Epidemiology

Among patients undergoing primary decompressive lumbar spine surgery, incidental durotomy occurs at an incidence of 9–12%, making it the most frequently reported complication of lumbar spine surgery.^6,7 After cauda equina syndrome, dural tears are the most frequent underlying diagnosis in malpractice litigation involving lumbar spine surgery.⁸

Pseudomeningoceles are a rare consequence of dural injury. In a review of 1700 laminectomies, Swanson and Fincher reported pseudomeningocele formation in four patients.⁹ In a group of 400 postlaminectomy patients with persistent back pain and/or radicular symptoms, Teplick et al. found pseudomeningoceles in 8 patients (2%).³

Factors that predispose to a dural tear in surgery are prior surgery or irradiation, congenital spinal malformations, unrecognized dural adhesions, dural calcification contiguous with overlying lamina, and the absence of epidural fat.^8,10,11 Severe spinal stenosis and large herniated discs make root dissection and dural retraction more difficult and can predispose to dural tears. Dural injuries and pseudomeningoceles are understandably more frequent with posterior than anterior approaches.

Pathogenesis

Pseudomeningoceles generally develop following an intraoperative rent in the dura and arachnoid, but can occur following dural needle puncture procedures, especially after multiple punctures.⁹ The cerebrospinal fluid leaks dorsal to the lamina, into a space artificially created by dissection of the paravertebral musculature. The fluid cavity is lined by flattened connective tissue cells resting on loose, areolar connective tissue.^2,4 An encysted pseudomeningocele refers to a CSF-filled herniation of an arachnoid lined cyst through a small rent in the dura. The narrow opening in the dura acts like a one-way valve leading to an increase in size of the CSF collection and formation of a pseudomembrane.²

The size of the dural defect, pressure of the inflowing CSF, and the resistance of the surrounding soft tissues all influence the size of a pseudomeningocele. Postoperative weakness of paraspinal muscles may be a reason contributing to expansion of the pseudomeningocele.¹² Spinal fluid pressure keeps the dural defect open and prevents healing. Herniation of nerve filaments into the original dural defect may be another factor responsible for keeping the dural opening patent (Fig. 105.1).⁴

Fig. 105.1 Meningeal layers involved in the formation of a pseudomeningocele, with schematic representation of nerve root entrapment at the site of the dural defect.

Puncture of the dura can result in escape of large volumes of CSF, leading to intracranial hypotension and a demonstrable reduction in CSF volume. The adult subarachnoid pressure of 15 cmH₂O can be reduced to 4 cmH₂O or less. The rate of loss of CSF through a dural perforation following puncture with a 25-gauge or larger needle is generally greater than the rate of CSF production.¹³

Clinical features

Cerebrospinal fluid leaks after needle procedures in the lumbar spine classically present with headache. There is evidence that CSF leaks may be inevitable following needle puncture of the thecal sac,¹⁴ and it is unclear why only some of these patients are symptomatic. The size of the dural leak has no direct relationship with the occurrence of headache.¹⁴ The actual mechanism of headache is also unclear. Some authors suggest that the lowering of CSF pressure following a leak causes traction on the intracranial structures in the upright position. These structures are pain sensitive, leading to the characteristic headache.¹³ Others authors report that the drop in intracranial pressure causes a compensatory intracranial venodilation, which causes the headache. The headache is generally localized to the frontal and occipital regions, with occasional radiation to the neck and shoulders. Other symptoms may include nausea, vomiting, tinnitus, and vertigo.

While pseudomeningoceles may be asymptomatic, many present with local swelling, symptoms of CSF leak, or nerve root impingement. Local swelling¹⁵ may occasionally be the only suggestion of a pseudomeningocele.¹² The swelling may vary in size, depending on whether the patient is standing or recumbent. There is no apparent correlation between the size of the pseudomeningocele, the size of the defect in the dura, and the degree of symptoms.^4,16 Postural headache frequently occurs in patients with pseudomeningocele. Miller and Elder⁴ reported one patient whose postural headache could be increased by manual pressure directly over a large fluctuant meningocele.

Back pain, with or without radicular leg symptoms, is commonly associated with pseudomeningocele. In a review of 10 patients with postoperative pseudomeningocele, Miller and Elder found that back pain was a common feature in all patients, with nine having radicular pain radiating into the lower limbs.⁴ Manual pressure over the paraspinal musculature on the side of the cyst may exacerbate the back pain.¹⁶ Coexisting worker’s compensation issues can lead to doubts being expressed as to the authenticity of such complaints in the postoperative patient.¹⁶

Radicular findings in patients with pseudomeningocele result from herniation and kinking of the nerve roots in the dural defect,¹⁷ adhesion of the nerve roots to the edges of the sac,¹⁸ or direct pressure of the pseudomeningocele on an adjacent nerve root. Eismont et al.¹⁹ reported a case of persistent postoperative radicular pain along with focal deficits following lumbar disc excision. Myelography revealed a pseudomeningocele involving the left fourth and fifth lumbar nerve roots. Intraoperatively, it was found that the L4 root had buttonholed through the site of the dural tear, while the pseudomeningocele was found to be compressing the L5 root. Ossification of a pseudomeningocele can rarely cause neural compression.²⁰ In many patients an obvious site of nerve root entrapment is not identifiable.⁴ Asymptomatic meningoceles can become symptomatic after a precipitating event leading to nerve entrapment.²¹ Radicular pain in a previously asymptomatic patient can be precipitated by maneuvers that increase intracranial and intraspinal pressure, such as coughing, sneezing, or jugular compression.¹⁶ Unusual presentations associated with pseudomeningoceles include bone erosion of the posterior vertebral elements²² and chronic meningitis.²³

Investigations

Myelography shows a characteristic pattern of dye extruding through the stalk of the pseudomeningocele, and a fluid level may be visible when the opening in the dura is large. Percutaneous injection of contrast material directly into the cyst cavity has been carried out when a palpable or visible mass is present at a surgical site.¹⁶ CT-myelography can effectively demonstrate the neck and outline of the pseudomeningocele although nerve root entrapment is not well visualized.¹⁸

MRI scans are very sensitive for CSF collections and pseudomeningoceles (Fig. 105.2). Some authors recommend a routine postoperative MRI scan following any repair of a durotomy, even in the absence of symptoms of a pseudomeningocele, in order to ensure that no further leakage is present.²⁴ Digital subtraction myelography may occasionally reveal a pseudomeningocele when plain myelography, CT scan, and MRI show a fluid collection posterior to the dural sac, but no obvious connection between the two.²⁵ In radionuclide cisternography, radionuclide contrast material is injected into the subarachnoid space followed by MRI evaluation. Areas filled with CSF show up as high-signal regions.²⁶

Fig. 105.2 (A) Parasagittal T2-weighted MRI of postoperative lumbar pseudomeningocele. (B) Axial T2-weighted MRI of the same patient, showing the pseudomeningocele communicating with the thecal sac.

Management of dural tear and pseudomeningocele

In an attempt to reduce the incidence of postdural puncture CSF leak, numerous modifications have been made to spinal needles. Newer needle designs with narrow cutting tips and atraumatic bevels have resulted in a lower incidence of postspinal headache, neural irritation,^13,27 and CSF leak.²⁸

Spontaneous recovery from postdural puncture headache (PDPH) occurs in over 85% of patients within 6 weeks. Many forms of conservative management have been suggested and practiced, mostly without any scientific evidence to support their efficacy. Bed rest has been historically recommended, but does not lower the incidence of PDPH although it may delay its onset and decrease its intensity.²⁹ Some authors suggest that bed rest may prevent further complications from traction and rupture of meningeal veins caused by intracranial hypotension.³⁰ For severe and persistent headaches epidural blood patch (EBP) injections have been successfully used. Autologous blood is injected into the epidural space, where it forms a clot and seals the dural hole. Although over 90% of PDPH respond initially to EBP injections, a sustained and durable response has been identified in only 61–75%.³¹

Intraoperative durotomy is managed by primary dural repair, with a watertight suture of the dural defect. The key to prevention of postoperative pseudomeningocele is a meticulous primary repair of dural tears detected intraoperatively. Fine suture material such as 5-0 or 6-0 monofilament, braided or coated synthetic is suitable, although leakage of CSF can occur through the needle holes in the dura. A continuous running suture is quickest to perform, but may result in unraveling of the entire suture line if the tension is not perfect or the suture breaks. The use of magnification and coaxial illumination is essential for adequate repair. Local application of fibrin glue greatly reinforces a repair site.^26,32

When a CSF leak is noticed postoperatively, surgical treatment is generally recommended for the immediate and definitive management of the lesion.¹⁹ Some authors recommend nonoperative management in a patient with a well-healed incision presenting with a soft subcutaneous bulge and no postural headache.³³ Waisman and Schweppe treated seven patients by reinforcing the skin suture line, bed rest in the Trendelenburg position, and repeated puncture and drainage of the subcutaneous CSF collection. The CSF leakage stopped immediately after the incision was resutured, the subcutaneous fluctuation disappeared in 10–28 days, and an 8-year follow-up (clinical and ultrasound examination) revealed no pseudocyst formation.³⁴ Eismont et al. reported one case where resuturing the skin stopped the CSF leakage through the incision. A symptomatic pseudomeningocele subsequently developed, and the authors recommends surgical repair of the dura for all dural leaks noted postoperatively.¹⁹

Closed subarachnoid drainage is frequently used in the management of CSF leaks^33,35 and occasionally for established pseudomeningocele.³⁶ The catheter is positioned in the subarachnoid space using fluoroscopic technique. Epidural catheter placement may be used for drainage of a pseudomeningocele.³³ The proximal end of the catheter is connected to a sterile drainage system. The basis of this external CSF drainage is that a reduction of CSF pressure and preferential escape of the CSF through the catheter for 4–5 days generally allows the dural fistula to heal itself.

Lumbar peritoneal shunting may play a role in the management of refractory CSF leaks.^33,37 The dural opening is surgically repaired when feasible. A catheter is then introduced into the subarachnoid space, and a second catheter left in the meningocele cavity; both catheters are subcutaneously tunneled into the peritoneal cavity, where they are positioned under the diaphragm using laparoscopic assistance. The advantages of this technique including immediate mobilization of the patient, avoidance of complications of external drains and repeated aspirations, and a high success rate. The procedure is not in widespread use due to technically simpler and less invasive options being available.

Occasionally, dural tears are lateral or ventral in position and not amenable to direct repair techniques. Thinned out or scarred dura adjacent to segments of thickened extradural scar tissue in the multiply operated back does not readily lend itself to direct repair. Some dural openings are irreparable after intradural procedures when the dura is abnormal or has to be partially resected. In such situations suturing fat, muscle graft, or autologous fascia lata graft over the repaired dural tear may reinforce the dural repair. Graft substitutes including cadaveric dura, silicone-coated Dacron, and bovine pericardium have been used, but are associated with the possibility of disease transmission and immunogenic tissue reaction.

In previously irradiated tissue a vascularized myocutaneous flap can be rotated from an uninvolved area to the poorly vascularized defect and sutured to well-vascularized surrounding tissue. Newer techniques such as laser tissue welding are being investigated. Preliminary results show that primary repair combined with laser welding produces a higher leak pressure and tensile tissue strength than either technique used alone, with no evidence of underlying thermal tissue injury.³⁸

Patients with established pseudomeningoceles who are not significantly disabled by their symptoms may be treated with observation alone. Follow-up MRI scans occasionally show complete resolution of small pseudomeningoceles.² Sustained local pressure has been found to help control some symptoms of pseudomeningoceles. Leis et al.¹⁵ reported this form of treatment in a patient who was advised surgery but declined. The patient used a wide belt worn tightly around the waist and a folded towel under the belt so that it applied direct pressure over the swelling. This relieved the postural headache symptoms immediately. Repeat MRI scans done 8 weeks and 6 months after surgery showed progressive decrease in the size of the pseudomeningocele, till there was only a tiny focus of fluid left. The authors suggest a trial of focal compression for symptomatic relief of postural headache from pseudomeningocele. If symptomatic improvement is obtained, a more prolonged trial of mechanical compression may promote dural closure.¹⁵

Symptomatic pseudomeningoceles generally require surgical intervention.^2,21 The operative procedure usually involves opening of the pseudomeningocele, followed by identification and closure of the dural defect with flaps from the cyst wall or a free graft. Augmentation of the repair with fibrin glue, muscle, fat, and fascial grafts is frequently necessary. Resection of overlying lamina is generally necessary for adequate visualization of the dural defect, and spinal fusion is indicated when excessive bone resection renders the spine unstable.⁴ Miller and Elder⁴ reported on the surgical treatment of 10 symptomatic pseudomeningoceles. The authors reported good results in seven patients, satisfactory results in two, and a poor result in one patient.

Apart from surgical repair, the only other surgical option for pseudomeningocele is CSF diversion, a technique that is not always successful, especially when spinal fusion implants prevent the normal reapproximation of paraspinal tissues around the collection. In addition, a ball-valve mechanism at the dural fistula site may prevent complete drainage of CSF from the pseudomeningocele. The cavity could persist and reaccumulate CSF after the drain is removed.³⁹

Complications of untreated pseudomeningocele include recurrence of radicular symptoms, the possibility of infection with resultant meningitis,²³ ossification of the cyst wall producing canal stenosis and claudication,⁴⁰ and rarely erosion of the bony vertebral elements from an expansile cyst.²²

Introduction

The term ‘hematoma’ when used in postoperative spine patients generally refers to a subcutaneous or subfascial incisional hematoma, and must be distinguished from a true spinal hematoma. The frequency of incisional hematomas depends on many factors, including the type and extent of surgery. An incidence of 0.08% (8.7 in 10 000) incisional hematomas was reported among 28 395 patients undergoing lumbar laminectomy,⁴¹ whereas rates of up to 12% have been reported following scoliosis surgery.⁴²

True spinal hematomas occur most frequently in the epidural space (Table 105.1). The overall incidence of symptomatic spinal hematomas is very low, with a recent meta-analysis finding only 613 symptomatic patients over the past 170 years.⁴³ Lawton et al.⁴⁴ reported a 0.1% incidence (12 in 10 500) over a 14-year period comprising 10 500 spine surgery cases at one institution. Smaller asymptomatic hematomas occur more frequently than recognized in postoperative patients. When postoperative MRI is obtained, the reported incidence is 100%, even with procedures such as microdiscectomies.⁴⁵

Table 105.1 Distribution of spinal hematomas according to anatomic localization⁴³

The incidence of spinal hematoma following epidural/subarachnoid injection procedures is low. In a meta-analysis of 613 patients with spinal hematoma formation, Kreppel et al. reported that 10.3% (63 in 613) of these incidents occurred following lumbar puncture or spinal anesthesia.⁴³ In a literature review of epidural hematomas after needle procedures, Adler et al. reported that 12 of 15 reported cases either had an abnormal bleeding profile or were on anticoagulant therapy.⁴⁶ A review of subdural hematomas after lumbar injections showed similar results, with 17 of 19 reported patients receiving some form of anticoagulant therapy.⁴⁶

Etiology

In an extensive search of the literature on spinal hematomas, Kreppel et al. found a total 613 case studies reported over the last 170 years. Most cases had a multifactorial etiology, and no etiologic factor could be pinpointed in 29.7% of patients. Idiopathic spinal hematomas represented the most common category among all different age groups analyzed.⁴³ Spontaneous epidural bleeds can arise following fibrinolytic therapy for coronary heart disease.⁴⁷ Spinal hematomas also spontaneously develop from the lumbar facet joint⁴⁸ and ligamentum flavum.⁴⁹ The mechanism in these cases is presumed to be minor injury in the setting of degenerative changes. Arteriovenous malformations are another reported cause of spontaneous hematomas.⁵⁰ Groen and Ponssen⁵¹ reported a spontaneous spinal epidural hematoma following rupture of the internal vertebral venous plexus. Spontaneous hematomas have also been reported in patients with liver disease and autoimmune conditions.⁵² Spinal hematomas occur rarely following spinal fractures, but are more prevalent with fractures in patients with ankylosing spondylitis.⁵³ Spinal manipulation therapy is known to produce hematomas with resultant neurological deficits.⁵⁴

Spinal injection procedures such as lumbar punctures, epidural and subdural anesthetic blocks, and therapeutic nerve root injections can lead to spinal hematoma formation. The hematomas are commonly epidural although there are a few reports of subdural hematomas. The common underlying factors in these reports are the coexistence of coagulation disorders, multiple punctures performed during the procedure, the insertion of a catheter, removal of a catheter, prior surgery, fibrosis, undetected spina bifida, and tethered cord syndrome. Preexisting coagulopathy may increase the likelihood of hematoma formation following spinal injection procedures, and the likelihood of back pain and neurologic deficit is also higher in these patients.

Spinal hematoma is a recognized cause of postoperative worsening of neurological status in the first 24–48 hours following surgery. Kou et al.⁵² reviewed factors that predisposed to epidural hematoma formation in 12 patients who underwent lumbar laminectomy over a 10-year period. Multilevel procedures (p=0.037) and a preoperative coagulopathy (p=0.001) were significant risk factors. Age, body mass index, durotomy, and the use of postoperative drains were not statistically significant risk factors. In multilevel laminectomies, the increased exposure needed may increase the risk for bleeding from the paravertebral muscles. The larger exposures of the epidural space also increase the risk of insidious bleeding from the prominent internal vertebral venous plexus.⁵² The lordotic posture of the spine postoperatively may have a role in accentuating the pressure effects of a spinal hematoma. Ko and Kakiuchi⁵⁵ reported on a patient who developed paraparesis 3 hours following decompressive surgery for lumbar spinal stenosis. Flexion of the spine was noticed to relieve the patient’s symptoms. The patient was positioned with the lumbar spine in sustained lumbar flexion and the deficits resolved within 5 days.

Subdural spinal hematomas occur less frequently than epidural hematomas, and again are more likely with preexisting coagulopathy or following injection procedures. Accidental subdural spinal injection can occur in patients with postsurgical fibrosis causing obliteration of the epidural space, and mimics the radiographic appearance of epidural hematoma. Spinal subdural hematomas can also track down from a cranial subdural hematoma. The presence of blood within the dura may produce a fibroproliferative reaction of the leptomeninges, resulting in arachnoidal fibrosis and poorer prognosis.

Clinical features

The presentation of spinal hematomas varies depending on the precipitating event, size of hematoma, canal dimensions, and location of the hematoma within the spinal column. Small hematomas are generally not evident clinically, and may be a universal phenomenon in the immediate postoperative period. Montaldi et al.⁵⁶ found computed tomographic changes within the spinal canal suggestive of postoperative spinal hematoma or scar tissue in 84% of asymptomatic patients 1 week after lumbar discectomy. Kotilainen found MRI changes representing hematoma in 100% of asymptomatic patients following lumbar microdiscectomy on the first postoperative day.⁴⁵

Neurologic deficits from larger spinal hematomas are infrequent. Patients with incomplete deficits fare better than those with complete paraplegia, and the location of the hematoma has a bearing on the clinical presentation as well as outcome. Spinal epidural hematomas at the lumbar cauda equina level usually fare better than hematomas at the spinal cord level. Spontaneous remission has been observed in hematomas at the cauda equina level with minor neurologic deficits.⁵⁷ Spontaneous resolution has also been reported in two cases where the presentation was rapidly progressive over minutes.⁵⁸ The symptoms resolved while investigations were being carried out over the next few hours. The authors suggested that while urgent decompression remains the treatment of choice in a symptomatic spinal epidural hematoma, conservative management may be indicated where there are definite signs of neurological improvement in the initial few hours.

In patients who develop spontaneous spinal epidural hematomas without preceding trauma, an acute onset of local pain may be followed by radicular paresthesias. Within hours, signs of spinal cord compression can appear, presenting as progressive paraplegia and loss of sensory function. In 61 patients who developed spinal epidural hematomas after epidural puncture for anesthesia, and had been on anticoagulants, Vandermeulen et al. noted muscle weakness was the first sign in 28 of 61 patients, back pain in 23, and sensory deficit in 9.⁵⁹ Paraplegia was recorded to occur within 14.5±3.7 hours after ending of the anesthetic. The most successful recovery following surgical decompression occurred in patients in whom surgical intervention was performed in less than 8 hours. In another review of five patients with spinal hematomas the authors found that four of five patients complained of acute back pain with paresthesias in both legs, followed by rapidly progressive paraplegia.⁵⁰ The fifth patient had an incomplete cauda equina syndrome. Slowly accumulating hematomas may result in delayed presentation up to a week following surgery. Long tract symptoms and findings may be seen if the hematoma impinges on the cervical spinal cord, following spinal manipulation or cervical epidural steroid injections. Although the clinical presentation of spinal epidural hematomas is relatively characteristic, other diagnoses need to be considered (Table 105.2)

Table 105.2 Differential diagnosis of spinal epidural haematomas

Investigations

Magnetic resonance imaging scanning is the diagnostic method of choice for evaluation of spinal hematomas (Fig. 105.3). MRI scans allow rapid evaluation of large sections of the spinal column and provide accurate information about the size and longitudinal extent of the lesion. Epidural hematomas have an isointense signal on T1-weighted images. On T2-weighted images, acute hematomas show low-intensity signal in the periphery with a central high-intensity signal region. Beginning 4–14 days after the hematoma, the signal intensity increases from the periphery to the center due to the presence of methemoglobin derived from erythrocytes. Subdural hematomas can mimic epidural hematomas on MRI, but generally show high-intensity signals on both T1- and T2-weighted images. Preservation of the posterior epidural fat and visualization of the dural outline on axial images provide further clues to the subdural location of the hematoma. Intraspinal compressive lesions such as tumors or abscesses may be differentiated from subdural hematomas by virtue of a rim or uniform enhancement on MRI, whereas subdural hematomas rarely enhance. Some authors describe a 3-pointed star or ‘inverted Mercedes star’ appearance in lumbar subdural hematomas.⁶⁰ CT scanning, myelography, and ultrasound have all been used to detect spinal hematomas, but do not have the sensitivity or specificity of MRI scanning.

Fig. 105.3 Axial T2-weighted MRI of 83-year-old male with a thoracic epidural hematoma causing significant anteroposterior flattening of the spinal cord.

Management

An awareness of predisposing factors may substantially lower the incidence of spinal hematoma formation. For patients on anticoagulant therapy and undergoing spinal/epidural anesthetic procedures, the following guidelines may help reduce the occurrence of spinal hematoma:⁶¹ (1) atraumatic puncture, (2) a time interval of at least 60 minutes between spinal anesthesia and heparinization (based on the half-life period of heparin), and (3) close monitoring of coagulation parameters. Since the removal of epidural catheters is associated with a high incidence of bleeding complications, Sage⁶² recommended against the removal of spinal/epidural catheters in the first 2 hours following heparin administration, as many patients develop temporary blood concentration of heparin at this time. Following spine surgery, most authors recommend at least 12 hours before restarting anticoagulant therapy.⁶³

Although the use of postoperative suction drains following spine surgery is widespread, there is surprisingly limited evidence supporting this practice. In a recent Cochrane review of 21 studies in orthopedic literature on this subject, the authors found no difference in the rate of hematoma formation or other wound complications whether drains were used or not.⁶⁴ Similar results were noted in a randomized trial on the efficacy of postoperative lumbar drainage on 200 patients after single-level nonfusion lumbar surgery. Patients were randomized into two groups of either receiving postoperative drains for 48 hours, or having no drain inserted. The authors found that no patient in either group developed a significant hematoma/seroma requiring surgical drainage.⁶⁵

The management of established spinal hematomas ranges from simple observation to emergency decompression surgery. The decision to treat a hematoma nonoperatively depends on the patient’s presenting symptoms, duration of symptoms, and evidence of clinical recovery in the early hours following onset of neurological deficit. In patients with no neurological deficits, nonoperative measures are reasonable. Even severe deficits have been treated with watchful expectancy and intravenous steroids when there are definite signs of neurologic improvement in the early hours following the onset of symptoms. The more spacious dimensions of the lumbar spinal canal, and the absence of spinal cord in this region lead to less damaging effects of a hematoma in the lumbar region. Younger patients with lumbar spine hematomas showing early evidence of neurological recovery can be considered for close observation.

Most spinal hematomas presenting with acute neurological deficits are treated by emergency surgical decompression. The need for decompression of extrinsic cord compression has been demonstrated in experimental as well as clinical studies. Using an animal model, Delamarter et al.⁶⁶ studied the effects of duration of spinal cord compression. Persistent pressure on the cord for 6 hours led to progressive necrosis within the cord, and neurologic recovery did not occur in this group following decompression. In a clinical review of 55 patients who developed spinal hematomas after spinal or epidural anesthesia and who were treated with emergency decompressive laminectomies, Vandermeulen et al found ‘good’ or ‘partial’ neurological recovery in 77% (10/13) of patients operated within 8 hours, 37.5% (3/8) of those treated within 8–24 hours, and only 17% (2/12) of those decompressed after 24 hours.⁵⁹ Exceptions to the rule of urgent decompression may be made when: (1) a patient presents 48–72 hours after onset of neurologic involvement, and (2) significant recovery is noted in the initial few hours after onset of the deficit.

Operative decompression should be carried out along the entire length of a hematoma to ensure adequate evacuation and maximize chances of neurologic recovery. In some cases of extensive spinal hematoma, a limited decompression may obtain similar results without the morbidity of an extensive decompression. Osmani et al. reported limited decompression for a patient who had a hematoma extending from T4 to L5. The authors performed laminectomies from T11 to L5, and then used an arterial embolectomy balloon catheter to extract the more proximally located epidural blood clots. The patient reportedly recovered ‘significant motor power’ at 2 months.⁶⁷ Another option that may play a role in the future in extensive spinal hematomas is the use of thrombolytic agents locally. Thrombolysis and evacuation of hematoma is achieved by intermittent irrigation of the subdural space with recombinant tissue plasminogen activator (rt-PA), followed by saline lavage.⁶⁸

Introduction

The incidence of seromas occurring after lumbar spine procedures varies from 2% to 16%, with higher rates after surgery in the elderly population.^69,70 Seromas develop after various posterior spinal procedures, including scoliosis/deformity correction surgery, posterior lumbar interbody fusion, instrumented lumbar fusion procedures, and after paraspinal procedures for the treatment of CSF fistulas. Seroma also occur after implantation of spinal cord stimulators and pain pumps.

Clinical presentation

In most cases, seroma formation presents with persistent clear drainage from the surgical suture line noticed in the immediate postoperative period that persists or increases in volume with time. In some cases the drainage appears from a small breakdown in an otherwise healed incision, 3–6 weeks following surgery. Complete wound dehiscence is a potential complication. Rarely, postoperative seromas can cause neural compression with recurrence or worsening of neurological deficits.

Management

Attention to surgical technique with prevention of dead spaces during closure is of primary importance in the prevention of seromas. Particular attention needs to be paid to closure of the subcutaneous layer. Interrupted sutures are preferred to a running closure, to avoid the entire closure coming apart if integrity is lost at one site. The skin should not be unnecessarily undermined during the approach, in order to avoid a potential seroma cavity. A postoperative suction drain does not appear to affect the incidence of postoperative seromas.⁶⁵ Wearing an abdominal binder may help prevent seroma formation after implantation of spinal cord stimulators or pain pumps.⁷¹

The diagnosis is self-evident in most cases. MRI signal characteristics of postoperative seromas include relatively well-defined lesions of low or intermediate signal intensity relative to adjacent muscle on T1-weighted images and very high signal intensity on T2-weighted images.

Seromas are frequently successfully managed by one or two percutaneous needle aspirations. Sterile technique needs to be used, and the risk of bacterial contamination of the surgical incision makes this option unattractive. Gram-staining of the aspirate is recommended to rule out the possibility of infection, and antibiotic therapy instituted if infection is detected. In situations where an infected seroma develops in the subcutaneous pocket housing a pain pump/spinal cord stimulator, antibiotic irrigation of the cavity is advised.⁷¹ Open debridement and implant removal need to be considered in all cases with recognized infection.