Failed back surgery (FBS) is a nonspecific term that implies the outcome of spine surgery did not meet the expectations established before surgery of the patient and the surgeon. It does not mean that the patient failed to get total pain relief or return to full function. Such expectations are not reliably attainable. It might be argued that when the patient is not satisfied, but the surgeon feels the outcome meets expectations, it is a failure of communication rather than the surgery.
There can be wide discrepancies between patient and surgeon with regard to expectations of outcome with patients usually having higher expectations than their surgeon.1 Factors associated with higher patient expectations of outcome include younger age, not widowed, prior chiropractic care, poorer function as measured by Oswestry Disability Index (ODI), and worse mental health score.2
A surgeon’s expectations for outcome in a specific patient should be based on published medical evidence, the nature of the structural problem, the number and types of prior surgeries, the psychological health of the patient, and the skills and experience of the surgeon. When the surgeon communicates reasonable expectations, there is greater chance the patient will be satisfied if those expectations are met.3
There are almost too many options for the treatment of patients with FBS. It seems logical that the best outcome will occur when treatment is matched to each specific patient’s structural, neuropathic, or psychological cause. This often requires a multidisciplinary approach. To be prepared to treat these challenging patients, physicians should know the reasons spine surgery might fail, the structural causes of FBS, and the best treatment options for each situation.
Causes of Failed Back Surgery
It has been said that the best surgical outcome occurs when the right surgeon does the right surgery for the right problem on the right patient at the right time. There is a significant risk for FBS when any of these rights goes wrong.
There are several studies that have looked at the structural and neuropathic causes of FBS.4,5,6,7 Clinically, it is most important to determine the structural cause for the pain whenever possible. If a structural etiology is identified, it is useful to decide whether the cause is a residual, recurrent, or new problem. Residual or recurrent structural pathology may be due to inadequate evaluation before surgery or mismatch of the surgery needed versus the surgery performed, unrecognized complications, or technical failure. New structural problems are those that developed after surgery, which might or might not be a consequence of the surgery itself.
Another model is based on the time course of the residual, recurrent, or new pain after surgery. Patients who never improve or who deteriorate in the first 4 weeks or so after surgery are likely to have residual pathology, a complication, or a technical failure. Most of these patients will still be under the care of their surgeon rather than a pain medicine specialist. Patients who get somewhat better initially but then deteriorate may have developed instability, instrumentation failure, recurrent disk herniation, or delayed infection. Those who get better but then deteriorate after 6 months or so might have new pathology at the same or adjacent segment, which can include for example facet or sacroiliac joint (SIJ) pain or pseudarthrosis.
MISMATCH: SURGERY NEEDED VERSUS SURGERY PERFORMED (“WRONG SURGERY”)
When a surgeon chooses an operation that is not appropriate for the structural problem, the operation is likely to fail. It might prove useful, in this model, to appreciate the distinction between surgery for leg pain as opposed to surgery for back pain. True radicular (rather than referred) leg pain is usually caused by neural compression. Surgeries aimed at improving leg pain include decompression of involved neural structures in the neural foramen, central canal, or by a disk herniation. Assuming facet and SIJ source of pain have been excluded, surgeries directed toward improving axial low back pain (LBP) include fusion for pain arising from a disk (discogenic pain) or instability, for example. It follows that if a patient had predominantly axial LBP but had a decompression or discectomy without fusion, it might not have been the “right surgery” for the clinical problem, and there is a high likelihood of poor result. In other words, there was a mismatch between the clinical problem and the surgery. The surgeon performed a “leg pain operation” for LBP.
A second type of mismatch occurs when the surgery does not address all the patient’s pathology. For example, there are patients who have pathology such as disk degeneration, disk herniations, spinal stenosis, or combinations at multiple motion segments. The surgeon might have elected to operate on only the worst segment or segments, thereby leaving the patient with residual problems at the adjacent segments. This type of patient might have required multilevel surgery or, if the number of levels is excessive, perhaps no surgery at all.
Another example is the patient with both central and foraminal stenosis. The surgeon achieves decompression of the central canal but does not achieve adequate decompression of the lateral canal. The patient may be left with leg pain due to remaining foraminal stenosis, which might incorrectly be thought to be neuropathic pain. Yet, another example is the patient with severe foraminal stenosis. A surgeon may decide to perform a limited decompression for fear of causing instability that might necessitate fusion. As a result, there is inadequate decompression and the patient does not get better. The patient needed decompression and fusion but only had decompression. This is another mismatch between the surgery needed and the surgery performed.
Another example of a mismatch between the surgery needed and the surgery performed occurs when the surgeon fails to fully consider the effect of surgery on the motion segment. For example, consider a patient with spinal stenosis and very slight spondylolisthesis and spinal stenosis who has leg pain. If the surgeon chooses decompression without fusion, many times, the spondylolisthesis progresses and the patient develops progressive LBP.
Finally, and not to be minimized, some spinal pathology is not treatable by surgery. An example would be four or five levels of painful disks.
If a patient has not been fully evaluated, it is more likely that the surgery will fail. Surgery performed after an incomplete evaluation might leave significant structural pathology unattended. Incomplete evaluation is often due to over reliance on imaging studies, particularly magnetic resonance imaging (MRI), without fully appreciating the information from the history and examination. Many spine specialists use confirmatory diagnostic spinal injections to complement the other information. Plain radiographs done standing with flexion and extension views can be very valuable to disclose spondylolisthesis or deformity that was not seen on supine MRI.8 Pain relief after transforaminal epidural injection would suggest any foraminal stenosis seen on MRI is in fact the pain generator.9,10,11 Discography had been used frequently but is rarely used now due to the possibility of creating long-term degeneration of otherwise normal disks coupled with the great improvements in the quality of MRI scans.12
Psychological health is one of the most important variables with respect to the outcome of surgery.13,14 Therefore, psychological evaluation is especially important for patients (“right patient”) with long-standing pain and impairment or when the surgeon has any inkling that there might be a significant psychological problem. Neither surgeons nor physiatrists perform well in identifying patients with psychological illness compared to objective psychological testing and so referral is often in the best interests of all.15
Most psychological illness in the postoperative patient was present before surgery but not considered or was underestimated. Many patients with long-standing chronic LBP have some degree of psychological illness, some mild but some more severe, which might have altered surgery decision making.13 The most common problems described are depression, anxiety disorder, and substance use disorder.14 Although these disorders may not be a contraindication for surgery, treatment before surgery and psychological follow-up afterward might prove useful to increase the chances of a good result. In addition to these familiar illnesses, there is growing evidence that a patient’s coping abilities, fear and fear-avoidance behavior, and history of sexual or physical abuse may play important roles in continuing pain, impairment, and disability and contributing to a poor outcome.16,17,18,19,20,21,22,23
COMPLICATIONS
Many of the complications of spinal surgery occur early in the postoperative period and should have been recognized by the surgeon. Infection usually occurs early but occasionally does not appear for weeks or months. Misplaced pedicle screws can cause new leg pain, usually in a single dermatome, and is often present immediately after surgery but can occur slightly later. Neural injury during surgery has similar symptoms and appears right after surgery. Complications that occur or are discovered later include facet or pedicle fractures; pedicle screw misplacement; and bone graft collapse, resorption, or dislocation (after interbody fusion). In addition, surgery might have been performed at the wrong level.
Pedicle screws can cause LBP by any of several mechanisms. There can be pain over the screws, sometimes attributed to chronic irritation of overlying soft tissues and bursa formation. This usually is seen months after surgery. Pedicle screw misplacement can cause leg pain if a screw breaches a pedicle and irritates or injures a nerve. This can be mistaken for neural injury due to surgical trauma. Therefore, if there is leg pain in a single dermatome, it is necessary to obtain computed tomography (CT) scan to see if there has been even slight breach of the pedicle.
TABLE 75.1 Differential Diagnosis of Some of the More Common Extraspinal Causes of Failed Back Surgery by Symptoms, Signs, Radiology, and Injections
Diagnosis
Symptoms
Signs
Radiology
Injections
Hip
Groin pain
Limp; limited hip internal rotation
Standing x-rays of pelvis + hips
Relief with intra-articular anesthetic
Trochanteric pain syndrome
Lateral thigh pain
Tenderness over GT and muscle insertions
Not helpful
Relief with injection of GT and muscle insertions
Nerve entrapment or neuroma
Pain in peripheral nerve distribution
Allodynia, + Tinel sign
MRI, EMG, and nerve conduction
Relief with anesthesia of affected nerve
Peripheral arterial insufficiency
Arterial claudication
Poor pulses
Positive arterial studies
N/A
EMG, electromyography; GT, greater trochanter; MRI, magnetic resonance imaging; N/A, not applicable.
Pseudarthrosis is a failure of fusion. Some patients with nonunions have pain, but others do not. Therefore, one cannot assume that the nonunion is the cause of the pain. Plain radiographs are not reliable to show if a fusion is solid. However, if standing films with sagittal flexion and extension views show motion, it would indicate nonunion. The most useful test is a CT scan that includes reformatted curved coronal sections that are taken out to the tips of the transverse processes in addition to the usual sagittal and axial images.24,25 CT also allows visualization of the anterior column to look for lucency surrounding an interbody fusion device.
Some patients undergo surgery and then develop instability, defined as greater than 3 mm translation on standing flexion/extension x-rays. It is not uncommon to see patients with very slight spondylolisthesis before surgery for disk herniation or spinal stenosis in whom no fusion was done because the slip was so slight. Then some months after surgery, the back pain worsens, and plain x-rays reveal progression of the slip.
TECHNICAL FAILURE
Even though the surgery was appropriate for the symptoms and pathology, the surgeon may not have accomplished the technical goals. Technical failure is usually apparent on good-quality imaging studies. Inadequate decompression of a foramen is not uncommon. There may be incomplete removal of a disk herniation, especially if it was a far lateral herniation. There may be misplacement of screws or incorrect connection of internal fixation.
RESIDUAL PATHOLOGY
Spinal Pathology
Residual pathology implies that surgery did not correct all the structural problems that had been causing pain. Problems such as pain arising from facet or SIJ might not have been recognized. There might have been pathology at other levels in addition to the operated segment.
Extraspinal Pathology
Residual pathology also includes extraspinal disorders that can mimic spinal problems.26,27,28,29,30 When an extraspinal disorder is the cause of pain, it might have been present prior to surgery and not recognized or occurred afterward. Some of the more common problems that can mimic motion segment spine pain are shown in Table 75.1 and include primary hip disorders, SIJ pain, greater trochanteric pain syndromes (GTPS), and peripheral nerve injury or entrapment.26,27,28,29,30
Painful disorders of the hip region most often are independent of spinal pathology but, at times, can coexist, especially in older patients.30 This is a challenging problem because there is considerable overlap between their respective symptoms and signs. Brown et al.31 studied a referral population with leg pain to determine if there were signs or symptoms that might differentiate between osteoarthritis of the hip and a spinal disorder. The factors that were suggestive of primary hip pathology were the presence of a limp, groin pain, or limited internal rotation of the hip. Factors more suggestive of spinal stenosis were lateral thigh pain, buttock pain, and pain below the knee, particularly in the absence of groin pain. Weight-bearing radiographs of the hip can serve as an initial screening tool.
Disorders of the structures near the hip that can mimic spine pain include GTPS, subtle sacral fractures, gluteal tendonitis, hamstring syndrome, ischial bursitis, or tendonitis. Patients with GTPS typically complain of pain in the proximal lateral thigh, often with radiation to the distal thigh and occasionally below the knee.26,29 Pain is usually increased by lying on the affected side, climbing stairs, and transition from sitting to standing. Pain may arise from the bursa itself or from the gluteal medius and minimus.32 On exam, there is tenderness over the greater trochanter, but in the variants of GTPS, there may be tenderness at the local muscle insertion sites. Diagnosis is confirmed by relief of pain with the injection of local anesthetic into the bursa and nearby muscles.
Peripheral nerve trauma or entrapment can mimic radiculopathy.27 In the patient with FBS, the most relevant problems are lateral femoral cutaneous nerve entrapment or injury (meralgia paresthetica), which presents with pain in the lateral or anterolateral thigh, peroneal nerve entrapment, and sciatic nerve entrapment.27,28 With entrapments, pain will be in the distribution of the peripheral nerve involved, not in a true lumbar dermatome. There may be a positive Tinel sign over the area of pathology. Diagnosis can often be confirmed by nerve conduction studies, and there will be temporary relief of pain after injection with local anesthetic in the area of presumed entrapment.
RECURRENT PATHOLOGY
Recurrent pathology can recur even after perfect surgery. Disk herniations recur in up to 15% of patients after discectomy. Foraminal stenosis can recur after if there is progressive degeneration of the same motion segment.
NEW PATHOLOGY
Pain after prior spine surgery may be totally unrelated to the index surgery. The usual structural causes of LBP can occur after an index surgery.
Structural Etiologies of Failed Back Surgery
Despite the many reasons for FBS, there are a limited number of structural etiologies for the residual or recurrent pain. The studies that looked at the most common structural causes of FBS had very similar findings (Table 75.2).4,5,6,7 The most common structural causes of pain after surgery are foraminal stenosis, recurrent or residual disk herniation, one or more painful disks, facet joints (FJs), SIJ, and neuropathic pain.
There are four studies that reported on the causes of FBS in individual series. Unfortunately, one is from 1981 and two from 2002. The most recent study was in 2010 but only looked at patients with prior fusion. It does not appear that there have been more recent descriptions despite FBS being fairly common. Because of the age of the data, some considerations are necessary. There has been much greater recognition of neuropathic pain and pain arising from the SIJ pain and FJs. Surgical techniques have improved. There is greater recognition of foraminal stenosis. It is too early to know all the causes in patients who have undergone total disk replacement. That said, the data we have is what we have to work with, coupled with experience.
TABLE 75.2 Most Common Spinal Causes of Failed Back Surgery in Three Reported Studies5,6,7
Burton et al.4 in 1981 reported an analysis of several hundred patients with FBS. About 58% had foraminal stenosis, 7% to 14% had central canal stenosis, 12% to 16% had recurrent (or residual) disk herniations, 6% to 16% had arachnoiditis, and 6% to 8% had epidural fibrosis. Other less common causes in their series included neuropathic pain, chronic mechanical pain, painful disk above a fusion, pseudarthrosis, foreign body, and surgery performed at the wrong level. They were unable to establish a diagnosis in less than 5% of their patients, even though their patients were evaluated early in the CT scan era and well before MRI scans. They did use discography. In 1981, surgeons were less aware of foraminal stenosis and discogenic pain.
In a retrospective review of 181 patients with FBS seen at a tertiary care spine center, Waguespack et al.5 could make a diagnosis in 94% of patients. Slipman et al.6 also could make a diagnosis in about 89% of patients. Some patients had more than one primary diagnosis. DePalma et al.7 looked at the etiology of pain in patients who had lumbar fusions. The details are shown in Table 75.2
In the following section, I have used functional definitions of structural abnormalities that are a composite of those proposed by the North American Spine Society.33 The differential diagnoses of some of the more common causes of FBS along with some helpful symptoms, signs, radiologic findings, and response to injections are shown in Table 75.3.
FORAMINAL STENOSIS
Foraminal stenosis was found in 15% and 35% of FBS patients in the studies of Waguespack et al.5 and Slipman et al.,6 half of what was seen 25 years ago.4 The lower prevalence may be due to increased awareness of the problem, improved imaging studies, and/or better understanding of the need for meticulous decompression. Patients with foraminal stenosis have pain that is predominantly in the leg or buttock region, often in the distribution of a single dermatome. Pain is usually worsened by standing and walking and relieved by sitting. MRI or CT scan shows narrowing of the canal at the index level or an adjacent segment.
There are no data regarding the utility of selective nerve root blocks in FBS. There are suggestions that when performed with excellent technique, they can provide some information, especially perhaps when there is no relief, which might suggest the targeted root was not the cause of pain.10 According to one systematic review, there is good evidence that lumbar selective nerve root blocks can aid with the identification of one or more symptomatic roots.9 There is moderate evidence that identifying or excluding a root as cause of pain improves surgical outcome.9 Other systematic reviews did not find the evidence convincing.11
TABLE 75.3 Differential Diagnosis of Common Causes of Failed Back Surgery by Symptoms, Signs, Radiology, and Injections
Diagnosis
Symptoms
Signs
Radiology
Injections
Lateral canal stenosis
Leg pain > LBP; relief with sitting
Loss of lumbar lordosis
MRI: foraminal stenosis
Relief with transforaminal epidural
Painful disk
LBP; ? worse with sitting
Restricted flexion in standing
MRI: degenerated disk(s)
No sustained relief
Neuropathic pain
Leg pain burning
Dysesthesia
Hypoalgesia
Allodynia
No alternative diagnosis
+/- relief with sympathetic block
Facet syndrome
Left or right LBP
? Facet tenderness
Not specific
Medial branch block relieves pain
Recurrent HNP
Vary with location; leg pain > LBP
Variable
HNP on MRI
Epidural may provide temporary relief
SI joint pain
Gluteal pain with referral to leg and groin
May have + provocative testing
Not helpful
SI joint injection relieves pain
HNP, herniated nucleus pulposus; LBP, low back pain; MRI, magnetic resonance imaging; SI, sacroiliac; +/-, may be helpful.
PAINFUL DISK (DISCOGENIC PAIN)
Pain that arises from within a disk is often referred to as discogenic pain. One or more painful disks was found to be the cause of FBS in 25% to 30% of patients.4,5,6,7 Painful disks can occur at the level of prior surgery, at an adjacent segment, or rarely at the index level despite prior posterolateral fusion.34,35 When there is a residual painful disk at the index surgical level, it is likely that it was not treated adequately by performing a fusion, especially an interbody fusion. When there is a painful disk at an adjacent segment, it was either present prior to surgery and not addressed or the disk degenerated after surgery.35 Risk factors for adjacent segment degeneration include body mass index >25 kg/m, preoperative disk degeneration, and superior FJ violation during surgery
Although there is no totally consistent symptom complex for discogenic LBP, especially after surgery, the diagnosis is more likely when there is dominant midline LBP that might radiate to the left and right of the midline, the gluteal regions, and often to the leg in a nondermatomal fashion. Pain is usually worse sitting and during transition from sitting to standing. It may improve with standing or walking. Physical examination is not specific. There may be decreased flexion in standing due to pain. There may be tenderness over the spinous processes but not over the FJs.
When the diagnosis of discogenic pain is suggested by history and examination, MRI shows a single degenerated disk, and other potential causes of chronic LBP have been excluded, it is likely that the diagnosis is correct.
DISK HERNIATION
Recurrent or residual disk herniation was seen in 7% to 12% of patients with FBS.4,5,6 There are two common presentations of pain from a disk herniation: radicular and axial. The topography of the pain is primarily due to the location of the herniation. A posterolateral herniation is more likely to compress or irritate a nerve root and therefore present with predominant leg pain. A midline herniation, unless very large, does not compress neural elements and presents with predominant LBP. A disk that is degenerated and herniated can cause both leg pain and LBP. In the presence of epidural or perineural fibrosis, a disk herniation may cause more leg pain than expected than if there were no fibrosis.
Only gold members can continue reading. Log In or Register to continue