Major vascular insults have been associated with both transperitoneal and retroperitoneal approaches to the intervertebral disc space for anterior lumbar interbody fusion (ALIF), particularly at the lower lumbar levels (L4-L5), because the vascular anatomy must be retracted across the disc space for adequate exposure.1–15 Previous literature suggested that the most common vascular injuries have resulted from tears and lacerations of such vessels as the aorta, iliac, and vena cava, and have ranged up to 18.4% with a large variation among those studies (0.08% to 18.4%).1,5 Typically, these complications permit immediate recognition and repair. In contrast, ischemic insult due to surgical retraction of these vessels or development of arterial thrombosis is not easily detected intraoperatively unless some type of continuous intraoperative monitoring is performed.4,6,7,10–15 Various modalities have been used, including palpation of leg vessels,6,13 pulse oximetry,13–15 and spinal cord monitoring.14,15
Although rare, the first case of thrombotic occlusion associated with an ALIF was reported by Marsicano et al4 in 1994. Since that original report, other publications have cited similar complications.6,7,10–13,15 Alarmingly in the majority of these cases, this type of insult went undetected intraoperatively because of an absence of adequate intraoperative monitoring, and even resulted in a lethal outcome in a few cases.4,6,7,10,11,13 Furthermore, several of these case reports shared the common postoperative complications of pain, accompanied by sensory and motor deficits, which were mistakenly interpreted as nerve damage, and not as a result of vascular in-sufficiency.4,6,7,10,13 It is now recognized that numerous factors predispose a patient to arterial thrombosis during an ALIF, including traumatic dissection from previous abdominal surgeries, obesity, prolonged operative time, hypotensive anesthesia, a history of smoking, a history of thrombosis, age, a low vascular bifurcation in front of the L5-S1 levels, and anatomic anomalies in vascular tortuosity.4,7,9,11,12
In this case report, intraoperative neuromonitoring using SSEPs, electromyography (EMG), and pedicle screw stimulation were used during an elective circumferential fusion. During closure of the ALIF, a complete loss of the left leg's SSEPs with no palpable pulses in that extremity indicated the development of a progressive thrombotic occlusion of the left iliac artery, attributable to retraction. This resulted in an emergent vascular intervention involving thrombectomies of the internal and external iliac arteries with patch angioplasty of the left common and proximal external iliac arteries. Postoperatively, the patient's neurologic insult resolved completely.
MATERIALS AND METHODS
The patient was a 46-year-old man with a preoperative diagnosis of degenerative disc disease involving herniated discs at L4-L5 and L5-S1 with accompanying spinal stenosis. Having failed nonoperative measures, the patient elected to proceed with a circumferential fusion consisting of an ALIF and posterolateral fusion with pedicle screw instrumentation.
Surgical exposure involved a “mini-open” technique using a left, para rectus, retroperitoneal approach via an incision of approximately 10 cm.9 The ALIF consisted of disectomies at the L4-L5 and L5-S1 levels, partial corpectomies at the L4-L5 levels, insertion of 15 mm bone dowels at the L4-L5 and L5-S1 levels, and fixation using an anterior tension band plate extending from the L4 to S1 levels (Femoral Ring Allograft System, SYNTHES Spine, Paoli, PA). Surgical procedures at the L5-S1 levels involved ligation of the middle sacral artery and vein with no vessel retraction. However, at the L4-L5 levels, the left ilio-lumbar vein was sacrificed, and the left common iliac artery and vein were retracted medially with Kitner retractors. There was intermittent release of the retractors during the 35-minute period required for disc space preparation, partial corpectomies, and insertion of the bone dowel.
Intraoperative neuromonitoring was performed to assess peripheral and central nervous system functions using 3 modalities: (1) posterior tibial nerve SSEPs (PTN SSEPs), (2) acoustic, free-run EMG recorded from lower extremity muscle groups innervated by the L4, L5, and S1 nerve roots (quadriceps femoris, tibialis anterior, and medial gastrocneumius), and (3) pedicle screw stimulation. Using the Cadwell Cascade (Kennewick, WA), interleaving transcutaneous stimulation of the PTN at the medial malleolus of each leg was delivered using standard stimulating procedures (stimulus rate: 4.13 Hz, stimulus intensity: 2× the motor threshold, stimulus duration: 0.3 ms). For each leg, recording electrodes were placed for 3 channels: (1) at the popliteal fossa for the peripheral site, (2) Fz was referred to C2S for the subcortical or brainstem waveforms, and (3) Cz′ was referred to Fz for the cortical waveforms (Fig. 1). Using aggressive filtering techniques (30 to 200 Hz) to eliminate excessive high-frequency artifact in order to obtain the most reliable and easiest interpretable waveforms with the least averaging, postinduction baseline waveforms were established (dashed lines) with active updates (solid lines) superimposed for comparison (Figs. 1A–C). Peak-to-trough amplitudes (μV) and absolute latencies (ms) of the dominant components of each waveform were monitored continuously, and documented before and after various routine and critical anesthetic and surgical events. For each trace in Figure 1, the absolute latency is shown above or below the dominant waveform component. Peak-to-trough amplitudes are shown within the parentheses at the top of the dominant components of the subcortical and cortical waveforms. In addition, vital signs, anesthetic levels, and any alerts based on persistent changes from baseline waveforms were documented on appropriate hardcopy “snapshots” of the waveforms.
During the ALIF, there were no transient or persistent changes in the waveform morphologies of the PTN SSEPs, and an absence of any mechanically elicited EMG discharges from individual nerve roots at the spinal levels involving the disectomies, partial corpectomies, distraction and insertion of the bone dowels, and instrumentation. However, during closure of the anterior incision, a gradual decline and eventual loss of the peripheral, subcortical and cortical waveforms to left PTN stimulation developed over a period of approximately 10 minutes with an absence of neurotonic EMG discharges (Fig. 1B, see arrows). This prompted an examination for palpable and Doppler pulses at the dorsalis pedis and posterior tibial locations, and femoral artery of the left leg that revealed an absence of pulses. In contrast, examination of the right leg revealed baseline-equivalent SSEP waveforms with all pulses intact (Fig. 1B). However, it should be noted that this patient was not monitored using great toe pulse oximetry, which may have provided earlier detection of an impending ischemic insult.14
As a result of the loss of the SSEP waveforms at all neural levels combined with the absence of pulses in the left leg, an emergent vascular consult was obtained. The elapsed time from the loss of the SSEP waveforms during the anterior closure to the incision for reexploration was approximately 45 minutes. Initially, the abdominal incision was reopened and the retroperitoneum was exposed. There was a palpable pulse in the left common iliac, but no palpable pulses in the internal and external iliac arteries. After heparinzation, clamps and vessels loops were placed about the 3 iliac vessels. A longitudinal arteriotomy of the proximal external iliac artery, which was extended distally into the common iliac artery to appreciate the orifice of the internal iliac, revealed thrombus throughout the area. After several passes of sizes 3 and 4 Fogarty catheters, thrombus was retrieved from the internal and external iliac arteries until excellent backflow bleeding was established. Saphenous vein was then harvested for a patch angioplasty of the left common and external iliac arteries. Postoperative, coagulopathic results that may have predisposed this patient to such an insult were negative. However, atherosclerotic plaque or vasospasm that may produce clot after retraction of the left iliac artery cannot be eliminated.2
With satisfactory perfusion restored about 2 hours after the initial degradation of the SSEP waveforms recorded after left PTN stimulation, complete recovery of the SSEP waveforms at each recording site occurred over a period of about 30 minutes (Fig. 1C). Despite the successful vascular intervention and complete recovery of the SSEP waveforms to previously established baselines, the planned posterolateral fusion with pedicle screw instrumentation was abandoned. A wake-up test was performed with the patient responding appropriately. Postoperatively, the patient presented with decreased sensation of the left thigh that resolved by the time the patient was discharged from the hospital. Subsequently during a routine office visit, the patient presented with a lack of radiculopathy and back pain, thus additional surgery involving a posterolateral fusion was not rescheduled.
On approach to the L4-L5 and L5-S1 levels during an ALIF, transient ischemic insults have commonly resulted from partial or complete occlusion of the left iliac artery due to retraction for adequate surgical exposure.14,15 More serious ischemic complications have involved the development of thrombosis within this vessel and its branches, also typically attributed to forceful or prolonged retraction.4,6,7,10–13 One simple-minded modality suggested for detecting thrombotic insults relied on intermittent palpation of lower extremity vessels.6,13 An additional inexpensive modality advocated for continuous vascular monitoring used great toe pulse oximetry.13–15 For example, as a clinical “red flag,” Kulkarni et al13 proposed an alarm criteria of desaturation to a value < 80%. Although these monitoring techniques are relatively inexpensive and noninvasive, they do not provide adequate assessment of sensory and motor functions of the leg after such an insult, nor for evaluating recovery of neural function if a vascular complication requires intraoperative intervention. However, because the time course of the thrombosis is progressive and may be delayed for 36 hours postoperatively,7 intermittent palpation of leg pulses or pulse oximetry would be beneficial to patients during the acute recovery period.
In a recent report designed to use neuromonitoring, Brau et al15combined pulse oximetry of the great toe and cortical SSEPs generated by stimulation of the appropriate sensory dermatome or PTN at the ankle for surgical monitoring during ALIFs from levels L2-L3 to L5-S1. Using an alarm criteria of desaturation below 90% in combination with “any” increase in latency and decrease in the amplitude of the cortical SSEPs, they reported that 57% of the patients were subjected to some degree of transient ischemia due to compression of the left iliac arteries. The time course of changes fell into 3 categories: (1) on average 3 minutes after desaturation, the morphology of SSEPs typically degraded from baseline waveforms, (2) if saturation dropped to 0, complete loss of the cortical SSEPs occurred on average after 6 minutes, (3) however, an abrupt and complete loss of SSEPs without a gradual change also occurred. On average, recovery of the SSEP waveforms occurred within 8 minutes after saturation returned to preevent baseline values. They further suggested that if the cortical SSEPs do not return to baseline within 15 to 20 minutes after release of retraction, then arterial thrombosis should be suspected. In conclusion, these authors contended that should desaturation occur, the fusion can be completed safely within 1 hour. Using a variation in this neuromonitoring protocol, Sato et al14 recorded both spinal cord and cortical evoked potentials combined with great toe pulse oximetry, and concluded that pulse oximetry demonstrated a greater sensitivity for detecting iliac arterial compromise than SSEPs (ie, complete desaturation occurred before there was any “significant” degradation of SSEPs waveforms). However, this preliminary report lacked any characterization of the time course of degradation and recovery of the variables that were monitored.
Although Brau et al15 reported a remarkable 100% correlation between great toe desaturation and degradation of cortical SSEPs, as well as in their recovery, there are several limitations related to the neuromonitoring protocol described above. Firstly, making clinical interventions based on changes in single-channel (cortical), SSEP neuromonitoring and failure to apply an established alarm criteria to the degradation of waveform morphology may lead to an unacceptable incidence of false positives, unlike the present case report.16–23 It is well known that cortical waveforms are most susceptible to nonsurgical factors (eg, anesthesia, age, blood pressure, and preexisting neurologic disorders), and are typically less reliable than spinal cord or subcortical evoked potentials for intraoperative neuromonitoring.16–23 Neuromonitoring protocols using mixed-nerve SSEPs during ALIF surgery should rely heavily on the replication of waveforms recorded at multiple sites that are both caudal and cephalic to the operative site, with special attention paid to the replication of all these waveforms (eg, in our case, the peripheral, subcortical, and cortical levels).16,18,21,23
Secondly, the interventional or warning criteria applied to the intraoperative degradation of SSEP waveforms should be based on established criteria such as the “50/10” rule or its variations (ie, the amplitude of the dominant waveform component decreases by >50% and/or a >10% prolongation in latency), and not any change in the replication of waveforms recorded at multiple levels.20–24 When Sato and colleagues14 applied these established criteria to changes in both spinal cord and cortical evoked potentials of the left leg, the incidence of vascular compromise was approximately half (26%) that reported by Brau et al,15 suggesting that there was a high incidence of false positives. However, it is generally agreed that if the waveforms present with well-formed morphologies which replicate consistently after relatively few stimulations, a persistent amplitude reduction and/or latency prolongation should merit an intraoperative alert to the surgeon(s), particularly if it is associated with a critical surgical maneuver like retraction of the iliac vessels.21
Timely detection of an ischemic insult due to retraction and thrombosis of the left iliac arteries during anterior, lumbo-sacral, reconstructive surgery is possible with adequate intraoperative monitoring. Surgical monitoring using multimodality neuromonitoring and pulse oximetry in combination with identifying patient risk factors (eg, peripheral vascular disease, the elderly, previous abdominal surgery, obesity, etc.) along with intraoperative precautions (eg, avoiding prolonged or forceful retraction of the iliac arteries) can lead to a lower prevalence of significant vascular complications. Our study represented an incidence of using spinal cord monitoring at multiple levels to optimize the detection of a progressive and clinically significant thrombotic insult to the left iliac arteries during an ALIF. Such neuromonitoring protocols will not only aid in timely recognition of an ischemic insult, but also provide an assessment of neural recovery if surgical salvage is required, as it did in our case.
Furthermore, because one of the primary surgical objectives during an ALIF is to decompress the offending herniated disc and subsequently restore the height of the disc space by using the tallest possible implant to maximize stability, surgical irritation and distraction of the neural elements are also potential iatrogenic insults. Thus, neurologic injury may potentially occur in the absence of vascular insults that can be easily detected by great toe pulse oximetry. To differentiate iatrogenic neurologic injury from vascular insults, multimodality neuromonitoring protocols which incorporate sensory and motor evoked potentials, and EMG of individual nerve roots may afford such determinations and guide subsequent intervention(s) different from that caused by vascular insults, as well as providing an assessment of recovery of neural function.21–23,25
This case report represented a relatively rare insult to neural function during an ALIF due to the development of thrombosis associated with retraction of the left iliac artery. Preoperative identification of risk factors that may predispose patients to thrombotic insult cannot be overemphasized. Inexpensive techniques like palpation of vessels and great toe pulse oximetry do not provide optimal surgical monitoring. Multilevel, multimodality, intraoperative neuromonitoring can provide a timely warning of both ischemic and direct neurologic insults, as well as an assessment of neural function if vascular intervention is required.
1. Westfall SH, Akbarnis BA, Merenda JT, et al. Exposure of the anterior spine. Technique, complications, and results in 85 patients. Am J Surg. 1987;154:700–704.
2. Watkins R. Anterior lumbar interbody fusion
surgical complications. Clin Orthrop. 1992;284:47–53.
3. Baker JK, Reardon PR, Reardon MJ, et al. Vascular injury in anterior lumbar surgery. Spine. 1993;18:2227–2230.
4. Marsicano J, Mirovsky Y, Remer S, et al. Thromboti occlusion of the left common iliac artery after an anterior retroperitoneal approach to the lumbar spine. Spine. 1994;19:357–359.
5. Faciszewski T, Winter RB, Lonstein JE, et al. The surgical and medical preoperative complications of anterior spinal fusion surgery in the thoracic and lumbar spine. Spine. 1995;20:1592–1599.
6. Raskas DS, Delamarter RB. Occlusion of the left iliac artery after retroperitoneal exposure of the spine. Clin Orthop Related Res. 1997;338:86–89.
7. Khazim R, Boos N, Webb JK. Progressive thrombotic occlusion of the left common iliac artery after anterior lumbar interbody fusion
. Eur Spine J. 1998;7:239–241.
8. Kuslich SD, Ulstrom CL, Griffith SL, et al. The bagby and kuslich method of lumbar interbody fusion. History, techniques, and 2-year follow-up results of a United States prospective, multicenter trial. Spine. 1998;23:1267–1279.
9. Mayer MH. The alif concept. Eur Spine J. 2000;9(suppl 1):S35–S43.
10. Hackenberg L, Liljenqvist U, Halm H, et al. Occlusion of the left common iliac artery and consecutive thromboembolism of the left popliteal artery following anterior lumbar interbody fusion
. J Spinal Disord. 2001;14:365–368.
11. Oskoulan RJ Jr, Johnson JP. Vascular complications in anterior thoracolumbar spinal reconstruction. J Neurosurg (Spine). 2002;96(suppl 1):1–5.
12. Chang Y-S, Guyer RD, Ohnmeiss DD, et al. Case report. Intraoperative left common iliac occlusion in a scheduled 360-degree spinal fusion. Spine. 2003;28:E316–E319.
13. Kulkarni SS, Lowery GL, Ross RE, et al. Aterial complications following anterior lumbar interbody fusion
. Report of eight cases. Eur Spine J. 2003;12:48–54.
14. Sato KS, Slepian D, Enoch D. Lower extremity pulse oximetry for anterior lumbar procedures. Proceedings of the American Society of Neurophysiological Monitoring, Las Vegas, NE, 2003.
15. Brau SA, Spoonamore MJ, Snyder L, et al. Nerve monitoring changes related to iliac artery compression during anterior lumbar spine surgery. Spine J. 2003;3:351–355.
16. Lüeders H, Gurd A, Hahn J, et al. A new technique for intraoperative monitoring of spinal cord function. Multichannel recording of spinal cord and subcortical evoked potentials. Spine. 1982;7:110–115.
17. Lubicky JP, Spadoro JA, Yuan HA, et al. Variability of somatosensory evoked potential monitoring during spine surgery. Spine. 1989;14:790–798.
18. Abel MF, Scott SJ, Wenger DR, et al. Brainstem evoked potentials for scoliosis surgery. A reliable method allowing use of halogenated anesthetic agents. J Ped Orthop. 1990;10:208–213.
19. York DH. A critical evaluation of the 50% criterion in sep monitoring. Proceedings of the American Society of Neurophysiological Monitoring, St Louis, Missouri, 1995.
20. Nuwer MR. Spinal cord monitoring with somatosensory techniques. J Clin Neurophysiol. 1998;15:183–193.
21. Schwartz DM, Drummond DS, Schwartz JA, et al. Neurophysiological monitoring during scoliosis surgery. A multimodality approach. Sem Spine Surg. 1997;9:97–111.
22. Owen JH. The application of intraoperative monitoring during surgery for spinal deformity. Spine. 1999;2:649–662.
23. Toliekis JR. Intraoperative monitoring using somatosensory evoked potentials
. A position statement by the American society of neurophysiological monitoring. J Clin Mon Comp. 2005;19:241–258.
24. Wagner W, Peghini-Halbig L, Maürer JC, et al. Intraoperative sep monitoring in neurosurgery around the brain stem and cervical spinal cord. Differential recordings of subcortical components. J Neurosurg. 1994;81:213–220.
25. Isley MR, Balzer JR, Pearlman RC, et al. Intraoperative motor evoked potentials. Am J END Technol. 2001;41:266–338.