Contemporary management and outcomes of acute type A aortic dissection: An analysis of the STS adult cardiac surgery database

1 INTRODUCTION

Acute type A aortic dissection (AAAD) remains a highly lethal disease. Without surgical therapy, mortality exceeds 50% at 30 days.1, 2 Despite significant improvements in intra-operative management and post-surgical critical care, mortality, and morbidity rate still remain high, with reported mortality rates of 25% or more in several large series.3, 4

There are currently no prospective randomized data to guide surgical strategies for AAAD. Key areas of uncertainty include choice of arterial cannulation site, temperature management on cardiopulmonary bypass, and strategies for brain protection during deep hypothermic circulatory arrest (HCA). We utilized the Society of Thoracic Surgeons Adult Cardiac Surgery Database (STS ACSD) to examine current patient characteristics, predictors and outcomes for AAAD, and describe contemporary surgical management strategies in North America.

2 MATERIALS AND METHODS

2.2 Study population

A total of 4262 patients underwent surgery for aortic dissection between July 2011 and September 2012. After excluding for prior percutaneous intervention and/or cardiac surgery, 2982 AAAD patients from 640 sites nationwide were examined in this study (Figure 1).

3 RESULTS

3.1 Patient characteristics

Demographics and baseline clinical characteristics are shown in Table 1.

Table 1. Patient characteristics

	Number (N = 2982)	%	Median	Mean ± Std. deviation	Interquartile range
Demographics
Age (years)	2982	100	60.0	59.68 ± 14.01	50.00-70.00
Gender: female	1025	34.4
Race
Other	80	2.7
Asian	102	3.4
Hispanic	139	4.7
Black	563	18.9
Caucasian	2067	69.3
Missing	31	1.0
Risk factors
Diabetes
Insulin	59	2.0
No Insulin	266	8.9
Hypertension	2375	79.6
Last pre-op creatinine (mg/dL)	2909	97.6	1.1	1.30 ± 1.05	0.90-1.40
Last pre-op creatinine >2.0 mg/dL	176	5.90
Renal failure (dialysis/creatinine >2.0 mg/dL)
Renal failure—dialysis	58	2.0
Renal failure—no dialysis	125	4.2
No renal failure	2716	91.1
Missing	83	2.8
CVD w/o CVA
CVD-History CVA	232	7.8
CVD-No History CVA	105	3.5
No CVD	2634	88.3
Missing	11	0.4
CVD w/o CVA-time
CVA ≤2 weeks	90	3.0
CVA >2 weeks	142	4.8
Chronic lung disease
Severe	45	1.5
Moderate	86	2.9
Mild	200	6.7
No	2632	88.3
Missing	19	0.6
Peripheral vascular disease	598	20.1
Body surface area	2982	100	2.0	2.0 ± 0.3	1.8-2.2
Body mass index	2958	99.2	28.3	29.5 ± 6.9	25.0-32.8
Pre-operative cardiac status
Congestive heart failure within 2 weeks	309	10.4
NYHA classification
IV	133	4.5
III	76	2.6
II	69	2.3
I	20	0.7
Missing	2684	90.0
Myocardial infarction and time
MI >21 days	118	4.0
MI 8-21 days	10	0.3
MI 1-7 days	68	2.3
MI >6 h but <24 h	60	2.0
MI ≤6 h	83	2. 8
No MI	2632	88.3
Missing	11	0.4
Unstable angina—no recent MI	566	19.0
Cardiogenic shock	472	15.8
Resuscitation	156	5.2
Atrial fibrillation	243	8.2
Ejection fraction	1420	47.6	58.0	55.8 ± 11.0	50.0-62.0
IABP—inotropes	279	9.4
Liver disease	92	3.1
Unresponsive neurologic state within 24 h of surgery	99	3.3

• CVA, cerebrovascular accident; CVD, cerebrovascular disease; IABP, Intra-aortic balloon pump; NYHA, New York Heart Association; MI, myocardial infarction.

The median time from hospital arrival to operating room entry was 2.6 (IQR 1.1-6.1) h, and 79% of patients underwent operation within 24 h of arrival (Figure 2). Cardiogenic shock was present in 16% (472/2982). In addition, 5% (156/2982) underwent active resuscitation prior to surgery.

3.2 Operative characteristics

Operative characteristics are shown in Table 2. Overall procedure times (median 5.8 h), and perfusion times (median 181 min) were long. The median cross-clamp time was 102 min. There was significant variation in the lowest temperature on bypass, with a median lowest temperature of 20°C (IQR 18-25) (Figures 3A and 3B).5

Table 2. Operative characteristics

	Number (N = 2982)	%	Median	Mean ± std. deviation	Interquartile range
Status of procedure
Emergent salvage	82	2.6
Emergent	2466	82.7
Urgent	329	11.0
Elective	103	3.45
Missing	2	0.1
Operative times
Total operative time (hours)	2967	99.5	5. 8	6.1 ± 2.3	4.7–7.3
Time to surgery (arrival to OR)	2516	84.4	2.6	6.6 ± 11.0	1.1 − 6.1
CPB details
Perfusion time (min)	2929	98.2	181.0	195.4 ± 80.2	143.0 − 233.0
Lowest temperature ( °C)	2817	94.5	20.0	21.4 ± 5.4	18.0 − 25.0
Lowest hematocrit	2780	93.2	23.0	23.0 ± 4.8	20.0 − 26.0
Arterial cannulation
Aortic	858	29.1
Femoral	1354	45.9
Axillary	914	31.0
Other	194	6.6
Cross clamp details
Aortic occlusion
Balloon occlusion	4	0.1
Aortic cross clamp	2909	97.6
None/partial cross clamp	63	2.1
Missing	6	0.2
Cross clamp time (min)	2889	96.9	102.0	112.6 ± 58.3	71.0 − 141.0
Circulatory arrest details
Circulatory arrest
No circulatory arrest	636	21.3
Circ. arrest/no cerebral perfusion	935	31.3
Cerebral perfusion:
Antegrade	709	23.8
Retrograde	586	19.7
Both	88	3.0
Unknown	18	0.6
Missing	10	0.3
Circulatory arrest time (min):
With no cerebral perfusion	917	30.8	25.0	27.2 ± 13.9	19.0 − 33.0
With cerebral perfusion	1302	43.7	5.0	16.3 ± 22.3	0.0 − 28.0
Cerebral perfusion time (min)	1390	46.6	30.0	34.3 ± 21.2	22.0 − 41.0
Circulatory arrest + cerebral perfusion time (min)	1296	43.5	40.0	50.4 ± 34.1	29.0 − 62.0
Operative indicators
Intra-Op TEE performed	1957	65.6
Aortic insufficiency (TEE based)
Trivial	917	46.9
Mild	512	26.2
Moderate	292	14.9
Severe	79	4.0
N/A	50	2.6
Missing	107	5.5
Concomitant procedures
Concomitant CABG	427	14.3
Concomitant aortic valve	1575	52.8
Valve sparing root remodeling (Yacoub)	5	0.3
Valve sparing reimplantation (David)	36	2.3
Homograft	1	0.1
Resuspension AV with replacement of ascending aorta	712	45.2
Resuspension AV without replacement of ascending aorta	60	3.8
Replacement and insertion of aortic non-valved conduit	84	5.3
Root reconstruction with valved conduit	387	24.6
Repair/reconstruction	84	5.3
Replacement	206	13.1
Concomitant mitral valve	43	1.4
Repair	25	58.1
Replacement	16	37.2
Missing	2	4.7
Concomitant endovascular procedure:
None	2922	98.0
No debranching	38	1.3
Debranching	10	0.3
Unspecified	1	0.1
Missing	11	0.3
Concomitant other procedure	352	11.8

• AV, aortic valve; CABG, coronary artery bypass graft; OR, operating room; TEE, transesophageal echocardiogram.

The arterial cannulation strategy used for AAAD repair is presented in Figure 4. While the STS ACSD is limited in the granularity of reporting of operative strategies, the results reported here suggest that among patients undergoing AAAD repair, a strategy of initial femoral cannulation was most commonly utilized, with axillary cannulation being the second most common approach.

3.2.1 Circulatory arrest

Circulatory arrest was not used in 21% of cases. Among the remaining 79% of patients, circulatory arrest was performed without brain protection in 40.3%, with antegrade cerebral perfusion in 30.6%, retrograde cerebral perfusion in 25.3%, and both antegrade and retrograde in 3.4% of cases.

For patients undergoing circulatory arrest without brain protection, median circulatory arrest time was 25 min (IQR 19-33). Among patients undergoing circulatory arrest with brain perfusion, median brain perfusion time was 30 min (IQR 22-41), and the median total body circulatory arrest time was 40 min (IQR 29-62).

3.2.2 Procedures performed

Concomitant CABG was performed in 14% of cases, and an aortic valve procedure in 53% of cases. Among those patients having aortic valve procedures, the following were performed: (1) root reconstruction with valved conduit (root replacement): 25.6%; (2) aortic valve resuspension: 54%; (3) isolated aortic valve replacement: 13%; (4) valve-sparing root replacement: 2.6%.

Endovascular procedures were performed in 2% of cases.

3.2.3 Blood product utilization

As expected, blood product utilization was substantial in this population. Total blood product requirement (intra- and post-operatively) included a median of 5 (IQR 2-9) units of packed red blood cells, 4 (IQR 2-8) units of fresh frozen plasma, 1 (IQR 0-3) unit of cryoprecipitate, and 3 (IQR 2-4) units of platelets.

3.4 Center experience with AAAD

During the study period, one or more primary AAAD repairs were performed at only 640/1140 (56%) of centers reporting data to the STS ACSD, with a median of three cases performed at each center during the study period. Only 11% of centers (n = 72) performed 10 or more cases (Figure 5). The higher the hospital volume, the lower the unadjusted odds ratio (OR) for mortality (hospital volume ≥3, OR = 0.67, P = 0.004; hospital volume ≥10, OR = 0.68, P = 0.0001), and OR = 0.98 with each additional case increase (P < 0.0001).

4 DISCUSSION

This study represents a comprehensive and contemporary review of patient characteristics, operative management, and outcomes of patients with AAAD in North America. Key findings include a mortality rate of 17%, a stroke rate of 11%, and large variations in operative approaches with significant differences in cannulation strategy, cerebral protection strategy, and temperature management. In addition, only a few centers have significant experience with AAAD.

The overall mortality in this report of 17% is lower than the IRAD experience of 25%3 and the National Inpatient Sample analysis of 26%.4 However, the lower mortality rate could also be due to the exclusion of re-operative patients from this analysis. Mortality in the German Registry for Acute Aortic Dissection Type A (GERAADA) was similar at 20%.6 In contrast, single-center series have reported mortality rates of 2.8% to 9%.7-9 This discrepancy in outcomes between large registries and single-center series suggests opportunities for improvement in outcomes if best practices are identified and implemented.

This study highlights a remarkably large variation in practice and techniques used to treat AAAD in North America. The site used for arterial inflow was most commonly the femoral artery, with a substantial number of cases reporting axillary and direct aortic cannulation. Femoral cannulation enables rapid institution of cardiopulmonary bypass and has a long history of use in AAAD. However, this is in contrast to the Europeans where only 25% cannulate the femoral artery,10 and the Canadians where only 17% cannulate the femoral artery.11 Femoral cannulation can cause extension of dissection into the iliofemoral vasculature, which can lead to pressurization of the false lumen and malperfusion at the onset of cardiopulmonary bypass.12 There is some evidence that femoral cannulation is associated with increased embolic events,13, 14 although other studies have shown that there is no difference in stroke risk with femoral cannulation.15, 16 Axillary artery cannulation has more recently gained popularity as a cannulation site for aortic and arch surgery.17-20 This is especially so in Europe and Canada.10, 11 It simplifies antegrade cerebral perfusion during systemic circulatory arrest. There is some evidence that axillary artery cannulation improves outcomes for AAAD.20, 21 Other groups have proposed that direct aortic cannulation can be safe.22-24 Other potential sites of cannulation include the left ventricular apex, innominate artery,25 and carotid artery. However, the utilization of these approaches was unable to be determined in this study.

Perfusion management in this study was strikingly heterogeneous. Utilization of circulatory arrest in operations for AAAD is common as it allows for examination of the aortic arch for tears and a more complete resection of diseased ascending aorta. However, 22% of cases in this study were not performed under circulatory arrest. Among patients undergoing operation with hypothermic circulatory arrest, cerebral perfusion strategies varied greatly with upwards of 40% without any adjunctive cerebral perfusion during the period of circulatory arrest compared with only 29% in the GERAADA registry and 18% in the Canadian survey.10, 11 A large retrospective study from GERAADA found that for circulatory arrest times less than 30 min, hypothermic circulatory arrest alone compared with antegrade cerebral perfusion yielded similar results.26

Among those with adjunctive cerebral perfusion in this study, slightly more had antegrade cerebral perfusion (44%) as compared to retrograde cerebral perfusion. This is in contrast to the Europeans where 69% had antegrade cerebral perfusion and the Canadians where 72% had antegrade cerebral perfusion.10, 11 There is growing evidence to suggest that antegrade cerebral perfusion is associated with better long-term survival.15, 27 Others have reported use of combined antegrade and retrograde cerebral perfusion with good results as well.28 However, the necessity for and optimal route of cerebral perfusion is still controversial.

The minimum temperature achieved during circulatory arrest also varied greatly. Lansman et al and Griepp et al popularized profound hypothermia for organ protection during aortic arch surgery.29, 30 There is emerging evidence to suggest that mild to moderate hypothermia is comparable to deep hypothermia in terms of operative risk,31-33 and the optimal temperature during circulatory arrest remains unclear. One has to weigh the risks of potentially increased coagulopathy, increased length of operation, and increased systemic inflammatory response with deeper degrees of hypothermia versus the potential benefits of better organ protection.

The type of procedures performed to treat AAAD is largely dictated by the extent of the dissection. In this experience, more than half of patients required management of the aortic valve either with replacement or repair via resuspension. More than a quarter of patients required root replacement. Valve-sparing root replacement was performed very infrequently. Treatment of the aortic arch and descending thoracic aorta cannot be determined in this study as the information is not available in the database.

Stroke remains a devastating complication after AAAD repair and is a major cause of mortality. The rate of 11% in this report is similar to other large studies of AAAD despite significant differences in cerebral perfusion strategies. This is considerably higher than most other cardiac surgical procedures. Cerebral protection via antegrade or retrograde cerebral perfusion has been shown to decrease neurologic complications in surgery involving the aortic arch.34 However, in this study, cerebral perfusion was not correlated with lower neurologic injury.

Out of more than 1000 centers participating in STS ACSD, only 640 centers reported more than one case during the study period. Among those centers with experience in the treatment of AAAD, the majority performs fewer than three cases a year. This case distribution implies that few surgeons (and few centers) are able to accumulate substantial experience with the treatment of these complicated cases. There is a correlation between hospital experience and lower mortality. However, it is still unclear whether individual surgeon experience, rather than hospital experience, affects mortality.

5 CONCLUSIONS

Acute type A aortic dissection remains a highly lethal and morbid condition. Surgery to treat the disease is a relatively uncommon procedure with only about 3000 cases per year in North America and individual center and surgeon experience is limited. There are large variations in surgical practice, particularly with regards to temperature management on cardiopulmonary bypass, approach to brain protection, and arterial cannulation strategy. Future studies should examine intra-operative characteristics that might affect outcome including type of cannulation, lowest temperature on cardiopulmonary bypass, type of cerebral perfusion, and management of the aortic arch. Large multicenter trials will be required to definitively answer these important questions. Given the high mortality and morbidity of surgical treatment of AAAD and current heterogeneous treatment strategies, these trials are imperative.