Asian Cardiovascular and Thoracic Annals http://aan.sagepub.com/
Spinal cord injury following thoracic and thoracoabdominal aortic repairs Nirmal Panthee and Minoru Ono Asian Cardiovascular and Thoracic Annals published online 1 September 2014 DOI: 10.1177/0218492314548901 The online version of this article can be found at: http://aan.sagepub.com/content/early/2014/09/01/0218492314548901
Published by: http://www.sagepublications.com
On behalf of:
The Asian Society for Cardiovascular Surgery
Additional services and information for Asian Cardiovascular and Thoracic Annals can be found at: Email Alerts: http://aan.sagepub.com/cgi/alerts Subscriptions: http://aan.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav
>> OnlineFirst Version of Record - Sep 1, 2014 What is This?
Downloaded from aan.sagepub.com at COLUMBIA UNIV on November 30, 2014
XML Template (2014) [19.8.2014–12:56pm] //blrnas3/cenpro/ApplicationFiles/Journals/SAGE/3B2/AANJ/Vol00000/140190/APPFile/SG-AANJ140190.3d
(AAN)
[1–12] [PREPRINTER stage]
Review
Spinal cord injury following thoracic and thoracoabdominal aortic repairs
Asian Cardiovascular & Thoracic Annals 0(0) 1–12 ß The Author(s) 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0218492314548901 aan.sagepub.com
Nirmal Panthee and Minoru Ono
Abstract Objective: To discuss the currently available approaches to prevent spinal cord injury during thoracic and thoracoabdominal aortic repairs. Methods: We carried out a PubMed search up to 2013 using the Medical Subject Headings: ‘‘aortic aneurysm/surgery’’ and ‘‘spinal cord ischemia’’; ‘‘aortic aneurysm, thoracic/surgery’’ and ‘‘spinal cord ischemia’’; ‘‘aneurysm/surgery’’ and ‘‘spinal cord ischemia/cerebrospinal fluid’’; ‘‘aortic aneurysm/surgery’’ and ‘‘paraplegia’’. All 190 original articles satisfying our inclusion criteria were analyzed for incidence, predictors, and other pertinent variables related to spinal cord injury, and we compared the results in recent publications with those in earlier reports. Results: The mean age of the 38,491 patients was 65.3 4.9 years. The overall incidence of paraplegia and/or paraparesis was 7.1% 6.1% (range 0%–32%). The incidence of spinal cord injury before 2000, from 2001 to 2007, and 2008–2013 was 9.0% 6.7%, 7.0% 6.1%, and 5.9% 5.2%, respectively (p ¼ 0.019). Various predictors of spinal cord injury were identified, extent of disease being the most common. Modification of surgical techniques, use of adjuncts, and better understanding of spinal cord perfusion physiology were attributed to the decrease in postoperative spinal cord injury in recent years. Conclusions: Spinal cord injury after thoracic and thoracoabdominal aortic repair poses a real challenge to cardiovascular surgeons. However, with evolving surgical strategies, identification of predictors, and use of various adjuncts over the years, the incidence of spinal cord injury after thoracic/thoracoabdominal aortic repair has declined. Embracing a multimodality approach offers a good insight into combating this grave complication.
Keywords Aortic aneurysm, cerebrospinal fluid, evoked potentials, hypothermia, paraplegia, spinal cord ischemia
Introduction
Methods
Spinal cord injury (SCI) remains a devastating complication after thoracoabdominal aortic aneurysm (TAAA) repairs. Its incidence ranges from 1% to 32%;1–190 2.3% to 32% with open surgery,1–15 and 1% to 19% with thoracic endovascular aneurysm repair (TEVAR).16–35 This large difference among series is due to varying proportions of high-risk patients, use of different surgical adjuncts for spinal cord protection, and selection bias, which is the rule in current clinical practice. Despite so many years of experience, we have not been able to completely eliminate this devastating complication, although we have reduced the incidence by using various adjuncts and with better understanding of cord perfusion physiology.
We carried out a PubMed search of publications listed up to 2013, using the Medical Subject Headings: ‘‘aortic aneurysm/surgery’’ and ‘‘spinal cord ischemia’’; ‘‘aortic aneurysm, thoracic/surgery’’ and ‘‘spinal cord ischemia’’; ‘‘aneurysm/surgery’’ and ‘‘spinal cord ischemia/ cerebrospinal fluid’’; ‘‘aortic aneurysm/surgery’’ and ‘‘paraplegia’’. By excluding review articles, case reports, letters to the editor, commentaries, editorials,
Department of Cardiac Surgery, University of Tokyo, Tokyo, Japan Corresponding author: Minoru Ono, MD, PhD, Department of Cardiac Surgery, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8655 Japan. Email: [email protected]
Downloaded from aan.sagepub.com at COLUMBIA UNIV on November 30, 2014
XML Template (2014) [19.8.2014–12:56pm] //blrnas3/cenpro/ApplicationFiles/Journals/SAGE/3B2/AANJ/Vol00000/140190/APPFile/SG-AANJ140190.3d
(AAN)
[1–12] [PREPRINTER stage]
2
Asian Cardiovascular & Thoracic Annals 0(0)
guidelines, author replies, surgical techniques, original articles with acute traumatic rupture of the aortic isthmus, articles in languages other than English, reports of animal experiments, and those without the full text available through the medical library system of the University of Tokyo, we analyzed 190 original articles with a patient pool of 38,491. Spinal cord injury has been reported after repair of acute traumatic aortic isthmus rupture, but these patients are usually young, often have associated polytrauma, and lack the atherosclerotic risk factors associated with TAAA; therefore, we excluded them from the analysis. We divided the publications into 3 broad groups based on the year of publication: before 2000, between 2001 and 2007, and between 2008 and 2013. This grouping was based on an arbitrary consideration. We looked for SCI in the last 6 years (2008–2013) to obtain the most recent overview of this dreadful complication. We were also interested to find out what it was like before 2000. In doing so, obviously, the remaining patients fell into the group 2001–2007. We looked for incidence and predictors of SCI (paraplegia and/or paraparesis), extent and complexity of the disease, surgical techniques, and use of adjuvant measures in these 3 groups. We focused on the percentages reported for the variables of interest in each individual article. We borrowed the technical terms from the authors and did not define any variable ourselves. The authors of the original papers defined clamp-and-sew as a technique in which they replaced the diseased segment of the aorta by simple clamping; whereas they defined distal perfusion as any maneuver that perfused the aorta distal to repair. Distal perfusion included left heart bypass (left inferior pulmonary vein-to-distal descending aorta bypass using a centrifugal pump without the use of an oxygenator), atrial-femoral bypass (left atrium-to-femoral artery bypass using a centrifugal pump without the use of a pump oxygenator), partial cardiopulmonary bypass (cardioplegia not utilized), full cardiopulmonary bypass (with cardioplegia), and deep hypothermic circulatory arrest. Although left heart bypass and atrial-femoral bypass essentially achieve the same purpose, we mentioned them separately in our analysis to recognize the work by different authors utilizing left inferior pulmonary vein-to-descending aorta bypass or left atrium-to-femoral artery bypass. Data were analyzed using SPSS version 20 (SPSS, Inc., Chicago, IL, USA). One-way analysis of variance was utilized for statistical tests, with a p value < 0.05 being considered statistically significant. Data are presented as mean standard deviation of the percentages of each variable, accompanied by the number of studies mentioning that variable. These data represent the
means of percentages obtained from studies that mentioned the variable in their analysis, and do not represent studies that did not analyze that particular variable.
Results The results are summarized in Table 1. The mean age of the patients was 65.3 4.9 years. Patients undergoing TAAA repair before 2000 were significantly younger compared to the other 2 groups. The overall incidence of SCI in this pooled analysis was 7.1% 6.1% (range 0%–32%). Subgroup analysis revealed that the incidence of SCI following TAAA repairs before 2000, 2001–2007, and 2008–2013 was 9.0% 6.7%, 7.0% 6.1%, and 5.9% 5.2%, respectively (p ¼ 0.019). Further analysis of studies reporting delayed-onset SCI revealed that 12.4% 22.8% of all cases of SCI were delayed-onset. The incidence of delayed-onset SCI in the 3 time periods was not significantly different. The extent and complexity of repairs among groups were comparable except for TAAA type I and the presence of dissection. The percentage of type I aneurysm declined from 32.4% 13.7% before 2000 to 18.1% 13.0% between 2008 and 2013. The incidence of dissection (acute and chronic) increased over time; however, the incidence of acute dissection was similar in all 3 time periods. Although it did not reach statistical significance, arch involvement was more common between 2008 and 2013. Use of the clamp-and-sew technique of repair was not significantly different among groups. Distal perfusion was utilized in 69.1% 32.2%, 87.9% 21.7%, and 86.4% 27.4% in the 3 time periods, respectively (p ¼ 0.014). Use of left heart bypass, atrial-femoral bypass, partial cardiopulmonary bypass, full cardiopulmonary bypass, and deep hypothermic circulatory arrest were not significantly different among the 3 groups. A shunt was used in 41.5% 38.2% of cases treated before 2000, but there was no shunt use reported thereafter. The rate of intercostal artery reimplantation was not significantly different among the 3 time periods, nor were the use of TEVAR, hybrid procedures, or CSF drainage. Motor-evoked potentials (MEP) and somatosensoryevoked potentials (SSEP) were monitored in similar percentages of patients in all 3 groups. Epidural cooling and naloxone use was highly variable but the differences between groups were not significant. Preoperative artery of Adamkiewicz identification (magnetic resonance imaging, computed tomography, or spinal angiography) was reported in 24 studies but the differences among time periods were not significant. We looked for predictors of SCI identified by multivariate and/or univariate analysis. Extent of the disease was the most commonly described predictor of
Downloaded from aan.sagepub.com at COLUMBIA UNIV on November 30, 2014
XML Template (2014) [19.8.2014–12:56pm] //blrnas3/cenpro/ApplicationFiles/Journals/SAGE/3B2/AANJ/Vol00000/140190/APPFile/SG-AANJ140190.3d
(AAN)
[1–12] [PREPRINTER stage]
Panthee and Ono
3
Table 1. Summary of pooled analysis of 38,491 cases of thoracic and thoracoabdominal aortic repair. Variables* Age (years) SCI Late SCI Extent and complexity of disease Proximal DTA Distal DTA Entire DTA DTA; extent not specified TAAA type I TAAA type II TAAA type III TAAA type IV TAAA type not specified Arch involvement Pararenal Dissection Acute dissection Chronic dissection Rupture Emergency/urgent Surgical techniques and adjuncts Clamp-and-sew Distal perfusion Left heart bypass AF bypass Shunt Partial CPB Full CPB DHCA CSF drainage MEP SSEP Epidural cooling Naloxone ICA reimplantation Preoperative AKA identification TEVAR Hybrid procedure
Overall (190 studies)
Before 2000 (50 studies)
2001–2007 (61 studies)
2008–2013 (79 studies)
p value
65.3 4.9 7.1% 6.1% (190) 12.4% 22.8% (20)
63.1 5.1 9.0% 6.7% (50) 3.1% 2.2% (2)
65.4 3.9 7.0% 6.1% (61) 15.7% 26.0% (10)
66.4 5.2 5.9% 5.2% (79) 10.7% 22.6% (8)
0.001 0.019 0.767
36.8% 24.4% (19) 22.7% 22.4% (24) 33.9% 14.4% (22) 69.4% 32.2% (111) 25.6% 15.7% (94) 31.1% 20.6% (95) 20.3% 9.9% (76) 16.8% 16.1% (58) 35.5% 28.1% (17) 51.4% 37.3% (20) 48.4% 28.4% (6) 34.9% 23.6% (124) 17.8% 24.6% (55) 29.4% 15.1% (60) 19.2% 22.8% (48) 23.6% 17.8% (69)
41.6% 27.2% (6) 23.9% 19.1% (5) 34.4% 14.8% (7) 59.5% 36.1% (25) 32.4% 13.7% (28) 34.8% 13.7% (27) 21.2% 8.7% (20) 17.4% 8.1% (14) 28.3% 12.8% (7) 30.0% (1) 40.5% 9.1% (2) 25.3% 12.2% (36) 10.1% 11.2% (12) 25.2% 15.6% (11) 14.8% 9.7% (19) 23.6% 13.1% (16)
41.1% 30.7% (7) 23.4% 21.2% (7) 36.0% 15.6% (8) 65.4% 31.0% (34) 27.6% 17.1% (32) 32.4% 21.5% (31) 19.1% 9.6% (26) 14.2% 10.7% (19) 44.5% 31.5% (6) 39.9% 33.2% (13) 100% (1) 34.7% 20.4% (40) 15.3% 24.6% (19) 29.7% 9.5% (23) 17.5% 25.0% (14) 22.1% 24.6% (20)
27.2% 10.9% 21.8% 25.5% 31.0% 14.5% 76.9% 29.8% 18.1% 13.0% 27.1% 20.6% 20.7% 11.1% 18.3% 22.1% 34.6% 44.3% 79.9% 35.4% 36.5% 19.5% 42.1% 29.7% 23.7% 28.5% 30.9% 18.5% 26.2% 31.1% 24.6% 15.2%
(6) (12) (7) (52) (34) (37) (30) (25) (4) (6) (3) (48) (24) (26) (15) (33)
0.532 0.982 0.812 0.057 0.001 0.299 0.741 0.705 0.614 0.073 0.095 0.005 0.259 0.578 0.345 0.890
64.1% 35.4% 79.5% 29.1% 84.7% 26.3% 58.4% 33.5% 41.5% 38.2% 77.4% 36.2% 85.2% 33.6% 78.4% 32.7% 79.6% 29.0% 90.6% 23.1% 91.4% 21.2% 79.8% 22.6% 74.8% 22.8% 76.2% 31.1% 89.2% 20.8% 86.8% 26.1% 86.5% 29.2%
70.1% 35.1% (19) 69.1% 32.2% (37) 54.1% 50.4% (6) 57.1% 31.7% (12) 41.5% 38.2% (8) 62.2% 41.0% (4) 60.9% 50.4% (3) 64.4% 33.6% (3) 66.7% 30.7% (18) 89.0% 24.5% (5) 84.2% 23.2% (8) 100% (2) 62.2% 9.5% (2) 65.4% 34.1% (11) 0 (0) 100% (1) 0 (0)
58.3% 36.3% (6) 87.9% 21.7% (28) 62.3% 36.5% (10) 100% (2) 0 (0) 100% (5) 99.5% 0.7% (2) 83.7% 29.4% (9) 82.9% 29.4% (32) 100% (9) 92.0% 25.2% (10) 75.2% 22.1% (3) 0 (0) 87.8% 21.8% (21) 83.1% 27.2% (6) 80.1% 35.7% (16) 100% (1)
44.2% 36.5% (4) 86.4% 27.4% (22) 84.7% 26.3% (6) 24.8% 5.9% (2) 0 (0) 67.2% 44.9% (5) 100% (3) 77.5% 39.3% (7) 85.1% 24.8% (23) 86.8% 27.7% (18) 100% (6) 53.0% (1) 100% (1) 64.8% 37.2% (11) 95.5% 10.9% (6) 89.3% 21.3% (39) 85.0% 30.6% (9)
0.389 0.014 0.374 0.067
(29) (87) (22) (16) (8) (14) (8) (19) (73) (32) (24) (6) (3) (43) (12) (56) (10)
0.232 0.331 0.699 0.090 0.382 0.404 0.234 0.191 0.053 0.321 0.451 0.655
*Numbers in brackets are the no. of studies. AF: atrial-femoral; AKA: Adamkiewicz artery; CPB: cardiopulmonary bypass; CSF: cerebrospinal fluid; DHCA: deep hypothermic circulatory arrest; DTA: descending thoracic aorta; ICA: intercostal artery; MEP: motor-evoked potentials; SCI: spinal cord injury; SSEP: somatosensory-evoked potentials; TAAA: thoracoabdominal aortic aneurysm; TEVAR: thoracic endovascular aortic repair.
postoperative spinal cord dysfunction, with extents I and II carrying the highest risk.1,3,6,8,15,19,36–47 In cases of TEVAR, the extent of distal location and extent of aortic coverage with respect to the celiac artery, and the number of stent grafts used during surgery also predicted SCI.24,28 Other factors identified as
predictors of postoperative SCI included longer aortic clamp time,9,12,15,36,40,46–50 lack of distal perfusion,9 no use of adjuncts,1,40 oversewn T9-T12 intercostals,41,47 no use of atrial-femoral bypass,15 left subclavian artery covering without revascularization,30 concomitant open abdominal aortic surgery,30 extent of
Downloaded from aan.sagepub.com at COLUMBIA UNIV on November 30, 2014
XML Template (2014) [19.8.2014–12:56pm] //blrnas3/cenpro/ApplicationFiles/Journals/SAGE/3B2/AANJ/Vol00000/140190/APPFile/SG-AANJ140190.3d
(AAN)
[1–12] [PREPRINTER stage]
4
Asian Cardiovascular & Thoracic Annals 0(0)
segmental artery sacrificed,53 failure to cool actively with bypass,12 level of the distal aortic clamp,54 longer lower spinal cord ischemic time,55 and increasing red blood cell infusions.56 Preoperative patient-related factors predicting SCI included age,36,57 rupture,36,44,58 proximal aortic aneurysm,36 history of renal dysfunction,36 acute presentation,3,8,38,58 previously repaired abdominal aortic aneurysm,1 history of cerebrovascular disease,1,43 nonelective operation,28,41,46 false aneurysm,9 inflammation,9 medial degeneration,9 comorbidities (hypertension, smoking, dyslipidemia, renal insufficiency),25 absence of post-dissection aneurysm,59 acute dissection,56 non-dissecting aneurysm,60 chronic renal insufficiency,28,30,33,34,61 and preoperative shock.55 Hypotension was also a predictor of SCI; both intraoperative hypotension,46,47 and postoperative hypotension.12,15,62,63 No use of CSF drainage,3,12,38,59 a decrease in cardiac index with aortic occlusion,3 normothermia,39 irreversible changes in MEP and SSEP,64 and lack of epidural cooling46,48 were also identified as predictors of SCI by multivariate and/or univariate analysis.
Discussion Spinal cord injury remains a devastating complication after TAAA repair. Its incidence is reported to range from 1% to 32%. Despite various adjuncts and surgical techniques, complete elimination of this complication is virtually impossible. Our analysis of the data of over 38,000 patients showed the overall incidence of SCI to be 7.1% 6.1%. Because this result is an average of a varying pool of patients in different series, the actual incidence of SCI was different in each patient pool (0%–32%). An important finding demonstrated by this analysis is the decreasing trend of SCI over the last few years. As we analyzed the publications, the extent of disease remained almost the same in the 3 time periods, except for a higher proportion of patients with type I TAAA before 2000. Nevertheless, the incidence of SCI decreased significantly from 9.0% 6.7% to 5.9% 5.2%. Use of adjuncts such as distal perfusion, hypothermia, CSF drainage, epidural cooling, naloxone, and MEP and SSEP monitoring has contributed to reducing SCI over recent years. Significantly more patients received distal perfusion after 2000, and although it did not reach statistical significance, a clear trend of increased use of CSF drainage was found after 2000. Neuromonitoring with MEP and SSEP, and preoperative artery of Adamkiewicz identification with magnetic resonance angiography and computed tomography-angiography were also increasingly utilized after 2000. ICA reimplantation showed an increasing trend between 2001 and 2007 compared to before 2000, and every effort was made to reimplant each
pair of ICA. However, after 2007, surgeons were concerned about the longer operative time needed to reimplant ICA, and they started to sacrifice intercostals that were not crucial to spinal circulation by monitoring MEP. Monitoring MEP helped surgeons to identify crucial ICA to reimplant, thus decreasing the total aortic clamp time. With the knowledge of the above mentioned predictors of paraplegia, a large number of strategies have been utilized in various centers to combat SCI following TAAA surgery. Because the extent of the aorta replaced is an important predictor of postoperative paraplegia, careful preoperative planning and use of multiple adjuncts can reduce SCI in these high-risk patients. Distal perfusion is an important strategy to reduce postoperative SCI. The simple clamp-and-sew technique results in more SCI. However, if combined with selective epidural cooling to a CSF temperature of 26.4 C, the simple clamp-and-sew technique also offers spinal cord protection.41 Distal perfusion has routinely been included in TAAA repairs in recent years.4,9,10,65–67 Several authors have shown that aortic clamp time is a predictor of postoperative SCI. Biglioli and colleagues49 reported that in descending thoracic aortic aneurysm repair, spinal cord perfusion can be adequately maintained without reimplantation of segmental vessels or use of atrial-distal bypass when the aortic crossclamp time is short (30 min was associated with SCI,9,40 and >60 min was associated with an increased incidence of SCI.48 Use of left heart bypass has been shown to lower the risk of SCI, mainly because it unloads the proximal circulation during aortic occlusion while maintaining adequate distal perfusion, allowing a longer time for all distal intercostal reimplantation.54 However, Coselli and colleagues56 found that left heart bypass was not a predictor of SCI. Use of cardiopulmonary bypass (partial or full) has also been shown to lower the risk of SCI.68 Use of shunts (Gott shunt) significantly reduced SCI during descending aortic surgery.69 The first prospective randomized study to examine the role of CSF drainage in preventing SCI after highrisk TAAA surgery showed negative results.62 However, these findings were later refuted by the same authors in another randomized clinical trial.5 The main difference between these two findings might be because in the first study, the authors limited the CSF drainage volume to a maximum of 50 mL; however, in the randomized clinical trial, their aim was to decrease CSF pressure to
Spinal cord injury following thoracic and thoracoabdominal aortic repairs Nirmal Panthee and Minoru Ono Asian Cardiovascular and Thoracic Annals published online 1 September 2014 DOI: 10.1177/0218492314548901 The online version of this article can be found at: http://aan.sagepub.com/content/early/2014/09/01/0218492314548901
Published by: http://www.sagepublications.com
On behalf of:
The Asian Society for Cardiovascular Surgery
Additional services and information for Asian Cardiovascular and Thoracic Annals can be found at: Email Alerts: http://aan.sagepub.com/cgi/alerts Subscriptions: http://aan.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav
>> OnlineFirst Version of Record - Sep 1, 2014 What is This?
Downloaded from aan.sagepub.com at COLUMBIA UNIV on November 30, 2014
XML Template (2014) [19.8.2014–12:56pm] //blrnas3/cenpro/ApplicationFiles/Journals/SAGE/3B2/AANJ/Vol00000/140190/APPFile/SG-AANJ140190.3d
(AAN)
[1–12] [PREPRINTER stage]
Review
Spinal cord injury following thoracic and thoracoabdominal aortic repairs
Asian Cardiovascular & Thoracic Annals 0(0) 1–12 ß The Author(s) 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0218492314548901 aan.sagepub.com
Nirmal Panthee and Minoru Ono
Abstract Objective: To discuss the currently available approaches to prevent spinal cord injury during thoracic and thoracoabdominal aortic repairs. Methods: We carried out a PubMed search up to 2013 using the Medical Subject Headings: ‘‘aortic aneurysm/surgery’’ and ‘‘spinal cord ischemia’’; ‘‘aortic aneurysm, thoracic/surgery’’ and ‘‘spinal cord ischemia’’; ‘‘aneurysm/surgery’’ and ‘‘spinal cord ischemia/cerebrospinal fluid’’; ‘‘aortic aneurysm/surgery’’ and ‘‘paraplegia’’. All 190 original articles satisfying our inclusion criteria were analyzed for incidence, predictors, and other pertinent variables related to spinal cord injury, and we compared the results in recent publications with those in earlier reports. Results: The mean age of the 38,491 patients was 65.3 4.9 years. The overall incidence of paraplegia and/or paraparesis was 7.1% 6.1% (range 0%–32%). The incidence of spinal cord injury before 2000, from 2001 to 2007, and 2008–2013 was 9.0% 6.7%, 7.0% 6.1%, and 5.9% 5.2%, respectively (p ¼ 0.019). Various predictors of spinal cord injury were identified, extent of disease being the most common. Modification of surgical techniques, use of adjuncts, and better understanding of spinal cord perfusion physiology were attributed to the decrease in postoperative spinal cord injury in recent years. Conclusions: Spinal cord injury after thoracic and thoracoabdominal aortic repair poses a real challenge to cardiovascular surgeons. However, with evolving surgical strategies, identification of predictors, and use of various adjuncts over the years, the incidence of spinal cord injury after thoracic/thoracoabdominal aortic repair has declined. Embracing a multimodality approach offers a good insight into combating this grave complication.
Keywords Aortic aneurysm, cerebrospinal fluid, evoked potentials, hypothermia, paraplegia, spinal cord ischemia
Introduction
Methods
Spinal cord injury (SCI) remains a devastating complication after thoracoabdominal aortic aneurysm (TAAA) repairs. Its incidence ranges from 1% to 32%;1–190 2.3% to 32% with open surgery,1–15 and 1% to 19% with thoracic endovascular aneurysm repair (TEVAR).16–35 This large difference among series is due to varying proportions of high-risk patients, use of different surgical adjuncts for spinal cord protection, and selection bias, which is the rule in current clinical practice. Despite so many years of experience, we have not been able to completely eliminate this devastating complication, although we have reduced the incidence by using various adjuncts and with better understanding of cord perfusion physiology.
We carried out a PubMed search of publications listed up to 2013, using the Medical Subject Headings: ‘‘aortic aneurysm/surgery’’ and ‘‘spinal cord ischemia’’; ‘‘aortic aneurysm, thoracic/surgery’’ and ‘‘spinal cord ischemia’’; ‘‘aneurysm/surgery’’ and ‘‘spinal cord ischemia/ cerebrospinal fluid’’; ‘‘aortic aneurysm/surgery’’ and ‘‘paraplegia’’. By excluding review articles, case reports, letters to the editor, commentaries, editorials,
Department of Cardiac Surgery, University of Tokyo, Tokyo, Japan Corresponding author: Minoru Ono, MD, PhD, Department of Cardiac Surgery, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8655 Japan. Email: [email protected]
Downloaded from aan.sagepub.com at COLUMBIA UNIV on November 30, 2014
XML Template (2014) [19.8.2014–12:56pm] //blrnas3/cenpro/ApplicationFiles/Journals/SAGE/3B2/AANJ/Vol00000/140190/APPFile/SG-AANJ140190.3d
(AAN)
[1–12] [PREPRINTER stage]
2
Asian Cardiovascular & Thoracic Annals 0(0)
guidelines, author replies, surgical techniques, original articles with acute traumatic rupture of the aortic isthmus, articles in languages other than English, reports of animal experiments, and those without the full text available through the medical library system of the University of Tokyo, we analyzed 190 original articles with a patient pool of 38,491. Spinal cord injury has been reported after repair of acute traumatic aortic isthmus rupture, but these patients are usually young, often have associated polytrauma, and lack the atherosclerotic risk factors associated with TAAA; therefore, we excluded them from the analysis. We divided the publications into 3 broad groups based on the year of publication: before 2000, between 2001 and 2007, and between 2008 and 2013. This grouping was based on an arbitrary consideration. We looked for SCI in the last 6 years (2008–2013) to obtain the most recent overview of this dreadful complication. We were also interested to find out what it was like before 2000. In doing so, obviously, the remaining patients fell into the group 2001–2007. We looked for incidence and predictors of SCI (paraplegia and/or paraparesis), extent and complexity of the disease, surgical techniques, and use of adjuvant measures in these 3 groups. We focused on the percentages reported for the variables of interest in each individual article. We borrowed the technical terms from the authors and did not define any variable ourselves. The authors of the original papers defined clamp-and-sew as a technique in which they replaced the diseased segment of the aorta by simple clamping; whereas they defined distal perfusion as any maneuver that perfused the aorta distal to repair. Distal perfusion included left heart bypass (left inferior pulmonary vein-to-distal descending aorta bypass using a centrifugal pump without the use of an oxygenator), atrial-femoral bypass (left atrium-to-femoral artery bypass using a centrifugal pump without the use of a pump oxygenator), partial cardiopulmonary bypass (cardioplegia not utilized), full cardiopulmonary bypass (with cardioplegia), and deep hypothermic circulatory arrest. Although left heart bypass and atrial-femoral bypass essentially achieve the same purpose, we mentioned them separately in our analysis to recognize the work by different authors utilizing left inferior pulmonary vein-to-descending aorta bypass or left atrium-to-femoral artery bypass. Data were analyzed using SPSS version 20 (SPSS, Inc., Chicago, IL, USA). One-way analysis of variance was utilized for statistical tests, with a p value < 0.05 being considered statistically significant. Data are presented as mean standard deviation of the percentages of each variable, accompanied by the number of studies mentioning that variable. These data represent the
means of percentages obtained from studies that mentioned the variable in their analysis, and do not represent studies that did not analyze that particular variable.
Results The results are summarized in Table 1. The mean age of the patients was 65.3 4.9 years. Patients undergoing TAAA repair before 2000 were significantly younger compared to the other 2 groups. The overall incidence of SCI in this pooled analysis was 7.1% 6.1% (range 0%–32%). Subgroup analysis revealed that the incidence of SCI following TAAA repairs before 2000, 2001–2007, and 2008–2013 was 9.0% 6.7%, 7.0% 6.1%, and 5.9% 5.2%, respectively (p ¼ 0.019). Further analysis of studies reporting delayed-onset SCI revealed that 12.4% 22.8% of all cases of SCI were delayed-onset. The incidence of delayed-onset SCI in the 3 time periods was not significantly different. The extent and complexity of repairs among groups were comparable except for TAAA type I and the presence of dissection. The percentage of type I aneurysm declined from 32.4% 13.7% before 2000 to 18.1% 13.0% between 2008 and 2013. The incidence of dissection (acute and chronic) increased over time; however, the incidence of acute dissection was similar in all 3 time periods. Although it did not reach statistical significance, arch involvement was more common between 2008 and 2013. Use of the clamp-and-sew technique of repair was not significantly different among groups. Distal perfusion was utilized in 69.1% 32.2%, 87.9% 21.7%, and 86.4% 27.4% in the 3 time periods, respectively (p ¼ 0.014). Use of left heart bypass, atrial-femoral bypass, partial cardiopulmonary bypass, full cardiopulmonary bypass, and deep hypothermic circulatory arrest were not significantly different among the 3 groups. A shunt was used in 41.5% 38.2% of cases treated before 2000, but there was no shunt use reported thereafter. The rate of intercostal artery reimplantation was not significantly different among the 3 time periods, nor were the use of TEVAR, hybrid procedures, or CSF drainage. Motor-evoked potentials (MEP) and somatosensoryevoked potentials (SSEP) were monitored in similar percentages of patients in all 3 groups. Epidural cooling and naloxone use was highly variable but the differences between groups were not significant. Preoperative artery of Adamkiewicz identification (magnetic resonance imaging, computed tomography, or spinal angiography) was reported in 24 studies but the differences among time periods were not significant. We looked for predictors of SCI identified by multivariate and/or univariate analysis. Extent of the disease was the most commonly described predictor of
Downloaded from aan.sagepub.com at COLUMBIA UNIV on November 30, 2014
XML Template (2014) [19.8.2014–12:56pm] //blrnas3/cenpro/ApplicationFiles/Journals/SAGE/3B2/AANJ/Vol00000/140190/APPFile/SG-AANJ140190.3d
(AAN)
[1–12] [PREPRINTER stage]
Panthee and Ono
3
Table 1. Summary of pooled analysis of 38,491 cases of thoracic and thoracoabdominal aortic repair. Variables* Age (years) SCI Late SCI Extent and complexity of disease Proximal DTA Distal DTA Entire DTA DTA; extent not specified TAAA type I TAAA type II TAAA type III TAAA type IV TAAA type not specified Arch involvement Pararenal Dissection Acute dissection Chronic dissection Rupture Emergency/urgent Surgical techniques and adjuncts Clamp-and-sew Distal perfusion Left heart bypass AF bypass Shunt Partial CPB Full CPB DHCA CSF drainage MEP SSEP Epidural cooling Naloxone ICA reimplantation Preoperative AKA identification TEVAR Hybrid procedure
Overall (190 studies)
Before 2000 (50 studies)
2001–2007 (61 studies)
2008–2013 (79 studies)
p value
65.3 4.9 7.1% 6.1% (190) 12.4% 22.8% (20)
63.1 5.1 9.0% 6.7% (50) 3.1% 2.2% (2)
65.4 3.9 7.0% 6.1% (61) 15.7% 26.0% (10)
66.4 5.2 5.9% 5.2% (79) 10.7% 22.6% (8)
0.001 0.019 0.767
36.8% 24.4% (19) 22.7% 22.4% (24) 33.9% 14.4% (22) 69.4% 32.2% (111) 25.6% 15.7% (94) 31.1% 20.6% (95) 20.3% 9.9% (76) 16.8% 16.1% (58) 35.5% 28.1% (17) 51.4% 37.3% (20) 48.4% 28.4% (6) 34.9% 23.6% (124) 17.8% 24.6% (55) 29.4% 15.1% (60) 19.2% 22.8% (48) 23.6% 17.8% (69)
41.6% 27.2% (6) 23.9% 19.1% (5) 34.4% 14.8% (7) 59.5% 36.1% (25) 32.4% 13.7% (28) 34.8% 13.7% (27) 21.2% 8.7% (20) 17.4% 8.1% (14) 28.3% 12.8% (7) 30.0% (1) 40.5% 9.1% (2) 25.3% 12.2% (36) 10.1% 11.2% (12) 25.2% 15.6% (11) 14.8% 9.7% (19) 23.6% 13.1% (16)
41.1% 30.7% (7) 23.4% 21.2% (7) 36.0% 15.6% (8) 65.4% 31.0% (34) 27.6% 17.1% (32) 32.4% 21.5% (31) 19.1% 9.6% (26) 14.2% 10.7% (19) 44.5% 31.5% (6) 39.9% 33.2% (13) 100% (1) 34.7% 20.4% (40) 15.3% 24.6% (19) 29.7% 9.5% (23) 17.5% 25.0% (14) 22.1% 24.6% (20)
27.2% 10.9% 21.8% 25.5% 31.0% 14.5% 76.9% 29.8% 18.1% 13.0% 27.1% 20.6% 20.7% 11.1% 18.3% 22.1% 34.6% 44.3% 79.9% 35.4% 36.5% 19.5% 42.1% 29.7% 23.7% 28.5% 30.9% 18.5% 26.2% 31.1% 24.6% 15.2%
(6) (12) (7) (52) (34) (37) (30) (25) (4) (6) (3) (48) (24) (26) (15) (33)
0.532 0.982 0.812 0.057 0.001 0.299 0.741 0.705 0.614 0.073 0.095 0.005 0.259 0.578 0.345 0.890
64.1% 35.4% 79.5% 29.1% 84.7% 26.3% 58.4% 33.5% 41.5% 38.2% 77.4% 36.2% 85.2% 33.6% 78.4% 32.7% 79.6% 29.0% 90.6% 23.1% 91.4% 21.2% 79.8% 22.6% 74.8% 22.8% 76.2% 31.1% 89.2% 20.8% 86.8% 26.1% 86.5% 29.2%
70.1% 35.1% (19) 69.1% 32.2% (37) 54.1% 50.4% (6) 57.1% 31.7% (12) 41.5% 38.2% (8) 62.2% 41.0% (4) 60.9% 50.4% (3) 64.4% 33.6% (3) 66.7% 30.7% (18) 89.0% 24.5% (5) 84.2% 23.2% (8) 100% (2) 62.2% 9.5% (2) 65.4% 34.1% (11) 0 (0) 100% (1) 0 (0)
58.3% 36.3% (6) 87.9% 21.7% (28) 62.3% 36.5% (10) 100% (2) 0 (0) 100% (5) 99.5% 0.7% (2) 83.7% 29.4% (9) 82.9% 29.4% (32) 100% (9) 92.0% 25.2% (10) 75.2% 22.1% (3) 0 (0) 87.8% 21.8% (21) 83.1% 27.2% (6) 80.1% 35.7% (16) 100% (1)
44.2% 36.5% (4) 86.4% 27.4% (22) 84.7% 26.3% (6) 24.8% 5.9% (2) 0 (0) 67.2% 44.9% (5) 100% (3) 77.5% 39.3% (7) 85.1% 24.8% (23) 86.8% 27.7% (18) 100% (6) 53.0% (1) 100% (1) 64.8% 37.2% (11) 95.5% 10.9% (6) 89.3% 21.3% (39) 85.0% 30.6% (9)
0.389 0.014 0.374 0.067
(29) (87) (22) (16) (8) (14) (8) (19) (73) (32) (24) (6) (3) (43) (12) (56) (10)
0.232 0.331 0.699 0.090 0.382 0.404 0.234 0.191 0.053 0.321 0.451 0.655
*Numbers in brackets are the no. of studies. AF: atrial-femoral; AKA: Adamkiewicz artery; CPB: cardiopulmonary bypass; CSF: cerebrospinal fluid; DHCA: deep hypothermic circulatory arrest; DTA: descending thoracic aorta; ICA: intercostal artery; MEP: motor-evoked potentials; SCI: spinal cord injury; SSEP: somatosensory-evoked potentials; TAAA: thoracoabdominal aortic aneurysm; TEVAR: thoracic endovascular aortic repair.
postoperative spinal cord dysfunction, with extents I and II carrying the highest risk.1,3,6,8,15,19,36–47 In cases of TEVAR, the extent of distal location and extent of aortic coverage with respect to the celiac artery, and the number of stent grafts used during surgery also predicted SCI.24,28 Other factors identified as
predictors of postoperative SCI included longer aortic clamp time,9,12,15,36,40,46–50 lack of distal perfusion,9 no use of adjuncts,1,40 oversewn T9-T12 intercostals,41,47 no use of atrial-femoral bypass,15 left subclavian artery covering without revascularization,30 concomitant open abdominal aortic surgery,30 extent of
Downloaded from aan.sagepub.com at COLUMBIA UNIV on November 30, 2014
XML Template (2014) [19.8.2014–12:56pm] //blrnas3/cenpro/ApplicationFiles/Journals/SAGE/3B2/AANJ/Vol00000/140190/APPFile/SG-AANJ140190.3d
(AAN)
[1–12] [PREPRINTER stage]
4
Asian Cardiovascular & Thoracic Annals 0(0)
segmental artery sacrificed,53 failure to cool actively with bypass,12 level of the distal aortic clamp,54 longer lower spinal cord ischemic time,55 and increasing red blood cell infusions.56 Preoperative patient-related factors predicting SCI included age,36,57 rupture,36,44,58 proximal aortic aneurysm,36 history of renal dysfunction,36 acute presentation,3,8,38,58 previously repaired abdominal aortic aneurysm,1 history of cerebrovascular disease,1,43 nonelective operation,28,41,46 false aneurysm,9 inflammation,9 medial degeneration,9 comorbidities (hypertension, smoking, dyslipidemia, renal insufficiency),25 absence of post-dissection aneurysm,59 acute dissection,56 non-dissecting aneurysm,60 chronic renal insufficiency,28,30,33,34,61 and preoperative shock.55 Hypotension was also a predictor of SCI; both intraoperative hypotension,46,47 and postoperative hypotension.12,15,62,63 No use of CSF drainage,3,12,38,59 a decrease in cardiac index with aortic occlusion,3 normothermia,39 irreversible changes in MEP and SSEP,64 and lack of epidural cooling46,48 were also identified as predictors of SCI by multivariate and/or univariate analysis.
Discussion Spinal cord injury remains a devastating complication after TAAA repair. Its incidence is reported to range from 1% to 32%. Despite various adjuncts and surgical techniques, complete elimination of this complication is virtually impossible. Our analysis of the data of over 38,000 patients showed the overall incidence of SCI to be 7.1% 6.1%. Because this result is an average of a varying pool of patients in different series, the actual incidence of SCI was different in each patient pool (0%–32%). An important finding demonstrated by this analysis is the decreasing trend of SCI over the last few years. As we analyzed the publications, the extent of disease remained almost the same in the 3 time periods, except for a higher proportion of patients with type I TAAA before 2000. Nevertheless, the incidence of SCI decreased significantly from 9.0% 6.7% to 5.9% 5.2%. Use of adjuncts such as distal perfusion, hypothermia, CSF drainage, epidural cooling, naloxone, and MEP and SSEP monitoring has contributed to reducing SCI over recent years. Significantly more patients received distal perfusion after 2000, and although it did not reach statistical significance, a clear trend of increased use of CSF drainage was found after 2000. Neuromonitoring with MEP and SSEP, and preoperative artery of Adamkiewicz identification with magnetic resonance angiography and computed tomography-angiography were also increasingly utilized after 2000. ICA reimplantation showed an increasing trend between 2001 and 2007 compared to before 2000, and every effort was made to reimplant each
pair of ICA. However, after 2007, surgeons were concerned about the longer operative time needed to reimplant ICA, and they started to sacrifice intercostals that were not crucial to spinal circulation by monitoring MEP. Monitoring MEP helped surgeons to identify crucial ICA to reimplant, thus decreasing the total aortic clamp time. With the knowledge of the above mentioned predictors of paraplegia, a large number of strategies have been utilized in various centers to combat SCI following TAAA surgery. Because the extent of the aorta replaced is an important predictor of postoperative paraplegia, careful preoperative planning and use of multiple adjuncts can reduce SCI in these high-risk patients. Distal perfusion is an important strategy to reduce postoperative SCI. The simple clamp-and-sew technique results in more SCI. However, if combined with selective epidural cooling to a CSF temperature of 26.4 C, the simple clamp-and-sew technique also offers spinal cord protection.41 Distal perfusion has routinely been included in TAAA repairs in recent years.4,9,10,65–67 Several authors have shown that aortic clamp time is a predictor of postoperative SCI. Biglioli and colleagues49 reported that in descending thoracic aortic aneurysm repair, spinal cord perfusion can be adequately maintained without reimplantation of segmental vessels or use of atrial-distal bypass when the aortic crossclamp time is short (30 min was associated with SCI,9,40 and >60 min was associated with an increased incidence of SCI.48 Use of left heart bypass has been shown to lower the risk of SCI, mainly because it unloads the proximal circulation during aortic occlusion while maintaining adequate distal perfusion, allowing a longer time for all distal intercostal reimplantation.54 However, Coselli and colleagues56 found that left heart bypass was not a predictor of SCI. Use of cardiopulmonary bypass (partial or full) has also been shown to lower the risk of SCI.68 Use of shunts (Gott shunt) significantly reduced SCI during descending aortic surgery.69 The first prospective randomized study to examine the role of CSF drainage in preventing SCI after highrisk TAAA surgery showed negative results.62 However, these findings were later refuted by the same authors in another randomized clinical trial.5 The main difference between these two findings might be because in the first study, the authors limited the CSF drainage volume to a maximum of 50 mL; however, in the randomized clinical trial, their aim was to decrease CSF pressure to
