Does competitive flow reduce internal thoracic artery graft patency?

CARDIOVASCULAR

Does Competitive Flow Reduce Internal Thoracic Artery Graft Patency? Joseph F. Sabik III, MD, Bruce W. Lytle, MD, Eugene H. Blackstone, MD, Mohammad Khan, MD, Penny L. Houghtaling, MS, and Delos M. Cosgrove, MD Departments of Thoracic and Cardiovascular Surgery and Biostatistics and Epidemiology, The Cleveland Clinic Foundation, Cleveland, Ohio

Background. In coronary arteries with moderate stenosis, competitive flow may lead to internal thoracic artery (ITA) graft occlusion. The goals of this study were to determine if competitive flow reduces ITA patency, and if there is a degree of coronary stenosis below which ITAs should not be used. Methods. From 1972 to 1999, 50,278 patients underwent primary coronary artery bypass grafting (CABG). Of these, 2,002 had at least one ITA graft and postoperative angiography before coronary reintervention; 2,999 angiograms of 2,121 ITAs were made. Time-related ITA occlusion was modeled using longitudinal analysis to identify its risk factors while accounting for lack of independence introduced by repeated angiography and multiple ITA anastomoses per patient. Proximal coronary stenosis (maximum preoperative stenosis between ITA anastomosis and aorta) was the surrogate for competitive flow.

Results. Unadjusted ITA patency was 93%, 89%, 90%, and 92% at 1, 5, 10, and 15 years after CABG. Risk factors associated with ITA occlusion were lesser degree of proximal coronary stenosis (p < 0.0001); longer time from CABG in grafts to non-left anterior descending coronary arteries (p < 0.0001); female sex (p ⴝ 0.0003); later date of CABG (p ⴝ 0.01); right ITA (p < 0.0001); and smoking (p < 0.0001). In all arteries, as preoperative proximal coronary stenosis decreased, ITA patency declined; however, at no degree of stenosis was there a sharp decline. Conclusions. Internal thoracic artery patency decreases as coronary competitive flow increases. However, the nature of this relationship indicates ITAs should not be abandoned at moderate grades of stenosis.

B

ative angiographic data were available describing location and severity of native coronary arteriosclerosis; (3) at least one ITA was used as a bypass graft; (4) at least one postoperative angiogram was performed before any repeat coronary intervention; and (5) detailed information was available on patency and stenosis of ITA grafts. During this period, 50,278 patients underwent primary CABG, 2,002 of whom (Table 1) met the above criteria. Preoperative, operative, and postoperative variables (Appendix 1) were retrieved from the Cardiovascular Information Registry, and use of these data for research was approved by the Institutional Review Board.

ecause of its excellent long-term patency, beneficial effect on survival, and resistance to arteriosclerosis, the internal thoracic artery (ITA) is the preferred conduit for coronary artery bypass grafting (CABG) [1, 2]. Controversy exists, however, as to whether ITAs should be used to bypass coronary arteries with noncritical stenoses, because competitive blood flow through them may result in ITA graft constriction and occlusion [3–11]. The objectives of this study were to determine if competitive blood flow reduces ITA graft patency, and if there is a degree of coronary artery stenosis below which ITA graft patency is so severely reduced that ITA grafts should be avoided.

Patients and Methods Patients Using our computerized Cardiovascular Information Registry at The Cleveland Clinic Foundation, patients were identified and included in this study if the following criteria were met: (1) they had primary CABG at the Cleveland Clinic from 1972 to 1999; (2) detailed preoperPresented at the Thirty-ninth Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Jan 31–Feb 2, 2003. Address reprint requests to Dr Sabik, Department of Thoracic and Cardiovascular Surgery, The Cleveland Clinic Foundation, 9500 Euclid Avenue, Desk F24, Cleveland, OH 44195; e-mail: [email protected].

© 2003 by The Society of Thoracic Surgeons Published by Elsevier Inc

(Ann Thorac Surg 2003;76:1490 –7) © 2003 by The Society of Thoracic Surgeons

Postoperative Angiography Patients underwent postoperative coronary angiography for a variety of reasons. Early in the series the procedure was often for routine monitoring of grafts; however, by far the most common indication was recurrence of ischemic symptoms. There were 2,999 angiographic studies from 1972 to 2001; 1,367 patients had one postoperative angiogram, 472 had two, 128 had three, 25 had four, and 10 had five. These angiograms provided assessment of 2,476 ITA grafts to the left anterior descending coronary artery (LAD), 291 grafts to diagonals, 30 grafts to the right coronary artery (RCA), 83 grafts to the posterior descending artery (PDA), and 119 grafts to marginal circumflex branches. Angiography was performed as early as the 0003-4975/03/$30.00 doi:10.1016/S0003-4975(03)01022-1

SABIK ET AL ITA GRAFT PATENCY

1491

same day as CABG to as late as 24.9 years afterward (Fig 1).

Internal Thoracic Artery Grafts A total of 2,121 ITAs were used in these 2,002 patients: 2086 in situ and 35 free grafts. Of the 2,121 grafts, 2,024 were left ITA grafts (2006 in situ and 18 free) and 97 were right ITA grafts (80 in situ and 17 free). The ITA was used to bypass the LAD in 1,757 patients, diagonals in 198 patients, obtuse marginal branches of the circumflex artery in 86 patients, RCA in 24 patients, and PDA in 56 patients. No ITA grafts studied were performed to the main circumflex artery.

Competitive Flow Six general sites for ITA graft anastomoses were defined: LAD, diagonal, circumflex, marginal circumflex, right coronary, and posterior descending. Maximum preoperative proximal coronary stenosis between ITA anastomosis at these sites and aorta was the surrogate for competitive flow. For example, for the PDA in a right dominant system, maximum stenosis in either the RCA or PDA was used as the surrogate for competitive flow (proximal

Table 1. Patient Characteristics Characteristics Demography Age, mean ⫾ SD (years) Male Noncardiac medical history Diabetes Peripheral vascular disease Smoking history Hypertension Left ventricular dysfunction Normal/none Mild Moderate Severe Not recorded History of myocardial infarction Coronary artery diseaseb Left main LAD LCX RCA Number of diseased systemsb 0 1 2 3

n 53.8 ⫾ 8.8 1,765

%

88

Fig 1. Number of ITA grafts observed angiographically each year after CABG. (CABG ⫽ coronary artery bypass grafting; ITA ⫽ internal thoracic artery.)

coronary stenosis for the PDA site accounted for dominance). Similarly, for the LAD site, maximum stenosis in either the left main or LAD was used. Because in these patients no ITA was sited directly on the circumflex coronary artery, this site was not available for analysis. Internal thoracic artery grafts were used to bypass 63 (3%) coronary arteries with less than 50% proximal stenosis, 536 (25%) with 50% to 70% stenosis, 1,124 (53%) with 71% to 99% stenosis, and 394 (19%) totally occluded coronary arteries.

Patency

164/1,630a 148/1,995 840/1,662 632/1,178

10 7 50 54

An ITA graft was defined as patent as long as it was not occluded. Graft occlusion was the study end point for analysis.

1,106 590 238 64 4 920/1,967

55 30 12 3

213 1,915 1,460 1,550

11 96 73 78

6c 228 595 1,169

0.3 11 30 58

Identification of risk factors for ITA occlusion is challenging because it must account for lack of independence on two hierarchical levels of aggregation: multiple angiograms per patient (repeated measures) and possibly multiple ITA anastomoses in each patient (within-patient variability) [12]. To address this challenge, multivariable analysis of ITA occlusion used longitudinal logistic repeated-measures analysis (SAS PROC GENMOD) (Appendix 1). Variables considered in the multivariable analyses of predictors of graft occlusion are listed in Appendix 2. Because of the limited capability of PROC GENMOD to explore multivariable relations, variables were initially screened by multivariable logistic regression using liberal inclusion criteria. Then these candidate variables were analyzed simultaneously, eliminating one by one those not significant at the p less than 0.05 level. Because graft patency generally declines with longer time from CABG, specific investigation of the manner by which this temporal phenomenon was hastened or slowed was required. Investigation of modulation of temporal rates was achieved by forming interactions of all variables with interval between CABG and angiogra-

a

Denominator is different from 2,002 because these variables were not included in the Registry during the entire study period. Sporadic missing b data are also evident. Defined as 50% lesion in the coronary system. c Six patients had left main trunk disease only. These are not coded anatomically as multisystem disease. LAD ⫽ left anterior descending coronary artery system; LCX ⫽ left circumflex coronary artery system; RCA ⫽ right coronary artery system; SD ⫽ standard deviation.

Statistical Analysis

CARDIOVASCULAR

Ann Thorac Surg 2003;76:1490 –7

CARDIOVASCULAR

1492

SABIK ET AL ITA GRAFT PATENCY

phy. Because of the complexity of such an investigation, initially data were modeled for each ITA anastomotic site separately; these results were then combined to formulate the overall model. Transformations of scale were investigated for ordinal and continuous variables to calibrate risk factors with outcome. To investigate in particular whether there was a degree of proximal coronary stenosis below which ITA patency was severely reduced, we sought a possible nonlinear (steep decline) relation between degree of proximal coronary stenosis and graft occlusion. Graphical presentation of results is for mean effects, after considering the two sources of within-patient variability. Graphical solutions of the longitudinal equation are risk adjusted by holding values for variables in the model constant, as described in figure legends.

Results Overall Internal Thoracic Artery Patency Unadjusted ITA graft patency was 93%, 89%, 90%, and 92% at 1, 5, 10, and 15 years after CABG, respectively. At the time of the 2,999 assessments of ITA patency, 82% of ITA grafts had no stenosis; 3.1% had 1% to 50% stenosis; 2.2% had 51% to 70% stenosis; 3.3% had 71% to 99% stenosis; and 9.8% were occluded.

Risk Factors for Internal Thoracic Artery Occlusion Multivariable analysis identified six factors associated with ITA graft occlusion: (1) degree of preoperative proximal coronary stenosis, (2) time from CABG in nonLAD arteries, (3) sex, (4) date of surgery, (5) laterality of ITA, and (6) smoking status (Table 2).

Influence of Competitive Flow on Internal Thoracic Artery Patency Lesser degree of preoperative proximal coronary stenosis was associated with increased risk of ITA graft occlusion (Table 3 and Fig 2). This was true for all coronary arteries. However, at least down to a stenosis of 40%, there was no degree of proximal coronary stenosis below which ITA graft patency sharply declined (see Fig 2). Internal thoracic artery to LAD grafts had better patency than those to any other coronary artery (Tables 2 and 3 and Fig 3). Similar ITA patency was found in grafts to diagonal, obtuse marginal branches of the circumflex, posterior descending, and right coronary arteries (p ⫽ 0.4 compared with LAD), although few grafts were anastomosed to the RCA. Longer interval from CABG to angiography was associated with increased ITA occlusion in grafts performed to all coronary arteries except the LAD (Table 3 and Fig 2). This effect of time was magnified in ITA grafts to coronary arteries with lesser degrees of proximal coronary stenosis (see Table 3 and Fig 2), whereas the effect of time on ITA to LAD grafts was negligible (p ⫽ 0.5). This difference in graft occlusion behavior was unlikely to be due to chance (p ⬍ 0.0001).

Ann Thorac Surg 2003;76:1490 –7

Table 2. Risk Factors Identified by Multivariable Analysis as Associated With Internal Thoracic Artery Graft Occlusion Risk Factor

Estimate ⫾ SE

p

Degree of proximal stenosis Time from CABG Non-LAD LAD Female Right ITA Date of CABG Smoking

⫺ 2.0 ⫾ 0.23

⬍.0001

0.084 ⫾ .018 ⫺.010 ⫾ 0.015 0.65 ⫾ 0.18 1.13 ⫾ 0.23 0.22 ⫾ 0.088 0.21 ⫾ 0.15

⬍.0001 .5 ⬍.0003 ⬍.0001 .01 ⬍.0001

Direction Lesser Longer

Later

⫺2.07

Intercept

CABG ⫽ coronary artery bypass grafting; ITA ⫽ internal thoracic artery; LAD ⫽ left anterior descending coronary artery; SE ⫽ standard error.

Influence of Other Factors on Internal Thoracic Artery Patency Internal thoracic artery patency in grafts to both LAD and non-LAD coronary arteries was lower in women than men (Fig 4). Also, right ITA grafts had decreased patency to both LAD and non-LAD coronary arteries compared with left grafts (Fig 5). Earlier date of CABG was associated with increased ITA patency (Table 2), whereas smoking was associated with increased ITA occlusion (Fig 6).

Comment Background Controversy exists as to whether ITAs should be used to bypass coronary arteries with moderate stenosis. Multiple reports document that ITAs bypassing coronary arteries with high competitive flow have a tendency to constrict and fail [3–11]. This observation was first made by Barner in 1974, who found that 11% of ITA grafts failed due to high competitive flow; he introduced the term “disuse atrophy” to characterize the phenomenon [3]. Geha and Baue also reported failure of ITA grafts in coronary arteries without severe stenosis, referring to the Table 3. Time-Related Change in Internal Thoracic Artery Graft Patency According to Degree of Proximal Coronary Stenosis in LAD and Non-LAD Arteriesa Degree of Proximal Coronary Stenosis Year After CABG 1 5 10 15

LAD

Non-LAD

50%

70%

90%

100%

50%

70%

90%

100%

92 92 93 93

95 95 95 95

97 97 97 98

98 98 98 98

91 88 83 76

94 92 89 84

97 98 94 91

98 97 96 93

a

Values are percent patency. Percentages are based on longitudinal model presented in Table 2. Data are for male nonsmokers with left ITA grafts; date of operation ⫽ 1974.

CABG ⫽ coronary artery bypass grafting; ITA ⫽ internal thoracic artery; LAD ⫽ left anterior descending coronary artery.

SABIK ET AL ITA GRAFT PATENCY

1493

Fig 3. Estimates of ITA patency of left ITA to LAD and non-LAD coronary arteries at 10 years after CABG, according to degree of proximal coronary stenosis. Solid lines are point estimates; dashed lines are 70% confidence limits. Graph depicts male nonsmokers; date of operation ⫽ 1974 (see Table 2). (CABG ⫽ coronary artery bypass grafting; ITA ⫽ internal thoracic artery; LAD ⫽ left anterior descending coronary artery.)

chronic canine model, demonstrated that ITA flow was not affected by competitive flow, even when the ITA was grafted to a completely patent coronary artery.

Principal Findings

Fig 2. Estimates of 1-, 5-, 10-, and 15-year patency of ITA grafts according to degree of proximal coronary artery stenosis. (A) ITA to LAD. (B) ITA to non-LAD coronary arteries. Graphs depict male nonsmokers with left ITA grafts; date of operation ⫽ 1974 (see Table 2). (ITA ⫽ internal thoracic artery; LAD ⫽ left anterior descending coronary artery.)

angiographic appearance of constricted ITA grafts as the “distal thread phenomenon” [4]. They concluded that a marginal lesion in a coronary artery should be a contraindication to ITA grafting [4]. From angiographic studies in 99 patients, Ivert and colleagues reported that ITA graft failures were due to either technical errors during operation or grafting the ITA to coronary arteries with low-grade stenosis [5]. However, others have not observed an association between degree of proximal coronary stenosis and graft patency [13–19]. In a previous study from our institution, Cosgrove and colleagues reported that ITA grafts performed to coronary arteries with less than 50% stenosis had similar patency to those with greater stenosis [13]. Urschel and colleagues reported that intraoperative angioplasty of proximal stenoses did not affect ITA patency, despite increased competitive flow [14]. In addition to these clinical observations, Spence and colleagues [15] in an acute canine model, and Lust and colleagues [16] in a

In this study, ITA graft patency decreased slightly as proximal coronary stenosis between aorta and anastomotic site decreased. Down to a stenosis of 40%, there was no degree of proximal coronary stenosis below which ITA patency showed a marked decline. Too few ITA grafts were performed to coronary arteries with less than a 40% stenosis to determine with certainty the effect of competitive flow on ITA grafts to these vessels.

Fig 4. Estimates of ITA patency in men and women at 10 years after CABG for grafts to LAD and non-LAD coronary arteries, according to degree of proximal coronary artery stenosis. Graph depicts nonsmokers with left ITA grafts; date of operation ⫽ 1974 (see Table 2). (CABG ⫽ coronary artery bypass grafting; ITA ⫽ internal thoracic artery; LAD ⫽ left anterior descending coronary artery.)

CARDIOVASCULAR

Ann Thorac Surg 2003;76:1490 –7

CARDIOVASCULAR

1494

SABIK ET AL ITA GRAFT PATENCY

Ann Thorac Surg 2003;76:1490 –7

in 9.5% (14/147) of ITA grafts studied 16 days to 62 months after CABG [7]. Two (14%) of these failed ITA grafts bypassed LADs with more than 50% stenosis, whereas 12 (86%) bypassed LADs with 50% or less proximal stenosis.

Secondary Findings

Fig 5. Left (L) and right (R) ITA patency at 10 years after CABG to both LAD and non-LAD coronary arteries, according to degree of proximal coronary stenosis. Graph depicts male nonsmokers; date of operation ⫽ 1974 (see Table 2). (CABG ⫽ coronary artery bypass grafting; ITA ⫽ internal thoracic artery; LAD ⫽ left anterior descending coronary artery.)

These findings are consistent with the physiology of arterial grafts. Internal thoracic arteries are able to autoregulate their size and blood flow in response to demand [6 –11]. As proximal coronary stenosis decreases, competitive flow increases, and demand for ITA graft flow falls. This cascade of events results in ITA constriction and, over time, increased risk of atrophy and occlusion. These findings of a mild effect of competitive flow on ITA graft patency have been reported by others [6 –11, 19]. Kawasuji and colleagues performed angiography 1 month after CABG in 100 patients with ITA to LAD grafts and observed all grafts were patent, but 15% (2/13) of those performed to coronary arteries with 50% or less stenosis were severely constricted [19]. Seki and colleagues observed either severe constriction or occlusion

Internal thoracic artery patency was highest and most durable in grafts performed to the LAD. These findings are consistent with previous reports on vein grafts, in which patency of grafts to the LAD exceeded that to the left circumflex artery and RCA [20 –23]. This may be because it is easier technically to graft anterior coronary arteries, and the amount of myocardium being supplied by the LAD is greater than that supplied by other coronary arteries, resulting in a larger blood flow demand being placed on ITA grafts to the LAD. Internal thoracic artery grafts with greater blood flow demand are less likely to fail [6 –11]. Earlier date of CABG was associated with better ITA patency. Early in this series, patients were more likely to have undergone postoperative angiography for routine monitoring of graft status, whereas later in the series the reason was almost exclusively recurrence of ischemic symptoms. Previous reports have demonstrated lower graft patency in symptomatic patients [24, 25]. Left ITA grafts had better patency than right ITA grafts, perhaps due to the course that right ITA grafts must take through the mediastinum to reach the heart. Because of their smaller size, women have smaller coronary arteries than men. Technical difficulties associated with grafting small arteries are believed to be one possible cause of higher operative risk observed in women, and may also be responsible for lower graft patency [26, 27]. Smoking is strongly associated with progression of coronary artery disease, and patients who continue to smoke after CABG have a higher risk of return of angina, myocardial infarction, and coronary reintervention [28 – 32]. In addition, Hashimoto and colleagues found by multivariable analysis that a history of smoking decreases ITA graft diameter [6]. These effects on both coronary arteries and ITA grafts probably account for the lower graft patency we observed in smokers.

Limitations Patients underwent postoperative angiography for a variety of indications, but mostly for recurrence of ischemic symptoms. This selectivity is expected to result in lower estimates of graft patency than if all patients had been studied. Evidence for this assumption is the high patency of grafts among patients who underwent routine angiography early in this study. However, although absolute graft patency reported is probably lower than actual, the factors found to influence patency should not be affected. A prospective study with routine angiography at specified intervals would be the best method of determining actual graft patency. Although in some prospective studies high proportions of patients have undergone early ANGIOGRAPHIC FOLLOW-UP.

Fig 6. Left ITA patency at 10 years after CABG in male smokers and nonsmokers, according to degree of proximal coronary stenosis. Graph depicts date of operation ⫽ 1974 (see Table 2). (CABG ⫽ coronary artery bypass grafting; ITA ⫽ internal thoracic artery; LAD ⫽ left anterior descending coronary artery.)

angiography, by 5 years patient dropout due to death, reoperation, and refusal alter the characteristics of the remaining population in nonrandom ways [20, 33–35]. Despite this study not being prospective, unadjusted ITA patency was similar to that previously reported [35–38]. Patency in this study (and all others) represents a form of cross-sectional assessment. We do not know when between CABG and angiography, or between angiographic studies, graft occlusion actually occurred. COMPETITIVE FLOW. We used maximum degree of proximal coronary stenosis as a surrogate for competitive flow. However, other factors, such as collateral blood flow from other coronary arteries, diameter and distal run-off of the grafted coronary artery, and serial proximal stenoses affect competitive flow as well. In the absence of direct measurements of coronary blood flow, degree of proximal coronary stenosis was the only means available to us to estimate competitive flow. We related a dichotomous end point, ITA graft occlusion, to competitive flow. If a wide range of stenotic ITA values had been observed, a more powerful study would have been of degree of ITA stenosis. As our results indicated, however, grafts were either completely open or occluded in 92% of patients; thus, variability in degree of stenosis was small, resulting in little information loss from dichotomizing the end point.

SABIK ET AL ITA GRAFT PATENCY

9.

10.

11. 12. 13. 14.

15. 16. 17.

18.

Conclusion Internal thoracic artery graft patency decreases as coronary artery competitive flow increases. However, the effect of competitive flow on ITA graft patency is mild, and at no degree of proximal coronary stenosis is there a severe decline in ITA patency. This finding suggests that ITAs should not be avoided when bypassing coronary arteries with moderate degrees of stenoses.

19.

20.

21.

References 1. Loop FD, Lytle BW, Cosgrove DM, et al. Influence of the internal-mammary-artery graft on 10-year survival and other cardiac events. N Engl J Med 1986;314:1–6. 2. Lytle BW, Loop FD, Cosgrove DM, et al. Fifteen hundred coronary reoperations. Results and determinants of early and late survival. J Thorac Cardiovasc Surg 1987;93:847–59. 3. Barner HB. Double internal mammary– coronary artery bypass. Arch Surg 1974;109:627–30. 4. Geha AS, Baue AE. Early and late results of coronary revascularization with saphenous vein and internal mammary artery grafts. Am J Surg 1979;137:456 –63. 5. Ivert T, Huttunen K, Landou C, Bjork VO. Angiographic studies of internal mammary artery grafts 11 years after coronary artery bypass grafting. J Thorac Cardiovasc Surg 1988;96:1–12. 6. Hashimoto H, Isshiki T, Ikari Y, et al. Effects of competitive blood flow on arterial graft patency and diameter. Mediumterm postoperative follow-up. J Thorac Cardiovasc Surg 1996;111:399 –407. 7. Seki T, Kitamura S, Kawachi K, et al. A quantitative study of postoperative luminal narrowing of the internal thoracic artery graft in coronary artery bypass surgery. J Thorac Cardiovasc Surg 1992;104:1532–8. 8. Pagni S, Storey J, Ballen J, et al. ITA versus SVG: a compar-

22. 23.

24.

25.

26.

27.

28.

1495

ison of instantaneous pressure and flow dynamics during competitive flow. Eur J Cardiothorac Surg 1997;11:1086 –92. Shimizu T, Hirayama T, Suesada H, Ikeda K, Ito S, Ishimaru S. Effect of flow competition on internal thoracic artery graft: postoperative velocimetric and angiographic study. J Thorac Cardiovasc Surg 2000;120:459 –65. Pagni S, Storey J, Ballen J, et al. Factors affecting internal mammary artery graft survival: how is competitive flow from a patent native coronary vessel a risk factor? J Surg Res 1997;71:172–8. Nasu M, Akasaka T, Okazaki T, et al. Postoperative flow characteristics of left internal thoracic artery grafts. Ann Thorac Surg 1995;59:154 –62. Diggle PJ, Liang KY, Zeger SL. Analysis of longitudinal data. New York: Oxford University Press, 1994. Cosgrove DM, Loop FD, Saunders CL, Lytle BW, Kramer JR. Should coronary arteries with less than fifty percent stenosis be bypassed? J Thorac Cardiovasc Surg 1981;82:520 –30. Urschel HC Jr, Razzuk MA, Miller E, Chung SY. Operative transluminal balloon angioplasty. Adjunct to coronary bypass for extended myocardial revascularization of more than 3000 lesions in 1000 patients. J Thorac Cardiovasc Surg 1990;99:581–9. Spence PA, Lust RM, Zeri RS, et al. Competitive flow from a fully patent coronary artery does not limit acute mammary graft flow. Ann Thorac Surg 1992;54:21–6. Lust RM, Zeri RS, Spence PA, et al. Effect of chronic native flow competition on internal thoracic artery grafts. Ann Thorac Surg 1994;57:45–50. Gaudino M, Bruno P, Piancone FL, Pragliola C, Manzoli A, Possati G. Internal mammary artery grafts and competitive flow: controversies persist. J Cardiovasc Surg (Torino) 1999; 40:553–4. Barron DJ, Livesey SA. Patency of an internal thoracic artery graft despite maximal competitive flow. Ann Thorac Surg 1995;59:1556 –7. Kawasuji M, Sakakibara N, Takemura H, Tedoriya T, Ushijima T, Watanabe Y. Is internal thoracic artery grafting suitable for a moderately stenotic coronary artery? J Thorac Cardiovasc Surg 1996;112:253–9. Frey RR, Bruschke AV, Vermeulen FE. Serial angiographic evaluation 1 year and 9 years after aorta– coronary bypass. A study of 55 patients chosen at random. J Thorac Cardiovasc Surg 1984;87:167–74. Gohlke H, Gohlke-Barwolf C, Sturzenhofecker P, et al. Improved graft patency with anticoagulant therapy after aortocoronary bypass surgery: a prospective, randomized study. Circulation 1981;64:II22–7. Lytle BW, Loop FD. Elective coronary surgery. Cardiovasc Clin 1981;12:31–47. Higginbotham M, Hunt D, Stuckey J, Sloman G. Prospective angiographic assessment of factors affecting early patency of saphenous vein– coronary artery bypass grafts. Aust N Z J Med 1980;10:295–9. Lawrie GM, Morris GC Jr, Chapman DW, Winters WL, Lie JT. Patterns of patency of 596 vein grafts up to seven years after aorta– coronary bypass. J Thorac Cardiovasc Surg 1977; 73:443–8. Lytle BW, Loop FD, Thurer RL, Groves LK, Taylor PC, Cosgrove DM. Isolated left anterior descending coronary atherosclerosis: long-term comparison of internal mammary artery and venous autografts. Circulation 1980;61:869 –74. Loop FD, Golding LR, MacMillan JP, Cosgrove DM, Lytle BW, Sheldon WC. Coronary artery surgery in women compared with men: analyses of risks and long-term results. J Am Coll Cardiol 1983;1:383–90. Fisher LD, Kennedy JW, Davis KB, et al. Association of sex, physical size, and operative mortality after coronary artery bypass in the Coronary Artery Surgery Study (CASS). J Thorac Cardiovasc Surg 1982;84:334 –41. Voors AA, van Brussel BL, Plokker HW, et al. Smoking and cardiac events after venous coronary bypass surgery. A 15-year follow-up study. Circulation 1996;93:42–7.

CARDIOVASCULAR

Ann Thorac Surg 2003;76:1490 –7

CARDIOVASCULAR

1496

SABIK ET AL ITA GRAFT PATENCY

29. Weintraub WS, Klein LW, Seelaus PA, Agarwal JB, Helfant RH. Importance of total life consumption of cigarettes as a risk factor for coronary artery disease. Am J Cardiol 1985;55: 669 –72. 30. FitzGibbon GM, Leach AJ, Kafka HP. Atherosclerosis of coronary artery bypass grafts and smoking. CMAJ 1987;136: 45–7. 31. Solymoss BC, Nadeau P, Millette D, Campeau L. Late thrombosis of saphenous vein coronary bypass grafts related to risk factors. Circulation 1988;78:I140 –3. 32. Cavender JB, Rogers WJ, Fisher LD, Gersh BJ, Coggin CJ, Myers WO. Effects of smoking on survival and morbidity in patients randomized to medical or surgical therapy in the Coronary Artery Surgery Study (CASS): 10-year follow-up. CASS Investigators. J Am Coll Cardiol 1992;20:287–94. 33. Campeau L, Enjalbert M, Lesperance J, Vaislic C, Grondin CM, Bourassa MG. Atherosclerosis and late closure of aortocoronary saphenous vein grafts: sequential angiographic studies at 2 weeks, 1 year, 5 to 7 years, and 10 to 12 years after surgery. Circulation 1983;68:II1–7. 34. Chesebro JH, Fuster V, Elveback LR, et al. Effect of dipyridamole and aspirin on late vein-graft patency after coronary bypass operations. N Engl J Med 1984;310:209 –14. 35. Singh RN, Sosa JA, Green GE. Long-term fate of the internal mammary artery and saphenous vein grafts. J Thorac Cardiovasc Surg 1983;86:359 –63. 36. Grondin CM. Late results of coronary artery grafting: is there a flag on the field? J Thorac Cardiovasc Surg 1984;87: 161–6. 37. Tector AJ, Schmahl TM, Janson B, Kallies JR, Johnson G. The internal mammary artery graft. Its longevity after coronary bypass. JAMA 1981;246:2181–3. 38. Barner HB, Swartz MT, Mudd JG, Tyras DH. Late patency of the internal mammary artery as a coronary bypass conduit. Ann Thorac Surg 1982;34:408 –12. 39. Blackstone EH. Breaking down barriers: helpful breakthrough statistical methods you need to understand better. J Thorac Cardiovasc Surg 2001;122:430 –9. 40. Lytle BW, Cosgrove DM 3rd. Coronary artery bypass surgery. Curr Probl Surg 1992;29:733–807.

Appendix 1. Data and Analysis Two sources of variability, accounted for in our analysis, have rarely been considered in previous analyses of graft patency: (1) repeated angiographic follow-up in the same patient, and (2) multiple anastomoses within a single patient. Modern methods of longitudinal analysis appropriately manage these levels of aggregation [12, 39]. We elected not to address yet another level of aggregation: one or more coronary reinterventions within a patient. Among the original group of 50,278 patients, 9,519 underwent follow-up angiography; however, only 4,333 angiograms were performed before coronary reintervention. Thus, in the past we have reported more individual angiographic assessments than in this study because of including postreintervention studies [40]. In our deliberations about the current study, we were uncertain whether graft patency after reintervention should be managed analytically as a time-varying covariable or as a modulated renewal process (restarting patients anew at time of reintervention), or as an even more complex entity. Thus, we elected to deal cleanly with the primary operation and censor patients at reintervention.

Ann Thorac Surg 2003;76:1490 –7

Analytically, current methods of longitudinal analysis do not readily incorporate separate modeling of time to each observation, as is available for analysis of timerelated events. Instead, identification of differences in rate of occlusion must be managed by possibly oversimplified variable–time interactions. As the methodology develops, its versatility in this regard is likely to improve.

Appendix 2. Patient Variables Considered in Analyses Demographics Age, sex, height, weight, body surface area, body mass index.

Symptoms Functional class (highest angina or functional limitation, I-IV), emergent.

Left Ventricular Function Left ventricular dysfunction grade (none; mildly, moderately, or severely impaired), history of prior myocardial infarction.

Cardiac Comorbidity Atrial fibrillation.

Noncardiac Comorbidity Smoking history, carotid artery disease, peripheral vascular disease, hypertension, diabetes (insulin-treated, oral hypoglycemic), renal disease by history, chronic obstructive pulmonary disease, family history of coronary artery disease.

Coronary Disease Preoperative maximal degree of proximal coronary artery stenosis between aorta and six general sites of internal thoracic artery graft anastomoses: left anterior descending, diagonal, circumflex, circumflex marginal, right and posterior descending coronary arteries.

Preoperative Laboratory Data Hematocrit level, concentrations of creatinine, cholesterol, triglycerides, blood urea nitrogen, bilirubin.

Operation Year of surgery, emergent/nonemergent, free internal thoracic artery used, right internal thoracic artery used.

Graft Site Left anterior descending, diagonals, right coronary artery, circumflex, circumflex marginals, posterior descending arteries.

Angiography Time from coronary artery bypass grafting to angiography.

SABIK ET AL ITA GRAFT PATENCY

1497

DISCUSSION DR SARA J. SHUMWAY (Minneapolis, MN): So if you have a patient with a proximal LAD lesion that’s 40%, would you use an internal thoracic artery (ITA) graft? DR SABIK: Some of these data are from early in our experience with coronary artery bypass surgery. We do not presently bypass coronary arteries with less than a 50% stenosis. DR SHUMWAY: But at 50% you would use it? DR SABIK: Yes. We would use an ITA to bypass a left anterior descending (LAD) coronary artery with a 50% proximal stenosis. DR JAMES TATOULIS (Melbourne, Australia): I enjoyed your presentation very much, and in general terms our group would concur. We found in more than 2,000 ITA graft angiograms that target vessel stenosis does affect the degree of patency. We have found that a 60% native coronary artery stenosis acts as the threshold for ITA grafts. Where greater than 60% stenosis is noted, the patency for LITA is 95%, and below 60% native coronary stenosis, the patency for LITA was 92%. These results were established during 15 years of follow-up. We have noticed a big difference in patency for the right internal mammary used in different territories. When we use the right internal mammary to the LAD, the patency is identical to the left internal mammary to the LAD. But when we use the right internal mammary to the right coronary system, the patency is much reduced, down to the 83% level. My question to you would be similar to Dr Shumway’s: if you have a young patient with a 50% stenosis in a 5-mm right coronary artery, what do you do? DR SABIK: As you know, our findings differ slightly from yours. We found that RITA grafts have a lower patency than LITA grafts, when used to graft both LAD and non-LAD coronary arteries. We also did not identify a degree of proximal coronary

artery stenosis below which there was a marked decrease in ITA graft patency. Internal thoracic artery graft patency decreased gradually as proximal coronary artery stenosis decreased. As to what conduit to use to bypass a right coronary artery with a moderate stenosis in a young patient, we prefer an arterial graft, such as a RITA. We do realize, however, that no perfect conduit exists for this situation. Because of competitive flow, long-term arterial graft patency will be lower, but probably still better, than late vein graft patency. DR TATOULIS: I would agree. Our philosophy has turned to using both of the mammaries to the left-sided vessels and an alternate graft for the right coronary system. DR SABIK: We agree with that grafting strategy as well. In the previous question, I thought you were referring to a patient with an isolated right coronary lesion. In patients with multivessel coronary artery disease, we prefer to use the two ITAs to bypass the left-sided coronary arteries prior to the right coronary artery. DR PIOTR BEDNARSKI (Nuernberg, Germany): Congratulation for an interesting and important presentation. Could you please tell us, what kind of anastomoses were investigated and what is the importance and difference of the end-to-side anastomosis versus the sequential graft? Also, did you observe any relation between the flow or the diameter of the internal mammaries used, the grade of the coronary artery stenosis, and the occlusion rate? DR SABIK: The vast majority of the ITA to coronary artery anastomoses in this series were performed as end-to-side anastomoses. However, there were some sequential grafts in the series. We do not have the data on ITA flow or diameter, and therefore I do not know how they affect ITA graft patency.

CARDIOVASCULAR

Ann Thorac Surg 2003;76:1490 –7