Headache & Pain
Neuroimaging of headache
Dec. 02, 2023
At vero eos et accusamus et iusto odio dignissimos ducimus qui blanditiis praesentium voluptatum deleniti atque corrupti quos dolores et quas.
Back pain remains an evolving and controversial topic in modern medicine and a multidisciplinary field that involves specialists in neurology, neurosurgery, orthopedic surgery, rheumatology, physiatry, and occupational medicine. In this update, the author highlights the emergence of facet replacement.
• Lumbar pain is multifactorial and entails great societal costs.
• Conservative care including antiinflammatory therapy and physical therapy should be the first treatment unless trauma, infection, or tumor is suspected.
• Surgical approaches to back pain are controversial unless overt instability is apparent.
Spinal disease was first diagnosed in The Edwin Smith Surgical Papyrus, which dates from the seventeenth century BC. The papyrus refers to lumbar fractures and the pain associated with them (13). Back pain, or low back pain, remains a common complaint of many patients. The annual incidence of back pain is estimated at 15% to 20% along with a prevalence of up to 60% (05). Pain is complex, with symptoms being variable and nondescript. Back pain is currently the fifth most common cause of patient visits, associated with costs approximating 38 to 50 billion dollars a year (05). Within the group of patients with back pain, 1% have been diagnosed with nerve root symptoms, and it has been found that the vast majority of patients with back pain will find resolution of this pain without any treatment (41). A patient’s symptoms may flare up periodically, becoming intolerable at times, causing the patient to seek treatment (27). Patients usually present to their primary care physician but are commonly referred to see a neurologist, neurosurgeon, or orthopedic surgeon. When back pain persists, and if neurologic symptoms develop, the physician must delve into the cause and begin a thorough diagnostic workup. Persistent low back pain should never be ignored, as it may be the first sign of a serious underlying process, ie, cancer.
In early years, pain was treated differently, and the management of acute pain is a newer concept (39). In fact, back pain may not originate from the spine but may be referred from hips or abdominal viscera. These characteristics of back pain may make the true diagnosis and treatment difficult at times.
When discussing back pain, specific terms need to be defined to help understand the etiology, diagnosis, and treatment. First, back pain may be mechanical or simply stated as musculoskeletal pain. A patient might have radiculopathy, which refers to pain in the distribution of a nerve root with associated weakness and loss of reflexes. In some cases, a patient may state that they have “sciatica,” which to the physician often refers to radiculopathy. Myelopathy refers to compression of the spinal cord; these patients often have upper motor neuron signs like spasticity and hyperreflexia. Myelopathy will clue the physician in to a cervical or thoracic lesion because the spinal cord ends at L1/2; these patients may still present with back pain. A combination of mechanical back pain with radicular and myelopathic symptoms is not uncommon. Spondylosis is a generic term for stenosis secondary to bony and ligamentous hypertrophy. Spondylolisthesis refers to anterior subluxation of one vertebral body on another. Finally, spondylolysis is failure of the pars interarticularis to form, resulting in a spondylolisthesis, classically at L5/S1 level.
Thorough history taking is key during the initial evaluation because patients will often describe their pain differently. The pain may be described as sharp and shooting or dull and aching. Symptoms may persist at all times without relief or may wax and wane. The pain may also be exacerbated by certain activities or just lying down. Patients may explain that certain positions relieve the pain or anti-inflammatory medications are required for relief. The onsets of symptoms are important in making a diagnosis. Low back pain that has occurred after an accident requires a different treatment algorithm as opposed to back pain that has gradually worsened over months. Pain often has triggers that are usually related to position or activity. Patients will often be able to localize the pain, but many will tell their physician that they have “sciatica.” The physician must clearly define what the patient means by “sciatica.” Generic terms should be avoided and specific terms are always preferred (pain in big toe or over the paraspinal muscles at a specific vertebral level). A history of steroid use, skin infections, and IV drug use is important to elicit because it helps identify possible etiologies. Back pain isolated to the spine often implies musculoskeletal pain, whereas back pain combined with pain down one or both legs implies nerve root compression. Some symptoms are pathognomonic. Patients who describe low back pain that is worse with walking and relieved in a sitting or recumbent position may have a diagnosis of neurogenic claudication (lumbar stenosis) (80). As patients age, there will be a certain degree of inevitable aches and pains. This alone is normal and patients should be reassured and treated expectantly.
Some patients may present with symptoms that are often associated with a progressive course that is widespread and fails to respond to conservative treatment. These symptoms include inconsistent straight leg raise test, superficial or widespread tenderness, axial loading pain, sensory and motor findings without a specific anatomic distribution, or overacting during examination (84). Any patient that has back pain with bowel and bladder dysfunction deserves immediate imaging, MRI being your first choice and myelogram a reasonable second option. Cauda equina syndrome, if diagnosed early, is readily treatable with patients having excellent outcomes. Problems with bowel or bladder dysfunction should always be elicited and never ignored when evaluating for back pain.
Back pain is typically a self-limiting disease. The vast majority of back pain patients, more than 95%, will have resolution of symptoms by 4 weeks. Symptoms that persist beyond this time frame deserve further evaluation. Patients often need reassurance that their pain will improve and likely resolve. As with any medical condition, education remains one of the most effective treatment tools. The physician should dispel rumors and present facts clearly. When in doubt, referral to a neurosurgeon, neurologist, or orthopedic spine surgeon should be done.
An obese 55-year-old woman experienced a sudden onset of severe lower back pain while picking up heavy boxes. She began taking antiinflammatory medications over the following week with minimal symptomatic relief. Therefore, she decided to go to the local emergency room for evaluation of her persistent pain. AP and lateral lumbar x-rays were taken, which failed to reveal any compression or fracture. Dynamic x-rays with flexion and extension views were also taken but failed to demonstrate any movement. On physical examination, she had full strength in all four extremities along with sensation intact to light touch and pinprick. Her reflexes were normal, with toe flexion bilateral and absence of clonus. Pain was elicited with palpation of the paraspinal muscles. She was discharged with instructions to continue anti-inflammatory pain medications and limit activity for 3 to 4 days. Over the next 3 weeks, her symptoms continued to bother her and she went to see her primary care physician in the office. Her examination was unchanged from her prior emergency room visit, and an MRI was obtained that demonstrated mild lumbar stenosis and disc bulges without any nerve root compression. The patient was eager to have a complete workup and underwent lumbar discography without any reproduction of her symptoms. Based on the negative workup, she was reassured and instructed to continue her medications as before. At this time, she was also fitted for a lumbar corset that she could wear as needed for comfort. Over the next 4 months, her pain improved and she was able to resume all of her daily activities by 6 months postinjury.
The physical examination will only compliment the history but often helps localize the source of pain. The patient should change into a gown and be standing in order to evaluate spinal symmetry. Careful attention to evaluating for fasciculation is important. Palpation of the paraspinal muscles and spinous process must be performed. The patient’s range of motion must be tested in all directions: flexion, extension, lateral bending, and rotation. The examiner should try to elicit a Patrick sign. This involves internally rotating the symptomatic leg to differentiate hip pain from radicular pain (06). A positive test, pain with rotation, localizes the source of pain to the hip. Attention should then be turned to examining muscle bulk and tone, with signs of atrophy being noted. Each muscle group should be examined individually. It is best to start proximal and work distally, going through each compartment’s flexor and extensor muscles. Sensation should be checked to both light touch (dorsal column) and pin prick (anterolateral column). The examiner must try to distinguish sensory changes related to a nerve root distribution from a peripheral neuropathy. Reflexes at the knee and ankle are important. Brisk reflexes with clonus and a Babinski sign help localize problems above the level of the lumbar spine, whereas absent reflexes on one side will help localize to a specific lumbar nerve root. Gait should also be tested, which is an excellent way to look for a foot drop.
To perform the straight leg raise test (“Lasegue sign”), the clinician lifts the extended leg of the supine patient (06). A negative test demonstrates no pain up to 90 degrees. The test is positive if pain is elicited at 30 to 45 degrees. A positive test is consistent with nerve root irritation or sciatica. If the contralateral leg is raised and similar symptoms occur, the specificity of the test is much higher, confirming nerve root irritation. Sometimes the exam and history are disconcordant and malingering may be suspected, ie, unilateral leg paralysis. The Hoover sign helps sort this situation out. The patient lies supine, and as they raise their normal leg, the other leg presses down. Any downward pressure in the affected leg is normal and rules out leg dysfunction (06). This reliable test should be used in concordance with a good physical exam and history.
Sacroiliac joint pathology may be diagnosed by eliciting tenderness over the joint to deep palpation (Fortin finger test). Other confirmatory tests include the FABER test in which the involved leg is flexed, abducted, and externally rotated and the Gaenslen test in which the involved thigh is extended while the contralateral leg is placed in forced flexion, pelvic distraction, thigh compression, and thigh thrust. A positive Fortin test as well as at least three of the above five provocative tests is strongly predictive of clinical sacroiliac disease. Steroid injections can complete the confirmation of the diagnosis.
Anatomy. There at five lumbar vertebrae, each separated by a fibrocartilaginous disc. The disc has two parts: the soft inner portion called the nucleus pulposus, and the sturdy outer portion known as the annulus fibrosus. Innervation to the disc is provided via the nerve roots once they exit their respective foramen (branch to the annulus fibrosus). There are five sacral vertebrae, which are fused without an interposed disc. The lumbar vertebral bodies are the largest in the body, with L5 being the largest. The vertebral bodies articulate with each other anteriorly via the disc and posteriorly via the facets. The superior facets are directed posteromedially and inferior facets are directed anterolaterally. The facet joints are also known as the zygapophyseal joint. Innervation to the joint is provided via the articular branches of the dorsal nerve root once it exits the nervous foramen. The anterior longitudinal ligament (ALL) runs anterior to the vertebral bodies, whereas the posterior longitudinal ligament runs anterior to the thecal sac and posterior to the vertebral bodies. Posterior to the thecal sac is the ligamentum flavum, or yellow ligament, which extends from lamina to lamina. Between each spinous process extends the interspinous ligament and supraspinous ligament. The main muscles of the back are the erector spinae that lie lateral to the vertebral column and the transversospinal muscles. The erector spinae, the main extensors of the back, are divided into three columns: iliocostalis-lateral column, longissimus-intermediate column, and spinalis-medial column. The multifidus, the main transversospinal muscle, helps stabilizes the spine during local movements. Sensation of these muscles comes from dorsal rami of spinal nerves at multiple levels (55).
Pathogenesis. Pain must be differentiated between acute, subacute (more than 6 weeks), and chronic (more than 3 months) pain. Acute pain often occurs early on, within days of a disease process, and may alleviate just as quickly with or without treatment. Chronic pain occurs over months to years and may never fully resolve even with proper treatment. Acute pain often responds well to pain medications, whereas chronic pain may be intractable.
The intervertebral disc will also age along with the patient. A healthy disc is noncompressible and will resist most forces. As the disc ages, it will lose water content and height. Normal wear-and-tear of the disc may lead to annular fissures and possible rupture of the inner nucleus pulposus, resulting in compression of nerve roots. Along with disc disease, the bone and ligaments may also be affected. Bones form osteophytes and ligaments hypertrophy, both of which may compress the nerve roots. Typically, a small amount of epidural fat is present, but in morbidly obese patients, significant amounts of epidural fat may be deposited (epidural lipomatosis), leading to nerve root compression and sometimes cauda equina syndrome. Once a nerve becomes compressed, demyelination may occur. If untreated, the nerve cells may die, leaving the patient with a permanent sensory and motor deficit (59).
Bacterial infection. Some reports have suggested a relationship between back pain and bacterial infection. Albert and associates performed a clinical study of 61 patients with no history of steroid infection or prior surgery undergoing lumbar disc surgery and found positive cultures for anaerobic infection in 43% of cases, with another two patients having aerobic or mixed positive cultures (02). Eighty percent of the patients with positive anaerobic cultures developed degenerative endplate changes after surgery compared to 44% of patients without these findings. As Modic changes are associated with back pain, the authors speculated a role in developing back pain with chronic colonization. Chou and colleagues looked specifically for Propionibacterium DNA in patients undergoing surgery and found a statistically significant increased incidence of infection seen in discs with an annular tear and loss of disc height (95). Two subsequent systemic reviews noted this relationship and concluded that there was moderate evidence of a relationship, but that larger prospective studies were necessary before the relationship could be confirmed (36; 81). Propionibacterium was the most common organism cultured.
Cytokines. Some reports have investigated the potential role of inflammatory cytokines in the development of inflammatory Modic changes and back pain. Altun evaluated 49 patients with an average age of 38 years with tissue obtained during discectomy for either acute or chronic disc herniations (04). He found higher levels of cytokines such as tumor necrosis factor alpha and interleukins in acute rather than chronic disc herniations. Conversely, Schroeder and colleagues looked at inflammatory cytokines in patients undergoing fusion procedures for degenerative disc disease and found a significant increase in interleukin 1 beta, granulocyte macrophage colony-stimulating factor, and ENA-78 in patients with active Modic change (71). Wang and colleagues and Yang and colleagues postulated a potential therapeutic role in inhibitors of tumor necrosis factor alpha and interleukin 1 beta for patients with degenerative disc disease (92; 86).
The socio-economic impact of lumbar spine disease is well recognized. Healthcare costs are only one aspect of this impact, which includes time loss from work and disability claims both appropriate and fraudulent. The overall cost is shared between time loss/work impact costs (56%) and medical costs (44%), with medical costs fairly equally divided between primary care, specialty care, inpatient costs, pharmacy costs, and physical therapy (23). Weir and colleagues looked at a large group of over 10,000 adults undergoing spine surgery in the United Kingdom. Approximately 20% of patients met diagnostic criteria for postoperative back pain issues at an added cost per patient of £5383 over 2 years, £10,195 over 5 years, and £14,318 over 10 years. This was felt to cost the United Kingdom Health System over £70 million pounds in 10 years (88).
Socio-economic factors certainly influence the success of treatment modalities as well. In a meta-analysis of 31 studies, Cheriyan and associates found a tremendous impact of work status on outcome in spine surgery (20). Patients without a secondary gain issue returned to work 83% of the time, whereas the secondary gain patients returned to work only 57% of the time. The overall risk ratio for unsatisfactory outcome was 2.12 and was similar when comparing American studies to papers from Europe and Australia.
Back pain has a multitude of causes. No clear-cut understanding of how and why back pain occurs is known. Diagnostic imaging, though, has offered up a greater understanding of the specific cause of a patient’s symptoms. Torn muscles and ligaments will often give acute back pain that resolves. Disc degeneration may cause symptomatic back pain with nerve root symptoms. Arthritis of the lumbar spine may narrow nerve root foramen as well as the canal itself, causing radicular or myelopathic symptoms. The following list attempts to break down possible causes of back pain (41):
- Disc degeneration or black disc disease
• Soft tissue abscess
• Extradural (55%)
- Sarcoma (bone or soft tissue; 3% to 4%)
• Intradural, extramedullary (40%)
• Intradural, intramedullary (5%)
• Dural AV-fistula (often antecedent trauma, progressive)
• Ankylosing spondylitis (HLA B27 in 95% of patients)
Labs. Routine lab work in the diagnosis of back pain usually is unnecessary, but patients who also have fever and weight loss will need a CBC and chemistry panel. ESR and C-protein levels are extremely helpful in following patients with infections, with decreasing levels as a sign that the infection may be clearing. Urine studies will detect Bence-Jones proteins and aid in the diagnosis of multiple myeloma. Any patient that is going to have surgery should have preoperative labs including: complete blood count, chemistry, coagulation studies, chest x-ray, and EKG.
Plain films. Standard AP and lateral lumbar spine x-rays are an excellent way to diagnose fractures and evaluate the bony anatomy. Dynamic lumbar spine films, which include flexion and extension views, may demonstrate abnormal movement (subluxation). These should only be obtained after a detailed history is obtained, because routine use will have a low yield. Unexpected findings occur in approximately one in 2500 adults (58). Plain films are not an appropriate study to evaluate soft tissue.
Myelography and CT-myelography. These are more invasive diagnostic tests but yield significantly more information. The studies allow evaluation of nerve roots and their relation to bony anatomy that may not be seen with routine MRI. A myelogram involves injection of contrast into the subarachnoid space. This is usually done with a lumbar puncture at either the L4/5 or L5/S1 interspace. Nonionic contrast is injected and plain lumbar spine films are taken, with many different views obtained (AP, lateral, oblique, flexion and extension). The examiner is looking for blockage of contrast within the subarachnoid space. The addition of a CT scan will further increase the diagnostic yield of the myelogram, with details of the cord and roots seen well as compared to an MRI. Risks of these procedures include allergy to contrast, infection, and bleeding. Some patients may develop postlumbar puncture headaches that require the patient to lay flat but often resolve within a day or two. Also, the contrast may occasionally be injected into the subdural space, decreasing the diagnostic yield of the study.
MRI. Lumbar magnetic resonance imaging has become the gold standard for the evaluation of lumbar disc disease. It is noninvasive and allows for excellent visualization of the soft tissue and discs. MRI has allowed the physician to also evaluate the spinal cord itself and detect tumors, demyelination, edema, and infection. The addition of contrast allows the physician to distinguish between postoperative scar and recurrent disc herniation. MRI has few risks, the main being an allergy to the contrast. Patients with claustrophobia, morbid obesity, or metal implants should avoid MRI scanners. MRI has allowed the examiner to see anatomy once not thought possible, and the technology continues to improve. One major downside to MRI is too much information, leading to a false-positive study and possibly unnecessary interventions. Clinical correlation with a thorough history and physical exam is extremely important with MRI scans.
Discography. The goal of lumbar discography is to obtain two important pieces of information: the anatomy of the disc and the pain it refers. Research has suggested that the disc itself may generate pain if damaged without direct nerve root compression (54). It has also been demonstrated that distention of the zygapophyseal joint (facet joint) may refer pain beyond the immediate area (30). The physician performs the study by injecting contrast into the disc space. This will delineate the anatomy whereas the introduction of fluid simultaneously increases the pressure within the disc. The increase in pressure will often cause pain and hopefully reproduce the patient’s symptoms. During the same procedure, the physician can also inject asymptomatic discs to ensure there are no false positives. A positive discogram should reproduce symptoms at the involved disc and no pain at uninvolved discs (at least two other levels).
Some studies have called the safety and validity of discography into question. Carragee and associates examined success rates for lumbar fusion in cases with discogram concordance and found clinical success in only 27% of patients compared to 72% of patients with spondylolisthesis (17). When less rigid success measurements were employed, the differential (91% vs. 43%) still held. In another study, Carragee and colleagues found significantly higher rates of disc herniation and disc degeneration compared to matched controls after lumbar discography (16). Chen and colleagues looked at MRI and discography and concluded that MRI could almost always predict discogram concordance (19). Finally, Resnick and associates, in a joint spine section review of the topic, concluded that discography as a sole criterion for lumbar fusion was a poor predictor of success and that discography should not be used as a standalone test (66).
For sacroiliac pain, steroid injections may be used for both diagnostic purposes to establish the cause of pain and to provide therapeutic benefit. However, the reliability of injections to accomplish either remains controversial (44; 76).
EMG. Electromyography is a neurodiagnostic study that has two parts. A neurologist or physiatrist usually performs the study. The first part is the nerve conduction study and the second is the needle examination of the muscle. The nerve conduction study consists of two parts: motor and sensory. Any motor or sensory nerve can be studied, but certain nerves lend themselves to be studied more often (peroneal and tibial). During the nerve conduction study, the examiner looks at amplitudes, conduction velocity, and distal latency. Amplitude is a reflection of the number of axons and will decrease in any disease process that damages the axons. Conduction velocity reflects the myelination of the axons. Demyelinating diseases will decrease the conduction velocity. Finally, distal latency (time) is affected by entrapment neuropathies and increases with a diseased nerve. Most nerve diseases will have changes to all 3. It is the degree that clues in the diagnosis. EMG is a needle exam of the muscle and consists of two main parts. The first part involves insertion of a needle into the muscle to evaluate for insertional activity. Increased insertional activity is abnormal, which may reflect muscle disease. The examiner also evaluates for spontaneous activity within the muscle, known as fibrillations and positive waves. Fibrillations are seen in neurodegenerative diseases and are spontaneous discharges of a muscle. The second part of the study evaluates the voluntary activation of the muscles. The examiner looks for normal recruitment patterns and motor unit potentials. Patients with isolated low back pain often have a normal EMG; any abnormal finding must be pursued (49).
Conservative. Bed rest was routinely used as the mainstay treatment for back pain in the 1950s (83). Patients would often lay flat for weeks. As expected, many complications occurred, and this fell out of favor in the 1980s. Bed rest for 2 days is fine but greater than 4 days results in worse outcomes (28). Encouraging ambulation and activities that avoid pain triggers may also be considered. Physical therapy and spinal manipulation, along with exercise, have demonstrated benefits in low back pain treatment (74). Visits to the physical therapist three times a week for 4 to 6 weeks and maintained after the sessions have helped prevent the recurrence of back pain. Often a patient’s symptoms improve and more aggressive treatment can be avoided. Physicians must communicate with the therapist about the patient’s symptoms and specific goals. Care should be made to avoid over-exercising in the acute phase, which may only aggravate symptoms (52).
Cognitive behavioral therapy has been proposed not only as a management strategy for chronic low back pain but also as a requirement prior to more aggressive intervention both for some forms of conventional surgery as well as spinal cord stimulation. Multiple trials have investigated the utility of these pain psychology interventions either on their own or as an adjunct to other modalities. Zgierska and associates conducted a trial of adults with chronic low back pain on high doses of narcotics (93). Introducing cognitive behavioral therapy in addition to conventional conservative care statistically significantly reduced pain severity ratings. Morone and colleagues specifically targeted older patients with chronic low back pain comparing a cognitive behavioral program with an educational program (56). Over the course of the intervention, a statistically significant improvement in pain scores was seen, although the effect was not durable after cessation of the program. Hajihasani and associates performed a Cochrane review comparing physical therapy alone and physical therapy with cognitive behavioral treatment and found that the majority of studies did document superior results with both modalities (42). However, other studies disagreed and it was unclear whether depression was treated successfully concurrently.
Antiinflammatory medications and acetaminophen remain the mainstays of treatment for back pain. Both are equally effective in treating arthritis but no randomized clinical trial for back pain has been performed (12). Muscle relaxants and narcotics have not demonstrated any benefit over anti-inflammatory or acetaminophen and have more side effects (50). If narcotics or muscle relaxants are used, a short course is ideal.
Duloxetine, a selective reuptake inhibitor of serotonin and norepinephrine, was approved by the FDA in 2010 for treatment of chronic musculoskeletal pain (chronic low back pain and chronic pain due to osteoarthritis) at a starting dose of 30 mg daily and a target dose of 60 mg daily. Non-depressed patients with nonneuropathic chronic low back pain may benefit (77).
Physicians often refer their patients for epidural steroid injections or nerve root blocks. Short-term improvement may occur but long-lasting results are rare. No clear-cut benefit has been demonstrated of placebo over steroid injections, regardless of their placement (epidural, facet, trigger point, or ligaments) (18). Results are controversial at best for the use of steroids in back pain.
Many other treatments have been proposed for the treatment of back pain. These include acupuncture, oral steroids, electrical stimulation, and traction. None of these have demonstrated any benefit in the treatment of acute low back pain but may carry a placebo effect and, therefore, be of benefit to patients. These should only be attempted with chronic back pain patients and are not without risks.
Another treatment advocated for chronic myofascial back pain is selective injection of botulinum toxin type A based on anecdotal reports of clinical success. Initial trials were based on properties of botulinum toxin as a muscle relaxant and antinociceptive agent. Göbel and colleagues performed a randomized, double-blind trial of this agent with placebo control into 10 trigger points, but without EMG guidance (40). There was a statistically significant reduction of pain in the treatment patient compared to placebo controls. Further studies looking at optimal dosing and timing of treatment were advised.
Regenerative biological treatments have been proposed as a potential therapeutic solution for chronic back and joint pain. These regenerative treatments include prolotherapy, which involves injection of a chemical irritant, generally dextrose, to engender an inflammatory process, stem cell preparations, and plasma-rich platelet or platelet-rich plasma injections. These treatments are controversial owing to often astronomical billing, variable biological preparations that limit actual delivery of the promised substance, as well as no more than generally anecdotal reports of efficacy or, at best, small clinical trials.
Tuakli-Wosornu and associates published a small randomized trial of platelet-rich plasma injections into degenerative disc spaces (79). In a series of 47 patients, significant pain improvement with platelet-rich plasma injections was maintained up to a year. Ruiz-Lopez and colleagues performed a small series comparing epidural injections with platelet-rich plasma versus steroids and found that although steroids provided slightly better pain relief at 1 month, platelet-rich plasma injections were superior at 6 months (68). Wu and associates found a similar 6-month advantage of platelet-rich plasma over steroids with facet injections in another small prospective trial (91). Objective reviews of these treatments, including an attempt at meta-analysis, conclude that there is an abundance of small anecdotal trials, but the overall suggestion of efficacy is clouded and needs to be replicated with large-scale controlled studies. Although enthusiasm is expressed, this is tempered by a healthy skepticism until better data are available (69; 26).
Alternative or Eastern medical strategies have been advocated as a potential treatment for chronic low back pain. Traditional skepticism in these modalities has been tempered by evidence-based data with randomized controlled trials. Lin and associates performed a meta-analysis of randomized controlled trials for meditation-based treatments for low back pain. There were 12 randomized controlled trials in the literature, 10 of which showed significantly reduced intensity of chronic low back pain compared to standardized therapies. Quality of life, however, improved in only three of these 12 trials (51). Zhang and colleagues performed a similar meta-analysis of tai chi in the treatment of chronic back pain (94). There were fewer eligible studies, but in a meta-analysis, an overall benefit was seen using tai chi with standard nonoperative care.
However, skepticism remains inevitable. Acupuncture is a relatively excepted treatment for multiple maladies including chronic pain. In a Cochrane database review, Mu and associates reviewed randomized controlled trials for acupuncture in chronic specific low back pain in adults (57). Thirty-three studies were included involving over 8000 subjects, although only 16 studies had a reasonable control group. There was no more than low certainty evidence of immediate benefit compared to sham intervention. One trial showed no benefit in terms of short-term quality-of-life measurements. There was more evidence of improved back function immediately after treatment and with quality of life scales. The authors concluded that acupuncture may not play more of a clinically meaningful role for chronic back pain than placebo treatments in the long-term, although some immediate relief was seen with moderate to very low certainty secondary to small sample size and high risk of bias in multiple trials (57).
Surgical. Patients should only be considered surgical candidates if they have failed conservative treatment. Exceptions should be made for patients with rapidly deteriorating neurologic deficits with back pain. Any patients with cauda equina syndrome must be rapidly decompressed because the patient’s outcomes are directly related to the time of intervention from symptom onset (01).
The difficulty in defining the proper surgical treatment after failed conservative management involves distinguishing the pain generator in the patient. If the source of pain comes from the facet joints with resulting instability, then the pain is best treated with a posterolateral fusion (89). On the other hand, when the disc becomes the pain generator, removal of the disc becomes the ideal treatment option (32). These beliefs lead to specific surgical treatment.
Class I evidence exists that supports lumbar fusion for carefully selected patients with intractable low back pain without appreciable lumbar canal or foraminal stenosis or spondylolisthesis. Class III evidence also recommends that a course of intensive cognitive and physical therapy may be an appropriate alternative for these patients (61). Wood and associates examined data for spinal fusion versus multidimensional supervised rehabilitation specifically for patients with isthmic spondylolisthesis (as opposed to other forms of spondylolisthesis) and found improved pain and functional outcomes for patients with this condition who had failed rehabilitation (90).
When there is no discogenic component of the pain and fusion is considered, the safest and often the quickest surgical treatment option involves a simple posterolateral fusion. Lumbar fusion for low back pain has played an increased role as a treatment option but remains controversial (31). This involves placing bone between adjacent decorticated transverse processes and allowing the fusion to occur. This has been done for over 50 years with variable success; radiographic fusion rates approach 86% with clinical success ranging from 16% to 95% (38). With the development of pedicle screws to help supplement the posterolateral fusion, radiographic fusion rates increase to 93% but clinical benefits remain stable (38). Pedicle screws help prevent motion of the operative segment, which helps promote fusion, but surgical risks increase. The screws may penetrate their bony confines and injure exiting nerve roots or the thecal sac. It is rare, but some patients may develop new radicular symptoms after a lumbar fusion with pedicle screws.
There is conflicting evidence from primarily class II and class III data regarding the use of pedicle screws. Pedicle screw fixation is recommended as an option for patients with low back pain who have received a posterolateral fusion and are at high risk for fusion failure due to the ability of pedicle screws to improve fusion rates. Resnick and colleagues, in their comprehensive report on lumbar fusion, have not recommended that pedicle screws be used routinely as an adjunct to posterolateral fusion because of the conflicting evidence regarding any beneficial effect on functional outcome and the higher costs and complications associated with its use (62). However, more and more surgeons will routinely employ pedicle screws to augment posterolateral fusion.
Lumbar fusion surgery requires a standard approach to the lumbar spine with subperiosteal dissection. Exposure is carried laterally to expose the transverse processes above and below the affected disc level. Patients who have lumbar stenosis may also undergo total laminectomies with foraminotomies. The authors do not advocate the use of lumbar fusion in patients without radiographic canal compromise, spondylolisthesis or instability, or a clear discogenic pain generator. Rarely, a patient with facet disease who gets clear-cut temporary relief from facet injections might be considered for such an approach, but the clinical benefits are limited.
A total discectomy may be performed when the disc is felt to be a significant pain generator or has collapsed from degenerative disease. The decision to perform disc removal can be based on severe radiographic disc collapse with or without confirmatory provocative discography. Unless the pain generator is clear cut on MRI - in other words, a 1-level abnormality with completely normal-appearing levels above and below - the authors advocate a positive discogram before performing anterior or posterior total discectomy and fusion. The disc may be replaced with a composite, metal, or bone cage, wedge, or dowel often packed with autograft or BMP. Pedicle screws can then be placed above and below the affected level (some of the cage manufacturers regard supplementary fusion as an option, but the authors feel that without some form of posterolateral fusion, standalone, posteriorly placed cages have an excessively high failure rate). The transverse process would then be decorticated to allow for bony fusion between the levels. The surgeon may use autograph and allograph for the fusion, often augmented with BMP-2. Fusion rates currently approach 100% when internal fixation is combined with BMP-2 (11). Increasingly, transforaminal interbody fusion (TLIF) has replaced posterior interbody fusion (PLIF) for posterior spinal discectomy and fusion. TLIF allows a unilateral exposure of the spine with less nerve root retraction and may be done in a minimally invasive fashion. Anterior fusions can also be performed in a minimally invasive fashion but should be avoided if there is significant canal or foraminal compromise. These procedures also are associated with significant cost and the potential for complications. The debate continues on whether lumbar fusion for patients with low back pain without deformity or neurologic deficit is appropriate. Complications from posterior fusion include spinal fluid leaks, extensive disruption of paraspinal muscles, and future segmental degeneration above and below the fused level.
In the context of degenerative disc disease at one or two levels underlying low back pain, options include both posterolateral fusion and the various interbody fusions for patients. The placement of an interbody graft is recommended to improve fusion rates and functional outcome, but it should be recognized that the marginal improvement achieved is associated with increased complication rates, especially when combined approaches such as 360 degrees are used. The use of multiple approaches (anterior and posterior) in attempts to accomplish lumbar fusion is not recommended as a routine option for treatment of low back pain without deformity (63).
Anterior surgical approaches are the most controversial surgical treatment for back pain. Chronic low back pain is the main indication for anterior lumbar interbody fusion of discogenic origin. The surgeon may fuse the affected lumbar segments (ALIF) or place an artificial disc (CHARITE). A rational theory behind total disc replacement is the prevention of adjacent-level disc disease secondary to motion preservation. This has yet to be demonstrated in long-term follow-up. The anterior lumbar spine may be approached transperitoneal, retroperitoneal, or laparoscopically. When a fusion is performed, the entire disc segment is removed along with the cartilaginous endplates. The posterior longitudinal ligament is left intact to protect the dura and maintain the posterior tension band. A bony fusion is placed with autograph, allograph, and metal implants with recombinant human BMP-2. A total disc replacement has been made available in the United States. A multicenter FDA trial demonstrated equivalent outcomes between ALIF and disc replacement with higher satisfaction in the CHARITE group (10). Regardless of the preferred procedure, associated risks exist. Injuries to major blood vessels carry a 2% to 4% risk. Retrograde ejaculation is another serious complication, occurring most often at the L5/S1 level where the autonomic nerves may be damaged, with an overall risk of 0.55% to 2% (08).
The anterior surgical fusion has typically been combined with posterior fusion secondary to low ALIF rates. Posterior fusion involves the placement of pedicle screws to provide internal fixation of the lumbar spine. This technique addresses pain generators, the disc, and facet joints. As expected, this surgery takes longer, has higher complication rates, and costs the most to both patients and hospital. Posterior fusion alone has been advocated for isolated posterior element disease (33). This would include facet joint disease or significant canal stenosis.
Videbaek and associates examined 95 patients undergoing either posterior fusion alone versus a combined anterior and posterior approach. They prospectively evaluated adjacent segment degenerative disc changes and stenosis over time after fusion. They found, during an 8- to13-year follow-up, an overall rate of 26% of participants with adjacent segment disease. They found better clinical outcomes in patients with anterior and posterior fusion, but no higher risk of adjacent segment disease. This group suggested that the 26% risk of adjacent segment disease was similar to the risk of degenerative changes in an age-matched population and, thus, did not advocate for disc replacement or dynamic fusion (82). However, it is possible that with disc replacement and BMP to augment anterior or posterior stand-alone fusion, combined approaches should rarely be necessary at this point.
As an adjunct in treatment of patients with low back pain, lumbar braces have been used as a preoperative diagnostic tool and for postoperative support to assist in fusion. Potential mechanisms of action of brace therapy include limiting spinal range of motion, correcting posture and deformity, and gross trunk motion. As a guideline, short-term use of a rigid lumbar support (1 to 3 weeks) is recommended as a treatment for low back pain in which duration has been less than 6 months. It is not recommended for patients with low back pain of greater than 6 months to use a brace, as no pertinent evidence exists of any long-term benefit with its use. If a brace is used, rigid braces are found to offer some benefit over soft braces. Mainly class III evidence suggests that the use of lumbar braces used prophylactically does not reduce the incidence of low back pain. There is limited information regarding the benefit of bracing for improving fusion rates or clinical outcomes following lumbar fusions with instrumentation for degenerative disease (64).
Although lumbar spinal fusion has become a common procedure, it is important to find the best assessment tools in order to determine which fusion technique results in the greatest functional outcome for patients undergoing surgery. That said, it is also important to find a valid and reliable assessment tool in order to accurately measure functional outcomes in patients. Outcome instruments that have been supported by class I and class II evidence in the evaluation of low back pain include the Spinal Stenosis Survey of Stucki, Waddell-Main, RMDQ, DPQ, QPDS, and Million Scale (65).
These issues remain immensely controversial within the spinal surgery community. A randomized clinical trial in 2002 compared different fusion techniques. As expected, radiological fusion rates increase with the addition of pedicle screws or doing a 360 fusion. The most interesting finding was that clinical outcomes were equal, which draws the conclusion that less is more (33). This article is by no means the definitive standard of the treatment of back pain, but it offers to the readers the idea that new technology makes pictures better but not necessarily the patient.
Berg and associates presented a randomized trial of total disc replacement compared to standard lumbar fusion in individuals with chronic low back pain (09). Although the operative complication rate was similar for both groups, the short-term disability indices and the 2-year pain-free scores were better after lumbar arthroplasty. The authors concluded that the short-term benefits of arthroplasty were superior to fusion; however, long-term effects were not clear. Hellum and associates examined 173 patients with chronic back pain randomized to arthroplasty versus rehabilitation and found modest gains in general with arthroplasty, but no improvement in return to work (45). They concluded that the modest gains afforded by arthroplasty had to be carefully weighed against surgical risk.
Delamarter and colleagues examined Pro-Disc L disc replacement surgery over two adjacent levels as opposed to 2-level fusion and found modest gains in patient satisfaction, pain scores, and functional scores as well as a statistically significant reduction in narcotics usage favoring arthrodesis (25).
Another newer technology used as an adjunct to spinal decompression is dynamic stabilization. As opposed to rigid stabilization, these systems are designed to maintain some motion and prevent adjacent segment disease without a disc replacement. Schaeren and colleagues investigated one of these commercially available systems with a 4-year follow-up in 26 patients, all of whom had degenerative mild-grade spondylolisthesis (70). This procedure was done without a traditional fusion; nonetheless, none of the patients had progressive spondylolisthesis. However, adjacent segment disease progressed in 47% of patients. Whether this represents the inevitable consequences of the underlying degenerative process or a failure in the concept of dynamic stabilization is not understood. Surgery for low back pain will continue to remain controversial, but, as more clinical trials get published, the surgeon will be able to make a more informed decision.
Surgical versus conservative treatment. A randomized trial comparing nonoperative management with surgery for lumbar disc herniations was published. The SPORT trial, Spine Patient Outcomes Research trial, randomized 501 patients between conservative management or surgery between March 2000 and November 2004. Inclusion criteria were symptoms classic of lumbar disc herniations: positive straight leg raise, decreased reflexes, and sensory or motor weakness that correlated with imaging studies. Exclusion criteria for the SPORT trail included: cauda equina syndrome, instability, infection tumor, and contraindication to surgery. Conservative management often involved antiinflammatory medications, narcotics, and physical therapy. Surgery, if performed, was a standard lumbar microdiscectomy. Primary outcomes looked at SF-36 scores and Oswestry disability index at 6 weeks, 3 months, 6 months, and 1 and 2 years. Secondary measures included patient self-reported improvement, work status, and symptom relief. The study ultimately found that both surgery and nonoperative management had similar outcomes at 2-year follow-up although this was not statistically significant. The major problem with this study was that a significant number of patients in both groups cross over to the other treatment group. Because such large numbers of patients changed their initial assigned treatment strategies it was difficult to draw conclusion and demonstrate statistical significance. Also, the nonoperative management was not standardized. Certain nonoperative strategies may be more effective and, thus, more patients may have stayed in the nonoperative group. The article suggests that nonoperative management is equivalent to surgery but due to significant limitations clinicians should not discourage patients from undergoing surgery. The physician needs a practical protocol that incorporates both nonoperative and operative strategies into their decision algorithm. It should be noted that no patient with cauda equina syndrome or acute weakness was denied surgery (87). The frequent deviation from the assigned treatment groups with a relatively straight-forward problem such as lumbar disc herniation underscores the difficulty in replicating this process for a problem such as chronic back pain for which the surgical strategies and radiographic correlations are less precisely definable.
Surgery versus nonsurgical care for discogenic back pain. A particular area of controversy is treatment of axial back pain associated with common chronic degenerative changes. Brox and colleagues performed a randomized study and followed patients with common degenerative changes for 4 years after surgery or nonoperative care, consisting of structured spinal rehabilitation and cognitive/behavioral therapy (14). They found similar rates of unplanned operation or reoperation and return to work in the two groups. Follow up at 9 years showed no advantage of surgery (34). Wang and associates performed a meta-analysis of published trials comparing surgery to nonoperative care and found no advantage as well (85). Conversely, patients with single level severe degenerative changes with Modic endplate changes and failure of conservative care have been shown to have excellent resolution of pain, although quality of life may have less benefit (67).
Vertebral compression fractures. Vertebral body fractures from osteoporosis often result in varying degrees of back pain. Osteoporosis afflicts 200 million women worldwide, subsequently leading to 1.5 million new fractures per year, half of which are vertebral body fractures (47). In the late 1980s, Deramond and associates in France percutaneously injected polymethylmethacrylate (PMMA) into the compressed vertebral body (35). This procedure does not restore height but stabilized the collapse vertebral body, which helped alleviate pain. The procedure is called vertebroplasty and has been used to treat vertebral hemangiomas, neoplasms, and compression fractures since its inception. In 1998, kyphoplasty was introduced to restore height to vertebral compression fractures prior to injecting PMMA for stabilization (21).
A retrospective review in 2000 on vertebroplasty demonstrated a 95% ability to provide some degree of pain relief with 63% of the patients having complete pain relief (07). They treated all types of fractures and had especially good results from those with malignancies with the added benefit of preventing further compression and possible subsequent canal compromise. A literature review of vertebroplasty and kyphoplasty could not demonstrate a clear benefit of one procedure over the other; both offered similar pain relief (46).
Complications of these procedures include superficial and deep infection, pulmonary embolus, pneumothorax, and canal compromise from PMMA leaks out of the body into the spinal canal. Vertebroplasty has a reported complication rate of less than 10% and kyphoplasty has a rate of 1.2% (37).
Some randomized trials have called into question the true benefit of these procedures (vertebroplasty specifically) over conservative care. Buchbinder and associates performed a multicenter double-blinded and placebo-controlled trial of vertebroplasty (15). There was no benefit either early or up to 6 months postprocedure afforded by vertebroplasty in this trial. Kallmes and colleagues performed a similar study comparing vertebroplasty to simulated (sham) procedures (48). They found no statistically significant difference between the groups at 1 or 3 months, although there was a trend toward better pain control with vertebroplasty that did not reach statistical significance.
It remains to be seen if vertebroplasty and kyphoplasty will continue to be primary management tools for back pain from compression fractures. It is unclear if restoration of height by kyphoplasty has any added benefit, but kyphoplasty does carry significantly higher costs than vertebroplasty. A randomized clinical trial of vertebroplasty versus kyphoplasty, or at this point kyphoplasty versus control l or sham treatment, would be useful to determine if kyphoplasty and height restoration offer superior benefit to vertebroplasty or conservative care, validating the higher cost.
Sacroiliac joint fusion. Attempts to address sacroiliac joint injection through arthrodesis or fusion across the joint have been attempted over the years to address joint pain. Schütz and Grob reported successful open fusion in only three of 17 patients (72). Newer percutaneous fusion procedures have been developed and are purported in smaller anecdotal studies to be of far greater benefit, with improvement and patient satisfaction rates of 90% published with minimal complications (22; 78). However, evidence-based reviews remain skeptical given the lack of class 1 data in these small class 3 studies (44; 73).
Eventually, randomized trials were performed comparing sacroiliac minimally invasive fusion to nonsurgical care. Polly and colleagues reported 2-year data on a study of 148 patients randomized to surgery or control and found over 80% improvement in pain scores and over 65% improvement in Oswestry Disability Index scores compared to less than 10% improvement in controls (60). Another randomized trial was performed by Duhon and associates, with 172 subjects at 26 centers (29). In this study, there was a statistically significant improvement in pain scores, Oswestry Disability Index scores, and quality-of-life measures as well as a decrease in opioid utilization at 2 years. Data now extends out as far as 4 years demonstrating sustained efficacy (24).
Facet replacement. The desire for motion preservation versus fusion has now expanded into a third surgical strategy. Although cervical arthroplasty has become the standard of care and rather universally accepted by surgeons, patients, and carriers, lumbar arthroplasties remains more controversial as outcomes mentioned above have not matched those in the cervical spine. Dynamic fixation devices have not provided a significant improvement in the prevention of adjacent segment disease as was hoped for. This lead to the development of facet replacement devices in the lumbar spine. McAfee and associates reported the first small series in 2007 with the TOPS system as a multicenter pilot study of 29 patients with promising results (53). Anekstein and colleagues reported a 7-year follow-up of 10 patients enrolled in a prospective trial of facet replacement and found significant improvement in leg pain scores and Oswestry scores with one device failure (03). Haleem and associates reported a 5-year follow-up in 10 patients and found clinical improvement in pain and quality of life scores with no revisions (43). Sielatycki and colleagues compared 52 patients undergoing lumbar facet replacement with 156 individuals undergoing fusion for multiple etiologies, including degenerative spondylolisthesis, recurrent disc herniation, and adjacent segment disease. Significant improvements were seen with these devices compared to fusion at 1 year in regard to both back and leg pain (75). It remains to be seen whether this technology effectively reduces adjacent segment disease over longer-term follow-up, but initial clinical successes appear promising.
All contributors' financial relationships have been reviewed and mitigated to ensure that this and every other article is free from commercial bias.
Richard S Polin MD
Dr. Polin of George Washington University has no relevant financial relationships to disclose.See Profile
Randolph W Evans MD
Dr. Evans of Baylor College of Medicine received honorariums from Abbvie, Amgen, Biohaven, Impel, Lilly, and Teva for speaking engagements.See Profile
Nearly 3,000 illustrations, including video clips of neurologic disorders.
Every article is reviewed by our esteemed Editorial Board for accuracy and currency.
Full spectrum of neurology in 1,200 comprehensive articles.
Listen to MedLink on the go with Audio versions of each article.
Headache & Pain
Dec. 02, 2023
Headache & Pain
Dec. 02, 2023
Headache & Pain
Dec. 02, 2023
Dec. 02, 2023
Neuro-Ophthalmology & Neuro-Otology
Nov. 28, 2023
Headache & Pain
Nov. 20, 2023
Neuro-Ophthalmology & Neuro-Otology
Nov. 20, 2023
Nov. 14, 2023