Despite contemporary surgical advances, patients with acute type A aortic dissection still present significant operative challenges both in terms of short- and long-term outcomes. Reported operative mortality rates are high (15%–28%) with stroke rates of 2% to 15%.1–7 Operative approach varies among surgeons and has been implicated in outcomes.5,8–11 The original technique to treat acute type A aortic dissection entailed aortic cross-clamping at the level of the innominate artery with simultaneous repair of the proximal and distal ascending aortic pathology.1,2,12 Open distal anastomosis with hypothermic circulatory arrest (HCA) has since been advocated,8,13,14 with retrograde cerebral perfusion (RCP) introduced as an adjunct in the mid-1990s.15–17 In more recent years, antegrade cerebral perfusion has been used for elective aortic arch procedures,9,18 but to a lesser degree in acute dissection.19 Multiple preoperative and operative factors have been implicated in 30-day survival, yet there is a paucity of studies examining long-term outcomes related to treatment and modalities that may contribute to the risk of reoperation following successful acute repair. The aim of this study was to evaluate factors that impacted both short-term and long-term outcomes following repair of acute type A aortic dissection.
The purpose of this study was to evaluate factors that impact outcome following repair of type A aortic dissection. Over 25 years (1984–2009), 252 patients underwent repair of acute type A dissection. Mean follow-up for reoperation or death was 6.9±5.9 years. Operative mortality was 16% (41 of 252). Multivariate analysis identified one risk factor for operative death: presentation malperfusion (P=.003). For operative survivors, 5-, 10-, and 20-year survival was 78%±3%, 59%±4%, and 24%±6%, respectively. Late death occurred earlier in patients with previous stroke (P=.02) and chronic renal insufficiency (P=.007). Risk factors for late reoperation included male sex (P=.006), Marfan syndrome (P<.001), elevated systolic blood pressure (SBP, P<.001), and absence of β-blocker therapy (P<.001). Kaplan-Meier analysis demonstrated at 10-year follow-up that patients who maintained SBP <120 mm Hg had improved freedom from reoperation (92±5%) compared with those with SBP 120 mm Hg to 140 mm Hg (74%±7%) or >140 mm Hg (49%±14%, P<.001). At 10-year follow-up, patients on β-blocker therapy experienced 86%±5% freedom from reoperation compared with only 57%±11% for those without (P<.001). Operative survival was decreased with preoperative malperfusion. Long-term survival was dependent on comorbidities but not operative approach. Reoperation was markedly increased in patients not on β-blocker therapy and decreased with improved SBP control. Strict control of hypertension with β-blocker therapy is warranted following repair of acute type A dissection.
This is a retrospective review of 252 consecutive patients who underwent surgery for acute type A aortic dissection between June 1984 and December 2009 at Washington University School of Medicine (Barnes-Jewish Hospital) by 26 different surgeons. The study was approved by the Washington University institutional review board. Preoperative, intraoperative, and postoperative variables were entered into a database prospectively and data were reviewed retrospectively. Long-term data were obtained by a combination of chart review and direct patient contact or telephone interview.
Mean age was 60.2 years (range 18–88 years) and 158 patients (63%) were men (Table I). Rates of hypertension, diabetes, chronic obstructive pulmonary disease, peripheral vascular disease, renal failure, coronary artery disease, and smoking were in the expected high range for this patient population (Table I). Marfan syndrome was evident in 14 patients (6%). Malperfusion was present preoperatively in 73 (29%) patients, including coronary in 7 patients, peripheral in 35 (lower extremity in 23, upper extremity in 12), renal in 8, bowel in 2, spine in 7, and cerebral in 14. Aortic regurgitation was severe in 64 patients (25%) and moderate in 20 patients (8%). One hundred and ten patients (44%) presented with shock, and second-time surgery requiring redo sternotomy was required in 11 patients (4%).
|Mean age at presentation, y||60.2±15.5 (range 18–88)|
|None or mild||168 (67)|
|Cardiogenic shock||110 (44)|
|Previous cardiac surgery||11 (4)|
|CAD or previous MI||56 (22)|
|Renal failure||30 (12)|
|Marfan syndrome||14 (6)|
All patients had surgery for acute Stanford type A dissection, 187 (74%) were classified as DeBakey type 1, while 65 (26%) were DeBakey type II (Table II). The intimal tear was identified in the proximal ascending aorta in 221 (88%) patients, and the tear involved the arch in 20 (8%); in 11 (4%) patients the location of the tear was not visualized. The primary tear was entirely resected in 227 patients (90%). Three different perfusion strategies were used including aortic cross-clamping without HCA in 88 (35%), HCA alone in 88 (35%), and HCA with supplemental RCP in 76 (30%). The specific techniques used varied by surgeon preference and were similar to those previously described.5,7,10
|Aortic valve intervention|
|No valve procedure||182 (72)|
|Composite valve graft||49 (19)|
|Separate valve and graft||14 (6)|
|Valve-sparing root||7 (3)|
|Extent of distal resection|
|Ascending aorta only||179 (71)|
|Hemi-arch replacement||68 (27)|
|Full arch replacement||5 (2)|
|Tear resected||227 (90)|
|Coronary artery bypass graft||46 (18)|
|Antegrade cooling||43 (17)|
|Retrograde cooling||209 (83)|
|Antegrade rewarming||146 (58)|
|Retrograde rewarming||106 (42)|
|Deep hypothermic circulatory arrest||164 (65)|
|Retrograde cerebral perfusion||76 (30)|
|Mean pump time, min||184±71|
|Mean cross-clamp, min||110±50|
|Hypothermic circulatory arrest, min||35±15|
|Operative death||41 (16)|
Valve replacement was required in 63 (25%) patients, including separate aortic valve prosthesis and ascending graft in 14 (6%), composite valve graft root replacement with coronary reimplantation in 49 (19%), and valve-sparing root replacement was performed in 7 patients (3%). Coronary bypass grafting was required in 46 (18%) patients. Mean cardiopulmonary bypass time for the entire group was 184±72 minutes, while it was 168±78 minutes for the 88 patients who had the ascending aorta replaced with a cross-clamp application, and was 193±67 minutes in the 164 patients who underwent HCA. In those who had the ascending aorta replaced with a cross-clamp application, the mean cross-clamp time was 102±45 minutes. For those who underwent HCA, the time of circulatory arrest was 35±16 minutes, while cross-clamp time was 114±52 minutes. Cooling was done in an antegrade fashion in 43 patients (17%) and retrograde in 209 patients (83%). Patients were rewarmed with antegrade flow in 146 (58%) and retrograde flow in 106 (42%).
Blood pressure (BP) control at late follow-up was categorized as follows: systolic BP (SBP) <120 mm Hg, 120 to 140 mm Hg, and >140 mm Hg; diastolic BP (DBP) ≤80 mm Hg or >80 mm Hg. Operative mortality was defined as death within 30 days or prior to hospital discharge.
Continuous data are reported as mean±standard deviation and compared using analysis of variance. Results were expressed as frequencies and percentages for categorical variables. Categorical variables were analyzed using the chi-square test or Fisher exact tests as appropriate. Odds ratios (ORs) are reported with 70% confidence intervals (CIs). Multivariate analysis (stepwise backward regression) was used to determine preoperative and intraoperative risk factors that were significant independent predictors of increased operative mortality, increased need for reoperation (long-term), and impaired long-term survival (SigmaStat 2.03; SPSS Inc, Chicago, IL).
Twenty-eight variables were analyzed to determine risks for operative mortality: age, year of operation, sex, hypertension, diabetes mellitus, coronary artery disease, pulmonary disease, cerebrovascular disease, peripheral vascular disease, chronic renal insufficiency, smoking history, Marfan syndrome, previous cardiac operation, cardiogenic shock, aortic insufficiency, preoperative flow complications, tamponade/rupture, DeBakey classification, intraoperative dissection, HCA, RCP utilization, initial perfusion antegrade/retrograde, rewarming antegrade/retrograde, primary tear location, primary tear resected, aortic valve preservation vs replacement, ascending only vs hemiarch replacement, concomitant coronary artery bypass grafting (CABG), use of antihypertensive medications (β-blockers, angiotensin-converting enzyme [ACE] inhibitors, calcium channel blockers, diuretics), and use of a statin. Multivariate analysis to determine risks for reoperation and long-term survival also included late SBP and late DBP (as described above). Kaplan-Meier analysis was employed to evaluate the effect of BP and β-blocker therapy on survival and need for reoperation. P values <.05 were considered statistically significant and all P values reported are two-sided.
Operative mortality was 16% (41 of 252). Multivariate analysis to determine factors that influenced operative mortality identified only one independent predictor: initial presentation with branch-vessel malperfusion (P=.01; OR, 2.5; CI, 1.8–3.6). Postoperative complications included permanent stroke in 17 patients (7%), myocardial infarction in 3 (1%), reoperation for bleeding in 34 (13%), respiratory failure (requiring >24 hours of ventilation) in 70 (28%), and renal failure requiring hemodialysis in 17 (7%).
The 88 patients who had the ascending aorta replaced with a cross-clamp application were compared with the 164 patients who underwent HCA. The cardiopulmonary bypass time was significantly shorter in the cross-clamp alone group (168±78 minutes) when compared with those who underwent HCA (193±67 minutes, P=.008), and there was a trend toward shorter cross-clamp times, but this was not significant (102±45minutes vs 114±52 minutes, P=.06). The rate of reoperation for bleeding was 7% (6 of 88) in the group who underwent cross-clamp alone vs 17% (28 of 164) in those who underwent HCA (P=.04).
Late Survival and Reoperation
Mean follow-up was 6.9±5.9 years in patients discharged from the hospital. Completeness of follow-up was 100% for reoperation and survival, 89% for late usage of β-blocker therapy (188 of 211 operative survivors), and 84% for late BP control (178 of 221 operative survivors). In the 211 operative survivors, 5-, 10-, and 20-year survival was 78%±3%, 59%±4%, and 24%±6%, respectively (Figure 1). Multivariate analysis revealed that late death was increased with earlier operative year (P<.001), advanced age (P<.001), history of cerebrovascular accident (P=.02; OR, 2.4; CI, 1.7–3.3), and chronic renal insufficiency (P=.005; OR, 2.4; CI, 1.7–3.3), but was independent of operative approach (ascending vs hemiarch [P=.12], cross-clamp vs circulatory arrest [P=.42], and aortic valve repair vs replacement [P=.38]). The 1-, 5-, 10-, and 20-year survival was 86%, 80%, 70%, and 37% for patients younger than 55 years at the time of dissection, while it was 72%, 61%, 44%, and 0% for patients who were 55 to 70 years and 72%, 55%, 31%, and 0% for patients who were older than 70, respectively (P<.001). Kaplan-Meier analysis of patients who did not present with renal failure showed that the 1-, 5-, 10-, and 20-year survival was 77%, 67%, 50%, and 24% while it was 73%, 53%, 36%, and 0% for those who presented with renal failure (P=.05).
Twenty-seven (13%) of 211 operative survivors underwent 30 late reoperations 59±50 months postoperatively (range, 1–170 months), with a reoperative mortality rate of 7%. Eight proximal reoperations required an anterior approach to replace the residual ascending aorta and arch in 5 patients or repair an ascending aortic pseudoaneurysm in 3 patients. Fifteen reoperations involved descending (n=5), thoracoabdominal (n=7), or infrarenal (n=3) aortic replacement, of which 6 were accomplished using endovascular techniques. Additional reoperations included innominate artery bypass (n=3), superior mesenteric bypass (n=1), and aortic valve replacement for senile stenosis (n=1), residual regurgitation (n=1), or failed homograft (n=1). Risk factors for late reoperation included male sex (P=.006; OR, 2.9; CI, 1.9–4.3), Marfan syndrome (P<.001; OR, 7.8; CI, 4.2–14.2), SBP >120 mm Hg at late follow-up (P<.001; OR, 8.9; CI, 4.6–17.), and absence of β-blocker therapy at late follow-up (P<.001; OR, 6.1; CI, 3.8–9.7).
Evaluation for presence of β-blocker usage at late follow-up revealed that 40 of 188 (21%) patients were not receiving β-blockers. Reoperation was performed in 12 of 148 (8%) on β-blocker therapy and 14 of 40 (35%) without β-blocker therapy (P<.001). Kaplan-Meier analysis demonstrated freedom from reoperation at 10 and 15 years of 86%±5% and 83%±6% with β-blocker therapy, but only 57%±10% and 37%±14% without β-blocker therapy (P<.001, Figure 2).
In regards to late death, there was no relation to β-blocker (P=.315), calcium channel blocker (P=.517), ACE inhibitors (P=.326), and statins (P=.955) in univariate analysis. There was increased late death rate in patients who were taking diuretics (P<.001; OR, 4.7; CI, 3.1–7.0). When all medications at late follow-up were included in the multivariate analysis of late death, diuretics remained a significant predictor (P<.001).
Evaluation for BP control at late follow-up revealed that 85 of 178 patients (48%) had SBP <120 mm Hg, 63 (35%) had SBP of 120 to 140 mm Hg, and 30 (17%) had SBP >140 mm Hg. Reoperation was performed in 3 of 85 patients (4%) with SBP <120 mm Hg, 13 of 63 patients (21%) with SBP 120 mm Hg to 140 mm Hg, and 10 of 30 patients (33%) with SBP >140 mm Hg (P<.001). Kaplan-Meier analysis demonstrated freedom from reoperation at 10 and 15 years of 92%±5% and 92%±5% when SBP was <120 mm Hg, 74%±7% and 66%±10% when SBP was 120 mm Hg to 140 mm Hg, and 49%±14% and 30%±14% when SBP was >140 mm Hg (P<.001, Figure 3).
Since the initial reports of surgical repair for acute aortic dissection described by DeBakey and Cooley’s group in the mid-1960s,20,21 mortality and morbidity has remained high despite the many medical and surgical advances made during this period.1,22,23 A report using information from the International Registry of Acute Dissections (IRAD) demonstrated that in 1334 consecutive patients between 1996 and 2003 at 18 large tertiary care centers, operative mortality was 24%,24 which although high, is substantially better than medical management alone, which carries a short-term mortality rate of 60%.23,25 The current report, spanning 25 operative years among 26 surgeons at a single academic center, demonstrated an operative mortality of 16%, which compares favorably to the IRAD report and other previous reports.1,12,22,24 The IRAD study found several factors that influenced operative survival, including age older than 70 years, prior cardiac surgery, hypotension or shock at presentation, migrating pain, cardiac tamponade, electrocardiogram evidence of myocardial ischemia, or any pulse deficit.24 The current report found that only malperfusion on presentation was predictive of impaired short-term outcomes. Some argue that elderly patients with type A dissection pose a prohibitive risk to operation.26 The current report and others have demonstrated acceptable mortality rates in elderly patients,27–29 and we argue that age alone should not be used as criteria to recommend medical therapy alone. Likewise, shock at presentation and myocardial ischemia decreased operative survival in the IRAD report but did not influence survival in this current study. Although this discrepancy could be attributed to the smaller numbers in the current study, caution must be exercised concerning refusal of surgery based solely on preoperative risk factors.
The literature has conflicting reports on the influence of different operative techniques on short-term outcomes.3,5,8–11,15–18,30,31 The current study identified no difference in short- or long-term survival based on operative technique, including cross-clamp vs HCA, extent of distal resection, antegrade vs retrograde perfusion, antegrade rewarming vs retrograde rewarming, or the use of retrograde cerebral perfusion. Today, several operative techniques continue to be employed in this high-risk patient population, and we feel that surgeons should use the technique with which they are most comfortable.
Long-term survival has been scarcely reported for patients who have undergone repair of acute type A dissection. The IRAD investigators reported a rate of 91% 3-year survival among operative survivors.32 To the best of our knowledge, the current report is the first to include long-term follow-up of patients out to 25 years. In this report, 5-, 10-, and 20-year survival was 78%, 59%, and 24%, respectively, which can be used as a benchmark with which future long-term studies following newer treatment modalities (eg, central cannulation or simultaneous endovascular stenting of the descending aorta during distal aorta anastomosis)33,34 can be compared.
Few reports have been able to elucidate characteristics that impact long-term mortality.4,5,10,27,35 Chiappini’s group found that only type II diabetes impacted late survival.35 Tsai and associates found that patients who died within 3 years of surgery had increased SBP compared with those who had not died (130 mm Hg vs 122 mm Hg, P<.01) and that vasodilator use was more common in nonsurvivors at 3 years (32.6% vs 17.3%, P=.02).32 The current study demonstrated that long-term survival was impaired for patients who had previous stroke or chronic renal insufficiency and the need for reoperation was improved for patients with better BP control at late follow-up. The increased death rate in patients taking diuretics potentially represents them functioning as a surrogate for late congestive heart failure (CHF), more so than hypertension. In fact, only 8% of patients without CHF symptoms at late follow-up were taking diuretics compared with 45% who were in higher New York Heart Association classes at late follow-up.
Patients who received β-blocker therapy also had a marked improvement in need for reoperation. Patients without β-blockers had 46% freedom from reoperation rate at 10 years, while those on β-blocker therapy had >80% freedom from reoperation. We hypothesize that the decreased need for reoperation in the group receiving β-blocker therapy and good BP control was likely due to diminished stress on the diseased aortic wall and a concurrent decrease in dP/dT (impulse). β-Blockers decrease both the dP/dT as well as the maximal BP when compared with vasodilators, which lower BP at the expense of an elevation of dP/dT and have been previously associated with improved survival.32 Controlling BP and the aortic impulse plays a large role in decreasing the rate of continued damage to the already diseased aorta.36–39 Our laboratory has previously demonstrated that an increase in BP of 26 mm Hg has a stress equivalent to an increase of aortic diameter of 1 centimeter.40 We feel strongly that patients should be on β-blocker therapy with stringent BP control to minimize the reoperative risk over their lifetime. Although there was a lower rate of death in patients taking β-blockers (37% taking β-blockers compared with 48% not taking β-blockers) at late follow-up, it was not statistically significant. This study may have been underpowered to detect a statistical difference in late mortality based on β-blocker usage.
Acute type A dissection remains a formidable problem with relatively high mortality and morbidity rates despite advances in both medical and surgical therapy. Although operative approach and most preoperative variables were not found to influence short-term survival, operative survival decreased with branch vessel malperfusion at the time of presentation. Long-term survival was dependent on underlying comorbidities (preoperative stroke and renal failure) but not operative approach. Reoperation was markedly decreased in patients on β-blocker therapy and with improved BP control at late follow-up. Strict control of hypertension with β-blocker therapy for life is warranted following repair of acute type A dissection.
Acknowledgments and disclosure: The authors gratefully acknowledge the clinical contributions of Hendrick B. Barner, MD, R. Morton Bolman, MD, James L. Cox, MD, Thomas B. Ferguson, MD, T. Bruce Ferguson, Jr, MD, William A. Gay, Jr, MD, Charles B. Huddleston, MD, Scott H. Johnson, MD, Nicholas T Kouchoukas, MD, Jennifer S. Lawton, MD, Nader Moazami, MD, Michael K. Pasque, MD, Michael Rosenbloom, MD, Thomas L. Spray, MD, Thoralf M. Sundt, III, MD, I-wen Wang, MD, and Clarence S. Weldon, MD. The authors have no relationships with industry for disclosure.