Early and late results following repair of dissections of the descending thoracic aorta

Early and late results following repair of dissections of the descending thoracic aorta R. Kent Jex, M.D., Hartzell V. Schaff, M.D., Jeffrey M. Piehler, M.D., R. Michael King, M.D., Thomas A. Orszulak, M.D., G o r d o n K. Danielson, M,D., Peter C. Pairolero, M.D., James R. Pluth, M.D. and Duane Ilstrup, M.S.,

Rochester, Minn. Management of dissections of the descending thoracic aorta remains controversial, especially with regard to timing and method of repair. To clarify these and other issues we have reviewed our total experience with repair of descending aortic dissections between 1962 and 1983. The 44 men and 20 women had a mean (--. SEM) age of 59 -+ 2 years (range, 19 to 83 years), and in all patients the dissection originated in and was limited to the aorta distal to the left carotid artery (Stanford type B, DeBakey types IIIa and IIIb). Twenty-nine patients underwent operation within 2 weeks of the onset of symptoms (acute), and the remainder had later repair (chronic). During repair, circulation distal to the aortic cross-clamp was supported with cardiopulmonary bypass or shunt in two thirds of patients. Overall, 18 deaths occurred ~<30 days postoperatively (operative risk 28%), and risk was higher in acute (45%) than~m chronic (14%)dissections. Operative risk was not. significantly related to protection of the distal circulation: The most serious postoperative complication was spinal cord ischemia manifested by paraplegia in five patients (8%) and transient or permanent paraparesis in six patients (9%). Risk of spinal cord ischemia was significantly lower in patients who had protection of the distal circulation during operative repair (8% vs. 44%, p = 0.003). Late survival, including hospital deaths, was 49% --. 7% at 5 years after operation; 22 of the 46 patients who survived repair were found to have aneurysms involving the thoracic and/or abdominal segments of the aorta. Our results indicate that repair of chronic dissection of the thoracic aorta has a lower operative risk than repair of acute dissections, and initial medical management of acute dissection may be indicated if no early complications occur. Risk of spinal cord ischemia is significantly reduced by cardiopulmonary bypass or shunt and is preferred over aortic cross-clamping alone. Finally, these patients require careful long-term follow-up because of the high incidence of residual or recurrent aortic aneurysms. (J VAsc SURG 1986; 3:226-37.)

The variations in management strategies and operative techniques for patients with dissections of the descending thoracic aorta attest to the imperfect resuits currently attained. Wheat et al.~3 and Doroghazi et al. 4 recommend initial medical therapy for patients with acute aneurysms of the descending thoracic aorta and reserve early surgical intervention for patients with impending rupture, continued pain despite adequate control of blood pressure, and compromise or occlusion of a major arterial branch. Such a plan recognizes that early intervention for acute From the Section of Thoracic and Cardiovascular Surgery, Mayo Clinic and Mayo Foundation. Presented at the Thirty-third Scientific Meeting of the North American Chapter, International Society of Cardiovascular Surgery, Baltimore, Md., June 6-7, 1985. Reprint requests: Hartzell V. Schaff, M.D., Mayo Clinic, 200 First St., S.W., Rochester, MN 55905.

226

aortic dissection carries a significantly higher mo~,, tality risk than does repair of chronic dissection of the descending thoracic aorta and that ea-Lv medical • of the descending treatment Can limit dissections "-: ' thoracic aorta. Other groups have advised early repair ofacute aneurysms of the descending thoracic aorta, believing that overall life expectancy will be improved by reducing the risk of early complications o f progression of the dissection, s The value of partial bypass or shunting to maintain perfusion to the distal aorta during aortic occlusion is unproved. Recommended techniques include routine partial cardiopulmonary bypass, veno-arterial bypass, left atriofemoral bypass, shunts (aorticofemoral or ventriculofemoral), and simple crossclamping without adjuncts, sq2 To clarify these and other issues, we have reviewed our experience with treatment of acute and

Volume 3 Number 2 February 1986

Results after repair of descending thoracic aorta dissections 227

Table I. Preoperative clinical features of patients with acute or chronic thoracic aortic dissections Acute (n = 29)*

Systemic hypertension Cigarette smoking Angina pectoris Chronic obstructive lung disease Renal insufficiency Cerebrovascular accident Chronic congestive heart failure Previous myocardial infarction Marfan's syndrome Diabetes mellitus

Chronic(n = 35)~

No. of patients

%

No. of patients

%

19 13 5 5 4 3 1 1 1 0

66 45 17 17 14 10 3 3 3 0

27 19 4 0 2 1 4 4 4 1

77 54 1i 0 6 3 11 11 11 3

*Age (mean _+ SEM) of patients in acute dissection group was 64 _+ 2 years; male: female ratio was 20:9. ~Age (mean + SEM) of patients in chronic dissection group was 55 + 2 years; male:female ratio was 24:11. ~ronic descending aortic dissections. This investigation focuses on determinants of operative risk, spinal cor~ injury, and late survival. METHODS AND DEFINITIONS

We reviewed the medical records of all patients at the Mayo Clinic who underwent surgical repair of nontraumatic descending thoracic aortic dissections (Stanford type B, DeBakey types IIIa and IIIb) from 1962 through 1983. Attention was focused on preoperative risk factors, methods of repair, and late survival. Long-term follow-up was obtained by examination at the Mayo Clinic, correspondence with the patient, patient's family, or patient's physicians, and, in several cases, by telephone interview. Current status was obtained in all but two survivors (96% complete), and the duration of follow-up ranged from 2 months to 18 years (mean 4.1 years). Aortic dissections involved the descending tho"acic aorta, and the entry site, in most cases, was near the left subclavian artery. In some patients, the proximal intramural hematoma extended to the origin of the l~ft c'~irotid artery. Dissections were classified as acute if chest pain or other symptoms were present for 2 weeks or less before operation. In all cases, the acute dissections were confirmed intraoperatively~ by the presence of subadventitial blood or fresh hematoma. Dissections were termed chronic if symptoms were present for more than 2 weeks preoperatively. Also, patients were considered to have chronic dissections if they were previously known to have dissections of the descending thoracic aorta but had recent development of severe back or chest pain. Traumatic aortic transections were not included in this study. Operative mortality included all patients who died during or within 30 days of operation. Hyper-

tension was defined as systolic blood pressure i> 150 mm H g or diastolic blood pressure I>100 mm Hg; also included were patients with normal blood pressure who were receiving antihypertensive medications. Statistical analysis was performed to compare discrete and continuous variables among patients with acute and chronic dissections. In addition, we analyzed the influence of clinical variables on operative morbidity and mortality. Discrete variables were compared with a chi-square test or Fisher's exact test where necessary, and the continuous variables were analyzed by the Wilcoxon two-sample test (normal approximation). Survival was estimated by the Kaplan-Meier method 13 and was compared with the expected survival of a normal population matched for age, sex, and demography. The log rank test 14 and the Cox modeP s were used to compare survival curves, and a p value ~<0.05 was considered significant. Logistic regression 16 of multiple variables was used to assess the association of shunting and crossclamp time to the occurrence of paraplegia and/or paraparesis. All continuous variables are expressed as mean + the standard error of the mean. PATIENTS

Sixty-four patients (44 men and 20 women) had descending thoracic aortic dissections. Their ages ranged f r o m 19 to 83 years (mean 64 + 2 years). Twenty-nine patients had acute aortic dissections, and 35 patients had chronic aortic dissections. Mean ages of the patients in the acute and chronic disease groups were 64 -+ 2 and 55 + 2 years, respectively (p < 0.01). Additional preoperative clinical features are presented in Table I. Hypertension and cigarette smoking were common in both groups. There was no

228

Journal of VASCULAR SURGERY

Jex et aL

Table II. Contributing causes of death among patients undergoing repair of acute or chronic thoracic aortic dissections~

Hemorrhage MI/LCO Sepsis Renal failure Respiratory failure Extension of dissection Cerebrovascular accident Pulmonary embolism

Acute

Chronic

(n=Z3)

(n=5)

No,

%

No.

%

5 4 4 3 3 2 1 1

38 30 30 23 23 15 8 8

1 3 1 0 2 0 0 0

20 60 20 0 40 0 0 0

MI/LCO = myocardial infarction and/or low cardiac output. *Most patients had multiple complications leading to death.

statistically significant difference in the incidence of the clinical variables comparing patients with acute dissections vs. those with chronic dissections. With acute aortic dissections, 28 patients (97%) had severe chest or back pain, and two patients (7%) presented with circulatory collapse. One patient in the acute group had preoperative paraparesis, and another patient had evidence of acute visceral ischemia. With chronic aortic dissections, 24 patients (69%) had pain at the time of presentation, one patient had circulatory collapse, two patients had visceral ischemia, and one patient had preoperative paraplegia. Hoarseness and dysphagia were prominent presenting complaints in two other patients. Gross and/or histologic evidence of atherosclerosis was noted in the involved segment of aorta in 83% of patients with acute dissections compared with 74% of chronic dissections. Connective tissue disorders (e.g., cystic medial necrosis) were noted in 3% of acute dissections and 20% of chronic dissections. No primary pathologic condition of the aorta was identified in 14% and 9%, respectively. During the interval of this study, the management of patients with descending aortic dissection varied. In general, patients with evidence of acute dissection of the descending thoracic aorta were initially treated with antihypertensive agents and beta-blockers. The duration and intensity of medical treatment varied with the consulting physicians and surgeons. All patients whose symptoms were not initially relieved by medical therapy were offered operative repair. Other indications for operation included clinical signs of impairment of distal arterial perfusion (e.g., renal insufficiency, visceral ischemia, or peripheral arterial ischemia). F o r patients with chronic aortic dissections, operation was offered to those patients who

had recurrence of chest pain or evidence of increased size of the descending thoracic aorta on serial~[_~est x-ray films, aortograms, and/or computed axial tomograms (CT scans). Operation was undertaken on an urgent basis in 83% of patients with acute aortic dissections vs. only 12% of those with chronic aortic dissections. OPERATIVE METHODS Several methods of repair of aortic dissections were used during this 22-year experience. Repair was performed through a left thoracotomy in 63 patients, and in the remaining patient the aortic dissection originated between the left carotid and left subclavian arteries and was repaired through a median sternotomy with full cardiopulmonary bypass. Replacement of the proximal descending thoracic aorta wi~~" a prosthetic tube graft was performed in 63 patient~, and one patient had direct aortorrhaphy. The dissection was repaired in 21 patients (33 %) with proximal aortic cross-clamping and without adjuncts to support the distal circulhtion. Forty-three patients had support of the distal circulation during aortic cross-clamping by cardiopulmonary bypass (36 patients) or ventriculo- or aorticoaortic (Gott) shunts (seven patients). Proximal control of the aorta was obtained between the left carotid and left subclavian arteries in 38 patients (59%) and below the origin of the left subclavian artery in 17 patients (27%). The ascending aorta was cross-clamped during cardiopulmonary bypass in one patient, and in the remaining eight patients (13 %) the level of aortic cross-clamping was not indicated in the operative record. The duration of aortic occlusion averaged 59 -+ 5 minutes during repair of acute aortic dissection and 69 + 6 minut¢~ during repair of chronic aortic dissection. Among the patients who had cardior~;llmonary bypass, 25 (69%) had femorofemoral 'ffypass, 10 (28%) had left atriofemoral bypass, and one (3%) had complete cardiopulmonary bypass with cannulation of the right atrium for venous return. The average duration of bypass was 87 + 12 minutes for repair of acute aortic dissection and 88 _+ 7 minutes for chronic aortic dissections. RESULTS Early results. Eighteen operative deaths (28%) occurred overall; 13 deaths occurred with acute aortic dissection (operative mortality rate, 45%), and five occurred following repair of chronic aortic dissections (operative mortality rate, 14%). Causes of

Volume 3 Number 2 Februaq, 1986

Results after repair of descending thoracic aorta dissections

229

Table III. Relationship of operative mortality to methods of repair of acute or chronic thoracic ao~"',c dissection No protection

Protection Shunt

Acute Chronic Total

Bypass

No. of deaths

%

No. of deaths

%

No. of deaths

%

4/9 2/12 6/21

44 i7 2--9

0/4 0/3 0/---¢

0 0 0

9/16 3/20 12/36

56 15 3-3

Table IV. Postoperative complications of patients with acute or chronic thoracic aortic dissections Acute

Renal insufficiency Par~ 'zgia Respiratory" insufficiency Paraparesis Hemorrhage with reoperatlon Proximal dissection Distal dissection MI/LCO Cerebrovascular accident Pulmonary embolism Other*

Chronic

Survivors

Nonsurvivors

Survivors

Nonsurvivors

(n = 16)

(n = 13)

(n = 30)

(n = S)

No.

%

No.

. %

No.

%

No.

%

4 2 1 1 1

25 13 6 6 6

4 2 4 2 3

31 15 31 15 23

5 1 4 3 4

17 3 13 10 13

0 0 2 0 1

0 0 40 0 20

0 0 0 0 0 7

0 0 0 0 0 44

1 1 5 2 1 3

8 8 38 15 8 23

0 0 2 1 0 10

0 0 7 3 0 33

0 1 I 0 0 1

0 20 20 0 0 20

MI/LCO = Myocardial infarction and/or low cardiac output syndrome. *Atrial arrhythmias, vocal cord paralysis, psychoses, and minor gastrointestinal bleeding.

early death are listed in Table II. Excessive bleeding was the most common factor preceding death; this occurred in 36% of patients who died after repair of acute dissection and 20% of patients who died after chronic aortic dissections. Most patients had multiple 5omplications contributing to their death. The preoperative clinical features listed in Table I were ~,alyzed to determine their influence on operative rh0rtality. Patients with preexisting chronic renal insufficiency and chronic obstructive pulmonary disease had a higher operative mortality rate than the remaining patients (50% vs. 26%, 60% vs. 25%), but these differences were not statistically significant. The other factors did not appear to influence operative mortality. We grouped together patients who had support of the distal circulation with shunts or bypass (protection) and compared these patients with those who had aortic cross-clamping alone (no protection). The overall operative mortality rate was similar in both groups (29% in the no protection group and 28% in the protection group). As shown in Table III,

overall operative risk was lowest among patients who had repair with the Gott shunt (0%). Causes of serious postoperative morbidity are listed in Table IV. The most serious complications were renal failure, myocardial infarction or low cardiac output, and respiratory insufficiency. There was n O significant difference in the occurrence of these nonfatal complications when patients with acute dissections were compared with those with chronic dissections. As might bc expected, respiratory insufficiency and the presence of myocardial infarction or low cardiac output had significant (p < 0.05) influence on operative mortality. When either of these two risk factors were present, the operative mortality rate was 63%. There were no demonstrable effects of cardiopulmonary bypass or shunts on morbid complications, although acute renal failure was noted in increased frequency in those patients with acute dissections who underwent repair of the aorta without protection of the distal circulation (Table V). Paraplegia or paraparesis was assumed to be a manifestation of spinal cord injury caused by isch-

230

Journal of VASCULAR SURGERY

Jex et al.

Table V. Risk of postoperative acute renal failure among patients undergoing repair of acute or chronic thoracic aortic dissections~ No protection (clamp only)

Protection Shunt

Acute Chronic Total

Bypass

No.

%

No.

%

No.

%

3/9 1/10 4/19

33 10 2-1

0/4 0/3 0/--7

0 0 0

2/12 1/20 3/32

17 5 -9

~Acute renal failure defined as anuria during the interval between surgical repair and death, or the necessity for dialysis; excludes intraoperative deaths (five) and preoperative renal failure (one).

Table VI. Risk of postoperative spinal cord ischcmia among patients undergoing repair of acute or chronic thoracic aortic dissections~ No protection (clamp only)

Protection Shunt

Acute Chronic Total

Bypass

No.

%

No.

%

No.

%

p Value~

5/8 3/10 8/18

63 30 4--4

1/4 0/3 1/---7

25 0 1--4

1/12 1/19 2/31

8 6 6

0.02 0.08 0.003

~Excludes preoperative paraplegia (two), intraoperative deaths (five), and postoperative brain death (one). ?Comparison of protection group and no protection group.

emia. Two of our patients had evidence of spinal cord injury preoperatively. One patient was paraplegic preoperatively and died intraoperatively from hemorrhage. A second patient had paraparesis preoperatively and suffered a cerebrovascular accident intraoperatively. Because of postoperative coma, the stares of his spinal cord injury could not be assessed subsequently. New postoperative paraplegia developed in five patients (four with acute dissection and one with chronic dissection): Six patients (three with acute dissection and three with chronic dissection) had transient or permanent postoperative paraparesis. To calculate risk of spinal cord injury, we combined the occurrence of new postoperative paraparesis with paraplegia. Also, this calculation excluded patients in whom no assessment of spinal cord function could be made postoperatively; excluded were patients with intraoperative death (five), preoperative spinal cord injury (two), and postoperative coma (one). Overall risk of spinal cord injury was 19.6% (11 of 56 patients); risk of postoperative spinal cord injury was higher following repair of acute aortic dissection compared with repair of chronic aortic dissection (29% vs. 13%).

Table VI shows the risk of postoperative spinal cord ischemia for patients classified according to protection or no protection in both the acute and chronic groups. Risk of spinal cord injury was significantly (p = 0.02) lower in the protected group for acute aortic dissections. Following repair of chronic aortic dissections, risk of spinal ischemia was 30% in the unprotected group vs. 6% in the protected grouF.I (p = 0.08). The length of aortic occlusion duriny ,repair of aortic dissection and the presence or abseh'~e of protection of the distal circulation are highly interrelated. Table VII gives the mean aortic cross-clamp times for patients with and without spinal cord injury. There is no statistical differences between the crossclamp times comparing patients with and without spinal cord injury in the group that had protection, whereas there was a significant difference in the crossclamp times of patients who had spinal cord injury compared with those without spinal cord injury in the group of patients without protection in the distal circulation. The risk of spinal cord injury as estimated from multivariate logistic regression compared the length of aortic cross-clamp time for patients with protec-

Volume 3 Number 2 February 1986

Results after repair of descending thoracic aorta dissections 231

1.o

Unprotected ~

~ ~...~

0.8

Probability of spinal cord injury

0.6

0.4

0.2

0.0

"

0

10

20

30

40

50

60

70

80

9'~

Aortic cross clamp time (min) Fig. 1. Relationship between probability of spinal cord injury and length of aortic occlusion. Notice that risk of spinal cord injury increasessubstantiallyafter 45 minutes of aortic occlusion for patients without protection of distal circulation (unprotected). No such relationship between probability of spinal cord injury and length of aortic occlusion was noted for patients with protection of distal circulation. Shaded areas indicate the 70% confidence limits. Table VII. Duration of aortic occlusion during repair of dissections of the descending aorta* Protection?

No protection

Total

Para?

No para

Para

No para

59 -+ 5 23 30-124

61 _ 11 2 50-71

63 _+ 7 13 30-124

57 _+ 8 5 34-80

44 + 6 3 33-54

69 - 6 29 26-140

69 1

77 + 7 18 37-136

87 _+ 26 3 56-140

40 -+ 4 7 26-60

71 _+ 5 31 30-136

68 _+ 11 8 34-140

41 -+ 3 10 26-60

Acute

Mean -+ SEM (min) No. of patients Range (min) Chronic Mean _+ SEM No. of patients Range (rain) Combined Mean _+ SEM (min) No. of patients Range (min)

64 +_ 4 52 26-140

63 _~ 7 3 50-71

*Patients with preoperative paraplegia or paraparesis (two), postoperative coma (one), and intraoperative death (five) excluded. 2Cross-clamp times unavailable on four patients. ~Para includes paraplegia or paraparesis.

tion to those with no protection (Fig. 1). This analysis demonstrates that risk of spinal cord injury is significantly related to protection of the distal circulation (p < 0.01). Risk of spinal cord injury increases significantly with increasing cross-clamp time in the nonprotected patients (p < 0.05). There was no increase in risk of spinal cord injury with increasing cross-clamp time among patients who had protection of the distal circulation. Late results. Late patient survival is illustrated in Figs. 2 and 3. The overall survival rate for patients with descending aortic dissections was 49% + 7% at 5 years and 30% _+ 8% at 8 years after operation. For those patients who survived hospitalization, 5and 8-year survival rates were 70% +_ 8% and

42% + 10%, respectively. Even for patients who survived hospitalization, late survival was significandy poorer for those with acute dissections than for those with chronic dissections (30% _ 17% vs. 90% + 7% at 5 years, p < 0.01). During the follow-up period, which ranged between 2 and 215 months, 20 patients died (eight after repair of acute dissection, 12 after repair of chronic dissection). One patient in each group was lost to follow-up. Causes of late death are listed in Table VIII. Complications of cardiovascular diseases accounted for all known late deaths. Subsequent aortic disease (residual or recurrent aortk dissection, fusiform or saccular aneurysms) developed during the follow-up period in 33% of sur-

232

Journalof VASCULAR SURGERY

Jex et al.

loo

20 [-I'-"

o! 0

-

-

p E x p e c t e d survival

(8)

I

I

I

I

I

I

I

I

1

2

3

4

5

6

7

8

Postoperative year Fig. 2. Survival following repair of dissections of descending thoracic aorta. Numbers in parentheses indicate number of patients at risk at each interval.

Table VIII. Cause of late mortality in patients surviving surgical rcpair of acute or chronic thoracic aortic dissections Cause Acute (n = 14)t 1 Myocardial infarction 2 Myocardial infarction 3 Unknown 4 Cerebrovascular accident 5 Intraoperative, repair of thoracoabdominal dissection 6 Ruptured abdominal aortic aneurysm 7 Myocardial infarction 8 Cytomegalovirus pneumonia (S/ P renal allograft failure) Chronic (n = 29)f 1 Intraoperative, repair of descending aomc aneurysm 2 Intraoperative, repair of ascending and arch dissection 3 Unknown 4 Unknown 5 Ascending dissection 6 Unknown 7 Congestive heart failure 8 Congestive heart failure 9 Myocardial infarction 10 Postoperative, repair of ascending and arch aneurysm 11 Myocardial infarction 12 Cerebrovascular accident

Interval* (too) 65 52 18 8 3 2.5 2 2 175 125 109 104 101 93 85 73 69 68 61 10

*Interval between origin repair and death. fn = number of patients at risk for disease.

vivors of repair of acute aortic dissection and 59% of survivors following repair of chronic aortic dissection (Table IX). Among patients who had repair of acute descending thoracic aortic dissection, subsequent aortic disease developed with approximately equal frequency in the descending thoracic or thor-

acoabdominal aorta compared with the !-grarenal aorta. No patient had involvement of the a'~cending aorta. In a group of patients who survived repair of chronic descending thoracic aortic aneurysms; there was involvement of the ascending aorta, thoracoabdominal aorta, and infrarenal aorta with approximately equal frequency. Many patients had disease in multiple segments. DISCUSSION

Origin of a dissection in the descending thoracic aorta occurs less commonly than in the ascending aorta, but its management is more controversial. Both surgical repair and medical treatment of descending thoracic aortic dissections can substantially improve the chance of patient survival, but the optimal form of management for the individual patient is not clearly established. The landmark report b.,DeBakey, Cooley, and Creech6 in 1955 describe~ repair of acute and chronic aortic dissectio~ with an operative mortality rate of 20%. In a subs'~quent report by DeBakey et al., 7 operative risk was reported to be 21% for surgically treated patients with dissection of the descending thoracic aorta, and risk was said to be twice as high in the acute stage compared with the subacute or chronic stages. Other authors have also reported a relatively high risk for repair of acute aortic dissection.4,s,8-1° Because of the generally tmsatisfactory results of early surgical treatment, many centers have followed the policy described by Wheat et al. ~3 for aggressive early medical therapy aimed at rigorous control of hypertension (initially trimethaphan in addition to reserpine was advocated; more recently sodium nitroprusside coupled with beta-blockade has become

Volume 3 Number 2 February 1986

Results after repair of descending thoracic aorta dissections

233

100

80

Percent survival

60

40

20

0

:i::':;:':::,i:: 1

2

I

I

3

I

I

I

4

5

Postoperative year

Fig. 3. Late survival for patients dismissed alive following repair of acute or chronic dissection of descending thoracic aorta. Late survivalwas significantlybetter for patients following repair of chronic dissections compared with those who had repair of acute dissections (p < 0.01). Table L~[. Subsequent aortic disease in patients surviving surgical repair of acute or chronic thoracic aortic dissections Acute (n = 15) * Site of disease Ascending thoracic aorta Descending thoracic aorta _+ thoracoabdominal aorta Infrarenal abdominal aorta No. of patients Mortality

No.

Chronic (n = 29)* %

0 3

0 20

4 5 2 (40)t

27 33 13

No. 6 6 7 17 6 (35)~

% 21 21 24 59 21

*n = number of patients at risk for disease (one patient in each group lost to follow-up). tMortality in those patients with aortic disease.

standard therapy). 4 The rationale for this policy is that extension of the dissecting hematoma in the descending thoracic aorta can regularly be halted by .5lch therapy and, as opposed to dissections of the ascending aorta in which cardiac tamponade, aortic valve insufficiency, or coronary occlusion can develop, there is little risk of sudden death in patients who have adequate medical therapy. Indications for early repair of aortic dissections during intensive drug therapy would be evidence of extension of the dissection, compromise or occlusion of a major branch of the aorta, or impending rupture as indicated by extravasation of blood into the pleural space or inability to control chest pain. Patients whose dissections stabilize can be managed medically and repaired at a later time when tissues are less friable than in the acute period. With such a strategy, Doroghazi et al.4 reported a 20% mortality rate among medically treated patients compared with a 49% mortality rate for patients treated surgically. The surgical results were, of

course, influenced by the nmnber of patients who were referred because of complications of medical therapy. Other authors have emphasized the important influence of complicating features on risk of repair of acute aortic dissections. Miller et al. s report an overall operative mortality rate of 33% for repair of descending thoracic aortic dissections (45% for acute dissections and 22% for chronic dissections). In their experience, operative risk for patients with complications of acute dissection of the descending thoracic aorta was 73%. In our own experience, operative risk is substantially higher for acute compared with chronic aortic dissections and parallels that presented in the literature previously (Table X). Spinal cord injury is one of the most serious complications following repair of descending aortic aneurysms, and the mechanisms of injury are complex. Aortic cross-clamping reduced mean aortic pressure in the distal aorta. 12"17"19Use of a shunt from the proximal aorta or the ventricle to the distal aorta

234

Journal of VASCULAR SURGERY

Jex et al.

Table X. Comparison of operative risk for repair of acute and chronic descending thoracic aortic dissection Overall Authors DeBakey et al.7 Reul et al. 9 Doroghazi et alJ Miller et al.5 Present

Institution Baylor THI MGH Stanford Mayo

Acute

Chronic

No. of patients

Risk (%)

No. of patients

Risk (%)

No. of patients

Risk (%)

332 91 49 43 64

21 21 45 29 28

-49 43 20 29

~ 25 40 45 45

-42 6 23 35

17 17 22 14

T H I = Texas Heart Institute; M G H = Massachusetts General Hospital. ~Operative risk noted to be twice as great during acute stage o f dissection compared with subacute and chronic stages.

increases mean aortic pressure distal to the aortic cross-clamp and reduces afterload on the ventricle. Experimentally, partial cardiopulmonary bypass or shunt reduces the incidence of paraplegia. 12 Several clinical reports, however, have not confirmed the protective effects of partial cardiopulmonary bypass or shunts compared with simple aortic cross-clamping.7,9-11Many of these reports, however, have combined patients with chronic atherosclcrotic aneurysms of the descending aorta with acute and chronic aortic dissections and transections. It is noteworthy that in our own experience (King RM, Schaff HV, Jcx RK, Piehler JM, Orszulak TA, Danielson GK, Pluth JR, unpublished data, 1984), presence or absence of protection does not influence risk of paraplegia among patients undergoing operation of chronic atherosclerotic descending aortic ancurysms. Our results are similar to those of Katz et al., 2 who studied the influence of partial cardiopulmonary bypass or shunt on risk of paraplegia following repair of acute aortic transections. In their study, as in ours, risk of spinal cord injury increased significantly with increasing aortic cross-clamp times only in patients without protection of the distal circulation. It is interesting to note that the divergence in risk of spinal cord injury between the protected and unprotected patients occurred at approximately the same time in these two studies. Our data suggest that spinal cord injury in patients without protection of the distal circulation can Occur with as little as 34 minutes of aortic occlusion, and the risk of spinal cord injury increases steeply after 45 minutes of occlusion, and the data ofKatz et al.21 indicated a similar rise in risk of spinal cord injury at approximately 30 minutes o f cross-clamping in patients without protection of the distal circulation. Individual variation in anatomy of the anterior spinal artery may also contribute to variation in incidences of spinal cord injuryY Recent studies in-

vestigating the relationship between spinal cord fluid pressure and subsequent spinal cord injury during aortic cross-clamping may elucidate addS) tional pathophysiologic mechanisms.23 Somatosensory-evoked potentials to assess the vulner~2~ility of the spinal cord to ischemic injury during aortic crossclamping may be valuable in assessing the adequacy of shunts or bypass, 24-26but this technique was not used in our patients. The 5- and i0-year survival of patients with acute and chronic dissections of the descending thoracic aorta is similar to those noted by Miller et al. s and DeBakey et al.7 The high risk of recurrent or residual aneurysms and dissections in both the distal thoracic and abdominal aorta indicates need for careful follow-up of these patients and emphasizes that repair of the primary site of dissection is not always curative. The issue of timing of repair of acute descending aortic dissection is unsettled. Our surgical results suggest that the operative mortality can be reduced by delaying repair at least 2 weeks after the acute event. Firm conclusions regarding such an approach cann% be made until additional information on the outcome of all patients treated by nonoperative ~ethods is available. Miller et al.27 continue to advocate early surgical intervention for patients with acute aortic dissection and have noticed a substantial (but statistically nonsignificant) reduction in operative mortality for repair of acute dissections in their more recent experience. In summary then, our experience indicates that repair of chronic dissection of the thoracic aorta has a lower operative risk than repair of acute aortic dissections, and we believe that initial medical management of acute dissections is indicated if there are no early complications. Risk of spinal cord ischemia is significantly reduced by partial cardiopulmonary bypass or shunt and is preferred over aortic cross-clamping alone. The beneficial effects of protection of the

Volume 3 Number 2 February 1986

distal circulation are m a r k e d w h e n the p e r i o d o f aortic ~, "~clusion exceeds 35 to 45 minutes. P o s t o p e r a t i v e patients require careful a n d regular l o n g - t e r m followu p because o f the h i g h incidence o f residual o r recurrent aortic aneurysms.

REFERENCES 1. Wheat Jr MW, Palmer RF, Bartley TD, Seelman RC. Treatment of dissecting aneurysms of the aorta without surgery. J Thorac Cardiovasc Surg 1965; 50:364-73. 2. Wheat Jr MW, Harris PD, Maim JR, Kaiser G, Bowman Jr FO, Palmer RF. Acute dissecting aneurysms of the aorta: Treatment and results in 64 patients. J Thorac Cardiovasc Surg 1969; 58:344-51. 3. Wheat Jr MW. Acute dissecting aneurysms of the aorta: Diagnosis and treatment--1979. Am Heart J 1980; 99: 373-87. 4. Doroghazi RM, Slater EE, DeSanctis RW, Buckley MJ, Austen WG, Rosenthal S. Long-term survival of patients with treated aortic dissection. J Am Coil Cardiol 1984; 3:102634. 5..Mil~, i DC, Stinson EB, Oyer PE, Rossiter SJ, Reitz BA, Griepp RB, Shumway NE. Operative treatment of aortic dissections: Experience with 125 patients over a sixteen-year period. J Thorac Cardiovasc Surg 1979; 78:365-82. 6. DeBakey ME, Cooley DA, Creech Jr O. Surgical consideration of dissecting aneurysm of the aorta. Ann Surg 1955; 142:586-612. 7. DeBakey ME, McCollum CH, Crawford ES, Morris Jr GC, Howell J, Noon GP, Lawrie G. Dissection and dissecting aneurysms of the aorta: Twenty-year follow-up of five hundred twenty-seven patients treated surgically. Surgery 1982; 92:1118-34. 8. Appelbaum A, Karp RB, I~rldin JW. Ascending vs. descending aortic dissections. Ann Surg 1976; 183:296-300. 9. Reul Jr GJ, Cooley DA, Hallman GL, Reddy SB, Kyger III R, Wukasch DC. Dissecting aneurysm of the descending aorta: Improved surgical results in 91 patients. Arch Surg 1975; 110:632-40. 10. Livesay JJ, Cooley DA, Ventimiglia RA, Montero CG, Warrian RK, Brown DM, Duncan JM. Surgical experience in descending thoracic aneurysmectomy with and without adjuncts to avoid ischemia. Ann Thorac Surg 1985; 39:37-46. 11. Crawford ES, Walker III HSJ, Salwa SA, Normann NA. Graft replac4,,ent of aneurysm in descending thoracic aorta: Results without bypass or shunting. Surgery 1981; 89:73-85. 12. Kouchoukos NT, Lell WA, Karp RB, Samuelson PN. Hemodynamic effects of aortic clamping and decompression with a temporary shunt for resection of the descending thoracic aorta. Surgery 1979; 85:25-30. 13. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. Am J Stat Assoc 1958; 53:457-81.

DISCUSSION Dr. Joseph N. C u n n i n g h a m (Brooklyn, N.Y.). I congratulate the authors on their clear presentation of data in this retrospective study. Their observations on the rela-

Results after repair of desceuding thoracic aorta dissections 235

14. Peto R, Peto J. Asymptotically efficient rank invariant procedures (with discussion). J R Stat Soc, series A 1972; 135:185-207. 15. CoxDR. Regression models andlife-tables (with discussion). J R Stat Soc, series B 1972; 34:187-220. 16. Walker SH, Duncan DB. Estimation of the probability of an event as a function of several independent variables. Biometrika 1967; 54:167-79. 17. Frantz PT, Murray GF, Shallal JA, Lucas CL. Clinical and experimental evaluation of left ventriculoiliac shunt bypass during repair of lesions of the descending thoracic aorta. Ann Thorac Surg 1981; 31:551-7. 18. Roberts AJ, Nora JD, Hughes WA, Quintanilla AP, Ganote CE, Sanders Jr JH, Moran JM, Michaelis LL. Cardiac and renal responses to cross-clamping of the descending thoracic aorta. J Thorac Cardiovasc Surg 1983; 86:732-41. 19. Gelman S, Reves JG, Fowler K, Samuelson PN, Lell WA, Smith LR. Regional blood flow during cross-clamping of the thoracic aorta and infusion of sodium nitroprusside. J Thorac Cardiovasc Surg 1983; 85:287-91. 20. Symbas PN, Pfaender LM, Drucker MH, Lester JL, Gravanis MB, Zacharopoulos L. Cross-clamping of the descending aorta: Hemodynamic and neurohumoral effects. J Thorac Cardiovasc Surg 1983; 85:300-5. 21. Katz NM, Blackstone EH, Kirklin JW, Karp RB. Incremental risk factors for spinal cord injury following operation for acute traumatic aortic transection. J Thorac Cardiovasc Surg 1981; 81:669-74. 22. Brewer III LA, Fosburg RG, Mulder GA, Verska JJ. Spinal cord complications following surgery for coarctation of the aorta: A study of 66 cases. J Thorac Cardiovasc Surg 1972; 64;368-81. 23. Berendes JN, Bred& JJ, Schipperheyn JJ, Mashhour YAS. Mechanisms of spinal cord injury after cross-clanaping of the descending thoracic aorta. Circulation 1982; 66:(suppl 1): 112-6. 24. Cunningham Jr JN, Laschinger JC, Merkin HA, Nathan IM, Colvin S, Ransohoff J, Spencer FC. Measurement of spinal cord ischemia during operations upon the thoracic aorta. Ann Surg 1982; 196:285-96. 25. Coles JG, Wilson GJ, Sima AF, Klement P, Tait GA. Intraoperative detection of spinal cord ischemia using somatosensory cortical evoked potentials during thoracic aortic occlusion. Ann Thorac Surg 1982; 34:299-306. 26. Laschinger JC, Cunningham Jr JN, Nathan IM, Knopp EA, Cooper MM, Spencer FC. Experimental and clinical assessment of the adequacy of partial bypass in maintenance of spinal cord blood flow during operations on the thoracic aorta. Ann Thorac Surg i983; 36:417-26. 27. Miller DG, Mitchell RS, Oyer PE, Stinson EB, Jamieson SW, Shumway NE. Independent determinants of operative mortality for patients with aortic dissections. Circulation i984; 70:(suppl 1):153-64.

tionship o f postoperative paraplegia to adequacy of distal circulation during aortic cross-clamping are quite pertinent and confirm our previous observations that measures to perfuse the aorta distally when possible will decrease the

236 Jex et al.

incidence of subsequent neurologic injury. However, distal perfusion, when inadequate, cannot be expected to change the incidence of paraplegia associated with thoracic aortic cross-clamping. We have employed the modality of evoked corticospinal potential measurements in 33 consecutive patients since 1982 to assess the adequacy of distal circulation during aortic cross-clamping and to aid in identification of critical intercostal vessels for reimplantation both in the chronically dissected aorta and in aneurysms of the atherosclerotic variety. Several statistically significant observations have been made. First, loss of evoked potential indicates inadequate distal perfusion or inadvertent interruption of critical intercostal vessels. Second, loss of evoked potentials for periods greater than 30 minutes are associated with an incidence of paraplegia in excess of 70%. This latter observation is similar to results reported here today relative to the relationship of cross-damp time to incidence of paraplegia. Finally, we have never encountered a case of paraplegia or any neurologic injury postoperatively when loss of evoked potential was limited to a time period less than 30 minutes following aortic cross-clamping. I would like to ask the authors three questions. First, ar e you currently employing evoked potential monitoring during thoracic aortic surgery? Second, could you clarify whether there was a difference in the rate of paraplegia between aortic repairs performed with a Gort shunt or distal bypass with a pump circuit? Finally, could you better define for us what the limits of "adequate" distal circulation are following aortic cross-clamping? Dr. D. Craig Miller (Stanford, Calif.). This is an important report, and I am pleased to see a group of clinical investigators finally stand up and say what many of us have been thinking for a long time. These investigators have jumped right in the middle of two major controversies. (1) Whether one should protect arterial perfusion to the spinal cord during repair of descending thoracic aortic dissections, and (2) whether nonadjunctive medical treatment or medical plus surgical treatment is best for patients with descending (type B) aortic dissections. I would like to share some background information with you that will shed a little perspective on the clinical dilemmas that Dr. Jex and his colleagues have addressed. First, why is there so much controversy about the most optimal management for patients with acute type B aortic dissections today? As you see here, in this large composite series of patients with acute type B aortic dissections treated medically or surgically in the 1960s and 1970s, the overall surgical hospital mortality rate was 38% in contrast to 22% for those patients treated medically. On the basis of these unfavorable results, emergency surgical treatment of patients with acute type B dissections fell out of general favor. The management strategy at some of these prestigious medical centers, however, called for nonadjunctive medical treatment. Failure of medical treatment, unfortunately, was

Journal of VASCULAR SURGERY

frequently punctuated by aortic rupture or irreversible renal/visceral infarction or ischemia; surgical interv%;fion under these conditions, as you see, was associated with operative mortality rates in the range of 75%. Under these circumstances this low salvage rate is understandable, but certainly not optimal. On the other hand, the operative risk for these patients has changed over the last two decades. In our Stanford experience (1963 to 1982) with 175 patients with acute or chronic aortic dissections, 26 patients had acute type B and 28 had chronic type B dissections. Over the entire 20year interval, the operative mortality rate for these 54 patients with type B dissections was high--38%--as also indicated by the Mayo Clinic data presented here. On the other hand, if one focuses solely on the 1977 to 1982 experience at Stanford, you see that the contemporary operative risks are those patients with acute type B dissections; the risk is only 13% (11% for those with chronic type,,~, dissections). For all 67 patients operated on since 19:p~/ (including 50 patients with type A dissections), ~ e overall mortality rate was 9% -+ 4% (-+ 70% confidex,ee level). Therefore, the operative risk ,for patients with acute aortic dissections, and most importantly those with type B dissections, has fallen rather markedly. I do not believe, furthermore, that these more favorable results are unique to Stanford. Therefore, the question of whether nonadjunctive medical management remains the most optimal treatment for patients with acute type B dissections should probably be reassessed. On the basis of experience, multivariate analysis revealed three independent significant determinants of increased likelihood of hospital death for 54 patients with type B dissections: older age, rupture, and renal and/or visceral ischemia. As many of you have inferred, it is all too commonly one of these factors that constitutes (sudden) "failure of medical management." Thus, we are left with a somewhat illogical and paradoxical situation in which the exact indications for operation are identical to those factors that are known to portend a higher operative, risk. Second, a major message has been sounded here that one method or another should be used to ens~'e~continued perfusion to the spinal cord to minimize the incidence of paraplegia. Since many different shunt or perfnsion techniques were used at the Mayo Clinic during this period, these investigators can, albeit retrospectively, shed light on this vexing clinical problem. At Stanford, partial (femorofemoral) cardiopulmonary bypass with complete heparinization has always been used for patients with acute or chronic type B (descending) dissections and all other patients with descending thoracic aortic aneurysms. As previously reported, our incidence of new paraplegia is 3% for patients with degenerative or atherosclerotic aneurysms and 4% for those with aortic dissections. We realize that no method of spinal cord protection proffers complete immunity from the risk of paraplegia, but, at least in our

Volume 3 Number 2 February 1986

hands and at the Mayo Clinic, the incidence can be minimi~ ~l if some method of distal perfusion is employed. In closing, the obvious question is whether these results have altered the current clinical strategy at the Mayo Clinic? Is the initial treatment of all patients with acute type B dissections still medical therapy only? What are your current operative mortality rates? I suspect they have fallen, as seen in our experience at Stanford. Dr. Schaff (closing). Dr. Miller and his colleagues have led the way in analysis of results of repair of aortic dissections. He asked about our current strategy for treating patients with acute dissections, and I can only give you mine and that of my surgical colleagues. We believe that most patients with acute dissection of the descending aorta can be managed medically, with the initial use of nitroprusside in conjunction with beta-blockade. He mentioned the difficulty in operating urgently on such patients if com-,lications ensue. In our experience, the most common in-

Results after repair of descending thoracic aorta dissections 237

dication for proceeding with early operative repair in patients with acute dissection is not acute arterial occlusion or aneurysm rupture but rather the inability to control pain. Dr. Cunningham, we have followed your work and that of your colleagues at New York University with sensory-evoked potentials but have not used this technique in any of the patients in this series. We reemphasize that our data on spinal cord injury include both paraplegia and transient or permanent paraparesis. Spinal cord injury is a potential complication during repair of dissection of the descending thoracic aorta even when partial cardiopulmonary bypass and shunts are used. We were not, in this review, able to assess the adequacy of the distal circulation with either of these methods, and it is possible that technical problems with perfusion contributed to spinal cord ischemia in the few patients in this group in whom it occurred.