This discussion will review the overall approach to patients with back pain. An article in the New England Journal nicely reviews the primary care approach to low back pain. The key idea is that when you see a patient with back pain, you will think beyond the most common causes, consider the full spectrum of possible underlying issues, choose the most appropriate and cost-effective studies, and then select the safest and most effective therapy.
In the differential diagnosis of low back pain, the main point is that 97% of cases are mechanical (Figure 1). When a patient presents with a problem that appears likely to be mechanical, it may well be the case, but you need also to consider inflammatory conditions, infectious problems, fractures, referred and functional pain. It is important not to skip the step of considering these, since each has its own pathway of workup and distinct treatment options.
There are published guidelines for the diagnosis and treatment of low back pain. Note that these guidelines were developed for acute low back pain, and they may not apply to chronic low back pain or to recurrent low back pain. However, acute low back pain occurs frequently, and many of the conclusions of these guidelines remain relevant.
Another important issue in diagnosis of low back pain is: what are the red flags for serious treatable back pain? What are the findings that would require an x-ray? What are the findings that would make you think about getting an MRI or CT scan? A 1994 article in Arthritis & Rheumatism called "Who You See is What you Get" found great variations in diagnostic approach, depending on whether you went to an orthopaedist, a rheumatologist or a neurologist. In some cases almost all patients had MRI studies and in some cases essentially no one had an MRI study. The point made was that there should be a middle ground, where criteria of appropriateness are applied to each case. Note that a similar level of caution needs to be applied to interpreting the changes on x-ray of the low back. While the patient may have facet abnormalities, disc narrowing or spondylolisthesis, clinical correlation is always needed before assuming that these are actually the cause of the patient's pain.
The physical exam of the back, of course, remains critical to making a correct diagnosis of the source of back pain. See back exam video for review of the key elements.
The issue of CT scan and MRI is important, especially because these tests are very sensitive. There are often findings on an MRI that may be irrelevant to the patient's pain. The New England Journal of Medicine published a study where they looked at asymptomatic study volunteers who worked in a hospital, who specifically denied any history of back pain. Each of these volunteers had a lumbosacral MRI. The number who had abnormalities on MRI was dramatic -- at age 60, for example, the incidence of disc degenerative disease was about 92%. You have to be very careful when interpreting those abnormalities on MRI as being the cause of back pain, since these findings may be unrelated to the reason for their present pain. (See further discussion of this study in section IX).
Your next question is: does the differential diagnosis matter? For example, would the management of acute lumbar fracture be different than management of a strain? It would matter, especially because finding a lumbar fracture would stimulate a work-up for osteoporosis. In addition, one might consider prescribing calcitonin, in view of evidence that it decreased the pain of acute spinal compression fracture. For the patient with referred pain, as with an abdominal aortic aneurysm, endometriosis, or colon carcinoma, it would obviously be beneficial to make the specific diagnosis.
Near the end of this discussion, we will discuss management issues that go beyond the published guidelines, especially because the guidelines are specific to acute low back pain. The guidelines do not address treatment of chronic low back pain, and they do not always lend themselves to extrapolation.
In approaching low back pain, it's important to think anatomically. Recall how vulnerable the intervertebral foramen and nerve roots are is to disc herniation, spurring from the facet joints, and thickening of the ligamentum flavum, which are often additive factors in patients with spinal stenosis. The facet joints are true diarthrodial joints, and can develop all the changes of osteoarthritis that can be seen, e.g., in the knee - and can develop spurs large enough to cause significant nerve compression. Also, recall that L4-5 and L5-S1 are the most common levels of disc herniation, so that the L5 and S1 roots are the most commonly affected. It is always helpful to think about the involved dermatomes, and where a loss of reflex, strength or sensation is likely to occur.
Nerve Compression Levels and Exam Findings
What are some clues in the medical history to the differential diagnosis of low back pain? One issue is the character of the pain. The classic pain that we often consider is radicular pain. A patient presents with pain in the buttock, posterior thigh and calf, with a radiating, shooting, tingling quality, and you suspect lumbar disc disease. This is often of a different quality than the pain of, for example, a peripheral neuropathy, such as a femoral neuropathy, which tends to have a more burning quality. The other most characteristic type of pain is visceral or referred pain, as when a patient has a kidney stone. In that case, the pain is of a colicky type, with a crescendo and decrescendo quality.
The patient's age and sex can help in the differential diagnosis of low back. In a young patient, think of osteoid osteoma, a benign tumor that tends to respond to aspirin, or inflammatory bowel disease causing spondylitis and sacro-iliitis. In older people, consider conditions such as multiple myeloma and abdominal aneurysm. With regard to sex ratio, osteoporotic fractures and fibromyalgia are more common in women.
Age Issues in Back Pain
Sex Ratios in Back Pain
Mechanical low back pain, representing 97% of cases, deserves to be discussed first. It is helpful to determine which factors bring out the pain. We are trying to separate the various types of mechanical low back pain, including muscle strain, spondylolisthesis (slippage of one vertebral body on the next), herniated disc, osteoarthritis and spinal stenosis. Low back pain that gets worse with sitting may indicate herniated lumbar disc. If a patient comes into your office for low back pain, and prefers to stay standing while you take their history, this is probably someone who has lumbar disc disease. The acuteness of onset can also help separate types of mechanical low back pain. Acute onset suggests herniated disc or acute muscle strain, while insidious onset fits with osteoarthritis, spinal stenosis or spondylolisthesis.
An important point regarding factors that aggravate mechanical low back pain is the role of lumbar extension. Extension causes discomfort in particularly in patients with spinal stenosis. Ask patients to hyperextend the back for 20 to 30 seconds and see if that causes pain. This maneuver might be particularly helpful is when you have an older person who tells you that after walking a block and a half they routinely get pain in the buttocks, down to the posterior thighs and into the calves. The key differential diagnosis is between pseudoclaudication due to spinal stenosis and vascular claudication. Pedal pulse examination is helpful but certainly not definitive. It is thus valuable to have the patient extend their back for 20 to 30 seconds and see if this reproduces the pain that they are getting when they walk. If hyperextension of the spine reproduces their pain, this suggests the diagnosis of pseudoclaudication due to spinal stenosis.
What are the three things that you look for on x-ray for osteoarthritis of the spine? The three cardinal signs of osteoarthritis in any joint location is true for the spine: spurs, joint space narrowing and marginal sclerosis. In the disc space, a lucent line, called the "vacuum phenomenon" is helpful in diagnosing disc degeneration.
In spinal stenosis, recall that three processes are generally working together to lead to cord and root compromise. Posterior disc herniation, facet joint hypertrophy with spurring, and finally ligamentum flavum thickening, all play their role. It is important to appreciate the extent to which the cord and roots are vulnerable to compression, since this will help explain why we see a number of elderly patients undergoing lumbar decompression procedures. We have a high-risk population undergoing a long operation (usually more than one level needs decompression) and a procedure with only a fairly good outcome in many cases. Yet it is still done, because the degree of cord and root compromise can lead to neurologic findings, urinary or fecal incontinence or intractable back or extremity pain, which require the procedure.
Inflammatory low back pain represents only a small percentage of patients presenting with low back pain, but it is important because the onset is early and the problem is lifelong and often of great functional significance. In addition, we have treatments which can help essentially all patients, and can lead to very major improvements in a good percentage.
Inflammatory low back pain includes the group of diseases called the seronegative spondyloarthropathies, which begin at a young age, with gradual onset. Like other inflammatory joint diseases, these are associated with morning stiffness that gets better with exercise. There is a tendency to develop lumbar and cervical fusion with associated severe postural abnormalities. We are getting much more aggressive in the treatment of patients with seronegative spondyloarthropathy, but it is still not clear whether we can prevent fusion. We know that TNF-alpha blocking agents for rheumatoid arthritis, such as etanercept, infliximab and adalimumab, can markedly decrease erosion in that disease, but we don't have longitudinal studies yet to know whether this kind of drug can prevent fusion in the spondyloarthropathies (but they clearly provide significant improvement in motion, stiffness and discomfort). Etanercept (Enbrel®), adalimumab (Humira®), infliximab (Remicade®) and Golimumab (Simponi®) are anti-TNF agents presently approved by the FDA for use in this condition. The fifth presently-available anti-TNF agent, certolizumab (Cimzia®) has not as yet received FDA approval for ankylosing spondylitis. We presently encourage patients with spondyloarthropathy not to use pillows when they sleep, and to do neck extension exercises, with the expectation that if the cervical spine should fuse, it will be in a more functional position.
A physical diagnostic test that should be part of your armamentarium is measuring the Schober index. This simple test helps us to find out how well the patient's spine opens. The patient stands, and you make a mark at the "dimple level," approximately L3. Then make a mark 10 cm above it and another 5 cm below it, so that the two most distant marks are 15 cm apart. Then, you ask the patient to bend forward and touch their toes. Measure the distance between the two furthest apart marks. From their initial 15 cm apart, these two marks should be at least another 3 cm farther apart in the lumbar-flexed position, so that the minimal index to be considered normal would be 18/15. Less than that suggests that the spine is not opening up normally, as might be the case in a patient with significant spondyloarthropathy.
Another test that should be part of your routine exam on a patient with known or suspected spondyloarthropathy is the measurement of chest expansion. This is especially true if you are doing a preoperative evaluation on a patient who has inflammatory disease of the back. Measure the chest circumference at the nipple level on a man and under the breast on a woman, and the difference between maximum inspiration and maximum expiration is the chest expansion. The difference should be 3 cm or more. If the patient does not have 3 cm of expansion, then they are at a higher risk of postoperative pulmonary complications and referral for physiotherapy for instruction in chest expansion exercise is likely indicated.
It is worth remembering the strong genetic aspect of the spondyloarthropathy. We don't routinely do HLA-B27 testing because it is of limited help, because only 20% of patients with positive HLA-B27 go on to get one of these diseases. Present studies are looking at subsets of the B27 marker, in the hopes of identifying a subset at highest risk.
When the question of inflammatory back disease arises, one should consider the constellation of signs and symptoms of the reactive arthritis syndrome (Figure 4). Patients may have the characteristic skin picture of keratodermia blennorrhagica, a variant of pustular psoriasis. Diarrhea at onset raises the question of Salmonella, Shigella or Campylobacter infection leading to reactive arthritis. Several recent reports have also incriminated C. difficile as a cause of reactive arthritis. Preceding genitourinary symptoms raise the question of chlamydia infection. Other parts of the picture of reactive arthritis are mouth sores, peripheral joint involvement, and eye inflammation (conjunctivitis and iritis).
Inflammatory Back Disease- Clues re Reactive Arthritis
X-ray can show changes in the sacro-iliac joint, with narrowing and erosion, and the sacro-iliac joint can actually be widened in the earlier stages of the disease. The lumbar spine can show calcification of the anterior longitudinal ligament and "bright corners" in the anterior aspect of the vertebral bodies, reflecting the inflammatory process.
Infections of the spine are not common, but important not to miss. Most of the clues relate to identifying patients with a history consistent with increased access of bacteria to the circulation, or are the usual signs and symptoms of systemic infection. You need to rule out infection in patients with back pain accompanied by fever and/or chills, IV drug users, dialysis patients, those who have had recent surgery, trauma or skin infections.
Clues to Spinal Infections
Fractures of the spine are often very difficult pain problems and also give clues to the possible presence of osteoporosis. In patients with severe osteoporosis, spinal fractures can occur with no early warning and no significant trauma. Patients with spinal compression fractures generally have marked spasm and very high pain levels. Elderly people who develop compression fractures of the spine remain one of the few groups of people that we still admit for back pain. It is very hard to admit someone with back pain now due to utilization review issues, since their care can often be done at home. However, elderly patients with compression fractures of the lumbar spine, who really can't get around, and may have limited support systems at home, often need to be admitted.
In patients with low back pain where the cause is difficult to determine, especially if they are elderly or have osteoporosis, consider sacral fracture. This may not be seen on standard x-rays, and may at times even be difficult to pick up on bone scan. CT scan or MRI will generally reveal these fractures.
In patients who have had multiple spinal compression fractures, or refractory post-fracture pain, more recent procedures which may be considered are percutaneous vertebroplasty or open kyphoplasty. Kyphoplasty can actually increase the height of a vertebral body, but requires an incision, while vertebroplasty is percutaneous. The 2009 literature has raised some questions about the effectiveness of vertebroplasty and the indications for these procedures have to be looked at carefully in any individual patient.
Malignancy involving the lumbar spine is important to consider, particularly in patients who have particular risk factors, especially prior documented malignancy elsewhere. Night pain is characteristic in malignancy, and clues in the lab such as high sedimentation rate and anemia, suggesting myeloma, are worth seeking. Osteoid osteoma, which generally presents in a young person, is a benign tumor which causes pain that tends to respond well to aspirin.
Spine Malignancy Clues
Referred pain to the lumbar spine can be critical to diagnose. Abdominal aneurysm, endometriosis, tubal pregnancy, kidney stones, pancreatitis, penetrating ulcers, colon cancer -- all of these may present with back pain. It is important to be alert to clues to any of these masqueraders, such as colicky-type pain, weight loss, abdominal exam abnormalities, and microhematuria.
Functional low back pain is a consideration in patients who have compensation issues or patients who have psychiatric issues. The Waddell tests are a group of ways to help identify functional pain. Two of these tests are the ones most commonly used. One is the distraction test, where, for example, the patient takes a full minute to get on the examining table under your observation. If you subsequently are on the phone and looking away from the patient, they easily get off the table to get a drink. You do see this occasionally.
The second important Waddell test is to look for anatomic versus non-anatomic patterns of pain and neurologic abnormalities (especially subjective abnormalities). When the findings follow no dermatomal pattern, we need to consider functional causation. It is useful in situations like this to give the patient a body diagram, and have them mark the areas of pain by cross-hatching, and areas of paresthesia with dots. Patterns which do not fit anatomically raise the concerns for functional causation. Fibromyalgia with back pain is often thought of as being functional, at least in the sense of our inability to localize the abnormality anatomically or biochemically. Of course, as we learn more about pathophysiology of conditions like fibromyalgia, conditions previously called functional may have a more apparent biochemical or biophysical explanation.
The term "functional low back pain" need not imply a psychiatric disturbance or malingering, but simply helps to define pain for which we can provide no explanation and does not yield a pattern suggesting a particular anatomic structure as its cause. Some patients now called "functional" will later be re-classified as we learn more about the back pain syndrome.
A guideline on acute low back pain was published in December, 1994. Acute low back pain was very specifically defined, meaning that these guidelines must be used with great caution in back pain which does not meet their definition of "acute." Acute low back pain is defined as 0 to 3 months duration and of immediate onset. This is therefore a very specific group of patients.
Classification of Low Back Pain --Duration
This work was done by the Agency for Health Care Policy and Research (AHCPR), which at the time was active in guideline development. This was the 14th guideline from this federal agency. Why was this problem chosen to review? First, the frequency of the problem is very great - back pain is the #2 reason for primary care visits and the #1 reason for orthopaedic visits. Secondly, great variations in regional and specialty-to-specialty back pain diagnosis and management have been documented, so a service to patients could be provided by offering a more standardized approach. Third, there was a number of studies available to review, although some were deemed inappropriate for inclusion. The criteria used to define the quality of evidence in each case is defined:
Panel Ratings of Evidence
The guidelines were set up particularly to help patients improve their activity tolerance and get back to work. There has been some criticism of these guidelines, in that they may conflict with the patient's focus on pain relief. There was particular concern about the use of these guidelines to determine Workers' Compensation coverage.
The guidelines were aimed at primary care physicians, and panelists felt that only a small number of the patients likely needed to go to specialists. The guideline committee had 23 members -- physicians (including a physiatrist and a rheumatologist), nurses, physical therapists and occupational therapists, two chiropractors and a lay member. They reviewed over 10,000 abstracts and picked just under 4,000 articles for their review.
In interpreting these guidelines, it is important to know that none of their conclusions regarding diagnosis and management of acute low back pain was supported by "A" level data (multiple relevant and high-quality studies - see Figure 8). But there was a number of "B" level conclusions (one relevant high quality study or several adequate ones). It is not nearly the quality of evidence as we see in present-day guidelines for the management of stroke or myocardial infarction.
The impression of the guideline committee was that chronic low back pain represented less than 5% of patients presenting with back pain. In a rheumatology practice, of course, this percentage would be much higher - and these guidelines may not apply.
The committee chose to divide the patients with acute low back pain into three categories (Figure 9). They felt that for their recommendations to hold: (1) it did not matter if the patient had facet disease or spondylolisthesis, or a great variety of other anatomic abnormalities which have been described in acute low back pain; and (2) they could be separated into having or not having sciatica, and having or not having a red flag (described below) for a more serious cause of the back pain.
AHCPR Classification of Acute Low Back Pain -- 3 Categories
Felt these 3 more helpful than attempting to classify by pathophysiology (only 15% get definitive diagnosis)
Committee recommendations for diagnosis of acute low back pain: First, they advised that all patients need a careful history and physical exam. If the patient does not have red flags, then x-ray and other studies can be delayed for one month, during which time 90% of patients will be better. Red flags are indicators of potentially serious spinal or non-spinal pathology, or groups of patients at higher risk for these more serious causes. The red flags can relate to findings on vascular, abdominal or genitourinary exam.
Indications (Red Flags) for X-ray in Acute Low Back Pain
The red flags identify patients more likely to get infection, malignancy or fracture, i.e. the patients less likely to have simple muscle strain. Age greater than 50, history of malignancy, fever, weight loss, elevated sedimentation rate, trauma and motor deficit all seem logical in this regard. Litigation/compensation is listed because Worker's Compensation cases generally require x-ray.
Committee recommendations regarding CT Scan and MRI: The table below reviews the indications for CT scan or MRI in acute low back pain.
Indications for Imaging Studies Beyond X-Ray
If the patient has signs of the cauda equina syndrome, with perineal numbness, difficulty with urination and/or hyperreflexia, then an urgent MRI, or at minimum a CT scan, is needed. The presence of red flags for infection or fracture will also likely require an MRI or CT scan. If symptoms last longer than a month and surgery is being considered, imaging is needed. When a patient has had prior back surgery, it is reasonable to move more quickly in getting imaging beyond the x-ray. If you order an MRI in a patient with prior back surgery, it's helpful to order it with and without gadolinium, which can help in separating old scar tissue from a new disc herniation.
When considering ordering, and then interpreting, an MRI study in a patient with acute low back pain, it is wise to recall a study discussed in Section I, in which asymptomatic volunteers had MRI studies of their lumbar spine. To enter the study, the volunteers had to have no history of low back pain, and be between the ages of 20 and 65. Volunteers older than 60, for example, had a 90% occurrence of degenerative disc disease. In another example, 40% of volunteers at age 35 had evidence of herniated lumbar disc. There are obviously a lot of patients with abnormalities on lumbar MRI not associated with symptoms, so use caution in connecting their pain with findings on the MRI, which may be unrelated. If a patient has left-sided back pain and you do an MRI and they have a herniation to the right, that finding is irrelevant to their presentation.
Committee recommendations re blood testing in acute low back pain: CBC and sedimentation rate are clues to tumor or infection.
Committee recommendations re treatment in acute low back pain:
In addition to dating from 1994, other concerns have been expressed regarding these guidelines, including concerns which were raised very shortly after their release. These guidelines can help reduce the use of x-rays in very early back pain without red flags, but there was special concern that these guidelines, often based on imperfect literature support, would be used to determine payments by insurance companies. Materson did not think that the data were sufficiently robust to justify their use in determining which treatments would and would not be covered by insurance.
1994 Critique -- Back Pain Guidelines
Richard Materson, MD, Bulletin on the Rheumatic Diseases, 45:2, 1994
Several principles guide decisions for surgery in patients with back pain:
Surgery in Acute Low Back Pain
One principle is that 80% of patients with sciatica recover eventually without surgery. For this reason, in patients with sciatica and no significant neurologic abnormalities, observation is often the most prudent course. A second principle is that severe progressive neurologic deficits, neurogenic bowel or bladder dysfunction and the cauda equina syndrome make up the most clear-cut indications for back surgery. A third principle is that the patient's clinical signs and symptoms must correlate well with studies such as the MRI and electromyogram to expect a good outcome from disc surgery.
Another clue that surgery may be needed is if the neural canal is narrowed greater than 50% with free fragment, in the setting of neurologic deficit.
In view of the high spontaneous improvement rate, careful patient selection is clearly needed for optimal results of spinal surgery.
Patients presenting with acute lumbar compression can have extremely severe pain. They also require a workup for underlying osteoporosis, in all but the highest-impact fractures. The diagnosis and some details of the management of these cases were discussed under "fractures and low back pain" above.
Management of Acute Compression Fracture
Although the level of pain is frequently very intense, it is best to try to mobilize lumbar fracture patients early, especially since they are often elderly, and at especially high risk from immobilization. Opioids are often needed for pain control. Nasal calcitonin has been shown to help for the pain of acute lumbar spine fracture, and has some benefit in decreasing bone resorption as well. In patients with refractory pain following lumbar compression fracture, vertebroplasty and kyphoplasty are possible (see discussion above). A bone density study is needed in the great majority of patients with lumbar compression fracture, perhaps excluding those with no other osteoporosis risk factors and who had very high impact fractures. Studies have shown that many patients with fractures in the U.S. have been discharged from hospitals with no plans for bone density management, a trend that hopefully will be reversed.
Chronic low back pain should be managed with as low-risk a regimen as possible. For some patients, physiotherapy, with local heat or ice application (10-15 minutes on/10 minutes off), combined with a home exercise program and education in lifting techniques can make a major difference. It is important that patients learn to tolerate a certain degree of pain, lest they allow themselves to become more disabled than necessary. A recent study demonstrated that a physical therapy regimen agreed upon by the patient and physician, with an agreed-upon date for return-to-work, did in fact allow patients with low back pain to get back to their jobs sooner than standard therapy.
Patients who need analgesic treatment for low back pain can sometimes be managed with acetaminophen. Some do well with low-dose non-steroidal anti-inflammatory agents, such as ibuprofen, and some require full anti-inflammatory doses of non-steroidal anti-inflammatory agents. Analgesics, such as codeine, propoxyphene and oxycodone, have a role but patient selection and caution are required.
In attempts to avoid chronic narcotic therapy, chronic radicular pain, in the past, was often treated with amitriptyline. That drug, however, has many drawbacks, especially anticholinergic side-effects, which are especially troublesome in this often-elderly population. In those patients not doing well despite therapy, physical modalities, acetaminophen and non-steroidal anti-inflammatory agents, we now more often use gabapentin in those with persisting pain from nerve compression. Gabapentin does appear to be better tolerated than amitripytline, but can cause ankle edema, drowsiness and nausea. Gabapentin is not specifically indicated for back pain or sciatica in its FDA label. Other medications tried for refractory sciatica “off label” include carbamazepine and other antiepileptic agents, such as pregabalin (Lyrica®). Duloxetine (Cymbalta®), is an oral dual reuptake inhibitor that enhances the levels of both serotonin and norepinephrine, and has also been tried “off label” for refractory back pain and sciatica.
For inflammatory back pain, it is important for patients to begin stretching and strengthening exercises. If there is chest wall involvement, chest physiotherapy is important. Avoiding pillows under the neck when sleeping can help to have the cervical spine, if it fuses, fuse in a more functional position. Non-steroidal anti-inflammatory agents are useful in these cases. Sulfasalazine and methotrexate were the standard for more severe cases. As noted above under “seronegative spondyloarthropathies,” Etanercept (Enbrel®), adalimumab (Humira®), infliximab (Remicade®) and Golimumab (Simponi®) are anti-TNF agents presently approved by the FDA for use in psoriatic spondylitis and ankylosing spondylitis.
A review of key points in evaluating low back pain is provided below:
Review -- Approach to Back Pain
In summary, 97% of patients with back pain will have a mechanical cause and most will get better quickly -- but the other potential causes need to be considered early on. Many of these other types of back pain require very specific approaches to therapy, so they need to be ruled out early. The "red flags" for early x-ray and for early CT or MRI need to be sought at the time of the patient's presentation. It is important to note that many patients do not need x-rays, and many more do not need CT or MRI scan. In addition to their expense, these latter studies are overly sensitive, and often reveal abnormalities in the lumbar spine which are not the cause of the patient's pain. Treatment modalities for acute low back pain are often different from those used for those with chronic pain. Careful early attention to differential diagnostic possibilities, and careful patient selection for individual treatment modalities, will maximize therapeutic success.
 Deyo RA, Weinstein JN: Low back pain. N Engl J Med. 2001 Feb 1;344(5):363-70. Review.
 Bigos S, Bowyer O, Braen G, et al: Acute Low Back Problems in Adults, Clinical Practice Guideline No. 14. AHCPR Publication No. 95-0642. Rockville, MD: Agency for Health Care Policy and Research, Public Health Service, U.S. Department of Health and Human Services. December 1994.
 Cherkin DC, Deyo RA, Wheeler K and Ciol MA: Physician variation in diagnostic testing for low back pain. Who you see is what you get. Arthritis Rheum. 1994 Jan;37(1):15-22.
 Jensen MC, Brant-Zawadzki MN, Obuchowski N, et al: Magnetic resonance imaging of the lumbar spine in people without back pain. N Engl J Med. 1994 Jul 14;331(2):69-73.
 Materson RS: The AHCPR Practice Guidelines for Low Back Pain, Bulletin on Rheumatic Disease, 45:2, 6-8, 1994
 Staal JB, Hlobil H, Twisk JWR, Smid T, Köke AJA. van Mechelen W. Graded Activity for Low Back Pain in Occupational Health Care. Annals Int Med 2004 Jan;140(2):77-84.