Intraoperative Spinal Sonography in Thoracic and Lumbar Fractures: Evaluation of

Transcription

1 353 Intraoperative Spinal Sonography in Thorai and Lumbar Fratures: Evaluation of Harrington Rod Instrumentation Robert M. Quener' erta M. Montalvo' Frank J. Eismonf arth. Green 3 Thirty-seven patients with thorai and lumbar spine fratures were treated with Harrington rod instrumentation (HRI), and the progress and results of that surgery were monitored with intraoperative spinal sonography (loss). dequate neural tissue deompression and spinal olumn alignment was ahieved in less than one-half (1/31, 5%) of the patients in whom HRI was performed as the first step of the surgial proedure. s a result of these findings, further surgial maneuvers were performed whih, in most ases, resulted in adequate spinal realignment and neural tissue deompression. In six patients, diret surgial redution of displaed bone fragments was performed before HRI. Sine total deompression of neural tissue may be important in patients with spinal ord or auda equina injuries, it is reommended that loss be used in all ases of HRI for thorai and lumbar spine fratures. The need to perform additional surgial maneuvers to aomplish neural tissue deompression may be obviated if intraoperative sonography shows adequate deompression with HRI alone. This artile appears in the May/June 1985 issue of JNR and the ugust 1985 issue of JR. Reeived July 18, 198; aepted September 30, 198. Presented at the annual meeting of the merian Soiety of Neuroradiology, oston, June Department of Radiology, R 130, University of Miami Shool of Mediine, Jakson Memorial Medial Center, P.O. ox , Miami, FL ddress reprint requests to R. M. Quener. 2 Department of Orthopaedi Surgery and Rehabilitation, University of Miami Shool of Mediine, Jakson Memorial Medial Center, Miami, FL Department of Neurologial Surgery, University of Miami Shool of Mediine, Jakson Memorial Medial Center, Miami, FL JNR 6: , May/June /85/ merian Roentgen Ray Soiety Harrington rod instrumentation (HRI) and spinal fusion in the treatment of fratures and frature/disloations of the thorai and lumbar spine is performed in order to stabilize the spine and to eliminate bone impingement on adjaent neural strutures. Until reently, intraoperative assessment of the effets of Harrington rod distration on vertebral malalignment and ompression of the spinal anal and its ontents has been possible only with intraoperative myelography or plain radiographs taken during the surgial proedure. Diret visualization of the offending bone fragments is not possible without some resetion of the posterior elements, sine the fragments lie anterior to the theal sa. We have found, however, that with intraoperative spinal sonography (loss) the effiay of HRI an now be immediately and aurately assessed. If spinal deompression is shown to be inadequate and signifiant malalignment persists, further orretive surgial steps an then be taken. We report the results of loss in 37 patients with unstable thorai and lumbar fratures who underwent HRI. We demonstrate the importane of loss in monitoring this type of surgery and show how it aids in the surgial management of these patients. Subjets and Methods The diagnosis of a thorai or lumbar spine frature was established on the basis of plain radiographs in all 37 patients. Spine CT sans were obtained in 35 of the patients. When fratures ourred at more than one ontiguous level, we designated the level of frature as that level at whih the major malalignment and anal ompression ourred. ll patients were andidates for Harrington rod distration beause the fratures were onsidered unstable and/or beause of the presene of a bone fragment that had been retropulsed into the spinal anal. loss, with instrumentation and tehniques previously desribed [1], was used in all 37 ases. fter a laminotomy at the injury site, baseline loss at that level was performed before

2 35 QUENCER ET L. JNR:6, May/June 1985 TLE 1: Results of Harrington Rod Instrumentation Initial Harrington rod instrumentation, adequate deompression Initial Harrington rod instrumentation, inadequate deompression Subsequent proedures resulting in adequate deompression Inreased Harrington rod distration.. one fragment removal one fragment impation. one fragment impation and inreased Harrington rod distration Subsequent proedures resulting in inadequate deompression. one fragment inpation, but nothing further done beause: No. of Patients Patient neurologially intat. 2 Fragment ould not be redued. 1 Inreased Harrington distration, but nothing further done beause: Patient neurologially intat. ut nothing further done beause patient neurologially intat... 2 one fragments removed or impated before bilateral Harrington rod instrumentation Harrington rod instrumentation, adequate deompression 5 Harrington rod instrumentation, inadequate deompression Further bone fragment impation, adequate deompression HRI in 35 of the 37 ases. The size of the laminotomy needed to give a window for adequate sonography varied, but about 1.0 x 1.5 m was the minimal size required. eause the amount of bone thus removed is minimal and beause the assoiated ligaments have already been torn or injured, the laminotomy does not inrease spinal instability. In the two ases where baseline sonograms were not obtained, Harrington rods were already in plae when the sonographi equipment arrived in the operating room. Sonography was performed after eah additional surgial maneuver, whether that maneuver was Harrington rod instrumentation, bone fragment removal/impation, or inreased Harrington rod distration. t eah stage of surgery, the sonogram was evaluated for vertebral alignment and ompression of the spinal ord, nerve roots, or theal sa. The use of sonography did not signifiantly prolong these surgial proedures. Spinal deompression was onsidered adequate when there were no residual bone fragments or soft-tissue elements ausing deviation of the spinal ord or displaement of the nerve roots of the auda equina from their normal ourse. The visualization of CSF around the entire ord or auda equina and the restoration of the anal to a normal or near-normal onfiguration were additional signs indiating adequate deompression. Deisions on whether to attempt additional surgial maneuvers were based mainly on these sonographi riteria but also depended on the patients' preoperative neurologi status. Our ase material was divided into three major ategories: (1) patients in whom the original HRI adequately deompressed the anal, (2) patients in whom the original HRI failed to adequately deompress the anal, and (3) patients who had bone fragments removed or impated before HRI (table 1). dditional surgial proedures were performed in those patients whose anals had been inadequately deompressed by the Harrington rods, and the subsequent sonograms were analyzed in a manner TLE 2: Level and Inidene of Frature Level T T7 T10 T11 T12 L1 L2 L3 L Total No. of Patients similar to that desribed above. Using these data, we were then able to determine the perentage of ases in whih the routine appliation of Harrington rods resulted in either adequate or inadequate spine deompression and to show how loss influened further surgial management when the initial Harrington rod plaement inadequately deompressed the anal. Results The inidene of frature is shown in table 2, with fratures of the thoraolumbar juntion (T12 and L 1) onstituting over half of our ases. ll of the fratures were onsidered potentially unstable [2-6] and were assoiated with displaed bone fragments within the anal. Fifteen patients were neurologially intat or had minimal defiits at the time of surgery; 22 patients had signifiant neurologi defiits. Thirteen patients were operated on more than 8 hr after the injury. aseline loss, in all patients but one, showed vertebral malalignment and bone fragments ompressing the theal sa. That one patient had a T10 frature and was neurologially intat, and his baseline loss showed neither mal alignment nor abnormal bone fragments. Using the riteria desribed above, HRI resulted in adequate spine deompression in 1 patients (figs. 1 and 2). In 17 patients, however, inadequate deompression was observed on loss after initial Harrington rod insertion. In 15 of those 17 patients, additional surgial proedures or maneuvers were performed in hopes of reduing or removing the offending bone fragment. In 11 patients, the additional proedures resulted in adequate spinal deompression (figs. 3 and ); in four patients, those proedures did not totally deompress the anal (fig. 5). In two ases, no attempts were made to improve the suboptimally deompressed anal beause both patients were neurologially intat and it was deided to avoid any further surgial manipulations. In six patients bone fragments were either removed or impated before HRI. Five then showed adequate deompression, while one patient required additional bone impation (fig. 6) before deompression was onsidered adequate. Table 1 summarizes these data. Disussion Over the years, there have been different opinions on how best to manage patients with unstable thorai and lumbar fratures due to losed spinal injuries. These have ranged

3 JNR:6, May/June 1985 loss OF HRRINGTON ROD INSTRUMENTTION 355 Fig. 1.-dequate deompression of auda equina after HRI., Transverse intraoperative sonogram before HRI. Compression of auda equina, more marked on right, by displaed bone fragment (arrows ). Small amount of erebrospinal fluid (CSF) (urved arrow) is seen underneath dorsal dura., Longitudinal sonogram. Displaed bone fragment (straight arrow) from dorsal part of adjaent vertebral body (urved arrow). Transverse (C) and longitudinal (0) sonograms after bilateral HRI. dequate deompression of spinal anal, with CSF (arrows) seen around entire auda equina. In C and in following figures, typial sonographi appearane of Harrington rods (arrowheads) is seen. o from the onservative approah of postural redution [2, 7-9] to the use of Harrington rods and bony fusion [6, 10-12]. The primary goals in the treatment of these injured patients is to improve spinal alignment, establish spinal stability so that progressive kyphoti deformity does not our, and redue the bone fragments that have been displaed into the anal. With stability, the hanes of developing inreasing bak pain are lessened, while bone redution may improve the patient's neurologi status. The trend in reent years has been to treat these patients via open redution, Harrington rod instrumentation, and spinal fusion, beause this method of establishing stability is assoiated with less pain and spine deformity and earlier patient mobilization and rehabilitation [11-13]. To provide proper stability, two intat vertebrae below the frature and three vertebrae above the frature are spanned by the Harrington rods. fter distrating and seuring the rods, adequate redution of the frature, whih is loated at the apex of the kyphos, is hoped to be ahieved. posterior bony fusion is ommonly performed after HRI as a supplement to this stabilization/redution proedure. Overdistration of the spine is often avoided by the presene of an intat anterior longitudinal ligament; however, if overdistration does our, it an be easily deteted by plain intraoperative spine radiographs. On the other hand, it is diffiult for the surgeon to be ertain that, with these "blind" rod insertions, adequate distration and redution has ourred, beause the displaed bone fragments are loated anterior to the theal sa and therefore are not visible. s a result, postoperative spine radiographs [10] or CT [1-16] often show persistent residual bony fragments within the spinal anal after HRI. This may neessitate reoperation in order to remove bone fragments at the site of anal ompromise [1, 15]. We believe that most patients with unstable thorai and lumbar fratures ould benefit from an adequate realignment and stablization of the spinal olumn and deompression of the neural elements. Clearly, this deompression would preferably be ahieved at the time of the initial operation without having to subjet the patient to additional surgial proedures. Diret surgial redution of displaed bone fragments bak into the vertebral body following a posterolateral approah (i.e., partial laminetomy and resetion of the medial one-third of the ipsilateral pedile and parts of the faet joint) has been desribed [3, 6, 11, 1] as a useful proedure to be done before HRI. Monitoring the results of this proedure with intraoperative myelography has been advoated [6] as a

4 356 QUENCER ET L. JNR :6, May/June 1985 Fig. 2.-dequate deompression at onus level after HRI., Transverse intraoperative sonogram before HRI. Compression of theal sa at onus level by displaed bone fragment (straight arrows), whih is greater on left. Conus (between open arrows) is seen as relative hypodensity in middle of theal sa. Individual nerve roots surrounding onus an be seen. Small amount of CSF (urved arrow) is present underneath dorsal dura; no CSF is seen ventrally. Large amount of blood (bid) is seen layering on top of dura., Midline longitudinal sonogram. Displaed bone fragment (arrow) is lose to but does not ompress ventral surfae of onus (between open arrows). Intat entral eho in middle of onus is seen. Transverse (C) and midline longitudinal (0) sonograms after HRI. CSF (urved arrows) is seen around entire onus and adjaent roots. This is onsidered adequate deompression, even though some bone fragment displaement (straight arrow) and slight vertebral mal alignment is still evident. Conus is learly seen as hypoehoi struture in enter of spinal anal in C and is outlined by open arrows in O. Improvement in alignment is best appreiated when distane between displaed bone fragment and ventral surfae of onus is ompared in pre- () vs. post-hri (0) sonograrms. o Fig. 3.-lnadequate deompression after HRI, requiring unilateral bone impation to adequately deompress anal., Transverse intraoperative sonogram before HRI. Marked ompression of auda equina at L2 level by displaed bone fragment (arrows)., Transverse sonogram after bilateral HRI. Lessening of sa ompression by bone fragment (straight arrows), but signifiant ompression persists. In omparing and, note that Harrington rod plaement has resulted in some CSF (urved arrows) to now be seen under dorsal dura. eause of presene of persistent bone fragment, right Harrington rod was removed, and via right posterolateral approah, bone fragment on right was impated bak into vertebral body and right Harrington rod was reinserted. C, Dense ehogeni struture in right ventral part of anal is Gelfoam (G), not bone. Note that Gelfoam has hyperehoi appearane but, unlike bone, does not aoustially impede sound beam. Without intraoperative sonography, this patient would have been losed after initial HRI and would have been left with ompressed auda equina.

5 JNR :6, May/June 1985 loss OF HRRINGTON ROD INSTRUMENTTION 357 Fig..-lnadequate deompression after HRI, requiring bilateral bone impation and inreased Harrington rod distration. Initial transverse () and longitudinal (8) intraoperative sonograms after insertion of Harrington rods. ilateral ventral ompression of auda equina at L3 level by displaed bone fragment (, arrows). mount of bony displaement is best appreiated as distane from straight arrow to urved arrow on longitudinal image (8). eause of these findings, both Harrington rods were removed, and bilateral bone impation was performed; Harrington rods were reinserted. C, Final transverse sonogram. Inreased amount of CSF (arrow) in theal sa ompared to pre-bony impation sonogram (). There is no longer ompression of auda equina. Markedly improved bony alignment is best appreiated when final longitudinal sonogram (0) is ompared to amount of bone displaement seen in 8 (f. straight and urved arrows). Without intraoperative sonography, patient would have been losed after initial HRI. 8 means of assuring adequate deompression of the subarahnoid spae before HRI and bony fusion. Sine most thorai and lumbar fratures our from T11 to L 1, as shown by our data (23 of our 37 patients) and the data of others [11, 1, 15], retration of the dura, onus, and adjaent nerve roots may be neessary in order to aomplish the type of bony redution desribed above. Not only may suh retration be made diffiult by the presene of tethering and sarring at the frature site, but suh maneuvers ould have potentially deleterious effets on the spinal ord and roots and their subsequent funtion. It learly would be valuable to know if the Harrington rods alone were apable of reduing the bone fragments, beause then unneessary retration of the spinal ord and prolongation of the surgial proedure would be avoided. loss has afforded us the opportunity to diretly visualize the effets of HRI and to monitor and guide subsequent surgial maneuvers. This type of imaging is quiker and we believe more aurate in assessing redution and alignment than is intraoperative ross-table radiography, a proedure that was reommended [11] before the availability of realtime portable sonographi equipment. In addition to demon- strating the presene of displaed bone fragments and spinal malalignment, loss an rapidly identify and loalize the onus medullaris; sine oasionally the onus may be in an unexpetedly low position, knowledge of this fat an be extremely helpful to the surgeon when bone fragment removal or impation is being attempted. Certainly sonography is simpler than performing a lateral hemilaminetomy and a ostotransversetomy on the side opposite the initial Harrington rod and verifying visually the adequay of the deompression [17]. s table 1 indiates, 31 of our patients had HRI as the initial step of the operative proedure performed in order to ahieve anal deompression. Of these patients, only 1 (5%) had adequate anal deompression (figs. 1 and 2), while 17 patients (55%) were inadequately deompressed (figs. 3 and ). In the remaining six patients, we do not know, had H R I rather than bone fragment impation or removal been first proedure (table 1), whether the anal would have been adequately deompressed. In general, we believe that injuries more than 2 weeks old may not be redued by HRI alone. The proedure of removing or impating bone fragments before HRI (fig. 6) was done in the early phase of our experiene with loss. However, now with inreased onfidene in

6 358 QUENCER ET L. JNR :6, May/June 1985 Fig. 5.-lnadequate deompression after HRI; inreased Harrington rod distration performed, but anal still not adequately deompressed; nothing further done beause patient neurologially intat. Initial transverse () and longitudinal (6) intraoperative sonograms obtained after HRI for T12 frature. Compression of distal spinal ord by displaement bone fragment (arrows). In 6, bone an be seen reahing and distorting ventral surfae of spinal ord (open arrow). eause of these findings, Harrington rods were distrated further, whih resulted in only slight improvement in vertebral malalignment. Transverse (C) and longitudinal (0) sonograms. Some CSF (urved arrow) now seen dorsally, but distortion of ventral ord surfae (open arrows) by displaed bone fragments (straight arrows) persists. eause patient was neurologially intat, it was deided not to attempt to further deompress anal, despite this demonstration of vertebral mal alignment. D our interpretation of these images, we believe that Harrington rods an be applied first; then if loss shows inadequate deompression, Harrington rod removal followed by bone removal/impation and/or inreasing Harrington rod distration may be performed. Of our 17 ases where the initial HRI was inadequate, 11 patients had subsequent proedures that resulted in adequate neural element deompression (table 1, fig. 5). The ability of loss to immediately detet the inadequay of HRI is of paramount importane and deserves reemphasis, sine immediate orretive surgial steps may be taken beause of these sonographi findings and the patient an be spared reoperation. There were four ases where these additional proedures did not adequately deompress the anal (table 1). In one patient, despite attempts at bone impation at the L3 level, the displaed bone fragment ould not be ompletely redued and the surgery was terminated after Harrington rod reinsertion. Three patients were neurologially intat, so even though there was sonographi evidene of improved but still suboptimal deompression, it was deided not to try any additional surgial maneuvers (fig. 5). When all of our ases were onsidered together, final loss showed that spinal anal deompression was eventually adequate in 31 of 37 patients. Of the six patients who were inadequately deompressed, five had minimal defiits or were neurologially intat on admission to the hospital. eause of the desire to limit the amount of surgery in this group of patients, either nothing further was done after the initial HRI (two patients) or if additional surgial proedures were performed (the three patients desribed above), the surgeons did not persist in attempting to ahieve total anal deompression. If these five patients are reognized as patients in whom persistent efforts at deompression were not made, then there was only one patient in whom attempts to totally deompress the anal were truly unsuessful. The results of our study show that if loss is not used to monitor the results of HRI in patients with lumbar and thorai fratures, there will be a signifiant perentage of ases in whih the anal is inadequately deompressed. If, on the other hand, loss is used, ontinued spinal malalignment and displaed bone fragments an be immediately reognized. Surgial steps an then be taken in order to improve spinal alignment and to redue the bone fragments from the anal. Sine adequate neural tissue deompression may relate to the eventual postoperative neurologi outome, partiularly in patients with an inomplete neurologi deficit, we believe that loss is invaluable in the surgial management of unstable lumbar and thorai spine frature. We reommend the use

7 JNR :6, May/June 1985 loss OF HRRINGTON ROD INSTRUMENTTION 359 Fig. 6.-one fragment removal before HRI., Initial transverse intraoperative sonogram. Displaed bone fragment (arrows) ompresses sa at L 1 level. one fragment was removed after left posterolateral deompression. e, Follow-up transverse sonogram shows rounder-appearing sa, but bone fragment (arrow) persists on right side. HRI was then performed. C, Transverse sonogram. one fragment on right (arrow) still not ompletely redued. s a result, Harrington rod was removed, and bone fragment was impated bak into adjaent vertebral body. D, Transverse sonogram after reinsertion of Harrington rod. More symmetri-appearing bony anal and more CSF (arrows) around roots of auda equina. o of loss in all ases where HRI is used to stabilize and redue thorai and lumbar spine fratures. REFERENCES 1. Quener RM, Montalvo M. Normal interoperative spinal sonography. JNR 198;5 : , JR 198;13: edrok GM. Stability of spinal fratures and frature disloations. Paraplegia 1971;9: Mfee PC, Yuan H, Lasada N. The unstable burst frature. Spine 1982;7 : Holdsworth FW. Fratures, disloations and fratures of the spine. J one Joint Surg [r] 1963;5: Holdworth FW. Fratures, disloations and fratures/disloation of the spine. J one Joint Surg [m] 1970;52: Shmidek HH, Gomes F, Seligson D, MSherry JW. Management of aute, unstable thoraolumbar (T ll- T,) fratures with and without neurologial defiit. Neurosurgery 1980;7 : urke DC, Murray DD. The management of thorai and thorailumbar injuries of the spine with neurologial involvement. J one Joint Surg [r] 197;56: Frankel HL, Hanok DO, Hyslop G, et al. The value of postural redution in the initial management of losed injuries of the spine with paraplegia and tetraplegia. Part 1. Paraplegia 1969;7 : Guttman L. Spinal ord injuries: omprehensive management and researh. Oxford: lakwell Sientifi, 1973 : Dikson JH, Harrington PR, Ervin WD. Results of redution and stabilization of the severely fratured thorai and lumbar spine. J one Joint Surg [m] 1978;60: Flesh JR, Leider LL, Erikson DL, Chou SN, radford DS. Harrington instrumentation and spinal fusion for unstable fratures and frature/disloations of the thorai and lumbar spine. J one Joint Surg [m] 1977;59: Yosipovith Z, Robin GC, Makin M. Open redution of unstable thoraolumbar spinal injuries and fixation with Harrington rods. J one Joint Surg [m] 1977;59: Lewis J, MKibbin. The treatment of unstable frature disloations of the thorao-iumbar spine aompanied by paraplegia. J one Joint Surg [r] 197;56: Durward QJ, Shweigel JF, Harrison P. Management of fratures of the thoraolumbar and lumbar spine. Neurosurgery 81981;8 : olimbu C, Firooznia H, Rafii M, Engler G, Delman. Computed omography of thorai and lumbar spine fratures that have been treated with Harrington instrumentation. Radiology 198;151 : White RR, Newberg, Seligson D. Computerized tomographi (:) assessment of the traumatized dorsolumbar spine before and after Harrington instrumentation. Clin Orthop 1980; 16 : Seljeskog EL. Comments. Neurosurgery 1981 ;8 :