grafting with the left internal thoracic artery (LITA) to the left anterior descending artery (LAD). Shown are LITA flow tracing (a), fast Fourier transformation (FFT) of the flow curve (b), and a post-operative angiogram of the LITA (c) which was patent well. F0: a power of the fundamental frequency, H1: a power of the first harmonic

Results Of 70 grafts, flows of four right gastroepiploic arteries could not be measured because they were too thick for successful placement of the probes. Typical recordings of a left internal thoracic artery (LITA) and saphenous vein graft (SVG), which were both demonstrated to be patent well in post-operative angiograms, were demonstrated in Figures 1 and 2. The patent graft flow waveform, whether in situ or aortocoronary, showed two phases of antegrade systolic and diastolic flow. As for graft flow and derived variables, there were significant differences in mean, maximal flow, pulsatility index, % insufficiency, and FFT ratio between the patent and

Figure 2 Data from a 62-yr-old female patient who underwent aortocoronary grafting with a saphenous vein (SVG) to the right coronary artery (RCA). Shown are SVG flow tracing (a), fast Fourier transformation (FFT) of the flow curve (b), and a post-operative angiogram of the SVG (c) which was patent well. F0: a power of the fundamental frequency, H1: a power of the first harmonic

CARDIOVASCULAR SURGERY

OCTOBER 2001 VOL 9 NO 5

Intra-operative flow analysis to evaluate CABG anastomotic: Yoshiyuki Takami and Hiroshi Ina Table 1 Flow and derived variables of the grafts which were revealed to be stenotic/occluded or patent well in post-operative angiography Stenotic or Patent grafts occcluded grafts (n = 57) (n = 9) Mean flow (ml/min) Maximal flow (ml/min) Minimal flow (ml/min) Pulsatility Index % Insufficiency FFTa ratio

P value

⬍0.001

7.6±10.3

51.1±32.4

41.6±23.6

107.0±72.1

⫺16.3±10.1

⫺9.0±29.2

0.856

30.1±26.8 40.4±34.1 0.68±0.27

2.8±2.0 2.9±7.4 3.08±2.27

⬍0.001 ⬍0.001 ⬍0.001

0.036

a

FFT=fast Fourier transformation of the flow waveform.

stenotic/occluded grafts (Table 1). Nine non-patent grafts included three LITA, three SVG, and three radial arteries. In a view of the above parameters such as mean flow, pulsatility index, and % insufficiency, it was impossible to precisely define a cut-off value to distinguish patent from stenotic/occluded grafts (Figure 3). While a LITA graft with a flow of 15 ml/min in the operating room was revealed to be occluded post-operatively (Figure 4), a LITA graft with a low flow of 6 ml/min in the operating room was patent well post-operatively (Figure 5). Among the variables calculated, only the FFT ratio appeared to have a

Figure 4 Data from a 62-yr-old female patient who underwent in situ grafting with the left internal thoracic artery (LITA) to the left anterior descending artery. Shown are LITA graft flow tracing (a), fast Fourier transformation (FFT) of the flow curve (b), and a post-operative angiogram of the LITA (c). Note that in spite of the intra-operative flow of 15 ml/min, the LITA graft was revealed to be occluded post-operatively. The FFT ratio was less than 1.0. F0: a power of the fundamental frequency, H1: a power of the first harmonic

cut-off value. All the stenotic or occluded grafts yielded an intra-operative FFT ratio of less than 1.0, while all patent grafts had a ratio of greater than 1.0 (Figure 3).

Figure 3 Shown are the distribution of the intra-operative mean flow and derived variables, including pulsatility index, % insufficiency, and FFT (fast Fourier transformation) ratio, of the grafts which were revealed to be stenotic/occluded or patent well in post-operative angiography. No parameteres but a FFT ratio could be a cut-off value (a ratio of 1.0) to distinguish patent from non patent grafts

CARDIOVASCULAR SURGERY

OCTOBER 2001 VOL 9 NO 5

501

Intra-operative flow analysis to evaluate CABG anastomotic: Yoshiyuki Takami and Hiroshi Ina

Figure 5 Data from a 68-yr-old male patient who underwent in situ grafting with the left internal thoracic artery (LITA) to the left anterior descending coronary artery (LAD). Shown are LITA graft flow tracing (a), fast Fourier transformation (FFT) of the flow curve (b), and a post-operative angiogram of the LITA (c). Note that in the LITA graft was patent well post-operatively although the intra-operative flow was low, 6 ml/min. The FFT ratio was greater than 1.0. F0: a power of the fundamental frequency, H1: a power of the first harmonic

Discussion It was traditionally common for a surgeon to determine the adequacy of CABG anastomosis based upon the palpation of graft pulsation, hemodynamic stability, and electrocardiographic changes, which are all unreliable and indirect. To increase the reliability, several methods have been developed for the intra-operative assessment of the anastomotic quality in CABG [6–10]. Among these, transit-time flow measurement is considered to be more convienient, less invasive, and less time consuming [12]. The transit-time method is based upon the principle that the time required for ultrasound to pass through blood is slightly longer upstream than downstream [10]. Since the ultrasound beam is wider than the diameter of the vessel lumen, it is neither necessary to know the vessel diameter or perform any complex calibrating procedures. Although the LITA flow of greater than 20 ml/min is considered to be normal as patent as Walpoth and colleagues demonstrated [10], we have experienced nine cases of patent LITA grafts, whose intra-operative mean flow was less than 20 ml/min (4, 6, 7, 8, 11, 13, 13, 15, 18 ml/min, respectively), as demonstrated in Figures 3 and 5. It is possible to judge an anastomosed graft with intraoperative flow of more than 20 ml/min as patent well. However, a surgeon can not assess a graft with a flow of less than 20 ml/min as either stenotic or occluded in the operating room. In short, it is impossible for a surgeon to evaluate anastomotic quality in CABG based upon the value of mean flow itself. As Jaber and associates demonstrated in their in vivo experiments [12], an optimal mean graft flow can not be defined because of the dynamic character 502

of the variables affecting graft flow, including blood pressure, heart rate, coronary resistance, and graft diameter. We also speculate that the degree of stenosis of the native coronary artery employed in anastomosis may affect flow values, particularly in the case of in situ arterial grafts rather than aortocoronary grafts. We do not have cut-off values for pulsatility index and % insufficiency to intra-operatively distinguish patent from stenotic or occluded grafts. Power spectral analysis of the graft flow tracing by FFT, which has been applied in several studies of hemodynamics to transform a waveform from the time domain to the frequency domain [13], has resulted in an interesting finding. It was possible in this study to define a cut-off value in the FFT ratio for distinguishing patent from stenotic or occluded grafts intra-operatively. As shown in Figure 3, the ratios of all the patent grafts enrolled in this study were greater than 1.0, while the ratios of all stenotic or occluded grafts were less than 1.0. The power spectral analysis of graft flow may reflect the flow tracing morphology. Based upon the specific physiology of coronary circulation, patent graft flow is predominantly diastolic, forming a trapezoid-shaped waveform, with a short systolic peak, as in Figures 1 and 2. In contrast, there is no diastolic flow in an occluded graft. As anastomotic stenosis increases, the predominance of the graft diastolic flow may decrease [12]. It is possible that the degree of the predominance of the graft diastolic flow is closely associated with the distribution of power spectral in FFT analysis of the flow curve. The patent graft flow tracing with a trapezoid-shaped diastolic flow may result from gradual decrease in power of the harmonics of the fundamental frequency in FFT analysis, which is represented by a FFT ratio of greater than 1.0. One limitation of the present study was that the correlation cited above between flow tracing and its power spectral analysis has not been fully investigated from the viewpoint of mathematics and physics. Another limitation was the difficulty in obtaining intra-operative flow profiles of a right gastroepiploic artery graft by using a flow probe of 3 or 4 mm. To generalize our findings, we must collect the data of gastroepiploic artery grafts by using a larger probe or by skeltonizing the graft. In conclusion, no parameters of the flow tracing itself can distinguish patent from stenotic/occluded grafts in intra-operative flow measurement. However, power spectral analysis of the flow waveform using FFT is considered to be quite useful for intraoperative differential diagnosis of the anastomotic quality in CABG.

References 1. Eagle, K. A., Guyton, R. A., Davidoff, R. et al., ACC/AHA guidelines for coronary artery bypass graft surgery: executive

CARDIOVASCULAR SURGERY

OCTOBER 2001 VOL 9 NO 5

Intra-operative flow analysis to evaluate CABG anastomotic: Yoshiyuki Takami and Hiroshi Ina

2.

3.

4.

5.

6.

7.

summary and recommendations. Circulation, 1999, 100, 1464– 1480. Yusus, F., Zucker, D., Peduzzi, P. et al., Effect of coronary artery bypass graft surgery on survival: overview of 10-year results from randomized trials by Coronary Artery Bypass Graft Surgery Trialist Collaboration. Lancet, 1994, 344, 563–570. Benetti, F. J., Naselli, G., Wood, M. et al., Direct myocardial revascularization without extracorporeal circulation: experience in 700 patients. Chest, 1991, 100, 312–316. Buffolo, E., de Andrade, J. C. S., Branco, J. N. R. et al., Coronary artery bypass grafting without cardiopulmonary bypass. Annals of Thoracic Surgery, 1996, 61, 63–66. Baumgartner, F. J., Gheissari, A., Capouya, E. R. et al., Technical aspects of total revascularization in off-pump coronary bypass via sternotomy approach. Annals of Thoracic Surgery, 1999, 67, 1653–1658. Louagie, Y. A. G., Haxhe, J. P., Buche, M. et al., Intraoperative electromagnetic flowmeter measurements in coronary artery bypass grafts. Annals of Thoracic Surgery, 1994, 57, 357–364. Canver, C. C. and Dame, N. A., Ultrasonic assessment of internal thoracic artery graft flow in the revascularized heart. Annals of Thoracic Surgery, 1994, 58, 135–138.

CARDIOVASCULAR SURGERY

OCTOBER 2001 VOL 9 NO 5

8. Oda, K., Hirose, K., Nishimori, H. et al., Assessment of internal thoracic artery graft with intraoperative color doppler ultrasonography. Annals of Thoracic Surgery, 1998, 66, 79–81. 9. Belboul, A., Radberg, G., Roberts, D. et al., Intraoperative assessment of coronary flow and coronary vascular resistance during coronary bypass surgery. Scandinavian Cardiovascular Journal, 1999, 33, 23–28. 10. Walpoth, B. H., Bosshard, A., Genyk, I. et al., Transit-time flow measurement for detection of early graft failure during myocardial revascularization. Annals of Thoracic Surgery, 1998, 66, 1097–1100. 11. Milnor, W. R. Hemodynamics, 2nd edn, Williams & Wilkins, Baltimore, 1989. 12. Jaber, S. F., Koenig, S. C., BhaskerRao, B. et al., Role of graft flow measurement technique in anastomotic quality assessment in minimally invasive CABG. Annals of Thoracic Surgery, 1998, 66, 1087–1092. 13. Chen, E. P., Bittner, H. B., Craig, D. M. et al., Pulmonary hemodynamics and blood flow characteristics in chronic pulmonary hypertension. Annals of Thoracic Surgery, 1997, 63, 806–813. Paper accepted 26 March 2001

503