Late clinical and arteriographic results in 500 coronary artery reoperations

J

THORAC CARDIOVASC SURG

81:675-685, 1981

Late clinical and arteriographic results in 500 coronary artery reoperations The incidence o] coronary artery reoperations averaged 2 .7o/c[rom 1967 through 1979. In a mean interval oj"51 months between operations. three-vessel disease increased from 24o/c to 63%. and 31 % o] these 500 consecutive patients lost previously normal left ventricular function. Three angiographic indication groups were identified: (I) progressive coronary atherosclerosis. 257 (51 %): (2) graft failure, 147 (29%); and (3) a combination ofprogressive coronary atherosclerosis and graftfailure, 96 (19%). Angina recurred earlier in patients with graft failure. mean 17 months compared with a mean oj"37 monthsfor the other groups. Twentv (4%) operative deaths occurred. The series is divided into 387 patients operated upon under normothermic anoxic arrest and 113 with systemic hypothermia and cold cardioplegia. In the cardioplegia group. perioperative myocardial infarction was 2.7% in comparison with 7.8%filr patients with anoxic arrest (p = 0.055). The number ofgrafts per patient increased from 1.0 to 1,9 and blood usage decreased from II units to 2.7 units. After a meanfollow-up oj" 42 months. angina was relieved or improved in 86%. Recatheterization o] 104 patients after a mean interval or 19 months showed a 79% vein graft patency rate and a 97% mammary artery graft patency rate, Grafting performedfor graft failure (47) yielded an 85% patency rate. Actuarial 5 year survival was 87.4% tor those with progressive atherosclerosis. 89.4% for patients with graft failure, and 91.5% for the combined indication group. Clinical improvement. graft patency. and long-term survival are nearly equal among the indication groups. Palliation derivedfrom these reoperations approaches that achieved after primary revascularitution,

Floyd D. Loop, M.D.,* Delos M. Cosgrove, M.D,,* John R. Kramer, M.D.,** Bruce W. Lytle, M.D.,* Paul C. Taylor, M.D.,* Leonard A. R. Golding, M.D.,* and Laurence K. Groves, M.D.,* Cleveland, Ohio

T

he clinical benefit derived from revascularization procedures depends on extent, rate, and distribution of progressive atherosclerosis and fate of the graft(s). After the first year, a seven percent annual rate of angina recurrence is reported for the next five years and is related mainly to progressive disease in ungrafted arteries, early graft closure, and late graft obstruction from atherosclerosis. J Variations in the natural history

From the Department of Thoracic and Cardiovascular Surgery and the Department of Cardiology, The Cleveland Clinic Foundation, Cleveland, Ohio. Presented at the Fifty-third Annual Meeting of the American Heart Association. Miami. Fla., Nov. 18, 1980. Received for publication Sept. 4, 1980. Accepted for publication Dec. 12, 1980. Address for reprints: Floyd D. Loop, M.D., Department of Thoracic and Cardiovascular Surgery, The Cleveland Clinic Foundation, 9500 Euclid Ave., Cleveland, Ohio 44106. *Department of Thoracic and Cardiovascular Surgery.

**Department

of Cardiology.

of patients with coronary atherosclerosis complicate attempts to compare results of medical and surgical treatment. The physician who manages recurring angina postoperatively may decide that a second attempt at symptom relief is outweighed by the risk of a more complicated operation. Many surgeons were reluctant to perform technically difficult, frequently prolonged, high-risk procedures and few reoperations were undertaken until 1972. Accidents during sternal re-entry or during mobilization of the heart from adhesions increase risk; a higher incidence of perioperative myocardial infarction is acknowledged and other forms of morbidity, notably bleeding and respiratory complications, are reported. Finally, the paucity of information about survival and symptom relief after reoperations accentuates the uncertain value of a second heart operation and causes the referring physician to adopt a conservative attitude. Our experience with 500 coronary artery reoperations is reported herein to provide a perspective on changing risks, influence of myocardial protection,

0022-5223/811050675+11$01.1010 © 1981 The C. V. Mosby Co.

675

The Journal of

676

Thoracic and Cardiovascular Surgery

Loop et al.

Table I. Incidence of reoperations Year

Primary operation

Reoperation

Percent

2.9

1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979*

35 146 366 991 1,321 1,406 1,549 1,888 1,999 2,363 2,383 2,317 1,427

1 0 1 10 14 41 50 59 63 65 58 74 64

0.3 1.0 1.1 2.9 3.2 3.1 3.2 2.8 2.4 3.2 4.5

Totals

18,191

500

2.7

*Through July. 1979.

reoperative graft patency, late clinical status, and 5 year longevity. Clinical experience Whereas this series spans nearly 12 years, most of the 500 patients were operated upon between 1972 and mid-1979. This experience includes bypass graft reoperations only and does not include bypass grafts associated with valve repair or replacement, ventricular aneurysmectomy, or simultaneous carotid endarterectomy procedures. Other than two patients who were operated upon within 30 days of the first revascularization* and six who underwent two or more reoperations (see addendum), this series is consecutive. The number of patients who underwent reoperation each year and the number in that year who received primary coronary bypass grafts solely are shown in Table I. Overall, coronary artery reoperations account for 2.7% of the isolated myocardial revascularization experience; however, from 1972 through mid-1979 reoperations averaged 3.1 %. In 1979, the incidence of reoperations in our yearly revascularization experience increased to 4.5%. Of these 500 reoperative patients, 386 (77 .2%) were operated upon originally in this institution. Of 500 patients, 452 (90.4%) were men and 48 (9.6%) women. Ages ranged from 28 to 74 years (median 49) at the initial operation and 34 to 77 years (median 54) at reoperation. The interval between first *Two patients, both alive today, were operated upon for graft stenosis within 30 days of their first operation. The timing of the second operation is not germane to this report and the patients' data have been deleted.

and second myocardial revascularizations ranged from 2 to 178 months (median 48 and mean 51 months). Women were slightly older (median 51 years) than men (median 49 years) at the first operation, but the interval between procedures was the same. Functional angina before reoperation was classified according to the New York Heart Association (NYHA) classification. All patients except six (1.2%) experienced angina. In five of the six without angina, reoperation was undertaken because of progression to threevessel disease; two of the five had severe left main stenosis (estimated >60% narrowing). The sixth patient, who had isolated right coronary artery stenosis, experienced atrial flutter 9 months postoperatively. Abnormal exercise response was elicited and coronary angiography revealed graft stricture proximally. The stricture was not atherosclerotic and an interposed graft was placed between the aorta and the distal graft. Before reoperation 128 (25.6%) experienced NYHA Class II angina, 232 (46.4%) Class III, and 134 (26.8%) Class I V. When angina class at reoperation was compared to symptoms before the first operation, angina had advanced one or more classes in 180 patients (36.0%); 223 (44.6%) had angina of similar severity, and 97 (19.4%) reported less severe or no angina compared with the severity of angina before the first procedure. Angina recurred within 1 to 150 months after the first operation (mean 31.3 months). Seventy-one (14.2%) patients were diabetic before reoperation and 138 (27.6%) were treated for hypertension. Whereas 354 (70.8%) gave a history of cigarette smoking before the first operation, only 141 (28.2%) admitted smoking prior to reoperation. Of the 500 patients, 462 had serum cholesterol and 386 had serum triglyceride levels measured before reoperation. Eighty (17.3%) had cholesterol values above 300 mg/dl and 158 (40.9%) had triglyceride values above 200 mg/dl. The mean cholesterol value before reoperation was 251 mg/dl and the mean triglyceride value, 222 mg/dl. The range, mean, and median values of serum cholesterol and serum triglycerides were essentially the same as those found in our survey of myocardial revascularization from 1967 through 1978. 2 Angiographic indications for reoperation have been classified into three groups: (I) progressive coronary atherosclerosis, (2) graft closure, and (3) combined indications (first two groups). Of the 500 patients, 257 (51.4%) were reoperated upon because of progressive atherosclerosis in previously ungrafted coronary arteries or because of atherosclerosis that had progressed beyond the previous graft anastomosis. This latter finding occurred in 13 vessels in 12 patients. Graft

Volume 81 Number 5 May, 1981

closure occurred next in frequency; 147 (29.4%) underwent reoperation for this reason. Eighteen patients had graft failure because of vein graft atherosclerosis (documented by microscopic examination). In 96 (19.2%) more than one indication led to reoperation. Angina recurred earlier in the graft closure group (mean 17 months) than in the progressive atherosclerosis (mean 37 months) and combined groups (mean 36 months). Patients with progressive atherosclerosis had the longest interval between first and second operations (mean 64 months). Those who had more than one angiographic indication, the combined group, had the next longest interval between operations (mean 53 months). The graft closure group had the shortest interval between procedures (mean 28 months). The differences in intervals for the graft closure group versus the progressi ve atherosclerosis and combined indication groups were significant (p < 0.001). The extent of coronary atherosclerosis has been classified into the traditional one-, two-, or three-vessel groups. Table II summarizes the number of vessels narrowed an estimated 60% or greater in each indication group. Multivessel disease was more prevalent in those reoperated upon for progressive atherosclerosis or combined indications than in those in the graft closure group (p < 0.001). In 63 patients (12.6%) the left main coronary artery was narrowed by more than an estimated 60%, and the differences in prevalence between indication groups were insignificant. Of special interest was the indication group classified as progression of disease. To qualify as a progressive lesion, the arterial narrowing progressed at least an estimated 25% and advanced beyond the 60% reduction in lumen diameter arbitrarily considered severe obstruction. We reviewed the serial arteriograms to determine the original status of the arteries that showed progressive lesions. Four groups were identified: (1) major coronary arteries greater than I mm in diameter narrowed at least an estimated 60% but not grafted at first operation; (2) a normal coronary vessel that later showed an estimated narrowing of 60% or greater; (3) a mild lesion, estimated 20% to 30% reduction in lumen diameter, that progressed to an estimated 60% or greater narrowing; (4) a moderate lesion, estimated 40% to 60% reduction in lumen diameter, that progressed to greater than 60%. Of the 257 patients, 206 were classified in one of the above groups: Group 1, 87 patients (33.9%), of whom 68 had received a Vineberg implant in that arterial distribution; Group 2, 37 patients (14.4%); Group 3, 35 patients (13.6%); Group 4,47 patients (18.3%). Fiftyone patients were classified in two or three groups. Thirty-eight of these 51 (14.8%) were classified in

Coronary artery reoperations

677

Table II. Extent of coronary atherosclerosis greater than estimated 60% obstruction

Group Progressive disease Graft closure Combined indications

No. of patients

Vessels obstructed One

I

Two

I Three

257

8

78

171

147 96

33 0

44 23

70 73

Table III. Status of left ventricular contraction Left ventricular contraction

Before first operation

Before reoperation

Normal One segment impaired Two segments impaired Diffuse impairment

447 41 10

294 124

2

19

63

Group I combined with one of the other groups. The 13 other patients (5.1 %) were classified in other multiple combinations. Since ejection fraction was not calculated in earlier years, left ventricular contraction prior to reoperation was assessed subjectively. Left ventricular performance was c1assified in one of four categories: (1) normal or only minor impairment; (2) one of six myocardial segments hypokinetic or akinetic; (3) two or more myocardial segments impaired, and (4) diffuse impairment. Table III compares the status of left ventricular contraction at the first operation with performance before reoperation. Whereas few patients had poor left ventricular contraction comparable to an ejection fraction of 30% or less, 153 (30.6%) lost normal ventricular contraction during the average of 4 years between the coronary artery procedures. The prevalence of hypokinesis or akinesis confined to one left ventricular segment ranged from 21% to 29% in these three indication groups, and those who had poor contractility, i.e., two or more myocardial segments impaired, were similarly distributed from 15% to 18% in these various groups (NS). Bypass graft operations only were performed as the first procedure in 324 of the 500 patients. Single artery grafts had been performed in 150 (46.3%), grafts to two arteries in 130 (40.1 %), and grafts to three or more arteries in 44 (13.6%). These 324 patients received an average of 1.6 grafts. Complete revascularization, defined as grafting of all major arteries greater than 1 mm with estimated lesions greater than 60%, was performed in 71.6% originally. Ten others had a patch

The Journal of Thoracic and Cardiovascular Surgery

Loop et al.

678

Table IV. Clinical characteristics and reoperative results in previous /MA implant and vein graft patients

Men Pre-reoperative angina NYHA III or IV Mean interval between operations Extent >60% One vessel Two vessels Three vessels Left main Left ventricular impairment Grafts per patient Incomplete revascularization Reoperative mortality Late NYHA Class I Five-year survival rate

Previous Vineberg implant (n = 165)*

Previous bypass grafts (n = 335)t

159 (96.4%) 130 (78.8%) 77.5 months

293 (87.5%) 236 (70.5%) 37.9 months

P = 0.002 P = 0.048 p < 0.001

4 (2.4%) 49 (29.7%) 112 (67.9%) 31 (18.8%) 75 (45.5%) 1.6 96 (58.2%) 14 (8.5%) 82 (60.3%) 82.9%

37 (11.0%) 97 (29.0%) 201 (60.0%) 32 (9.6%) 131 (39.1%) 1.6 121 (36.1%) 6(1.8%) 188 (59.7%) 93.9%

P = 0.002+

p Value

p = 0.003 NS NS P < 0.001 p = 0.003 NS p = 0.002

Legend: IMA. Internal mammary artery. NYHA, New York Heart Association. NS, Not significant.

* Includes 33 patients with vein grafts. t Includes II patients with patch grafts. Single versus multiple-vessel disease.

+

Table V. Grafts per patient, November, /967, to July, /979 Year

1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 Totals

No. of reoperation patients I

No. of grafts per patient

0 I 10 14 41 50 59 63 65 59 73 64

1 0 13 14 56 72 84 95 108 105 132 123

1.0 0 1.0 1.3 1.0 1.4 1.4 1.4 1.5 1.7 1.8 1.8 1.9

500

804

1.6

I

graft to the right coronary artery and one received patches to the right and anterior descending arteries in the first operation. Of the 500 patients, 165 had internal mammary artery (IMA) implants (Vineberg) performed as part of the first revascularization operation. Ninety-three (56.4%) of these operations were performed through a median sternotomy and 33 of these included a saphenous vein bypass graft. Thirty of 33 patients who had direct revascularization and Vineberg implants had grafting to only one vessel; grafting to two vessels was oerformed in two patients; one patient received three

grafts. Isolated left Vineberg implants were performed through a left lateral thoracotomy in 72 patients. Table I V lists the major angiographic and clinical differences between those who had IMA implants and those who had bypass grafts as the first operation. Early and late results will be addressed later. Nineteen patients who had 20 IMA grafts were reoperated upon because of IMA graft dysfunction. Seven of 19 underwent reoperation for graft stenosis and 12 of 19 had 13 completely occluded IMA grafts. The seven stenotic grafts were attributed to technique. One IMA graft was atherosclerotic; however, the reoperation was performed 2.5 months after the original operation and doubtless the IMA had been atherosclerotic at the first procedure. A review of the 13 occluded IMA grafts revealed that technical difficulties were sometimes compounded by a small recipient vessel. Nine of the 19 patients had the original operation in another hospital. The number of reoperation bypass grafts has increased almost steadily from 1.0 to 1.9 during this experience. A mean of 1.6 grafts was performed in these 500 patients. Complete revascularization was accomplished in 283 (56.6%) reoperations. Table V lists the average number of grafts per patient in reoperations each year from 1967 through July, 1979. Results Mortality. Twenty (4.0%) operati ve (hospital) deaths occurred during this 12 year experience. Yearly mortality ranged from 0% to 10%. Mortality was 0% in 41

Volume 81 Number 5

Coronary artery reoperations

May, 1981

Table VI. Quantitative and numerical variables analyzed for risk determination Age Sex NYHA functional class before reoperation Diabetes Hypertension Recurrence of angina interval Interval between operations Interval infarction Cholesterol level Trigl ycerides level Extent of disease Left main coronary artery lesion Left ventricular contraction Indication for reoperation Year of reoperation Cardioplegia Number of grafts Complete revascularization Perioperative myocardial infarction Respiratory insufficiency Low cardiac output Wound infection Late congestive hearr failure Late angina pectoris

patients with one-vessel disease, 2.8% (four deaths) in 145 with two-vessel disease, and 5.1% (16 deaths) in 314 with three-vessel disease. Of 263 patients who received one graft, nine (3.4%) died; of 180 patients with two grafts, eight (4.4%) died; of 57 patients with three or more grafts, three (5.3%) died. When the operation was undertaken for progressive atherosclerosis, 13 of 257 (5.1%) died, for graft closure in 147, three (2.0%) died, and when combined indications existed, four of 96 (4.2%) died. Operative mortality based on NYHA functional class preoperatively was as follows: Class I, no deaths; Class II, three (2.3%); Class III, nine (3.9%); Class IV, eight (6.0%). Of the 20 operative deaths, 17 were attributed to cardiac causes, i.e., ischemia/infarction or low cardiac output resulting from myocardial damage. Three of 20 died of neurologic complications. Contingency tables were used to examine the relationship between operative mortality and preoperative risk factors. Five variables demonstrated significant association with higher operative mortality: (1) presence of a left main stenosis of 60% or greater (p < 0.001), (2) poor left ventricular function (p = 0.02), (3) length of interval between first and second operations greater than the median interval (p = 0.03), (4) incomplete revascularization (p = 0.046), and (5) previous implant (p < 0.001).

679

Table VII. Variables in myocardial protection subsets Cardioplegia (n

Median age (yr) Extent of disease One vessel Two vessels Three vessels Normal left ventricle Grafts per patient Perioperative myocardial infarction* Operative mortality Two-year survival

= 113)

54

Ischemic arrest (n = 387) 54

6 (5.3%) 32 (28.3%) 75 (66.4%) 72 (63.7%) 2.0 3 (2.7%)

35 (9.0%) 114 (29.5%) 238 (61.5%) 222 (57.4%) 1.5 30 (7.8%)

3 (2.7%) 95.3%

17 (4.4%) 92.4%

*p = 0.055.

By means of a maximum likelihood estimation procedure," a logistic model was used to estimate the probability of operative mortality as a function of multiple variables (Table VI). Three of the five factors individually associated with operative mortality remained significant predictors in the logistic model: (1) left main trunk stenosis of 60% or greater, (2) incomplete revascularization, and (3) previous implant. In 1977 cold potassium cardioplegia was combined with systemic hypothermia for myocardial protection. The 113 patients operated upon between January, 1977, and July, 1979, with this method provide a subset for comparison with the conventional technique of normothermia and anoxic arrest (387 patients). Table VII compares the clinical and angiographic characteristics of the cold cardioplegia group with normothermic anoxic arrest patients. The operative mortality was slightly lower in the cardioplegia group (NS). However, the 2.7% incidence of perioperative infarction was lower with cardioplegia than the 7.8% found with anoxic arrest (p = 0.055). Morbidity. Definite perioperative infarction is defined as the appearance of new Q waves and was almost always associated with an abnormal rise of cardiac enzyme levels. Probable infarction is defined as decreased R voltage in two or more leads associated with elevated enzyme levels. As defined earlier, definite or probable myocardial infarction occurred in 33 (6.6%) of the 500 patients. Since 1967 the incidence of perioperative infarction ranged between 2.7% (1978) and 10.2% (1974); since 1977 it has been consistently less than 5%. Of 48 women, six (12.5%) sustained documented perioperative infarction in comparison with 27 of 452 men (6.0%) (NS). Thirty-three patients (6.6%) experienced hepatitis;

The Journal of

680

Thoracic and Cardiovascular

Loopetal.

Surgery

Table VIII. Symptom status for 451 late survivors Angina before reoperation

NYHA

Totals

I II III IV

Late NYHA angina class

I

I

II

6

5

I

117

88

211 117

124 53

24 67 40

451

270

132

III

I

IV

0 4

0

17

3

16

8

37

12

I

two of the 33 had hepatitis after the first operation and again after the second operation. Blood usage ranged from 9 to II units per patient per hospitalization through 1976. In 1977, 1978, and 1979 mean units transfused were 7.3, 3.0, and 2.7, respectively, and the incidence of clinical hepatitis averaged 2.6%. Postoperative bleeding which necessitated reoperation occurred in 51 (10.2%) patients but declined in 1979 to three cases in 64 (4.7%). Clinical status. After a mean follow-up of 42 months, 451 survivors were interviewed by telephone. When late death (29 patients), interval hospitalization, recurring symptoms, or postoperative catheterization occurred, the referring physician was contacted for further information. Follow-up was complete for late deaths and for all survivors. Functional Class 1 (no angina) was reported in 270 (59.9%); Class II, 132 (29.3%); Class 1II, 37 (8.2%); and Class IV, 12 (2.7%). A comparison of symptom status before and after reoperation is shown in Table VlII. Late angina status was improved by one or more classes in 198 of the progressive atherosclerosis group (88.0% of late survivors), in 112 of the graft closure group (82.3%), and in 78 of the combined indication group (86.6%). Late symptom status was equal among those classified as completely revascularized in comparison with patients incompletely revascularized. Of 262 completely revascularized survivors, 152 (58.0%) had no angina at follow-up in comparison with 118 of 189 survivors (62.4%) in the incomplete revascularization group. Overall, 223 of 262 late survivors (85.1 %) completely revascularized and 165 of 189 (87.3%) late survivors incompletely revascularized were improved by at least one NYHA functional class. Graft patency and left ventricular performance. In 104 patients studied a mean interval of 19 months (range 4 days to 103 months) after reoperation, four (3.8%) showed improvement in left ventricular contraction judged subjectively, and 90 (86.5%) had unchanged contraction. Left ventricular performance had

deteriorated in eight (7.7%); six of the eight patients had experienced perioperative infarction and two sustained late infarction between the reoperation and subsequent catheterization. Left ventriculograms were not available for comparison in two other patients. Graft status is taken from the most recent catheterization only. Of these 104 recatheterizations, 18 were done in asymptomatic patients routinely, 34 for angina Class II, 26 for Class 1II, and six for Class IV. Other indications for repeat angiography included dyspnea and angina in 19 and fatigue in one. Grafts to the anterior descending coronary artery had the highest patency rate (93.2%) followed by grafts to the circumflex (84.6%) and the right coronary artery (65.4%). The results in one patient who received umbilical vein grafts are not included. Two of three umbilical vein grafts were open at catheterization 3 months postoperatively; however, the patient died I year postoperatively and autopsy findings revealed closure of these grafts. Of 132 saphenous vein grafts studied, 104 (78.8%) were open. The anterior descending coronary artery had the highest patency rate, 39 of 43 (90.7%); next the circumflex, 32 of 38 (84.2%); and right coronary artery, 33 of 51 (64.7%). Thirty-two IMA grafts were recatheterized, 30 of which were anastomosed to the anterior descending, one to the right coronary, and one to the circumflex; one anterior descending graft was occluded. The patency of grafts performed for graft failure was surveyed. Of 129 anterior descending grafts performed for previous graft failure, 24 were restudied and all were patent. Of 68 replacements for failure of right coronary artery grafts, 15 were studied and 10 (66.7%) were open. Seven of 52 grafts performed for failure of circumflex grafts were studied and five of the seven (71.4%) were open. Only one diagonal graft was performed for graft failure and it was open postoperativeIy. These grafts performed for graft failure included those performed for previous stenotic grafts if they were replaced or revised at reoperation. Vessels previously not grafted and subsequently involved with progressive atherosclerosis were reviewed. Sixty-four of these grafted vessels were studied and 53 grafts (82.8%) were patent. Survival. After a mean follow-up of 42 months, 29 late deaths occurred of which 23 were judged cardiac and six noncardiac. The noncardiac deaths were caused by cancer in four cases, suicide in one, and fire in one. All deaths were included in actuarial computations. From the Kaplan-Meier method, 5 year actuarial survival rate for the entire series is 89.2% (Fig. 1). Those who were operated upon for progressive atherosclerosis

Volume 81

68 I

Coronary artery reoperations

Number 5 May, 1981

PERCENT

PERCENT

SURVIVAL

SURVIVAL

100

100

-'-e_..-"-e.e_ _ e

"

_ e_'_e'e

90

.

....- .-"~'=e-:90

e.

'h.. . _ .• __

-:::--::.~--~~=-=-&.,.. e

'e

- - _ e __•.•

91.5

~------.89.4

---e87.4

-e-.89.2

80

80

70 70

n

COMBINATION GRAFT CLOSURE

n

147

•

~~~~=6~~I~EEROSIS

n

257

60

REOPERATIONS1967 1979

60

500

12 12

96

• .6

24

36

48

60

MONTHS

Fig. I. Kaplan-Meier actuarial survival curve for the 500 patient reoperation series: 5 year survival rate is 89 ,2'7r, had an 87.4% survival rate compared with 89.4% for the graft closure subset and 91.5% for those operated upon for combined indications (NS) (Fig. 2). Complete myocardial revascularization is defined as grafting performed for all major vessels greater than I mm in diameter and obstructed greater than an esti mated 60%. Part of this complete attempt may have occurred at the first operation as long as the initial graft remained open. Complete revascularization was performed in 283 (56.6%), and their 5 year survival rate was 94.2% compared with 85.6% for those who received incomplete revascularization (p < 0.01). Patients who had two or more myocardial segments impaired before reoperation were arbitrarily classi fied as having poor left ventricular function. These 82 patients had an 84.3% 5 year survival rate compared with 91.0% for those with lesser impairment or normal contraction (NS). The 63 patients with severe left main stenosis had an 80.0% 3 year survival rate in contrast to a 94.0% 3 year survival rate for those without left main lesions (p < 0.001). For the 165 patients who had a previous implant, the 5 year survival rate was 82.9% compared with 93.9% for 335 patients who had had direct myocardial revascularization (p = 0.002). The same preoperative variables were analyzed to identify significant predictors of late survival among the 480 surgical survivors. As in the contingency table analysis and the maximum likelihood estimation, nineteen qualitative variables and three continuous (nu-

24

36

48

60

MONTHS

Fig. 2. Actuarial survival curves for each angiographic indication group. Those operated upon for combined reasons. i.e., graft failure and progressive atherosclerosis. had a 91.5'1< survival rate. Survival rate for those reoperated upon for graft closure was 89 A'7r, and for patients with progressive atherosclerosis. the 5 year survival rate was 8704'1<.

merical) variables were evaluated according to subgroups defined in terms of the individual risk factors. Kaplan-Meier survival curves were calculated by analyzing each variable separately, and the hypothesis of equality of survival between subgroups was tested using Cox's test (the log rank test). ~ Two variables, multiple grafts and previous implants, were identified as significant prognostic factors. More than one graft at reoperation was associated with improved 5 year survival rates (95.4% versus 89.4%; p = 0.04). The 5 year survival rate for 335 patients who had bypass grafts as the first procedure was significantly better than that for 165 patients who received Vineberg implants previously (93.9% versus 82.9%; p = 0.002). Intra-aortic balloon pump usage began in 1975. In the 2 I I patients subjected to normothermic anoxic arrest who have been reoperated upon since 1975, 14 (6.6%) patients had the balloon inserted; five of the 14 received balloon support preoperatively because of unstable angina or poor left ventricular performance. Eleven of the 14 survived and were discharged. In the 113 patient cardioplegia group, six (5.3%) had the balloon inserted, two preoperatively for unstable angina. Five of six survived and were discharged. Of the 16 discharged patients who received balloon support, all are alive after a mean follow-up of 15.6 months.

The Journal of

682

Loop et al.

Employment status of late survivors was obtained at the time of follow-up. One hundred seventy-seven (39.2%) are working full time, part time, or as a housewife, 224 (49.7%) survivors are active but not working, 42 (9.3%) are classified as inactive, and eight (1.8%) are completely disabled.

Discussion The incidence of reoperations for coronary atherosclerosis will be affected by factors that are presently in evolution. The surfeit of patients with single-vessel disease originally selected for operation has declined to less than 10% of revascularization patients today. Few "incomplete" operations are performed now and the number of grafts per patient has nearly doubled in the past decade. Graft patency has improved and today there is a tendency to graft the estimated 40% to 50% narrowed vessel, a practice which was unconventional only 5 years ago. Theoretically, these trends could limit the incidence of potential reoperation candidates in later years. Palliation derived from a coronary artery operation is, in most cases, satisfactory for the first 5 postoperative years. When symptoms recur, can another revascularization extend this palliation? We know that within this past decade primary coronary artery operations have become safer and the benefits attained are proportionally greater. 2 Are the same benefits likely for those operated upon for advanced atherosclerosis a second and even a third time? In contrast to a 1979 mean age of 56 years at first revascularization, the age at first operation in these 500 reoperation patients averaged 49 years. In earlier years, patients with one-vessel disease were selected frequently for primary revascularization, and one-vessel narrowing occurs more often in young patients. Barboriak and associates" concluded that their reoperation patients had the first bypass operation at a younger age, which indicates a predilection for advanced and possibly more progressive atherosclerosis. In contrast to their findings we could not document significantly higher cholesterol or triglyceride values in patients who had two operations. Angina recurred after a mean of 31 months and the mean interval between first and second operations was 51 months. In many patients coronary atherosclerosis progressed dramatically during the interim. Before the initial procedure, 172 (34.4%) had one-vessel disease compared with 41 (8.2%) at reoperation. Initially only 119 (23.8%) had three arteries severely involved compared with 314 (62.8%) before reoperation. Left ventricular performance declined similarly. Of 447 (89.4%) who had normal or nearly

Thoracic and Cardiovascular Surgery

normal function preoperatively, only 294 (65.8%) showed normal function before the second operation. Angiographic criteria for reoperation candidates have been classified into three indication groups. In our previous communication" incomplete revascularization was considered in the indication classification and included all patients previously having a Vineberg implant. In retrospect, almost all of them were reoperated upon for additional lesions outside the implant site. Hence, we modified angiographic indications from four to three groups. The 257 (51.4%) patients with progressive atherosclerosis constituted the major angiographic indication group; the 147 (29.4%) with graft closure were next in frequency. Patients who had evidence of progressive atherosclerosis combined with graft closure constituted the third group, combined indications, which occurred in 96 (19.2%). Three-vessel disease was found most frequently (p < 0.001) in those with combined indications or progressive atherosclerosis (Table II). Reul and associates! reporting the Texas Heart Institute experience, noted that progressive coronary atherosclerosis was the only reason for reoperation in 30% and a secondary reason in an additional 27%. Interestingly, three fourths of their patients had no visible lesion in those vessels at first catheterization. A survey of our patients operated upon for progressive coronary atherosclerosis included an analysis of arteries that had undergone progressive changes. A third of these vessels already had severe lesions, but the majority had been treated by a Vineberg implant. The other two thirds of the patients in the progressive atherosclerosis indication group were almost equally divided among those who had no perceptible lesion before the first operation, a mild lesion (estimated 20% to 30% narrowing) or a moderate lesion (40% to 60% narrowing), and those who had multiple progressive changes in vessels with lesions of different magnitudes originally. As shown in Table IV, patients who had IMA implants (Vineberg) previously had a longer interval between operations, more severe angina, a higher incidence of multivessel disease including left main stenosis, slightly more left ventricular impairment before reoperation, and a greater rate of incomplete revascularization in comparison with those who had bypass grafts initially. These differences account for the 8.5% operative mortality in the implant group versus 1.8% in those who had bypass grafts as the first operation. Angina recurred earlier in the graft closure patients (mean 17 months) than in other indication groups (mean 37 months). This finding coincides with the observation of Culliford and colleagues" that early reap-

Volume 81 Number 5 May, 1981

pearance of angina was most frequently attributed to technical factors and later angina to new native coronary lesions. Assuming that the patient is fit medically and has disabling angina, the major finding in favor of reoperation is technically favorable coronary anatomy, specifically adequate vessel size, and coronary arterial runoff. When the vessel in question is small or diffusely involved with distal atherosclerosis, a secondary graft is unlikely to be successful. Rarely do patients with two or more functioning grafts require reoperation. When patients with poor left ventricular function are considered for reoperation, angina should be the predominant symptom. Those with symptoms of heart failure resulting from diffuse myocardial fibrosis are at higher risk, especially when they have sustained cardiomegaly or ischemic cardiomyopathy. Patients with ejection fractions less than 20% are, in our opinion, rarely candidates for reoperation. Allen and associates" tested reoperative variables for prognostic significance and found that vessel size greater than 2.0 mrn and proximal plaque distribution only correlated with improved clinical status. Our technical practice in coronary artery reoperations has been reported. to Density of adhesions is unpredictable, but generally those who were operated upon 5 to 10 years earlier are less likely to have extensive or vascular adhesions than are those who are reoperated upon within a year of the first procedure. In multiple graft reoperations, dilated left ventricles, or when unusual technical difficulties are anticipated, we prefer bicaval cannulation and either a left atrioventricular or aortic vent. Not only is the heart better decompressed consistently, but effective myocardial cooling is prolonged and less cardioplegic infusion solution is used. Furthermore, the left ventricle can be mobilized more easily when systemic hypothermia and cold potassium cardioplegia provide a cold, flaccid heart. Table VII summarizes the clinical characteristics, mortality, infarction rate, and survival for the 113 patients who received cold cardioplegia as myocardial protection compared with 387 patients who were operated upon under normothermic conditions. The series are comparable clinically with the exception that more grafts per patient were performed in the cardioplegia group. In recent years the operative mortality with cold cardioplegia has been slightly lower but not statistically significant. However, the perioperative myocardial infarction incidence of 2.7% is less than the 7.8% (p = 0.055) reported for the normothermic anoxic arrest group. The 2 year survival rate for the smaller cardioplegia series is 95.3% compared with 92.4% for

Coronary artery reoperations

683

the larger subset without specific myocardial protection. Blood usage has improved considerably during the 12 years of experience. Until 1976, 10 units of blood were used per patient per reoperation. Blood conservation measures II have reduced the amount of blood used to 3 units per patient per hospitalization in 1978 and 1979. The incidence of hepatitis varied widely but averaged approximately 10% through 1976. Since then it has ranged from 1.6% to 4.1 % and is expected to decline further. Although postoperative bleeding cannot always be associated with excessive blood usage, our experience indicates that blood conservation measures have decreased the incidence of postoperative hemorrhage. Mortality has shown a favorable downward trend in recent years. Through the first 7 months of 1979 (end of the 500 patient series), 64 patients were operated upon without a hospital death. * Completeness of revascularization decreased when one compares the second operation (56.6%) with the first operation (71.6%). Largely this decline was due to a more conservative approach in earlier years, more diffuse atherosclerosis at the second procedure and, finally, more transmural scar which precluded revascularization of some major arteries. After an average follow-up of 42 months, most of the survivors were improved clinically. Of 45 I survivors, 388 were judged improved by one functional class and 59.9% claimed no angina. Angina was relieved or improved with equal frequency in each of the angiographic indication groups. This finding upholds the conclusion of Irarrazaval and associates" that patients with previous graft failure have the same clinical results as patients in other groups. Similar conclusions have been reached by others. 12. 1:3 The clinical results of symptom relief showed no difference between the complete and incomplete revascularization groups. Of 500 patients, 104 (21 %) underwent postoperative catheterization after an average interval of 19 months. Left ventricular performance was unchanged in 87%, improved in 4%, and worse in 8% compared with their status before reoperation. Graft patency rate was essentially the same as that in our primary revascularization cases. Vein graft patency rate averaged 79% and IMA patency rate was 97%. Patency rate for 47 grafts performed for previous graft failure was 85%. In those patients who have experienced closure of the anterior descending vein graft, use of the IMA has been advantageous especially for distal arterial grafts.:" In this reoperation series patency rates between IMA and vein *From January through December 1979, 126 bypass graft reoperations were performed and one hospital death occurred (0.8%).

The Journal of

684

Loop et al.

graft for the anterior descending artery are not different; however, the IMA is a useful technical alternative for small-vessel anastomoses or when saphenous veins are absent or unsuitable. The absence of late atherosclerosis in IMA grafts is interesting and bears watching. In this series, 18 patients had documented vein graft atherosclerosis; in four, one or more veins were narrowed or closed by the atherosclerotic process, and yet an IMA was patent and functioning well without angiographic evidence of atherosclerosis. The 5 year actuarial survival rate is 89.2%, a figure reasonably close to the 92.4% 5 year survival rate from the primary revascularization cases of 1971 to 1973. 2 Five-year survival rates by indication group are as follows: (I) progressive atherosclerosis, 87.4%; (2) graft closure, 89.4%; (3) combined indications, 91.5% (NS). Completeness of revascularization influenced longevity significantly. Statistical analyses of operative risk and predictors of late survival identified previous implant during the first operation as a significant variable. The composition of the implant group belies the statistics (Table 1V). As opposed to patients who had previous vein grafts, the implant group had significantly more angina of Class III or 1V, a greater interval between operations, more three-vessel disease, and a greater incidence of incomplete revascularization. The operative mortality of the implant cohort was 8.5% in comparison with 1.8% for previous vein grafts, which greatly influenced the 5 year survival rate. The therapeutic implications of this reoperation series are these. Technical gains, operative experience, improved myocardial protection, and blood conservation have reduced risk and morbidity. Graft patency is comparable to catheterization results attained after primary revascularization. Therefore, it is not surprising that the clinical response is toward general improvement and relief of angina for most patients who undergo reoperation. After reoperation a favorable 5 year survival rate of 89% is forecast. We conclude that patients who have recurring angina after an initial revascularization operation should be evaluated as candidates for a second coronary artery operation. When the procedure can be performed by surgeons who have experience with these technically more complicated operations and are adept at the procedure, patients who meet the indication criteria may receive further effective palliation.

Addendum Six patients, all men, underwent more than one reoperation. Three operations were performed in five

Thoracic and Cardiovascular Surgery

patients and four operations in one. The first operation was performed at a mean age of 46 years, the second at 49, and the third at age 56. The intervals between the first and second operations averaged 38 months and between the second and third, 77 months. The size of this subgroup precludes any statistical analysis. However, all these patients had symptoms of angina without congestive heart failure before each operation. Whereas four of the six had single-vessel disease before the first operation, all had triple-vessel disease at the third procedure. Five had either a Vineberg implant or patch graft as the first procedure and the sixth had a single bypass graft. The second operation consisted of a patch graft to the right coronary artery in one patient and an average of 1.5 aorta-coronary bypass grafts in the others. No complications were experienced after the first and second operations. Four of six patients experienced myocardial infarction in the interval between the second and third operations. Whereas the second operation was undertaken for progressive atherosclerosis, the third operation was performed because of graft closure and progressive atherosclerosis in four patients, graft closure in a fifth patient, and progressive coronary atherosclerosis in the sixth patient. The third operation consisted of bypass grafts only and an average of 2.5 grafts per patient were performed. A balloon pump was used preoperatively for stability in one patient. No perioperative infarctions occurred after the third procedure; one patient bled postoperatively. The patient who had a fourth operation was operated upon within a month of the third operation because of proximal graft stenosis. After the third operation, the mean follow-up has been 17 months. All patients are alive, although the man who had a fourth procedure is disabled by angina and another has moderate exertional chest pain. The other four experience no angina. Four are employed and the fifth is active but retired. We wish to express our appreciation to Judy Borsh, R.N., and Eric Christiansen, Research Associate, for their assistance in the collection and organization of these data, and to Richard L. Greenstreet, Ph.D., and George W. Williams, Ph.D., for their help with the statistical analysis. REFERENCES Campeau L, Lesperance J, Hermann 1, Corbara F, Grondin CM, Bourassa MG: Loss of the improvement of angina between 1 and 7 years after coronary bypass surgery. Correlations with changes in vein grafts and in coronary arteries. Circulation 60:Suppl 1:1-5, 1979 2 Loop FD, Cosgrove DM, Lytle BW, Thurer RL, Simpfendorfer C, Taylor Pc, Proudfit WL: An II-year evolution of coronary arterial surgery (\ 967-1978). Ann Surg 190:444-455, 1979

Volume 81 Number 5

Coronary artery reoperations

May, 1981

3 Jones RH: Probability estimation using a multinominal logistic function. J Stat Comput Simul 3:315-329, 1975 4 Thomas OJ, Breslow N, Gart 11: Trend and homogeneity analyses of proportions and life table data. Comput Biomed Res 10:373-381, 1977 5 Barboriak J1, Barboriak DP, Anderson AJ, Rimm AA, Tristani FE, Flemma RJ: Risk factors in patients undergoing a second aorta-coronary bypass procedure. J THORAC CARDIOVASC SURG 76:111-114, 1978 6 Irarrazaval MJ, Cosgrove DM, Loop FD, Ennix CL, Jr. Groves LK, Taylor PC: Reoperation for myocardial revascularization. J THORAC CARDIOVASC SURG 73:181188, 1977 7 Reul G1, Cooley DA. Wukasch DC, Ott DA, Coelho A, Chata L, Enterobic I: Reoperation for recurrent coronary artery disease. Causes, indications and results in 168 patients. Arch Surg 114:1269-1275, 1979 8 Culliford AT, Girdwood RW, Isom OW, Krauss KR, Spencer FC: Angina following myocardial revascularization. Does time of recurrence predict etiology and

9

10

II

12

13

14

685

influence results of operation? J THORAC CARDIOVASC SURG 77:889-895, 1979 Allen RH, Stinson EB, Oyer PE, Shumway NE: Predictive variables in reoperation for coronary artery disease. J THORAC CARDIOVASC SURG 75:186-192. 1978 Loop FD, Thurer RL, Lytle BW, Cosgrove DM: Reoperation for myocardial revascularization. World J Surg 2:719-729, 1978 Cosgrove DM, Thurer RL, Lytle BW, Gill CC, Peter M, Loop FD: Blood conservation during myocardial revascularization. Ann Thorac Surg 28:184-189, 1979 Kobayashi T, Mendez AM, Zubiate P, Vanstrom NR, Yokoyama T, Kay J H: Repeat aortocoronary bypass grafting. Early and late results. Chest 73:446-449, 1978 Vouhe P, Grondin CM: Reoperation for coronary graft failure. Clinical and angiographic results in 43 patients. Ann Thorac Surg 27:328-334, 1979 Loop FD, Carabajal NR, Taylor PC, Irarrazaval MJ: Internal mammary artery bypass graft in reoperative myocardial revascularization. Am J CardioI37:890-895, 1976