Background
A vast majority (up to 80%) of the US population are estimated to experience an episode of low back pain (LBP) during their lifetime. A subset of these individuals will go on to develop chronic LBP, which can limit their ability to perform activities of daily living (ADLs) and frequently, may hinder the ability to maintain gainful employment. This imposes a significant economic burden on the healthcare system via the increased utilization of healthcare resources. The extent to which chronic LBP affects the US population cannot be overstated. A cross-sectional survey of over 5000 American households, published in 2009, found the prevalence of chronic LBP to be 10.2% in 2006, a substantial increase from a similar analysis performed by the same authors in 1992 (3.9%). Of patients with debilitating back pain, an estimated 3% will require surgical intervention.
Although the index operation may portend excellent clinical outcomes for some, a subset of patients will develop either recurrence of their symptoms or new onset symptomatology (including either ipsilateral or contralateral back and/or leg pain) in the months to years following their initial operation. The rationale for why this occurs is grounded in multiple factors including the initial condition that was treated, whether a decompression was performed alone or in combination with a single-level or multi-level fusion, and the development of distinct pathological processes that can occur secondary to the index operation. These processes include adjacent segment degeneration/disease (ASD), lumbar pseudarthrosis, recurrent disc herniation, residual or recurrent spinal stenosis, instrumentation failure, wound-related complications, iatrogenic flatback syndrome, and cerebrospinal fluid (CSF) leak/pseudomeningocele development. Other factors that can influence the success of the index surgery include the duration of symptoms before surgery, the patient’s psychosocial health, and patient-specific factors such as other medical comorbidities, body mass index (BMI), smoking status, bone quality, and chronological age.
Revision Lumbar Spine Surgery
In patients with recurrent radicular pain and/or symptomatic spondylosis following surgery who present with radiographic evidence of pathological changes within the lumbar spine, revision lumbar spine surgery may be necessary. The incidence with which these patients are reoperated varies considerably in the neurosurgical literature, depending on the presenting pathology. The rates of recurrent lumbar disc herniation after discectomy, for example, can vary from 5% to 15%, with a subset of patients requiring repeat surgery to address persistent symptoms after failing conservative measures. Other pathologies, such as ASD following a lumbar fusion, can occur more frequently, necessitating that a greater fraction of patients are reoperated. Ghiselli et al., for example, documented a 27.4% reoperation rate (decompression with and without arthrodesis) in their series on 215 patients who had originally undergone posterior lumbar fusions because of symptomatic degeneration at adjacent segments. This rate, the authors observed, was higher in patients in whom the fusion was limited to the lumbar spine, and which did not extend to either the thoracic or sacral spines.
Regardless of the rationale for repeat lumbar surgery, the operating surgeon must be cognizant that a revision procedure is typically more technically challenging to perform and is associated with more inconsistent outcome results than the index surgery. The possibility for a compromised vascular supply during the revision surgery leading to wound healing-related complications, the potential for increased intraoperative blood loss, the presence of a distorted anatomy making it difficult to locate bony landmarks, and epidural scarring or fibrosis increasing the risk for durotomy are all aspects of a revision lumbar spine surgery that the surgeon must be aware of. The management of this last aspect, the possibility for epidural scarring/fibrosis which can predispose to durotomy, will be discussed in greater detail in this chapter.
History Taking and Physical Examination
Careful selection of patients in whom revision lumbar surgery is indicated is essential. Obtaining a thorough history and performing a meticulous neurological examination are the first steps in the work-up of patients presenting with either recurrent or new symptoms of back and/or leg pain who have previously undergone lumbar spine surgery. Key elements to gather from the history include any changes in motor function, any sensory abnormalities (i.e., numbness or tingling), any autonomic difficulties (bowel, bladder, or sexual dysfunction), and any reports of pain. Important components of the history include whether a CSF leak occurred during the index surgical procedure. This is best determined by asking the patient directly as well as by carefully reviewing the operative report from the initial surgery. Patients should also be asked whether a lumbar subarachnoid drain was used in the perioperative period. In addition, the patient should be queried as to whether they had persistent wound drainage or whether revision surgery was necessary to repair a CSF leak.
The physical examination should include a detailed general evaluation and a comprehensive neurological examination that include a full evaluation of motor and sensory function, assessment of deep tendon reflexes, and gait assessments. Additionally, the range of motion of the hip and knee joints should be evaluated to rule these out as occult causes of pain. A full assessment of the patient’s mental status and psychological health should be performed during this initial meeting. This is particularly important, as addressing any issues with regards to this before revision surgery has been shown to improve outcomes.
Using the combination of the clinical history and the physical examination is important in attempting to identify the pain generator/pathology at this stage, often before obtaining the necessary imaging studies. The presence of constitutional symptoms such as a fever, chills, weight loss, and/or night sweats in combination with back pain may point the practitioner toward infectious or neoplastic etiologies rather than toward a degenerative or mechanical one. The presence of radicular pain and signs of neurological compression in a patient who had previously undergone a lumbar discectomy and had been asymptomatic for several (6 or more) months may be indicative of a recurrent disc herniation. A pseudarthrosis, meanwhile, can present with recurrent mechanical back pain and neurological symptoms in a patient with a prior fusion attempt months to years after the index surgery. Importantly, in patients with prior posterior lumbar surgery, myofascial pain secondary to dissection of the paravertebral musculature can be present and is important to distinguish from true neuropathic pain.
Imaging Work-Up
Neuroimaging studies are essential in the evaluation of patients who may be candidates for revision lumbar surgery. Plain films should be obtained to evaluate for a successful fusion and any signs of instability. Routine anteroposterior (AP) plain film radiographs are particularly helpful to review at the time of revision surgery, as the anatomy is well visualized. Flexion and extension lateral plain radiographs are useful in the diagnosis of a pseudarthrosis, which can be made if pathological motion is present (usually greater than 3 degrees of angular motion or 3 mm of translational motion). However, there remains significant disagreement on the amount of motion that remains compatible with a successful fusion. Bono et al. have previously asserted that the amount of residual motion present on flexion-extension images varies with fusion type, and that a solid (but incomplete) fusion can occur in patients who have angular motion greater than 5 degrees on flexion-extension radiographs.
Computed tomography (CT) imaging is especially useful in the evaluation of bony anatomy. CT scans are the gold standard for identifying a successful bony fusion. In candidates for revision lumbar surgery, CT scans can be used to assess for screw orientation, whether the screws are properly set within the pedicle (or whether they are deviated from the optimal trajectory), and screw loosening (identified as lucencies around the screws). CT images can also guide surgeons with respect to whether the same pilot holes in which the screws were inserted should be used in the revision procedure or whether separate screw holes will need to be made. Screw width and length can also be assessed. These scans are also used to evaluate for the presence of bridging bone across motion segments, lucency around existing screws or cages, and graft subsidence, all of which can be indicative of an unsuccessful fusion attempt. Thin-section coronal and sagittal reconstructed CT scans in particular have been shown to be excellent for the early detection of pseudarthrosis. Kanemura et al. determined that a 1-mm radiolucent zone around the interbody cage in patients undergoing posterior or transforaminal lumbar interbody fusion is an early predictor of pseudarthrosis.
Magnetic resonance imaging (MRI) is especially valuable for identifying soft tissue structures such as nerve roots and disc abnormalities. MRI also has utility in patients without instrumentation, for the evaluation of soft tissue architecture and anatomy. MRI with gadolinium contrast administration promotes enhancement of vascularized tissues and can be used to differentiate epidural scar tissue (which uniformly enhances) from recurrent disc herniations (which has peripheral enhancement). MRI is valuable in assessing segments adjacent to prior instrumented fusions when determining if ASD (herniation or stenosis) has occurred next to the previously operated region. MRI can also be used to detect epidural abscesses and infections of the disc and/or vertebral bodies postoperatively.
The value of myelography in candidates for revision lumbar surgery should not be underestimated. Myelography is particularly useful for assessing neural compression and evaluating the anatomy of nerve root sleeves exiting the neuroforamina. Myelography combined with a postmyelogram CT scan, when using bone windows, can be useful to assess a successful fusion. Myelography is generally indicated in patients who are unsuitable candidates for MRI such as those in whom instrumentation is present or in patients with pacemakers. Additionally, when MRI has excessive artifact in the operative region, a CT myelogram can often provide superior visualization for patients who are candidates for revision surgery. We have found myelograms, with postmyelography CT scans, are valuable in the diagnostic work-up of patients for whom revision lumbar spine surgery is being considered.
Medical Optimization
Once both clinical and radiographic findings have identified a surgical candidate for revision lumbar surgery, it becomes important to ensure that he or she is medically optimized before being operated on. Factors that are carefully considered include smoking status (with an emphasis on adequate counseling and promotion of smoking cessation before surgery in current smokers), weight control (by starting patients on weight-control programs if they are overweight and/or morbidly obese and by encouraging physical activity and ambulation in the preoperative setting), and nutritional status. It is imperative to work with patients’ primary care providers to ensure that their other medical comorbidities (i.e., diabetes, opiate dependence, etc.) are adequately managed before surgery. Dual-energy x-ray absorptiometry (DXA) scanning is valuable for the assessment of bone mineral density. In certain subsets of patients (i.e., the elderly and female patients), DXA scanning of the lumbar spine and/or femoral neck can provide valuable information on bone quality and can be used to identify osteoporotic patients in whom the chances for a successful fusion may be lower. In addition, when patients have poor bone mineral density, the surgical plan is modified to maintain bony alignment rather than to attempt to correct deformity.
Dural Tears During Lumbar Spine Surgery
As previously noted, the potential for epidural scarring and fibrosis increases the risk for durotomy in revision lumbar surgery. An unrecognized dural tear can result in symptoms such as postural headaches, dizziness, photophobia, tinnitus, nausea, and vomiting. Additionally, CSF leakage through the dural defect can increase the risk for pseudomeningocele development, meningitis, arachnoiditis, and epidural abscess formation. Although data on the incidence of durotomy in revision lumbar surgeries are limited in the literature, there are a number of studies outlining the rates of dural tears during an index lumbar spine procedure. In a recent systematic review of the literature, Ghobrial et al. compared the rates of durotomy in minimally invasive and open lumbar degenerative spinal surgery procedures. The authors found the durotomy rate to be 8.11% in open procedures versus 6.78% in minimally invasive procedures. There is some concern that minimally invasive surgeries have less visualization of the operative field, and as such, the incidence of dural violations may be underestimated in the reported rates during these procedures. When Ghobrial et al. compared prospective versus retrospectively performed studies, the authors noted a durotomy rate in the prospective studies that was double that reported in the retrospective studies (9.57% vs. 4.32%). In an analysis of a multiinstitutional, prospectively maintained data registry, Adogwa and colleagues documented a 4% rate of incidental durotomy among 1741 patients who underwent primary lumbar fusion for LBP and/or radiculopathy.
Durotomy can result in neurological symptoms attributed to the presence of neural tissue in proximity to the dural sleeve, although studies differ in the reported rates of symptomatic complications as a result of durotomy. In the aforementioned study by Adogwa et al., no differences in the rates of symptomatic neurological damage, postoperative infection, or need for reoperation were observed in the group of patients in whom incidental durotomies occurred. Wang et al. also similarly noted that incidental durotomy conferred no increased risk of neural damage in their series on 641 patients who underwent lumbar decompressive surgery.
The risk for durotomy is reportedly higher in the literature in revision lumbar surgery cases in comparison with index procedures. This is no doubt owing to the presence of adhesions in the epidural space, as well as dural scarring and fibrosis that make intraoperative navigation and/or dissection more difficult during revision surgeries. In their series on 553 patients who underwent posterior and posterolateral decompressive and reconstructive lumbar spine surgeries, Kalevski et al. reported the rate of durotomy to be 28.6% in reoperation cases compared with 12.66% for all cases combined. In another study in the United Kingdom by Tafazal and Sell on prospectively collected data for 1549 lumbar spine surgery cases, the authors documented a 13.2% rate of durotomy following revision lumbar discectomy; the rate of durotomy in index cases, however, was substantially lower (3.5%). The management of durotomy reported in the literature varies widely with respect to whether primary repair of the defect is performed, whether a subfascial drain is placed, and whether early ambulation or mandatory bed rest is preferred in the postoperative period. In the next section, we outline our approach to the management of durotomy in revision lumbar surgery procedures, as well as our general considerations in these cases.
Our Approach to Revision Lumbar Surgery (and Management of Durotomy)
Proper patient positioning is an essential component of revision lumbar spine surgery. At our institution, we use the Jackson Spine Table (Mizuho Corp., Tokyo, Japan) to allow for maximal extension of the hips to induce as much lumbar lordosis as possible in these operations. In conjunction with muscle release during the surgical dissection, this positioning enables us to avoid flatback syndrome and achieve correct sagittal alignment postoperatively. Additionally, this limits the corrective force that must be applied to any instrumentation, thereby reducing the risk for screw and rod failures.
Sharp, up-biting curettes of various sizes are used to identify the bone-soft tissue interface at the edges of the prior decompressive surgery. Our approach is from the outside in (lateral to medial) rather than from top to bottom (rostral-caudal); we prefer to expose bone laterally and extend the dissection medially rather than to go in a rostral-to-caudal direction. We have found that this can be done more quickly, requires less soft tissue release (potentially limiting adjacent segment issues in the future), and enables smaller skin incisions. We recognize that some spine surgeons do not want to expose the disc space at levels that have been previously operated on. Our general strategy is that if there has been a previous interbody device placed, we often do not go back into that disc space, even if they are not fused at that level (unless there are dramatic degrees of lucency around the interbody device). If there are no devices in the disc space or if there was not a prior interbody fusion performed at all, we will make extensive use of interbody devices. Curettes and Kerrison rongeurs are used liberally to expose the involved disc spaces.
The previous surgery may have had a dural violation that may have become evident at the time of exposure or at the time of reviewing the preoperative imaging studies. Alternatively, there may not have been a previous violation, but we may cause one with our dissection when working in a heavily scarred region. Clearly, the ideal method is to avoid a dural violation with careful sharp dissection. When we do get a dural violation, we attempt to close it primarily with a watertight dural closure. In general, this means that we need to remove enough bone around the violation to get an adequate exposure to close the defect. We typically use the operating microscope for dural repairs. This has the benefit of improved lighting and visualization for the operating surgeon as well as the surgical assistant. Once we achieve a dural closure using a 4-0 Nurolon (Ethicon Inc., Somerville, New Jersey, USA) suture, we test the adequacy of the dural repair. We have the anesthesiologist perform a Valsalva maneuver to confirm that there is no continued leakage of CSF. We continue with the surgery once this has been accomplished. If we are changing instrumentation (old for new), or adding other levels, we will proceed with this. We typically secure the screws to each other using rods or plates. If we are placing interbody cages during the case, we do that at this time. In our practice, we routinely perform bilateral-lateral fusions in revision surgery cases. Before placing the bone graft, we copiously irrigate the operative field (with a minimum of 3 liters of normal saline) to improve hemostasis, remove debris, and decrease the risk of a postoperative infection. We do this after instrumentation has been secured and before placing the bone graft into the lateral gutters.
We then go back to midline after this and reassess dural closure with another Valsalva maneuver. If there is additional CSF leakage, we determine whether additional stitches can be placed to correct the defect. After we have done that, we repeat the Valsalva maneuver. We then place a dural graft (typically composed of a collagen matrix) such as DuraGen (Integra LifeSciences Corp., Plainsboro, New Jersey, USA) and then cover it with Evicel fibrin sealant (Ethicon Inc., Somerville, New Jersey, USA). Evicel has the added benefit of assisting with hemostasis.
Finally, a layered closure is performed; in our practice, we specifically close the layer of muscle below the fascia separately using interrupted 1-0 Vicryl (Ethicon Inc., Somerville, New Jersey, USA) sutures. We then perform interrupted figure-of-eight stitches using the 1-0 Vicryl suture to close the fascia. Twice as many stitches are placed when dealing with a CSF leak than in cases without one. A 2-0 Vicryl is then used for the subcutaneous and dermis layers followed by staples on the skin. In scenarios where we cannot achieve an adequate dural closure, we place a dural graft with Evicel and place a lumbar subarachnoid drain through which we divert CSF flow for 2 to 5 days. This drain is then clamped for 24 hours and if no clinical symptoms are present or CSF leakage occurs following clamping, the drain is removed. In our practice, we have found that the use of the lumbar drain in cases refractory to dural closure can be greatly beneficial. In addition, if we identify a CSF leak in the postoperative period that was not seen during the surgery, we make liberal use of CSF diversion using a lumbar drain as the initial treatment. If lumbar drainage is not successful in halting the egress of CSF, then a revision surgery may be necessary to attempt a direct dural repair primarily or, in rare cases, use of a dural graft may be employed to achieve a satisfactory dural closure.
Conclusion
Revision lumbar spine surgery may be necessary in a subset of patients following their index procedure depending on the initial pathology that was treated as well as the development of different pathological processes that can occur secondary to the initial surgery. Given that this second procedure is often more technically challenging to perform and may be associated with inconsistent clinical outcomes, careful selection of candidates for revision surgery is essential. A thorough history and physical examination should be obtained and performed, respectively, with an emphasis on attempting to identify the pain generator. A combination of conventional radiography, CT scanning, MRI, and myelography may be used in the diagnostic work-up of these patients. Once a candidate has been selected for revision surgery, medical optimization before surgery is a key component of the process. There is an increased risk for durotomy in the revision procedure, which has been described in the literature. In our practice, we emphasize careful closure of any dural defects encountered intraoperatively with the adjunctive use of dural grafts and fibrin glues. We have observed that a lumbar subarachnoid drain can be used postoperatively to assist in cases where an adequate dural closure could not be obtained intraoperatively. When the dural violation is successfully managed either intraoperatively or in the postoperative period, we have observed good long-term outcomes with rare complications.
Illustrative Case
A 58-year-old African-American male presented for a neurosurgical evaluation 5 years after a work-related injury left him with multiple extremity fractures and a low back injury (herniated L4–L5 disc). Two years after the injury, he underwent a minimally invasive surgery with an L4–L5 microdiscectomy without any relief in his symptoms. At the time of his neurosurgical evaluation, he complained of persistent LBP and right lower extremity (RLE) radicular pains as well as weakness and numbness in the RLE. Motor power was 4+/5 in the distal RLE with diminished pinprick sensation and a positive straight-leg raising test. A CT scan and an MRI from 3 months before this evaluation showed a right L4 laminotomy defect with degenerative disc changes. Plain film radiographs demonstrated modest (3 mm) motion at the L4–L5 level on dynamic flexion-extension films. Having undergone a full year of conservative treatment after his index surgery, he demonstrated no clinical improvements. A discussion about surgical and further continued conservative treatment was held. The patient had a previous surgical opinion from his original surgeon who recommended revision surgery. Following the author’s second opinion, which also suggested a revision surgery, he chose this option and requested that the author perform the surgery. An open L4–L5 bilateral transforaminal L4–L5 interbody fusion was performed. His surgery and postoperative course were unremarkable. In the postoperative period, because of persistent pain he sought additional treatment from a pain management specialist who performed an intervention that resulted in a CSF leak. An MRI demonstrated the leak in the midline at the L3–L4 level ( Fig. 6.1 ). Severe postural headaches followed which did not respond to two attempts at a blood patch. Additional imaging demonstrated findings consistent with a pseudomeningocele on CT myelogram ( Fig. 6.2 ). Lacking a response to alternative treatments, a revision surgery was planned and performed. A midline durotomy was identified at the L3–L4 level which was repaired primarily with a watertight dural closure. This revision surgery was successful with improvements in the symptoms related to the CSF leak and gradual continued improvements in his back and leg pains. At his 2-year follow-up evaluation, he had resumed gainful employment and was pleased with his clinical outcome. Plain film radiographs with dynamic flexion-extension views demonstrated a solid radiographic fusion with good sagittal alignment (lumbar lordosis [L1–S1] 50 degrees), with bridging trabecular bone in the intervertebral interspace and in the lateral gutters, no motion on dynamic imaging views, and no evidence of any lucencies or other instrumentation issues ( Fig. 6.3 ).
