Journal of Vascular Surgery
Volume 49, Issue 5 , Pages 1117-1124, May 2009

A modern theory of paraplegia in the treatment of aneurysms of the thoracoabdominal aorta: An analysis of technique specific observed/expected ratios for paralysis

  • Charles W. Acher, MD

      Affiliations

    • Corresponding Author InformationReprint requests: Charles W. Acher, MD, University of Wisconsin, University of Wisconsin Hospital, 600 Highland Ave, Room G4-317, Madison, WI 53792-7375
    • ,
  • Martha Wynn, MD

Department of Surgery, University of Wisconsin, University of Wisconsin Hospital, Madison, Wisc

Received 11 September 2008; accepted 30 October 2008.

Article Outline

Objective

To demonstrate that a modern theory of paraplegia prevention in thoracoabdominal aortic (TAAA) surgery is primarily non-anatomic and derives from experimentally validated interventions that prolong the ischemic tolerance, reduce reperfusion injury, and enhance the collateral perfusion of the spinal cord with or without assisted circulation.

Methods

Using an accurate predictive model (r2 > 0.95) for paraplegia risk we studied the effects of protective strategies in 82 clinical series reporting more than 15,000 patients treated from 1985 to 2008. The observed/expected (O/E) ratios were calculated for each series and the results were grouped by technique. The effect of interventions such as spinal fluid drainage (SFD), systemic hypothermia, epidural cooling, and naloxone on O/E ratios were studied. We analyzed changes in O/E ratios from Era 1 (1985 to 1997) to Era 2 (1997 to 2008) and within treatment techniques over time.

Results

The mean O/E ratio for paraplegia for all patients declined from 1.13 in Era 1 to 0.26 in Era 2. Adding SFD to patients treated with assisted circulation (AC) decreased the O/E ratio from 1.03 to 0.24 (P < .0001). Adding SFD to patients treated with aortic clamping without AC (XCL) decreased O/E from 0.91 to 0.23 (P = .0013). O/E for hypothermic arrest (HA) declined from 0.42 to 0.14 with SFD. The addition of SFD to AC, XCL, and HA accounted for most of the decline in O/E between Eras. Other factors which played a less defined but important role in the decline in O/E ratios were attention to higher mean arterial pressures (MAPs), more hypothermia, and neurochemical protection.

Conclusion

Paraplegia causation is anatomic but paraplegia prevention is physiologic (non-anatomic). We demonstrate that by using hypothermia, SFD, and increasing MAP, clinicians had similar improvements in paraplegia, reducing O/E deficit ratios from 1.03 to as low as 0.16, with or without intercostal reimplantation, and whether or not assisted circulation was used. Understanding the fundamental principles of paraplegia prevention and how to apply protective strategies leads to a reduction in paralysis in clinical series with or without the use of assisted circulation. This modern theory of paraplegia has significant implications for the rapidly advancing field of TAAA repair with branched endografts where the same principles apply.

 

Over the last 25 years, there has been a significant decline in paraplegia risk associated with surgical repair of the thoracoabdominal aorta. The dominate paradigm for thinking about paraplegia causation and prevention has remained anatomic with focus on the Artery of Adamkiewicz as described by Adams and van Geertruyden1 over 50 years ago. However, the major declines in paraplegia rates have not occurred from reattaching important radicular arteries but rather from the introduction of strategies that focus on maximizing effective collateral perfusion,2 reducing cord ischemia, and increasing the ischemic tolerance of the spinal cord during and after aortic replacement. What has emerged is a new paraplegia paradigm that is anatomic in causation, but primarily non-anatomic in prevention which is achieved by maximizing collateral circulation and increasing the ischemic tolerance of the spinal cord. Ironically this new paradigm has been thrown into sharp relief with the introduction of endovascular thoracic aortic aneurysm (TAA) and thoracoabdominal aortic aneurysm (TAAA) repair where an anatomic solution to paraplegia prevention is not possible. It is also clear that in endograft repair of the TAA, the more aorta covered the greater the paraplegia risk, and paraplegia rates in endograft repair have been reduced by applying non-anatomic strategies, such as spinal fluid drainage (SFD), to prevent paraplegia.

To study the clinical impact of protective strategies in spinal ischemia, we derived and published a paraplegia risk model to calculate risk index as observed/expected (O/E) ratios for paraplegia in our own and others' series.3 This model was highly accurate and accounted for more than 90% of the variability in our patients and in other clinical series. Using this model, we have previously demonstrated that by applying the non-anatomic strategies of increased collateral perfusion, hypothermia, SFD, and neurochemical protection we have reduced paraplegia risk by approximately 80% to a risk index (O/E ratio) of 0.20 without intercostal reimplantation or assisted circulation.3 More recently we demonstrated that the addition of intercostal reimplantation further decreased the O/E ratio to 0.05, for a 95% reduction in paraplegia risk without assisted circulation.4

In this report, we study clinical series from 1985 to 2008 and demonstrate that improved paraplegia outcomes arise primarily from applying interventions that maximize collateral perfusion and ischemic tolerance of the spinal cord during aortic replacement and that these same principles apply to endovascular aneurysm repair.

Back to Article Outline

Materials and methods 

We modeled 82 clinical series reporting more than 15,000 patients spanning 25 years and calculated O/E ratios for paraplegia for each series. We used paraplegia as a synonym for all spinal cord-related deficits (paraparesis or paraplegia, immediate or delayed), as we have done in all our previous reports. Patients were classified by treatment strategies: Crawford's technique of cross-clamp with intercostal reimplantation without assisted circulation or protective adjuncts (XCL); cross-clamp without assisted circulation plus SFD (XCL+SFD); assisted circulation without SFD (AC); assisted circulation with SFD (AC+SFD); hypothermic arrest without (HA) and with SFD (HA+SFD). Our own method of SFD plus naloxone, hypothermia, thiopental, and steroids was designated SFD/naloxone (SFDN). We evaluated treatment strategies from 1985 to 1997 (Era 1) and from 1997 to 2008 (Era 2).

Statistical analysis 

We used logistic regression and Fisher's Exact test to study significant variables using SAS JMP software for analysis (SAS Institute Inc, Cary, NC).

We evaluated patients using a paraplegia predictive model we derived and published previously. In the model, expected deficits = (0.1C1 + 0.2C2 + 0.05C3 + 0.02C4 + 0.01 TAA) + (0.3 [acute + dissection]).3 This model was validated by evaluating several thousand patients from the clinical literature treated without protective adjuncts and accounted for 99% (r2 = 0.99) of the variability among reports. The model compared the actual number of deficits to the predicted number of deficits for each series and was used to calculate O/E ratios. Deficit Risk Index (O/E ratio) = observed/expected deficits.3 We used the predictive model to study the impact of spinal cord protective strategies reported in clinical series totaling over 15,000 patients.

Back to Article Outline

Results 

Modeling clinical reports 

We modeled 82 clinical reports of 15,526 patients from 1985 to 2008. There was a significant decrease in O/E ratios for paralysis over the 25 years studied (P < .0001) (Fig 1). The best fit (highest r2 of 0.75) for this decline was logarithmic. We plotted expected deficits from our predictive model vs observed deficits for Era 1 and Era 2. The r2 for each regression line was >0.97, supporting the model's accuracy over time. There was a highly significant difference in mean O/E ratios between Era 1 and Era 2 (1.13 for Era 1 vs 0.26 for Era 2, P < .0001) (Fig 1). The predictive value of the model did not change significantly between Era 1 and Era 2 and the model continued to be an excellent measure of outcomes with the factors within the model accounting for greater than 97% of the variability between series. What did change between Era 1 and Era 2 was the O/E ratios.

  • View full-size image.
  • Fig 1. 

    A, Plot of O/E ratios for paraplegia in clinical reports by year shows a significant decline over the last 23 years. From the curve of this decline we divided our analysis into Eras: Era 1 from 1985 to 1997 and Era 2 from 1997 to 2008. B, Using our predictive model for paralysis risk we plotted observed vs estimated (EST) deficits. The regression plots for Era 1 and 2 showed a significant decline between Eras (P < .0001) and the model accounted for greater than 97% of the variability between reports for each regression line (r2 > 0.97).

We then compared reported results in the 82 series grouped by technique of repair and Era to evaluate whether the introduction of protective strategies accounted for the improved results between Era 1 and Era 2 (Table I, Table II).

Table I. Results for each technique by Era. 1 = Era 1 (1985-1997) 2 = Era 2 (1997-2008)
TechniqueC1C2C3C4TAADis/AcuteExpObsvO/EPTotal%Def
All AC15325552741725374643352460.81 207011.9
All AC218692019964924280177111022680.24<.000160564.4
AC+SFD1111195115107410584.3350.42 5326.6
AC+SFD217051842860890241490972.42210.230.18253214.2
HA 168722391259647.8210.44 2977.1
HA 2103143811416420696.4150.160.02575053
AC1 No Adj35328813656408263170.71901.11 124115.3
AC2 No Adj61342320927534.1320.940.559323013.9
All XCL1107112641106112817011147176060.84 473612.8
All XCL2198310158209149426218660.270.00139286.5
XCL+SFD151696182155437.2200.54 2787.2
XCL+SFD2N/AN/AN/AN/AN/AN/AN/AN/AN/A N/AN/A
XCL+SFDN1387732513610648.780.16 2343.4
XCL+SFDN2126244144209149349161.5290.180.79568723.3
XCL1 No Adj9821118101310411199546335780.91 427313.5
XCL2 No Adj726614007740.4370.920.807315224.3

AC, Assisted circulation; XCL, cross-clamp without assisted circulation; AC+SFD, assisted circulation plus spinal fluid drainage; XCL+SFD, cross-clamp without assisted circulation plus spinal fluid drainage; XCL+SFDN, cross-clamp without assisted circulation plus spinal fluid drainage/naloxone; HA, hypothermic circulatory arrest; HA+SFD, hypothermic circulatory arrest plus spinal fluid drainage; N/A, indicates that in Era 2 no series using SCL used SFD alone.

Table II. Results and O/E ratios for each technique, all reports, 1985-2008
TechniqueC1C2C3C4TAADis/AcuteExpObsvO/ETotal% DefP
AC41834715261408328206.12131.03138615.4
AC+SFD179720289539672815831048.32480.2457734.3<.0001
XCL10541184102710411191031673.36150.91442513.9
XCL+SFD all2153902373422005092447570.2313844.1.0013
XCL+SFD51696182155437.2200.542787.2
XCL+SFDN164321176260185455210.2370.1811063.3.017
XCL+EPIDC32518741117424394192.71140.59112110.2
HA738427914610653.64230.423396.7
HA+SFD98131771414319690.56130.1434632.8.0065
ENDO 26 15044722231.045314.3
ENDO+SFD234767662535332.15180.564563.9.3174
TOTALS398044192946260915724194 1309 155268.4

AC, Assisted circulation; AC+SFD, assisted circulation plus spinal fluid drainage; XCL, cross-clamp without assisted circulation; XCL+SFD, cross-clamp without assisted circulation plus spinal fluid drainage; XCL+SFDN, cross-clamp without assisted circulation plus spinal fluid drainage/naloxone; XCL+EPIDC, cross-clamp without assisted circulation plus epidural cooling; HA, hypothermic arrest; HA+SFD, HA+spinal fluid drainage; ENDO, endograft repair without spinal fluid drainage; ENDO+SFD, endograft repair plus spinal fluid drainage.

Assisted circulation (AC) 

The O/E ratios for all AC patients declined from 0.81 in Era 1 to 0.24 in Era 2 (P < .0001). The O/E ratios for AC without adjuncts (AC No Adj) did not change significantly between Eras (1.11 to 0.94, P = .5593) (Table I). However, when SFD was added to AC (AC+SFD) there was a significant decline in O/E from 1.03 to 0.24 (P < .0001) (Table II). The addition of SFD to AC clearly explains most of the decline in O/E between Era 1 and Era 2 and fits with the observation that most reports of AC No Adj are in Era 1 and most reports of AC+SFD are in Era 2. Within the AC+SFD group there was a decline in O/E between Eras from 0.42 to 0.23 (P = .1820) (Table I). The factors that account for this decline within the same technique are less clear. In hypothermic circulatory arrest patients (HA), the O/E declined from 0.44 to 0.16 (P = .0257) between Eras (Table I) and from 0.42 to 0.14 with SFD (HA+SFD) (P = .0065) (Table II). The most important factor in this decline between Eras was the introduction of SFD that occurred in Era 2. There was not a significant change in O/E ratios over time within AC or HA without SFD (Table I).

No AC 

For all patients that had aortic clamping without AC (XCL) the O/E ratios declined from 0.84 in Era 1 to 0.27 in Era 2 (P = .0013) (Table I). In XCL patients treated with no protective adjuncts (XCL No Adj) O/E ratios did not change significantly between Eras (0.91 to 0.92, P = .8073) (Table I). Adding SFD to XCL significantly decreased O/E ratios from 0.91 to 0.23 (P = .0013) (Table II). SFD clearly accounted for most of the reduction in XCL O/Es between Eras (Table I). The O/E ratio decreased from 0.60 for just SFD to 0.16 for SFDN (P = .017) (Table II). Within the SFD and SFDN groups there was no improvement over time, with risk remaining stable for those techniques.

Protective adjuncts across techniques 

When O/E ratios for all series (Era 1 and Era 2) that used protective adjuncts (SFD and/or hypothermia) were compared, there was no difference in mean O/E ratios for paraplegia between groups (P = .9942) whether it was AC+SFD (O/E = 0.263), XCL+SFD (O/E = 0.257) or HA (HA and HA+SFD) (O/E = 0.255) (Fig 2). Doing this same comparison just for Era 2 continued to show the groups had similar outcomes (P = .6004), but O/E ratios decreased for XCL+SFD (0.26 to 0.18, P = .2432) and HA (0.25 to 0.17, P = .1450) and remained unchanged for AC+SFD (0.26 to 0.24, P = .7422).

There were major treatment changes over time as surgeons adopted more effective protective strategies that account for the improved outcomes seen in Era 2. Fewer series in Era 2 used just AC, HA, or XCL as surgeons transitioned to AC+SFD, HA+SFD, and SFDN, with most using AC+SFD. Other factors changed over time that we were not able to quantify such as increased attention to hypothermia and perfusion pressures, fewer restrictions on volume of spinal fluid removed so spinal fluid pressure could be strictly controlled, and changes in neurochemical protection.

We analyzed eight recent reports (987 patients) of endograft treatment of TAA and TAAA. In those that used SFD, the O/E ratio was 0.56, and in those that did not indicate use of any protective strategies the O/E ratio was 1.04 (Table II).

Back to Article Outline

Discussion 

Any theory should be consistent with experimental data, real world observations, and, most of all, predictive of future observations. What we have observed from the experimental data and our own and others' analysis is that factors that maximize collateral blood flow5 such as cardiac index (CI) and mean arterial pressure (MAP) and SFD improve spinal cord perfusion.6, 7 Other interventions that increase ischemic tolerance of the spinal cord (barbiturates,8 hypothermia,9 and steroids10), reduce excitotoxicity from neuronal ischemia (hypothermia,11 naloxone,12 and steroids13), and attenuate reperfusion injury (steroids, hypothermia,14 and free radical scavengers15) also reduce spinal cord injury in TAAA repair. Intercostal reimplantation is not as important as these other factors, as indicated by Greipp et al7, 16 and our own reports.4, 5 The most recent report of our own results assessed the importance of intercostal reimplantation in a quantitative way using statistical modeling of O/E ratios for paraplegia in patients treated without and with intercostal reimplantation. We concluded that non-anatomic factors accounted for 80% of paraplegia risk and direct intercostal blood flow accounted for 20% of risk.4

Hemodynamics 

The experimental data tells us several important things about spinal cord ischemia. The primary experimental observation is that thoracic aortic occlusion causes a significant reduction in spinal cord blood flow and perfusion pressure (80%)17 and that the negative effect of this drop in perfusion increases with occlusion time,18 resulting in increasing risk of spinal cord infarction with longer occlusion times. Factors that change over time with thoracic aortic occlusion to the detriment of spinal cord perfusion are increasing spinal fluid pressure or central venous pressure (which decrease spinal cord perfusion pressure),19, 20 changes in hemodynamics such as arterial hypotension,21 decreased cardiac index, and decreased oxygen carrying capacity from blood loss. Optimizing these factors in experimental models improves ischemic tolerance of the cord by increasing cord perfusion and tissue oxygen delivery, thus increasing the allowable time of aortic occlusion before irreversible ischemia and infarction occur. Increasing collateral perfusion pressure increases blood flow and relative spinal cord perfusion pressure, draining spinal fluid decreases or prevents an increase in spinal fluid pressure thereby increasing blood flow and relative spinal cord perfusion pressure, and increasing or maintaining cardiac index improves cord blood flow and perfusion. Aggressively correcting anemia and hypovolemia and preventing coagulopathy improve tissue oxygen delivery.

There is ample clinical evidence that these factors are at play in human thoracoabdominal aortic replacement. In our own experience SFD, MAP, and cardiac index were important factors by univariate analysis and SFD and CI remained significant in multivariate modeling.4 Patients with lower CI and MAP were more vulnerable to paraplegia as were patients that did not have SFD. Some of the most compelling evidence of the importance of hemodynamic factors comes from human reports on the use of motor evoked potentials (MEPs) to assess cord function and intercostal arteries for reimplantation during aortic occlusion.22 Whether clinicians believed that reattaching intercostals was necessary or not, both camps have shown that increasing perfusion pressure during aortic occlusion returns evoked potentials to normal in the majority of patients and that keeping MAP >90 mm Hg recruits enough collateral circulation to adequately perfuse the spinal cord during aortic occlusion.7 What our data contributes to these observations is that collateral recruitment occurs whether the increase in perfusion pressure is proximal (no AC) or distal (AC) to the aortic clamp. This same MEP data illustrates one of the vulnerabilities of AC: the relative hypotension above the aortic clamp under perfuses the proximal collateral network. This proximal under perfusion during AC combined with normothermic perfusion may explain why historically the O/E ratios for AC were higher than the O/E ratios without AC.5 Draining spinal fluid and maintaining high perfusion pressures in this context improve paraplegia outcomes significantly, as demonstrated in the analysis of our own results. The importance of maximizing collateral hemodynamics cannot be overstated in cord protection. We and others have shown that focusing on enhancing collateral perfusion can protect all but the highest risk patients from paralysis.4, 7 We think some of the improvement between Eras in the AC+SFD patients demonstrated in the present study is accounted for by increased attention to these hemodynamic factors. Failure to control hemodynamics and spinal fluid pressure may account for less than optimal results seen in some reports, whether AC or XCL is used in open repair or repair is endovascular.

Ischemic tolerance 

Experiments have shown that the most effective method of protecting the nervous system from transient decreases in blood flow is hypothermia. Hypothermia prolongs the ischemic tolerance of spinal cord by reducing neuronal metabolism, decreasing oxygen demand in nervous tissue,9 and reducing the levels of neurotoxic excitatory neurotransmitters released11 during neuronal ischemia. Profound hypothermia (15°C) is protective for the brain and spinal cord for up to 60 minutes in aortic arch replacement. Moderate hypothermia (30 to 34°C) also significantly prolongs ischemic tolerance23 and experimentally each degree below 36°C prolonged allowable ischemic time by 50%.24 These temperatures can be accomplished without AC. Clinically, it is also apparent that in TAAA repair using hypothermic arrest, even with extensive intercostal artery reimplantation, the spinal cord is vulnerable during rewarming as oxygen demand increases but perfusion pressure may be low and flow is nonpulsatile. We believe the move away from normothermia is one of the major factors improving the results of AC or XCL when combined with SFD seen in the present study and is a factor in improvements between Eras for AC+SFD. Not achieving significant hypothermia may also be a factor in variations seen among series using similar techniques.

Experimental studies have shown the benefit of pharmacologic adjuncts in animal models of spinal cord ischemia. Thiopental has been shown to decrease nervous tissue metabolism and protect neurons during ischemia.8 The administration of high dose steroids has also been shown to reduce spinal cord injury in animal models of spinal cord ischemia.10 Neurotoxicity from excitatory neurotransmitter release has a less well-defined but experimentally significant negative effect on ischemic tolerance of spinal cord. Naloxone decreases this negative effect of ischemia in animal experiments25 and has also been shown to decrease the levels of neurotoxic excitatory neurotransmitters in spinal fluid during TAAA repair in humans.12 In our experience, the addition of naloxone to SFD had a significant protective effect, and in the present study, O/E ratios decreased from 0.60 in patients with SFD to 0.16 in patients treated with SFD plus naloxone. Our attention to hemodynamic factors, hypothermia, and controlling spinal fluid pressure may also contribute to this difference.

Intercostal arteries 

Perhaps one of the most contentious and misunderstood concepts in paraplegia causation and prevention in aortic surgery has been the role of the intercostal arteries in spinal cord blood flow during and after TAAA repair. Adam's landmark paper on the anatomic variations of spinal cord blood flow clearly defined the anatomic paradigm of paraplegia causation and in some ways, unfortunately, also defined the anatomic solution to paraplegia that surgeons have focused on for decades. However, it is now clear that although the interruption of intercostal artery flow is causative in paraplegia, most of the solutions for preventing paraplegia are non-anatomic, and paraplegia prevention depends on optimizing collateral circulation and prolonging ischemic tolerance with ‘non-anatomic' therapeutic interventions such as hypothermia, SFD, optimizing hemodynamics, and neurochemical protection. This collateral circulation solution is apparent from our own results that show an 80% reduction in paraplegia without intercostal reimplantation or AC5 and similar results reported by Griepp et al7, 16 using AC and no intercostal reimplantation.

The anatomic paradigm is further confused by the observation that in TAAA patients, 50% of the intercostals are chronically occluded26 and the anterior spinal artery is perfused by the axial collateral network and not directly from a greater radicular artery. Historically, attempts at identifying and selectively reattaching the Artery of Adamkiewicz have had poor results with O/E ratios >1 and as high as 4.26, 27 This, along with the observation that for many years intercostal arteries have been routinely reattached by most surgeons without paralysis prevention, highlight the failure of this strategy alone to prevent paralysis.28, 29 However, we have demonstrated that intercostal artery reimplantation when added to the strategies of hypothermia, SFD plus naloxone, and maximizing cardiac function improved our O/E ratio from 0.20 to 0.04, implying that direct intercostal blood flow accounts for 15 to 20% of the risk of paralysis in thoracoabdominal aortic replacement.4 Although most patients have adequate collateral circulation, a small number of patients do not and these patients require intercostal reimplantation to prevent paralysis.4

MEP data from clinical series has highlighted the importance of collateral perfusion pressure in maintaining cord function during aortic replacement whether the bias is toward intercostal ligation or reimplantation to sustain adequate cord perfusion. Greipp et al's7, 16 clinical reports are especially elegant in critically evaluating the limited role of direct intercostal blood flow for cord function during aortic replacement and ironically are confirmed by Jacobs et al,22 a strong advocate of selective intercostal reimplantation. Both observed the effect of correcting MEPs with higher perfusion pressures before reimplantation of intercostal arteries. It is clear that MEPs are overly sensitive in paraplegia prediction. It is also clear from the MEP data that, if AC is used, higher MAP during aortic repair is critical for cord protection. In fact, the importance of intercostal reimplantation may be increasing perfusion pressure in the collateral perfusion bed rather then directly reattaching the Artery of Adamkiewicz in the small number of patients where it makes a critical difference.30 The recent work of Backes et al31 using magnetic resonance imaging to identify collateral circulation to the spinal cord in TAAA patients, demonstrates the critical importance of collateral circulation in cord protection. Unfortunately, Backes only looked at distal collaterals, however, our own results without AC suggest proximal collaterals are equally important and the same principles of higher perfusion pressure and increased cardiac index apply.

Assisted circulation (AC) 

There is no experimental evidence that AC improves spinal cord blood flow in primates, pigs, or dogs when the experimental model simulates the distal perfusion conditions during TAAA replacement.17, 32 If, however, the thoracic aorta is perfused below a proximal aortic clamp in the area that gives rise to the Greater Radicular Artery, there is some protective benefit.33 Even in Meylaerts et al34 and Jacobs' experimental model, the controls were made hypotensive with blood loss to demonstrate a benefit from AC.

In spite of these experimental observations, many surgeons believe that AC is necessary to prevent paraplegia in TAAA repair. It is clear from our analysis that normothermic-AC without adequate perfusion pressure, even with intercostal reimplantation, is not protective and AC may increase paraplegia risk unless the principles of paraplegia prevention are understood and applied. These principles are: hypothermia is necessary to increase ischemic tolerance; adequate perfusion pressure is required to recruit collateral cord perfusion; applying these concepts with SFD reduces paraplegia risk even further; and a few patients with the most extensive aneurysms require intercostal artery reimplantation. So it is not the use of AC per se that makes a difference in reducing paraplegia, but rather learning how to use it effectively to accomplish the underlying principles of neuroprotection and collateral recruitment that reduce irreversible spinal cord ischemia. Critical clinical data supporting this comes from the effect of MAP on MEPs from both Griepp et al and Jacobs et al, who had comparable improved outcomes using AC with22 and without7, 16 intercostal reimplantation, and our own data showing comparable paraplegia outcomes without AC.4 The similar paraplegia results with or without AC illustrate that different techniques can be used to produce the same outcomes as long as the principles of paraplegia prevention are understood and applied in an effective way. This explains the improved outcomes using AC+SFD from Era 1 to Era 2.

The theory of paraplegia prevention 

What we have learned from our own experience and analyzing the experience of others with our predictive model is that the experimentally identified factors that optimize collateral cord perfusion (MAP, CI, and SFD), oxygen carrying capacity (circulating blood volume, CI, and hemoglobin), those that increase ischemic tolerance (hypothermia, steroids, naloxone, and thiopental), and reduce reperfusion injury (hypothermia and steroids) are the factors that make the most difference when applied clinically (80% risk reduction) whether or not AC is used. It is also apparent that direct intercostal blood flow is unnecessary in most patients but is necessary in a small number of patients, accounting for about 20% of overall paraplegia risk. This effect is most important in Crawford Type 2 patients. So although the theory of paraplegia causation is anatomic, the theory of prevention is primarily non-anatomic.

The question remains, can we use the principles of this theory to predict clinical observations including outcomes for endovascular procedures where direct intercostal perfusion is not possible because intercostals can not be reimplanted? As we have shown, we can predict the effect of current cord protection strategies and generate risk coefficients for those strategies. In the case of endovascular treatment of thoracic and thoracoabdominal aneurysms, there are currently just a few series to test our theory on. Some report hybrid procedures and some report branched or fenestrated endografts; some use protective adjuncts and others do not. The amount of aorta covered is not precisely defined in these reports but it is possible to extrapolate to identify C2 and C3 equivalent coverage by description in TAAA patients. If our theory is correct, without the use of protective adjuncts that maximize ischemic tolerance and collateral circulation, the O/E ratio should be around 1.0, and if protective adjuncts are used the O/E ratio should be no lower than 0.20 because of the effect of intercostal exclusion. We should therefore be able to test the predictive accuracy of our theory.

We studied eight endograft series comprising more than 900 patients, two reporting hybrid procedures with bypasses to the visceral vessels,35, 36 three reporting branched endografts,37, 38, 39 and three reporting just thoracic aneurysms.40, 41, 42 In the TAAA endograft series that used SFD prophylactically, the O/E ratio was 0.54;39, 40, 41 the mean O/E ratio for series reporting open TAAA repair using only SFD without hypothermia and neurochemical protection was approximately 0.55. In endograft series that used no protective adjuncts, the O/E ratio was 1.05.35, 36, 40, 41, 42 We recognize that most thoracic endovascular aortic repair (TEVAR) series have low-risk patients and those TEVAR series we included we modeled as thoracic aneurysms, which have a risk coefficient in the model of 0.01. The model assesses the risk of the population being studied whether it is high or low. We demonstrated, regardless of whether the risk in a series is high or low, that the risk is predictable and the series which used spinal cord protection had predictably less paralysis than those that did not. These O/E ratios confirm what we would predict based on our theory and support the view that we can understand and predict when and why we fail. These observations also support the theoretical limitations of spinal cord protection imposed by not being able to attach intercostal arteries.

So we have come full circle and demonstrated that factors affecting collateral circulation and ischemic tolerance of the spinal cord are the most important for paraplegia prevention whether treating TAAs with endografts or open surgery, and in fact endografts appear to carry the same risk of paralysis as open procedures if aortic coverage or replacement is equivalent. In addition, paraplegia risk conforms to the same rules in both procedures and much of the variability between series comes from how effective they have been in applying protective strategies. Understanding these principles of spinal cord protection will improve outcomes whether open or endograft aortic repair is done. Our theory also opens the door for other protective measures in endograft patients such as hypothermia with cooling blankets as suggested by Dr Randall Griepp, or the experimentally demonstrated ischemic preconditioning of the spinal cord with temporary balloon occlusion of the aorta.43 However, the inability to reattach intercostal arteries in those few patients that need direct intercostal flow puts a theoretical limit (O/E ratio of 0.20) on paraplegia reduction in endograft repair of TAAA, which is confirmed by our observations to date.

In open repair, it is the application of these same strategies that maximize collateral blood flow, reduce nervous tissue oxygen demand, prolong ischemic tolerance of the spinal cord, and reduce reperfusion injury and not the use of AC that has reduced paralysis risk.

Back to Article Outline

Author contributions 


Conception and design: CA, MW

Analysis and interpretation: CA, MW

Data collection: CA, MW

Writing the article: CA, MW

Critical revision of the article: CA, MW

Final approval of the article: CA, MW

Statistical analysis: CA

Obtained funding: Not applicable

Overall responsibility: CA

Back to Article Outline

References 

  1. Adams HD, van Geertruyden HH. Neurologic complications of aortic surgery. Ann Surg. 1956;144:574–610
  2. Griepp RB, Griepp EB. Spinal cord perfusion and protection during descending thoracic and thoracoabdominal aortic surgery: the collateral network concept. Ann Thorac Surg. 2007;83:S865–S869discussion S890-2.
  3. Acher CW, Wynn MM, Hoch JR, Popic P, Archibald J, Turnipseed WD. Combined use of cerebral spinal fluid drainage and naloxone reduces the risk of paraplegia in thoracoabdominal aneurysm repair. J Vasc Surg. 1994;19:236–246discussion 247-8.
  4. Acher CW, Wynn MM, Mell MW, Tefera G, Hoch JR. A quantitative assessment of the impact of intercostal artery reimplantation on paralysis risk in thoracoabdominal aortic aneurysm repair. Ann Surg. 2008;248:529–540
  5. Acher CW, Wynn MM, Hoch JR, Kranner PW. Cardiac function is a risk factor for paralysis in thoracoabdominal aortic replacement. J Vasc Surg. 1998;27:821–828discussion 829-30.
  6. Coselli JS, LeMaire SA, Schmittling ZC, Köksoy C. Cerebrospinal fluid drainage in thoracoabdominal aortic surgery. Semin Vasc Surg. 2000;13:308–314
  7. Etz CD, Halstead JC, Spielvogel D, Shahani R, Lazala R, Homann TM, et al. Thoracic and thoracoabdominal aneurysm repair: is reimplantation of spinal cord arteries a waste of time?. Ann Thorac Surg. 2006;82:1670–1677
  8. Nylander WA, Plunkett RJ, Hammon JW, Oldfield EH, Meacham WF. Thiopental modification of ischemic spinal cord injury in the dog. Ann Thorac Surg. 1982;33:64–68
  9. Michenfelder JD, Milde JH. The relationship among canine brain temperature, metabolism, and function during hypothermia. Anesthesiology. 1991;75:130–136
  10. Laschinger JC, Cunningham JN, Cooper MM, Krieger K, Nathan IM, Spencer FC. Prevention of ischemic spinal cord injury following aortic cross-clamping: use of corticosteroids. Ann Thorac Surg. 1984;38:500–507
  11. Rokkas CK, Cronin CS, Nitta T, Helfrich LR, Lobner DC, Choi DW, et al. Profound systemic hypothermia inhibits the release of neurotransmitter amino acids in spinal cord ischemia. J Thorac Cardiovasc Surg. 1995;110:27–35
  12. Kunihara T, Matsuzaki K, Shiiya N, Saijo Y, Yasuda K. Naloxone lowers cerebrospinal fluid levels of excitatory amino acids after thoracoabdominal aortic surgery. J Vasc Surg. 2004;40:681–690
  13. Wu GJ, Chen WF, Sung CS, Jean YH, Shih CM, Shyu CY, et al. Preventive effects of intrathecal methylprednisolone administration on spinal cord ischemia in rats: the role of excitatory amino acid metabolizing systems. Neuroscience. 2007;147:294–303
  14. Allen BT, Davis CG, Osborne D, Karl I. Spinal cord ischemia and reperfusion metabolism: the effect of hypothermia. J Vasc Surg. 1994;19:332–339discussion 339-40.
  15. Lim KH, Connolly M, Rose D, Siegman F, Jacobowitz I, Acinapura A, et al. Prevention of reperfusion injury of the ischemic spinal cord: use of recombinant superoxide dismutase. Ann Thorac Surg. 1986;42:282–286
  16. Griepp RB, Ergin MA, Galla JD, Lansman S, Khan N, Quintana C, et al. Looking for the artery of Adamkiewicz: a quest to minimize paraplegia after operations for aneurysms of the descending thoracic and thoracoabdominal aorta. J Thorac Cardiovasc Surg. 1996;112:1202–1213discussion 1213-5.
  17. Svensson LG, Rickards E, Coull A, Rogers G, Fimmel CJ, Hinder RA. Relationship of spinal cord blood flow to vascular anatomy during thoracic aortic cross-clamping and shunting. J Thorac Cardiovasc Surg. 1986;91:71–78
  18. Katz NM, Blackstone EH, Kirklin JW, Karp RB. Incremental risk factors for spinal cord injury following operations for acute aortic traumatic transection. J Thorac Cardiovasc Surg. 1981;81:669–674
  19. Blaisdell FW, Cooley DA. The mechanism of paraplegia after temporary thoracic aortic occlusion and its relationship to spinal fluid pressure. Surgery. 1962;51:351–355
  20. Marini CP, Levison J, Caliendo F, Nathan IM, Cohen JR. Control of proximal hypertension during aortic cross-clamping: its effect on cerebrospinal fluid dynamics and spinal cord perfusion. Semin Thorac Cardiovasc Surg. 1998;10:51–56
  21. Toung TJ, Chang Y, Williams M, Crain BJ, Traystman RJ, Bhardwaj A. Experimental spinal cord ischemia: model characterization and improved outcome with arterial hypertension. Crit Care Med. 2004;32:1346–1351
  22. Jacobs MJ, Mess W, Mochtar B, Nijenhuis RJ, Statius van Eps RG, Schurink GW. The value of motor evoked potentials in reducing paraplegia during thoracoabdominal aneurysm repair. J Vasc Surg. 2006;43:239–246
  23. Marsala M, Vanicky I, Yaksh TL. Effect of graded hypothermia (27° to 34°C) on behavioral function, histopathology, and spinal blood flow after spinal ischemia in rat. Stroke. 1994;25:2038–2046
  24. Strauch JT, Lauten A, Spielvogel D, Rinke S, Zhang N, Weisz D, et al. Mild hypothermia protects the spinal cord from ischemic injury in a chronic porcine model. Eur J Cardiothorac Surg. 2004;25:708–715
  25. Faden AT, Jacobs TP, Mougey E, Holaday JW. Endorphins in experimental spinal injury: therapeutic effect of naloxone. Ann Neurol. 1981;10:326–332
  26. Williams GM, Roseborough GS, Webb TH, Perler BA, Krosnick T. Preoperative selective intercostal angiography in patients undergoing thoracoabdominal aneurysm repair. J Vasc Surg. 2004;39:314–321
  27. Cunningham JN, Laschinger JC, Spencer FC. Monitoring of somatosensory evoked potentials during surgical procedures on the thoracoabdominal aorta. J Thorac Cardiovasc Surg. 1987;94:275–285
  28. Crawford ES, Crawford JL, Safi HJ, Coselli JS, Hess KR, Brooks B, et al. Thoracoabdominal aortic aneurysms: preoperative and intraoperative factors determining immediate and long-term results of operations in 605 patients. J Vasc Surg. 1986;3:389–404
  29. Svensson LG, Crawford ES, Hess KR, Coselli JS, Safi HJ. Experience with 1509 patients undergoing thoracoabdominal aortic operations. J Vasc Surg. 1993;17:357–368discussion 368-70.
  30. Uezu T, Koja K, Kuniyoshi Y, Miyagi K, Shimoji M, Arakaki K, et al. Blood distribution to the anterior spinal artery from each segment of intercostal and lumbar arteries. J Cardiovasc Surg (Torino). 2003;44:637–645
  31. Backes WH, Niejenhuis RJ, Mess WH, Wilmink FA, Schurink GW, Jacobs MJ. Magnetic resonance angiography of collateral blood supply to spinal cord in thoracic and thoracoabdominal aortic aneurysm patients. J Vasc Surg. 2008;48:261–271
  32. Maharajh GS, Pascoe EA, Halliday WC, Grocott HP, Thiessen DB, Girling LG, et al. Neurological outcome in a porcine model of descending thoracic aortic surgery. Stroke. 1996;27:2095–2100discussion 2101.
  33. Svensson LG, Von Ritter CM, Groeneveld HT, Rickards ES, Hunter SJ, Robinson MF, et al. Cross-clamping of the thoracic aorta (Influence of aortic shunts, laminectomy, papaverine, calcium channel blocker, allopurinol, and superoxide dismutase on spinal cord blood flow and paraplegia in baboons). Ann Surg. 1986;204:38–47
  34. Meylaerts SA, De Haan P, Kalkman CJ, Jaspers J, Vanicky I, Jacobs MJ. Prevention of paraplegia in pigs by selective segmental artery perfusion during aortic cross-clamping. J Vasc Surg. 2000;32:160–170
  35. Resch TA, Greenberg RK, Lyden SP, Clair DG, Krajewski L, Kashyap VS, et al. Combined staged procedures for the treatment of thoracoabdominal aneurysms. J Endovasc Ther. 2006;13:481–489
  36. Böckler D, Kotelis D, Geisbüsch P, Hyhlik-Dürr A, Klemm K, von Tengg-Kobligk H, et al. Hybrid procedures for thoracoabdominal aortic aneurysms and chronic aortic dissections - a single center experience in 28 patients. J Vasc Surg. 2008;47:724–732
  37. Roselli EE, Greenberg RK, Pfaff K, Francis C, Svensson L, Lytle BW. Endovascular treatment of thoracoabdominal aortic aneurysms. J Thorac Cardiovasc Surg. 2007;133:1474–1482
  38. Chuter TA, Rapp JH, Hiramoto JS, Schneider DB, Howell B, Reilly LM. Endovascular treatment of thoracoabdominal aortic aneurysms. J Vasc Surg. 2008;47:6–16
  39. Greenberg RK, Lu Q, Roselli EE, Svensson LG, Moon MC, Hernandez AV, et al. Contemporary analysis of descending thoracic and thoracoabdominal aneurysm repair: a comparison of endovascular and open techniques. Circulation. 2008;118:808–817
  40. Lepore V, Lönn L, Delle M, Bugge M, Jeppsson A, Kjellman U, et al. Endograft therapy for diseases of the descending thoracic aorta: results in 43 high-risk patients. J Endovasc Ther. 2002;9:829–837
  41. Rodriguez JA, Olsen DM, Shtutman A, Lucas LA, Wheatley G, Alpern J, et al. Application of endograft to treat thoracic aortic pathologies: a single center experience. J Vasc Surg. 2007;46:413–420
  42. Thompson M, Ivaz S, Cheshire N, Fattori R, Rousseau H, Heijmen R, et al. Early results of endovascular treatment of the thoracic aorta using the Valiant endograft. Cardiovasc Intervent Radiol. 2007;30:1130–1138
  43. Toumpoulis IK, Anagnostopoulos CE, Drossos GE, Malamou-Mitsi VD, Pappa LS, Katritsis DG. Early ischemic preconditioning without hypotension prevents spinal cord injury caused by descending thoracic aortic occlusion. J Thorac Cardiovasc Surg. 2003;125:1030–1036

 Competition of interest: none.

PII: S0741-5214(08)02000-4

doi:10.1016/j.jvs.2008.10.074

Journal of Vascular Surgery
Volume 49, Issue 5 , Pages 1117-1124, May 2009

Article Tools

* Image Usage & Resolution