Journal Logo

Original Clinical Science—General

Spinal Cord Ischemia in Pancreas Transplantation: The UK Experience

Phillips, Benedict L. MRCS1; Papadakis, Georgios MRCS2; Bell, Rachel MS, FRCS3; Sinha, Sanjay MS, FRCS4; Callaghan, Chris J. PhD, FRCS1; Akyol, Murat MD, FRCS2; Watson, Christopher J.E. MD, FRCS5; Drage, Martin PhD, FRCS1

Author Information
doi: 10.1097/TP.0000000000003028



Pancreas transplantation is a surgical treatment option for diabetes mellitus and is the most effective method of establishing euglycemia, with the potential to reverse or ameliorate some of the secondary complications of diabetes.1 Selected patients with renal failure may benefit from simultaneous pancreas and kidney (SPK) transplantation, or may undergo pancreas after kidney transplantation. While there are complications associated with pancreas transplantation, beyond the first year of transplantation there are survival and quality of life benefits.2,3

Spinal cord ischemia (SCI) is a rare but devastating condition, which can result in loss of lower limb function and impaired continence. Twenty eight percent of patients undergoing open thoraco-abdominal aortic aneurysm repair develop SCI.4 However, SCI is not widely acknowledged as a risk in transplantation, and we were able to find just 1 previous case in the literature.5

The anterior two-thirds of the spinal cord receive its blood supply from the anterior spinal artery, which arises from the vertebral artery and courses down the anterior aspect of the spinal cord (Figure 1). The anterior spinal cord receives further contributions from radicular arteries, which arise directly from the aorta and pass through the intervertebral foramina before forming anastomoses with the anterior spinal artery. The artery of Adamkiewicz is the most dominant radicular artery and typically supplies the anterior spinal cord between T9 and T12,6 although some anatomical variability exists.7 The thoracic spinal cord is a relative watershed area and is therefore dependent on augmented blood flow from these radicular arteries. The pathophysiology of SCI is likely to be multifactorial but is associated with systemic hypotension during and after surgery.8

Arterial supply of the spinal cord. The anterior spinal artery (ASA) arises from the vertebral artery and courses down the anterior aspect of the spinal cord. Radicular arteries arise from the aorta, pass through the intervertebral foramina, and form anastomoses with the ASA. The artery of Adamkiewicz is the most dominant radicular artery and supplies anterior spinal cord between T9 and 12, although anatomical variability exists.

Following a case of SCI after pancreas transplantation at Guy’s Hospital, we sought to determine the incidence of this unexpected complication. In this article, we report the United Kingdom experience of SCI in pancreatic transplantation (SCIIPT). The aim of this report is to quantify the risk of SCIIPT, to better inform patients during the consent process, and to raise awareness among transplant surgeons as to the possible etiology and treatment options.


Data Collection

The clinical lead in all 8 UK pancreas transplant units was contacted between May 2017 and May 2018 to identify cases of SCIIPT. Centers with cases were sent a structured questionnaire to collect detailed clinical and outcome data (supplementary digital content Table S1, SDC, Hypotension was defined as a systolic blood pressure of ≤ 90 mm Hg. The reporting period represented pancreas transplantation between January 2002 and December 2018 in the United Kingdom.

Data Analysis

As the incidence of SCIIPT is low, no comparative statistical analyses were applied. Descriptive statistics regarding demographic, operative, and patient outcomes were performed. The risk of SCIIPT was estimated using the number of radiologically-confirmed cases relative to the number of pancreatic transplants from UK registry data during the same time period.9


Survey Responses

All 8 pancreas transplant centers in the United Kingdom responded to the survey, with 4 centers reporting cases of SCIIPT. All 4 centers reporting SCIIPT completed the detailed structured questionnaire.

Risk Analysis and Baseline Demographics

Since 2002, there have been 6 cases of SCI following pancreatic transplantation (PT) in the United Kingdom (Table 1). According to the UK transplant registry, during the study period, 2633 pancreas transplants have been performed.9 Based on these figures, there is approximately a 1 in 440 risk of SCIIPT.

Summary of characteristics and findings in each patient with spinal cord ischemia in pancreas transplantation

Three cases of SCIIPT were reported from Cambridge and 1 case each from Oxford, Edinburgh, and London. SCI affected pancreas transplant recipients with a median (interquartile range) age of 41 (24–49) years, with 27 (21–42) years of exposure to type I diabetes mellitus. SCI occurred following SPK transplantation in 4 cases. One case occurred following pancreas after kidney transplantation and the final case following pancreas transplantation alone. Before transplantation, 5 recipients had established end-stage renal disease and were on renal replacement therapy.

Preoperative Assessment and Management

All patients had routine thrombophilia screening before transplantation. All thrombophilia screens were negative, aside from 1 patient. With regard to the 1 patient with an abnormal thrombophilia screen, dilute activated partial thromboplastin time analysis indicated transient lupus anticoagulant positivity, which was not demonstrated 26 days later when repeated. In view of the lack of thromboembolic events in the past, a multidisciplinary team deemed this patient suitable for SPK transplantation with routine thromboembolism prophylaxis. In addition to a thrombophilia screen, 4 recipients had preoperative thromboelastography testing, of which 1 recipient had high clot propagation (α-angle) and high maximum amplitude (clot strength). Given that this patient had a normal thrombophilia screen and no previous thromboembolic events in the past, routine prophylactic-dose enoxaparin 40 mg daily was planned. None of the 6 patients had known autonomic neuropathy. One patient had mild diabetic peripheral neuropathy. The risk of SCI was discussed and documented in 1 recipient, after SCI had occurred on 2 previous occasions at that center.

Operative Findings

During induction of anesthesia, central venous catheters were inserted in all patients. Continuous blood pressure monitoring was implemented, through insertion of an arterial cannula, in 5 patients.

Intraoperatively, none of the 6 cases of SCI had clamping of the aorta and no arterial branches were tied. Operative strategies to reduce the risk of SCI, such as epidural cooling, were not implemented in any case. Spinal cord neurophysiology was not recorded in any patient; therefore, none of the cases of SCI were identified intraoperatively.

During or after surgery, all patients experienced episodes of hypotension before the onset of neurological symptoms. Risk factors for hypotension were examined in detail, including hemorrhage, reoperation, epidural anesthesia, and the use of peripheral vasodilators such as epoprostenol.

Two recipients had significant intraoperative blood loss (500–2000 mL) and a third patient returned to theatre due to postoperative hemorrhage from the arterial anastomosis of the renal allograft. Before the onset of neurological symptoms, 3 patients required surgical reexploration (due to postoperative hemorrhage, suspected graft thrombosis, or suspected bowel leak). Intraoperatively, 5 patients experienced hypotension ranging from 10 minutes to 1 hour, either during organ implantation or subsequent surgical reexploration. Three patients had epidural anesthesia and 2 patients received an IV infusion of epoprostenol during and after pancreatic graft reperfusion.

Postoperative Findings

Postoperative hypotension occurred in 3 recipients, of which 2 had already experienced intraoperative hypotension. Two patients required renal replacement therapy in the early postoperative period for delayed kidney transplant function. Due to difficulties in gaining urgent central venous access, 1 patient underwent hemodialysis via an arteriovenous fistula, rather than hemofiltration. This may have further contributed to postoperative hypotension in this patient.

Before the onset of SCI, 5 of the 6 transplant recipients were given pharmacological venous thromboembolism prophylaxis in the form of unfractionated or low molecular weight heparin.

The onset of SCI, as defined by commencement of neurological symptoms, occurred within 48 hours of transplantation in 5 patients. Due to the timing of the patients’ complaints, it is likely that SCI occurred during the primary operation in 3 patients, during re-exploration in 1 patient, and postoperatively in intensive care in 1 patient. It was not possible to determine whether SCI occurred during the intraoperative or early postoperative period in 1 patient. The diagnosis of SCI was first documented on the first postoperative day in 3 patients and on the second postoperative day in 2 patients (missing notes in 1 patient). Loss of lower limb sensation was the primary symptom expressed by all patients. Epidural anesthesia appeared to confound the diagnosis of SCI in 1 case. Urgent MRI was arranged to confirm the diagnosis of SCI in all cases (an example of which is shown in Figure 2). MRI was reportedly normal in 1 case, but showed SCI affecting between T1 and T9 in 4 cases; the MR report was not available for 1 case. There was no evidence of epidural hematoma in the 3 patients who had epidural anesthesia. Cerebrospinal fluid drainage was implemented in 1 recipient after the diagnosis of SCI was made. The mainstay of early treatment for SCI for all cases was BP control and subsequent neurorehabilitation.

T2-weighted sagittal magnetic resonance imaging of the thoracic spine. An abnormal T2 signal is seen within the spinal cord between T1 (denoted by the presence of the first rib) and T5, suggestive of spinal cord ischemia (arrow). The spinal cord may also appear expanded, due to edema, in the acute phase (not demonstrated here).

All patients suffered significant neurological defects following SCI. At last follow-up, 1 patient has died from complications of bedsores, 3 patients are dependent on a wheelchair, and 1 patient requires a walking aid. Only 1 patient regained the ability to walk without aids following neurorehabilitation. Two patients continue to have chronic urinary and fecal incontinence. At last follow-up (February 2019), 1 patient had returned to work and 2 patients have died (11 ys and 6 ys posttransplantation).


There is approximately a 1 in 440 risk of SCI in pancreas transplantation. All 4 transplant centers in this report now discuss the risk of SCI with potential PT recipients (personal communications). Given the incidence and the potential impact on patients’ decision to proceed with PT, we believe all patients undergoing assessment should be informed of the risk of SCI. The potential risk factors for SCIIPT are summarized in Table 2.

Summary of potential risk factors for spinal cord ischemia following pancreas transplantation

The UK saw a significant shift in medicolegal practice in 2015 in which the landmark case ‘Montgomery v Lanarkshire’10 demonstrated that doctors have a duty to inform patients of risks that a patient may find significant. This marked a change from the previous “Bolam test,” in which doctors must act in accordance with a responsible body of medical opinion.11 Patients not only have a right to be informed of significant complications for their consent to be fully informed, but without adequate consent clinicians are at risk of litigation. The transplant community must therefore evolve its practice to better inform patients and to avoid medicolegal conflict. A nationwide consent process exists in the United Kingdom and is currently being revised to allow a more unified approach discussing risk with transplant recipients.

Hypotension during and after transplantation was probably the main contributing factor to the pathogenesis of SCIIPT. The thoracic spinal cord is a relative watershed area, which is susceptible to infarction in systemic hypotension. Diabetes mellitus, uremia, and long-term hemodialysis are independently associated with accelerated atherosclerosis,12,13 which possibly further increase the risk of SCI during relative hypotension. We suggest that a target blood pressure should be set during and after pancreas transplantation (eg, mean arterial pressure 80 mm Hg), with a clear plan if targets are not met. Continuous blood pressure monitoring, via arterial line, enables early recognition of hypotension, and provides opportunities to reverse hypoperfusion promptly. We recommend central venous access in all pancreas transplant patients as this provides fluid status monitoring, rapid fluid resuscitation, and access for hemodialysis in the event of delayed kidney transplant function in SPK transplant recipients. Euvolemia and an adequate starting blood pressure are necessary to withstand a transient drop in blood pressure that occurs immediately after reperfusion.14 Communication between the surgical and anesthetic leads is particularly important during this critical moment of the operation to ensure stable hemodynamic control.

Epoprostenol can be used as an antiplatelet agent to reduce the risk of graft thrombosis following PT but may also have contributed to systemic hypotension through its vasodilator effect. The use of prostaglandins in pancreas transplantation should therefore be weighed against the risk of vasodilation and consequent systemic hypotension. Nonvasodilatory agents are preferable for anticoagulation in pancreas transplantation.

Epidural anesthesia may also lower systemic blood pressure, further contributing to reduced blood flow to the spinal cord. Epidural catheterization may also increase pressure in the epidural space, reducing perfusion pressure in the spinal arteries, and is independently associated with risks of epidural hematoma formation. In addition, anesthesia of the lower limbs may mask the presentation of SCI and delay intervention as it did so in 1 case described in this report. The incidence of SCIIPT when these 2 interventions, epoprostenol and epidural anesthesia, were used together, was over 1%. Transversus abdominus plane or rectus sheath block are alternative forms of regional anesthesia in pancreas transplantation, which are not associated with hypotension.15,16 The authors would not recommend routine use of epidural anesthesia in pancreas transplantation.

Following SPK transplantation, between 13% and 35% of kidney transplants will have delayed graft function.17 In the immediate posttransplant period, when avoidance of hypotension is important, hemofiltration may be safer than hemodialysis in diabetic patients with hypotension.

Although hypotension has been identified as a probable risk factor for SCIIPT, the incidence of hypotension during pancreas transplantation in patients who have not developed SCIIPT has not been defined in the literature. Hypotension is often encountered during pancreas transplantation, related to bleeding,18 autonomic dysregulation secondary to diabetes, and postreperfusion syndrome.19 It is not clear what ultimately led to spinal cord infarction in these 6 patients. Significant bleeding in the intraoperative or postoperative period did not occur in 3 of the 6 patients. Whether the exact mechanism can be defined or not, SCIIPT appears to be a significant risk to patients. The apparent risk of SCIIPT is comparable to the risk of bile duct injury in laparoscopic cholecystectomy20—a complication routinely described to patients during informed consent for that procedure.

There is significant risk of thrombosis during PT, because diabetic patients are hypercoagulable, coupled with the transient hypercoagulable state of acute surgical stress.21 There is currently no unanimous approach to postoperative anticoagulation during PT in the UK. The current approach among the majority of the authors is that under certain circumstances, such as concerns regarding vascular anastomoses, blood flow through the transplant arteries or veins, or the macroscopic appearance of the organs, intraoperative intravenous unfractionated heparin (eg, 5000 IU bolus) may be used. The authors do not routinely give intraoperative anticoagulation. Thromboelastography-directed anticoagulation has been reported to be associated with a reduced incidence of thrombosis in SPK transplantation22 and may provide a more patient-tailored approach to monitoring perioperative23 and postoperative coagulation status.22

Aortic surgeons have adapted their practice to reduce the risk of SCI in their patients. Routine use of lumbar drains, with target cerebrospinal fluid pressures of 10–15 mm Hg, has reduced the rate of SCI in thoracoabdominal vascular surgery.24-28 Epidural cooling and spinal cord perfusion augmentation aim to maintain adequate spinal cord perfusion, while reducing the metabolic demand of the spinal cord.29-31 Permissive hypertension (target mean arterial pressure 80–100 mm Hg) has also been employed in aortic surgery to increase spinal cord perfusion.32 The transplant community should consider the risks and benefits of these strategies in pancreas transplantation to mitigate the risk of SCIIPT.

SCI may be suspected clinically in a patient complaining of lower limb motor or sensory deficit. However, real-time neurophysiology and spinal cord perfusion monitoring have the advantage of detecting SCI in the anesthetized patient.33-37 This may be pertinent in pancreas transplantation where the recipient may remain anesthetized in the early postoperative period in ICU, or if lower limb findings are being masked by epidural anesthesia. Early detection and treatment of SCI may rescue neurological function.32 A ‘spinal rescue protocol’ may provide clinicians with clear sequence of steps to restore spinal cord oxygenation and avoid long-term neurological deficit.32

There appears to be differences between the SCI in aortic surgery and pancreas transplantation. MRI findings in the pancreas cases suggested infarction of the spinal cord between T1 and T9, which does not represent the typical distribution of the Artery of Adamkiewicz (T9–T12), possibly suggesting that smaller radicular arteries are compromised during SCIIPT. Further research is necessary to understand the pathogenesis of SCIIPT, and whether interventions used in aortic surgery would confer neuroprotection in pancreas transplantation.

The limitations of this study are inherent in its design. Its retrospective nature and reliance on the accuracy of survey respondents meant that there was risk of recall and reporting bias. Correspondence with the UK pancreas transplant units was limited to the lead surgeon, which may have further introduced reporting bias. It is therefore possible that cases of SCIIPT were not reported and the incidence has been underreported in this study, although this is unlikely given the devastating nature of the complication. The survey design was semistructured. It is therefore possible that an unrecognized risk factor for SCIIPT was not considered in the survey’s questioning. An alternative method could have been unstructured interviews with all pancreas transplant surgeons in the UK, although logistically challenging. Furthermore, it was not possible to provide a meaningful control group. Although many pancreas transplant recipients have potential risk factors for SCI, not all of these patients develop SCI. However, the UK transplant registry does not record operative and postoperative observations such as blood pressure, or anesthetic strategies such as epidural anesthesia. We were therefore not able to use a comparator group to statistically interrogate the data for possible risk factors of SCIIPT, or determine which risk factors were potentially modifiable or nonmodifiable.

In conclusion, SCI is an underrecognized and underreported complication in pancreas transplantation, which is potentially avoidable with attention to perioperative and postoperative blood pressure. In this retrospective observational study, it was associated with systemic hypotension in all cases, and in no instance was the aorta clamped. We urge greater awareness among doctors and patients of this devastating condition.


UK pancreas transplant centers for their participation in identifying cases of spinal cord ischemia in pancreas transplantation. Dr Rohit Srinivasan, radiology registrar, for assistance with radiological interpretation. Sarah Cottee, transplant recipient coordinator, for collecting local clinical data.


1. Dholakia S, Mittal S, Quiroga I, et al. Pancreas transplantation: past, present, future. Am J Med. 2016; 129:667–673
2. Dean PG, Kukla A, Stegall MD, et al. Pancreas transplantation. BMJ. 2017; 357:j1321
3. Gross CR, Limwattananon C, Matthees BJ. Quality of life after pancreas transplantation: a review. Clin Transplant. 1998; 12:351–361
4. Messé SR, Bavaria JE, Mullen M, et al. Neurologic outcomes from high risk descending thoracic and thoracoabdominal aortic operations in the era of endovascular repair. Neurocrit Care. 2008; 9:344–351
5. Kiok MC. Neurologic complications of pancreas transplants. Neurol Clin. 1988; 6:367–376
6. Svensson LG, Klepp P, Hinder RA. Spinal cord anatomy of the baboon–comparison with man and implications for spinal cord blood flow during thoracic aortic cross-clamping. S Afr J Surg. 1986; 24:32–34
7. Etz CD, Kari FA, Mueller CS, et al. The collateral network concept: a reassessment of the anatomy of spinal cord perfusion. J Thorac Cardiovasc Surg. 2011; 141:1020–1028
8. Maniar HS, Sundt TM 3rd, Prasad SM, et al. Delayed paraplegia after thoracic and thoracoabdominal aneurysm repair: a continuing risk. Ann Thorac Surg. 2003; 75:113–9. Discussions 119
9. National Health Service Blood and Transplant. Pancreas transplantation annual reports. 2018. Available at Accessed July 1, 2019
10. Montgomery v Lanarkshire Health Board, [2015] SC 11 [2015] 1 AC 1430
11. Buttigieg GG. Re-visiting bolam and bolitho in the light of montgomery v lanarkshire health board. Med Leg J. 2018; 86:42–44
12. Zeadin MG, Petlura CI, Werstuck GH. Molecular mechanisms linking diabetes to the accelerated development of atherosclerosis. Can J Diabetes. 2013; 37:345–350
13. Vanholder R, Massy Z, Argiles A, et al.; European Uremic Toxin Work Group. Chronic kidney disease as cause of cardiovascular morbidity and mortality. Nephrol Dial Transplant. 2005; 20:1048–1056
14. Mittel AM, Wagener G. Anesthesia for kidney and pancreas transplantation. Anesthesiol Clin. 2017; 35:439–452
15. Hausken J, Rydenfelt K, Horneland R, et al. First experience with rectus sheath block for postoperative analgesia after pancreas transplant: a retrospective observational study. Transplant Proc. 2019; 51:479–484
16. Yeap YL, Fridell JA, Wu D, et al. Comparison of methods of providing analgesia after pancreas transplant: IV opioid analgesia versus transversus abdominis plane block with liposomal bupivacaine or continuous catheter infusion. Clin Transplant. 2019; 33:e13581
17. Kopp WH, Lam HD, Schaapherder AFM, et al. Pancreas transplantation with grafts from donors deceased after circulatory death: 5 years single-center experience. Transplantation. 2018; 102:333–339
18. Banga N, Hadjianastassiou VG, Mamode N, et al. Outcome of surgical complications following simultaneous pancreas-kidney transplantation. Nephrol Dial Transplant. 2012; 27:1658–1663
19. Dalal A. Intestinal transplantation: the anesthesia perspective. Transplant Rev (Orlando). 2016; 30:100–108
20. Bailey RW, Zucker KA, Flowers JL, et al. Laparoscopic cholecystectomy. Experience with 375 consecutive patients. Ann Surg. 1991; 214:531–40. Discussion 540
21. Muthusamy AS, Giangrande PL, Friend PJ. Pancreas allograft thrombosis. Transplantation. 2010; 90:705–707
22. Vaidya A, Muthusamy AS, Hadjianastassiou VG, et al. Simultaneous pancreas–kidney transplantation: to anticoagulate or not? Is that a question? Clin Transplant. 2007; 21:554–557
23. Burke GW 3rd, Ciancio G, Figueiro J, et al. Hypercoagulable state associated with kidney-pancreas transplantation. Thromboelastogram-directed anti-coagulation and implications for future therapy. Clin Transplant. 2004; 18:423–428
24. Coselli JS, LeMaire SA, Köksoy C, et al. Cerebrospinal fluid drainage reduces paraplegia after thoracoabdominal aortic aneurysm repair: results of a randomized clinical trial. J Vasc Surg. 2002; 35:631–639
25. Coselli JS, LeMaire SA, Schmittling ZC, et al. Cerebrospinal fluid drainage in thoracoabdominal aortic surgery. Semin Vasc Surg. 2000; 13:308–314
26. Khan NR, Smalley Z, Nesvick CL, et al. The use of lumbar drains in preventing spinal cord injury following thoracoabdominal aortic aneurysm repair: an updated systematic review and meta-analysis. J Neurosurg Spine. 2016; 25:383–393
27. Arora H, Ullery BW, Kumar PA, et al. Pro: patients at risk for spinal cord ischemia after thoracic endovascular aortic repairs should receive prophylactic cerebrospinal fluid drainage. J Cardiothorac Vasc Anesth. 2015; 29:1376–1380
28. Khan SN, Stansby G. Cerebrospinal fluid drainage for thoracic and thoracoabdominal aortic aneurysm surgery. Cochrane Database Syst Rev. 2012; 10:CD003635
29. Davison JK, Cambria RP, Vierra DJ, et al. Epidural cooling for regional spinal cord hypothermia during thoracoabdominal aneurysm repair. J Vasc Surg. 1994; 20:304–310
30. Ishikawa A, Mori A, Kabei N, et al. Epidural cooling minimizes spinal cord injury after aortic cross-clamping through induction of nitric oxide synthase. Anesthesiology. 2009; 111:818–825
31. Sinha AC, Cheung AT. Spinal cord protection and thoracic aortic surgery. Curr Opin Anaesthesiol. 2010; 23:95–102
32. Augoustides JG, Stone ME, Drenger B. Novel approaches to spinal cord protection during thoracoabdominal aortic interventions. Curr Opin Anaesthesiol. 2014; 27:98–105
33. Ertl M, Schierling W, Kasprzak P, et al. Optic nerve sheath diameter measurement to identify high-risk patients for spinal ischemia after endovascular thoracoabdominal aortic aneurysm repair. J Neuroimaging. 2015; 25:910–915
34. Keenan JE, Benrashid E, Kale E, et al. Neurophysiological intraoperative monitoring during aortic arch surgery. Semin Cardiothorac Vasc Anesth. 2016; 20:273–282
35. Luehr M, Mohr FW, Etz CD. Indirect neuromonitoring of the spinal cord by near-infrared spectroscopy of the paraspinous thoracic and lumbar muscles in aortic surgery. Thorac Cardiovasc Surg. 2016; 64:333–335
36. See RB, Awosika OO, Cambria RP, et al. Extended motor evoked potentials monitoring helps prevent delayed paraplegia after aortic surgery. Ann Neurol. 2016; 79:636–645
37. Etz CD, Di Luozzo G, Zoli S, et al. Direct spinal cord perfusion pressure monitoring in extensive distal aortic aneurysm repair. Ann Thorac Surg. 2009; 87:1764–73. Discussion 1773

Supplemental Digital Content

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.