Introduction

Reoperative thoracoabdominal aortic aneurysm (TAAA) repair is often necessary and brings with it a unique set of challenges; such repair is performed after a previous distal aortic repair and involves a history of prior open or endovascular repair to correct a defect of the descending thoracic aorta (descending thoracic aneurysm; DTA), TAAA, or abdominal aorta (abdominal aortic aneurysm; AAA). There are two distinct scenarios when the operative field may be reopened: reoperations necessitated by aortic aneurysmal disease progressing into previously healthy aortic tissue and expanding from a previously replaced section of the aorta (an extension of the previous aortic repair), or, to a lesser degree, because of a late complication of prior aortic repair (an open or endovascular repair failure).

Pioneering aortic surgeon E. Stanley Crawford described his emerging experience with reoperative TAAA repair in a 1972 report [1]. Here, three patients with prior replacement of the DTA all went on to develop extensive aneurysms in the remaining distal aorta within 3–10 years after the initial DTA repair. All patients survived reoperation—with the replacement TAAA graft anastomosed to the prior DTA graft—but one patient developed paraparesis, which never resolved. This anecdotal experience kept thoughts of postoperative spinal cord deficit at the forefront of developments, amid speculation that the development of collateral circulation may play a protective role in some. Regarding Crawford’s lifetime experience of 1509 TAAA repairs, nearly 25% had a prior distal aortic repair; these patients had similar rates of early death and paraplegia as compared to patients without prior repair [2].

Because late reoperation of prior aortic repair may occur months to years after the initial operation, it is difficult to assess the incidence of late events that necessitate reoperation. Reports from several experienced aortic institutions suggest that 10–30% of patients now requiring TAAA have had a prior distal open aortic repair [2,3,4,5]. In addition, although far less common, there is a growing experience of open TAAA repair after prior distal endovascular aortic repair (EAR) [6,7,8,9,10]. Information on whether reoperative repair yields poorer outcomes, specifically higher incidences of operative death and spinal cord deficit, appears to be contradictory [11, 12]. In light of these contradictory results, determination of whether reoperative repair should be undertaken after prior distal aortic repair is a complicated decision, especially in the most complex reoperative scenarios involving infection or fistula. To elucidate this decision, we have presented our experience [13] along with a review of other centers’ reoperative results.

Characteristics of reoperative patients

Because patients require reoperation for different reasons, there are distinctions between patients experiencing progressive distal aortic disease and those with late repair failure. In our own recently published results regarding prior open distal aortic repair [13], patients with progressive aortic disease tended to have higher rates of smoking and were slightly older than non-reoperative patients; those with late repair failure tended to have higher rates of connective tissue disorder and aortic dissection than did non-reoperative patients. Regarding reoperative patients with prior endovascular aortic repair, trends are more difficult to identify because of more varied reasons for reoperation and the relatively small numbers of patients needing further repair. In general, reoperative repairs are heterogeneous, and complicate the evaluation of risk of these patients.

Extension of repair for progressive aortic disease

Conceptually, we have defined this type of repair as an extension of repair necessitated by the progression of disease into aortic segments that are adjacent to and contiguous with the previously repaired segment; in most cases, the previously placed graft was left in place (Fig. 1). Afifi et al. described this group as reoperation within the same operative field because of the formation of an aneurysm in a new aortic segment not previously operated on [14]. Reoperation because of progressive aortic disease necessitating an extension of prior aortic repair is the most common reason for performing reoperative surgery [13,14,15]. In our experience, the majority of the previous distal open aortic repairs were infrarenal AAA repairs, and most of these were performed outside of our center [13]. Observations from Afifi et al. similarly indicate that the majority of prior repairs are to treat AAA [14].

Fig. 1
figure 1

Illustration depicting the progression of aortic disease. a Abdominal aortic aneurysm is shown. b Open replacement of the abdominal aorta is shown. c Over time, a nearby section of the thoracoabdominal aorta becomes aneurysmal and is adjacent to the previously replaced abdominal aortic graft. d Aortic repair is extended by performing an extent III thoracoabdominal aortic aneurysm repair; the two grafts are anastomosed together. Used with permission of Baylor College of Medicine

Full size image

In our experience, reoperative patients with progressive aortic disease tend to be older than non-reoperative patients. The extension of repair group also had a higher prevalence of coronary artery disease, cerebrovascular disease, and peripheral vascular disease. In addition, this group was more likely to be male, to have an aneurysm without dissection, and to be current or former smokers—these preoperative characteristics and older age may indicate that these patients face a greater atherosclerotic, chronic disease burden than other patients. Consistent with this hypothesis, we found increased rates of visceral or renal endarterectomy in this group of patients [13].

Repair failure: prior open repair

In the case of reoperative repairs undertaken for repair failure, the failure exists in only the region that was first repaired. The repair failure does not extend to any region, where the index graft replacement was not originally present. Within our group, we defined repair failure as a complication of prior open distal aortic repair that necessitated aortic reoperation; largely due to the following causes—pseudoaneurysm, patch aneurysm, and graft infection, rupture, or related fistula. Reoperation because of repair failure is relatively uncommon, in our series, this was nearly 3% of 3379 TAAA repairs [13]. In a series of 1900 DTA or TAAA repairs, repair failures accounted for roughly 15% of repairs performed—visceral patch aneurysm was most common (11%) followed by anastomotic pseudoaneurysm (8%), intercostal patch aneurysm (7%), and infection (5%) [14].

Patients with repair failure patients tend to be younger. This finding seems reflective of higher rates of connective tissue disorders and chronic aortic dissection [13, 15]. There is a sense that long-term postoperative surveillance protocols need better development in these younger patients with aortic dissection. We observed trends in having large aneurysms (> 7.5 mm) at the time of reoperation and also increased rates of rupture. Due to this, it is unsurprising that reoperative repairs for the repair failure group were more likely to be urgent or emergent repairs than non-reoperative repair [13, 14].

Repair failure: prior endovascular repair

Patients with completed endovascular aortic repair face continued challenges, because there is an incomplete understanding of how and why such repair fails. Part of the overall sense of these failures appears related to early off-label use; since then, the use of EAR has been greatly expanded, and the term “off-label” has lost meaning. There are several different types of late complications after EAR—only the most serious of these complications will necessitate open repair. In general, any continued aortic expansion that is resistant to secondary endovascular repair will necessitate open repair, as will endograft infection or related fistula or any rare events such as endograft-related device failure (e.g, fracture). In general, EAR in patients with heritable thoracic aortic disease, such as Marfan syndrome, is not supported by current practice guidelines [16]. Overall, the experience in reoperation because of EAR is relatively small and extremely heterogeneous—it is difficult to identify trends. However, that being said, evidence suggests that EAR for aortic dissection, particularly that for chronic aortic dissection, tends to have a higher rate of failure necessitating further aortic EAR repair as does repair in persons with heritable thoracic aortic disease [7, 17, 18].

Surgical techniques

Operative details for reoperative repairs often follow repair techniques for standard TAAA repairs, although each institution may vary the details of performing the procedure. For reoperative surgeries, our surgical approach broadly follows that of standard thoracoabdominal surgeries, with routine use of moderate systemic heparinization (1.0 mg/kg), mild passive hypothermia (32C–34C), and cold renal perfusion (4C) whenever the renal ostia are sufficiently exposed. For extensive aortic repair (extents I and II), we typically use left heart bypass and selective visceral perfusion. Because there is a sense of heightened risk of postoperative paraplegia in patients undergoing reoperative repair, we tend to use cerebrospinal fluid drainage more frequently than in patients undergoing non-reoperative repair; similarly, while we commonly strive to reattach intercostal and lumber arteries whenever possible, we tend to reattach additional pairs of these arteries in reoperative aortic repair.

Further care is taken regarding adhesions, as lysing adhesions remaining from the previous operations both lengthens and adds complexity to the current procedure. Exposure may be hampered by adhesions—sharp dissection or electrocautery at low-power settings may free adhesions. In addition, for open TAAA reoperative surgery, we attempt an off-set incision through a different intercostal space, which limits bleeding and unintentional injury of nearby structures when adhesions are dissected. During reoperative TAAA repair, we minimize lung retraction during anticoagulation—if separating the lung from the chest wall does result in parenchymal injury, before closing the chest, we apply a sealant to the affected area of the lung. In case of progressive aortic disease with a stable endograft, it is often possible to fully salvage the prior repair and incorporate it into the new suture line using full thickness bites of both the endograft and residual aortic wall. To treat visceral patch aneurysms, we commonly use a branched graft; in addition, in patients with known heritable thoracic aortic disease, we also use a branched graft to minimize residual native aortic tissue as visceral arteries are managed.

Because adhesions involving the spleen are common, we carefully inspect the spleen before closure. If a small capsular tear or other damage to the spleen is found, it may be possible to repair it using cautery and Surgicel; however, if the tear cannot be repaired, a splenectomy is performed. In our experience, where there are very dense adhesions, there is an increased risk to the spleen and removal is not uncommon. Patients undergoing splenectomy typically receive additional vaccination in the early recovery period. Reoperative aortic repair may warrant an aggressive use of blood products to minimize bleeding complications and fresh frozen plasma and platelets are particularly helpful in ameliorating postoperative bleeding.

Techniques for treating infection

When graft or endograft infection is suspected, repair is inherently complicated and greatly so in the presence of fistula—removal of the graft or endograft is commonly necessitated. In 1961, Blaisdell et al. described the first use of an extra-anatomic bypass approach to treat persistent graft infection in a patient who had undergone infrarenal abdominal aortic replacement with a Teflon graft [19]. Following postoperative treatment for low-grade fever, the elderly patient was discharged home; within a month, he returned to the hospital because of sudden pain and a pulsatile mass in his abdomen. Upon exploration, a small portion of proximal anastomosis of the graft was noted to have dehisced due to infection, which was subsequently replaced. Several weeks later, symptoms returned and upon exploration, the entire proximal anastomosis of the graft was dehisced due to gross infection. The graft was removed and the residual aortic stumps were oversewn. A bypass graft was anastomosed end-to-side at the midlevel of the descending thoracic aorta and routed well away from the infected field to the femoral arteries. Although the patient survived for several weeks, ultimately, he died of infection—however, the bypass graft was noted to remain patent and free of infection. For the next few decades, extra-anatomic bypass approaches were the preferred technique in cases of extensive or highly virulent graft infection. In 1987, Walker, Cooley, and others in Houston [20] described a series of 23 patients with prior abdominal aortic replacement grafts that became infected and with the development of fistula; remarkably, many of these patients survived after in-situ graft replacement.

Although surgery related to infection was uncommon in our experience, reoperative surgery due to the indication of infection requires the use of additional techniques. These may include replacing the entire graft, extensive debridement, using pedicled omentum or a muscle flap to cover the replacement graft [20,21,22], delivering antibiotics via irrigation catheters, or using a synthetic graft soaked in antibiotics (i.e., prepared tableside by soaking a polyester graft in rifampin) [23]. Although extra-anatomic bypass was the gold-standard approach in decades past, it is less commonly used today—drawbacks include catastrophic blow-out of the remaining aortic stumps. Although we rarely use a homograft as a distal aortic replacement because of some observed early degeneration and rupture, others have reported the successful use of homografts to treat graft infection in the distal aorta [24].

Techniques for heritable thoracic aortic disease

Even after successful aortic repair, patients with heritable thoracic aortic disease such as Marfan syndrome commonly face long-term surveillance for the development of late aortic complications or de novo aneurysm. Previously, we have reported our experience regarding 300 patients with Marfan syndrome; two-thirds of these patients underwent 2 or more aortic repairs, and many of these involved reoperation of the distal aorta [25]. Notably, endovascular aortic repair in such patients necessitates more stringent surveillance as the need for further reintervention in not uncommon [26]. However, there are situations where endovascular repair may be useful in such patients—when endovascular repair may be used to treat a late complication of prior open repair, namely, for pseudoaneurysm or patch aneurysms of the intercostal aorta, in which the endograft can be landed in previously replaced sections of the aorta that are now composed of synthetic graft [27].

Here, we have included figures of a complex aortic reoperation in a young patient with Marfan syndrome, to illustrate surgical decisions in a patient who required multiple reoperations (Fig. 2). This female patient first presented at 29 years old at an outside center with an acute type B dissection, for which a stent graft was placed in the descending thoracic aorta along with a bypass graft to maintain blood flow to the left subclavian artery, whose origin was obstructed by the endograft. One month later, due to a persistent endoleak and continued aortic expansion, the patient was sent to our service, at which point the stent graft was removed and replaced with a graft. Three years later, at age 32, the patient experienced an acute type A aortic dissection and had an emergent resection and replacement of the aortic root with a composite valve graft and concomitant ascending and proximal transverse arch aortic replacement as well as endovascular coverage of the remaining native aortic arch. Two years later, at age 34, the patient received an extent IV TAAA repair using a branched graft to reattach the visceral arteries and minimize the amount of residual native aortic tissue. Five years following her last procedure, the patient continues to do well.

Fig. 2
figure 2

Illustration depicting the case of a young woman with Marfan syndrome who underwent multiple open aortic repairs after initial stent-graft placement to treat acute DeBakey type III aortic dissection at age 29. At this time, a bypass graft was performed, because the origin of the left subclavian artery was obstructed by the stent graft. a One month later, because of persistent endoleak and expansion, the stent graft was removed and the descending thoracic aorta was replaced with a synthetic graft (b). Three years later, she experienced an acute DeBakey type I aortic dissection (c); the proximal aorta was repaired with an aortic root and hybrid total arch replacement (d). After an additional 2 years, the patient underwent extent IV thoracoabdominal aortic aneurysm repair using a four-branched replacement graft to treat progressive dilatation of the remaining native aorta. Nearly 5 years after her last procedure, she continues to do well. Used with permission of Baylor College of Medicine

Full size image

Outcomes

Consensus on outcomes after reoperative surgery of the thoracoabdominal aorta is lacking. Whether the operative mortality rate is better or worse compared to non-reoperative TAAA repairs is a matter of some debate, as different institutions have arrived at different conclusions. Afifi et al. [14] recently reported a higher incidence of operative death in reoperative repairs (22.9%), which has been supported by a prior report [28]. However, others have reported that reoperative repairs are not subject to worse outcomes [4, 15, 29, 30] this includes our own recent report on reoperation of the thoracoabdominal aorta [13]. A summary of key data for reoperative TAAA repairs, following previous descending thoracic, thoracoabdominal, or abdominal repairs, is shown above (Table 1); operative mortality ranges from 8.7 to 22.9% [4, 13,14,15, 29,30,31].

Table 1 Review of literature for reoperative repairs involving a prior open distal aortic repair
Full size table

Conclusion

Recent reports by aortic centers seem to corroborate that reoperative repairs are possible with good outcomes. It is impossible to determine the extent to which the fear of poor outcomes may play a role in inhibiting surgeons from performing reoperative repair [29]. However, this fear must then be weighed against the risks of inaction and the potential of a worse outcome, such aortic rupture. In addition, the role that endovascular technology may play in reoperative surgery has yet to be fully actualized, since these techniques remain largely experimental as applied to the thoracoabdominal aorta. Today, the use of a variety of techniques may be necessitated in the most complex reoperative thoracoabdominal aortic repairs—those involving infection or in patients with heritable thoracic aortic disease.