Outcomes of aortic arch repair with extended (≥ 90 minutes) antegrade cerebral perfusion

Abstract

Objectives. To analyze outcomes with extended duration of antegrade cerebral perfusion (ACP) during hypothermic circulatory arrest (HCA) for total arch repair (TAR). Design. Retrospective study of consecutive patients undergoing TAR with HCA and ACP. Patients were divided into group A (ACP ≥ 90 min, n = 12) and group B (ACP < 90 min, n = 17) and compared regarding in-hospital mortality and neurological complications (primary outcome measures) and major complications, biochemical markers of organ damage, and blood product use (secondary outcome measures). Complications were analyzed according to standards of the International Aortic Arch Surgery Study Group. Results. Overall in-hospital mortality was 4/29 (14%); 1/12 (8.3%) in group A versus 3/17 (18%) in group B, p = 0.62. No grade-V (lethal) neurological complication occurred, but five patients (all in group B) had grade-IV neurological complications: one general and two each focal and spinal neurological deficit (p = 0.047, overall). Prevalence of grade-II (temporary) general neurological deficit was 17% (group A) versus 27% (group B), p = 0.66. None of the patients suffered ≥ grade-IV respiratory or renal complications. Other complications, biochemical markers, and blood product usage were not significantly different. Conclusion. Outcomes in TAR with HCA and extended (≥ 90 min) three-vessel ACP were encouraging and could be contemplated with anticipated time-consuming TAR.

Key words::

• aortic arch
• cerebral
• organ protection
• outcomes
• surgery

Introduction

Total aortic arch replacement (TAR) is a surgical challenge with substantial early mortality and morbidity related to patient factors, surgical procedures, and organ, specifically brain, protection during the period of hypothermic circulatory arrest (HCA). Since its popularization by Kazui et al. in the early 1990s (1), antegrade cerebral perfusion (ACP) has become a widespread means of cerebral protection, perceived as prolonging the “safe” period of HCA. Indeed, arch repair outcomes with ACP have improved substantially (2–5). Yet, uncertainty and conflicting results remain regarding the upper time limit of ACP in combination with HCA (3–4,6–7). The aim of the present study is to report outcomes [early mortality and complications, as defined by the recently published standards of the International Aortic Arch Surgery Study Group (IAASSG) (8)] of patients undergoing TAR with extended (≥ 90 min) duration of ACP, as compared with shorter ACP duration, to help clarify if long-lasting three-vessel ACP during deep HCA appears safe. Based on previously reported outcomes (3,6–7), the study hypothesis was that very long ACP is associated with increased rates of early mortality and major complications, especially neurological.

Patients and methods

Patients and data management

All patients undergoing TAR from 2009 to 2014 were identified. TAR was defined as arch repair with reimplantation of all patent arch vessels. Two groups were defined: Group A (study group) included all patients with ACP lasting at least 90 min (n = 12); a single patient operated in 2004 was also included in group A. Group B (comparison group) consisted of all remaining patients undergoing TAR with ACP (n = 17) in the same time period. Demographic, clinical, operative, radiological, echocardiographic, and survival data were retrieved from medical records.

Outcome measures and definitions

In-hospital mortality and permanent neurological deficit were the primary study outcomes. All major complications, as identified and classified by the IAASSG (8), were secondary outcome measures. Postoperative levels of biochemical markers of organ damage (creatinine, creatine kinase-MB [CKMB], and alanine-amino transferase [ALAT]) and total blood product use were additional secondary outcome measures. The IAASSG separates general from focal cerebral dysfunction, and cerebral from spinal cord dysfunction. Neurological and other complications are graded from I (deviating from normal but self-limiting) to V (causing death). Grade-IV complications are generally similar to earlier arbitrary definitions of, for example, stroke, paraplegia, dialysis, tracheostomy, and re-exploration for bleeding, that is, potentially life-threatening complications requiring advanced treatment and/or potentially permanent. Preoperative patient characteristics, procedural variables, and postoperative variables were defined according to STS database definitions (9). Echocardiographic and radiological variables were obtained using standard techniques.

Surgical procedures

All operations were performed through median sternotomy. Arterial cannulation site varied: right axillary artery (with or without 8-mm side graft), ascending aorta or aortic arch, or femoral artery. The right atrial appendage or femoral vein was cannulated for venous return. The aortic arch was resected and a distal anastomosis was constructed: end-to-end or by creating an elephant trunk (ET), floating or frozen. Frozen elephant trunk (FET) was performed using a hybrid vascular prosthesis (Evita Open Plus, Jotec GmbH, Hechingen, Germany or Thoraflex, Vascutek Ltd., Inchinnan, Scotland). The arch vessels were reimplanted en bloc or separately to a bi- or trifurcated graft or using a branched arch graft. The proximal anastomosis was made to the native proximal aorta or to a replaced aortic root (xenograft, mechanical composite graft, or valve-sparing root replacement graft). The aortic valve was either preserved, with or without repair measures, or replaced, as part of a full root replacement. None of the patients underwent isolated aortic valve replacement. Additional surgical procedures were performed as indicated.

Perfusion and organ protection

After institution of cardiopulmonary bypass, patients were cooled to deep hypothermia, (core temperature < 20°C), except in a minority of group-B cases where moderate hypothermia (∼28°C) was employed. Arterial blood gases were managed, with few exceptions, with alpha-stat strategy throughout the procedure. Cold blood cardioplegia was used for myocardial protection, delivered in the aortic root, the coronary ostia, or retrogradely. During deep hypothermic cardioplegia was readministered every 30–45 min, during rewarming every 15–20 min. The left ventricle was vented through the right superior pulmonary vein. The surgical field was flooded with carbon dioxide: 10 l/min. ACP was delivered through separate 9F or 12 F perfusion cannulae (LeMaitre Vascular, Burlington, MA) introduced in patent arch vessel ostia. ACP flow, line pressure, and bilateral radial artery pressure were monitored. Flow was adjusted to restore regional cerebral saturation as measured by bilateral near-infrared spectroscopy using INVOS equipment (Covidien, Dublin, Ireland) to baseline. ACP flow was generally maintained at 5–600 ml/min and not exceeding 10 ml/kg/min. The left radial artery pressure during ACP was maintained at 40– 60 mmHg, with supplemental doses of vasoconstrictor if necessary. The ACP perfusate temperature was isothermic to core body temperature at the beginning of HCA. With few exceptions (all in group B), HCA and ACP were maintained for the duration of all anastomoses, that is, without a strategy of resuming distal body perfusion at an earlier stage.

Statistical methods

Data are presented as numbers with percentages or medians with interquartile ranges (IQRs) or ranges, as indicated. Groups A and B were compared using nonparametric statistical tests: Mann–Whitney U test (continuous variables) and Fisher's exact test (categorical variables). Due to few outcome events, multivariable statistical modeling was not employed.

The study was approved by the Regional Medical Research Ethical Committee, waiving the need to obtain individual written informed consent.

Results

As detailed in Table I, patients in group A (≥ 90 min ACP) were overall somewhat younger and had less comorbidities (except for history of cerebrovascular insult) compared with patients in group B (< 90 min ACP), but were more often subject to redo procedures. Etiology of aortic arch disease differed, with a high proportion of chronic dissection and no acute dissection in group A (Table II). Surgically, patients in group A more often had the cervical vessels anastomosed separately, while ET/FET and aortic valve or root procedures were similarly distributed between groups (Table III). Despite some cases of moderate HCA in group B, median lowest core temperatures did not differ between groups.

Table I. Patient demographics and clinical characteristics in group A and group B, respectively. Numbers with percentages or medians with interquartile ranges as applicable.

Variable	Group A ACP ≥ 90 min n = 12	Group B ACP < 90 min n = 17
Male	10 (83)	9 (53)
Age, years	63 (51–64)	69 (64–73)
BMI	25.8 (21.9–27.1)	25.4 (22.5–29.8)
COPD	0	1 (5.9)
Diabetes	0	2 (12)
Ischemic heart disease	0	3 (18)
Peripheral arterial disease	0	0
History of CVI	4 (33)	1 (5.9)
LVEF, %	60 (55–60)	55 (55–60)
Redo	9 (75)	5 (42)

ACP, Antegrade cerebral perfusion; BMI, Body mass index; COPD, chronic obstructive pulmonary disease, CVI, cerebrovascular insult; LVEF, left ventricular ejection fraction

Table II. Features of aortic pathology in group A and group B, respectively. Numbers with percentages; aortic diameter given as median with range.

Variable	Group A ACP ≥ 90 min n = 12	Group B ACP < 90 min n = 17
Atherosclerotic aneurysm	4 (25)	12 (70)
Acute dissection	0	4 (24)
Chronic dissection	8 (75)	1 (6)
Maximum diameter, mm	60 (44–78)	65 (40–90)

Table III. Intraoperative variables in group A and group B, respectively. Numbers with percentages or medians with ranges as applicable.

Variable	Group A ACP ≥ 90 min n = 12	Group B ACP < 90 min n = 17
Total arch replacement	2 (17)	7 (41)
Total arch replacement + ET	7 (58)	6 (35)
Total arch replacement + FET	3 (25)	4 (24)
En bloc arch vessel reimplantation	2 (17)	13 (76)
Separate arch vessel reimplantation	10 (83)	4 (24)
Aortic root/valve untouched	8 (67)	12 (70)
Aortic root replacement	4 (33)	3 (18)
Valve-sparing root replacement	0	2 (12)
Other concomitant procedure	0	2 (12)
AXC, min	124 (0–269)	100 (16–363)
ECC, min	310 (263–431)	225 (136–715)
HCA, min	122 (105–163)	70 (31–122)
ACP, min	115 (98–157)	68 (37–89)
Lowest core temperature,°C	20 (14–21)	20 (13–27)

ACP, antegrade cerebral perfusion; AXC, Aortic cross-clamp; ECC, extracorporeal circulation; HCA, hypothermic circulatory arrest

Overall, four patients died in hospital: one (8.3%) in group A versus three (18%) in group B, p = 0.62. The only death in group A resulted from mesenteric embolization, whereas two deaths in group B were intraoperative due to bleeding (one acute dissection and one redo TAR), and the third was caused by diffuse gut ischemia.

Totally, five patients, all in group B (0 vs. 33%, p = 0.047), experienced grade-IV neurological damage: one had repeated seizures and coma with prolonged intubation but recovered completely prior to discharge (0 vs. 6.7%, p = 1.0), whereas two patients had grade-IV focal neurological deficit (0 vs. 13%, p = 0.48) and another two had grade-IV spinal neurological deficit. Both the latter underwent FET, one for definitive treatment of a distal arch aneurysm which included coverage of a prominent intercostal artery; the other had the left subclavian artery ligated. Accounting for all spinal neurological deficit, there was a statistically significant trend toward worse outcomes in group B (p = 0.038, Table IV). The previously reported so-called “temporary neurological deficit” best translates to IAASSG global neurological deficit grade II, occurring in 17% in group A versus 27% in group B, p = 0.66 (Table IV).

Table IV. Distribution of perioperative complications in group A and group B, respectively (count).

	Grades I–II		Grades III–IV
Group	A	B	A	B	p for trend
Neurological
Global deficit	4	8	0	1	0.62
Focal deficit	2	0	1	2	0.085
Spinal deficit	0	0	3	2	0.038
Cardiovascular
Ischemia	2	2	1	1	1.0
LCOS	10	14	0	1	0.47
Arrhythmia	2	7	5	4	0.17
Pericardial eff.	4	10	0	0	0.091
Respiratory
Parenchymal	8	8	2	4	0.81
Pleural	4	9	0	2	0.039
Renal	5	4	1	4	0.73
Gastrointestinal
Gut	2	3	1	2	1.0
Hepatobiliary	3	4	0	1	0.75
Postop bleed	8	9	4	6	0.71
Wound complic.	0	2	0	1	1.0
Other infection	4	9	0	0	0.56

LCOS, Low cardiac output syndrome; Pericardial eff., Pericardial effusion; Wound complic., Wound complication

Definitions and grading of complications according to the IAASSG (8): Grades I–II, resolving spontaneously or requiring only pharmacological treatment. Grades III–IV, permanent and/or requiring interventional measures. p value for trend (including complication-free and grade V complications) using Fisher's exact test.

Complications, as defined by the IAASSG (8), were frequent in both groups (Table IV). Notably, 50% of all patients experienced low cardiac output syndrome; arrhythmia; and respiratory, renal, or bleeding complications of grade I–III, that is, not requiring mechanical circulatory support, permanent pacemaker, tracheostomy, dialysis, or surgical reexploration. Such grade-IV complications (apart from neurological complications) were uncommon in operative survivors: arrhythmia in 2/27 (7.4%), reexploration for bleeding in 2/27, and gut ischemia in 3/27 (11%). None of the patients had grade-IV respiratory or renal complications; overall, grade III–IV pericardial effusion, hepatobiliary complications, wound, or other infectious complications occurred in < 5% of patients.

Biochemical markers of organ damage or dysfunction showed no statistically significant intergroup differences regarding postoperative levels of (1) serum creatinine (group A: median, 116 μmol/L; IQR, 95–152 μmol/L vs. group B: 114 μmol/L (80–134 μmol/L), p = 0.31), (2) serum CKMB [24 μg/L (15–42 μg/L) vs. 25 μg/L (15–36 μg/L), p = 0.81], or (3) serum ALAT (0.34 μkat/L (0.53–0.78 μkat/L) vs. 0.45 μkat/L (0.34–0.81 μkat/L), p = 0.81).

Blood product use (cumulative for the entire hospitalization) was also equal: packed red cells 5 (IQR: 3–10) units in group A versus 7 (3–10) in group B (p = 0.64), fresh frozen plasma 4 (2–8) versus 6 (3–9) (p = 0.39), and thrombocytes 2 (1–3) versus 2 (1–3) (p = 0.93). Three patients in each group (25% vs. 18%, p = 0.67) received recombinant factor VII intraoperatively to improve primary hemostasis.

Discussion

For patients in group A (ACP ≥ 90 min), in-hospital mortality was 1/12 (8.3%). None of the patients in group A had grade-IV general or focal neurological deficit, or grade-IV spinal deficit. None of the patients in group A had grade-IV renal dysfunction (dialysis), respiratory parenchymal complication (prolonged ventilation or tracheostomy), or postoperative bleeding (requiring reexploration). Thus, the study hypothesis that extended ACP is associated with worse outcomes (primarily early mortality and neurological damage) was refuted. For secondary outcome measures, there were overall no statistically significant intergroup differences. The standards of IAASSG for reporting complications of arch repair (8) appear systematic and comprehensive. Their utility, and the influence of minor (grade I–II) complications on early and late outcomes, needs to be further delineated in larger study populations.

Comparatively large proportion, 12/29 (41%), of all TAR utilizing HCA and ACP during the study period, patients requiring extended ACP and was characterized by a high proportion of redo operations (75%), chronic dissection (75%), use of ET (83%), and separate arch vessel reimplantation technique (83%). Clearly, a strategy with earlier lower body perfusion could reduce this proportion, and is increasingly employed.

Outcomes of TAR have been ubiquitously reported (1–7). However, even large patient series fail to accrue significant number of patients with extended ACP, or to report their outcomes specifically. Yet, Hagl et al. had ACP > 80 min in 19/86 (22%) patients undergoing arch repair with ACP. The prevalence of permanent neurological deficit (PND) in this subgroup was 25–35% and temporary neurological deficit (TND) was ∼60%; they concluded that long total cerebral protection time was associated with PND and that ACP is not safe for prolonged periods (3). In one multicenter report from Di Eusanio et al., ACP times ranged up to 220 min, and 90/413 patients (22%) had ACP > 90 min. They found no relationship between duration of ACP and early mortality, PND, or TND (4). Zierer et al. reported results in 1002 patients undergoing arch operations with mild (28–30°C) HCA and ACP (6). Overall, early mortality increased with increasing duration of ACP to reach 11% (7/62) in the ACP 60–90 min group. Notably, in elective patients, no differences were found, and only 13/512 (2.5%) had ACP duration of 60–90 min. Thus, even from such a large cohort, it is difficult to draw conclusions regarding extended ACP. The German Registry for Acute Aortic Dissection Type A reported outcomes in 1558 patients (7). Early mortality roughly doubled when ACP duration exceeded 60 min but, again, only a total of 31 patients (2.0%) had ACP > 90 min, with a 30-day mortality of 8–24% and PND of 17–29%. Finally, the Hannover et al., on the other hand, reported increased early mortality, PND, and TND with increased duration of ACP (10). Interestingly, by propensity score analysis, early mortality (27% vs. 12%) and paraplegia (18% vs. 0%) were more common with moderate (25–28°C) compared to deep (< 25°C) HCA in the small (n = 27) subgroup of patients with the longest (> 60 min) period of HCA (10). Differences in outcome definitions, levels of hypothermia, conduct of ACP, and patient case mix may explain the varying results, but the present study findings do not indicate increased risk with very long ACP in terms of early mortality or major complications.

In the current study, the cornerstones of organ protection during ACP was deep (< 20°C) hypothermia with isothermic cerebral perfusate and selective perfusion of all three arch vessels. With this strategy, none of the patients in group A sustained any grade-IV complication save one case of ventricular fibrillation due to undiagnosed preexisting coronary artery stenosis and one case of embolic gut ischemia (Table IV), suggesting good organ protection and no perceived time limit for ACP (median: 115 min, up to 157 min). Moreover, biochemical markers of end-organ damage, as well as blood product use, were equal to that of the group with shorter ACP duration. Gradually adopting a strategy of moderate hypothermia and earlier lower body reperfusion in the most recent cases (group B), we would still consider a strategy of three-vessel ACP and deep hypothermia when anticipating a very time-consuming arch repair. The excellent results reported with arch repair under moderate hypothermia (6,10) are generally confined to much shorter durations of ACP (i.e., a large proportion of simple hemiarch procedures), while recent clinical (11) as well as experimental (12) studies lend support to superior protection by deep hypothermia as well as to three-vessel ACP (13).

Study limitations

This is a small, retrospective study of patients undergoing TAR during a limited time period, nevertheless involving differing details in cannulation, perfusion, and surgical procedures. Primary outcome events (early death and major neurological complications) were too few to allow multivariable statistical analysis.

In summary, TAR with ACP equaling or exceeding 90 min is seldom encountered but can be performed with satisfactory results regarding early and follow-up mortality, as well as neurological and other major perioperative complications, organ dysfunction, and blood product use when performed during deep hypothermia and with selective perfusion of all three arch vessels. This strategy could be contemplated in selected cases even when TAR in general is approached with less profound cooling or less elaborated cerebral perfusion setup.

Acknowledgment

Susanne Hylander, RN, for technical assistance.

Declaration of interest: The authors report no declarations of interest. The authors alone are responsible for the content and writing of the paper.

The study was financially supported by a grant from the Mats Kleberg Foundation.