Subscribe to eTOC

New Clinical Advice on Low Back Pain: When to Use Diagnostic Imaging

Most people arriving at their doctor's office with complaints of low back pain will be better within a few weeks to a month — with or without an imaging scan to figure out what is going wrong.

“Unless it is serious and unrelenting and the physician suspects a progressive neurological deficit or the patient has a history of cancer, the rule should be ‘wait,'” said Roger Chou, MD, a member of the Clinical Guidelines Committee for the American College of Physicians, which released a clinical guideline in the Feb. 1 Archives of Internal Medicine.

Dr. Chou, associate professor of medicine and scientific director of the Oregon Health & Science University Evidence-based Practice Center, and his colleagues reviewed a meta-analysis of six randomized studies, comprising a total of 1,804 patients with unexplained low back pain. They found no difference between routine lumbar imaging and usual care without imaging. The groups reported the same levels of pain, quality of life and overall improvement (as rated by the patients).

Most problems do not even show up on a CT, lumbar radiography or MRI, and routine imaging does not seem to improve clinical outcomes, Dr. Chou said. “What's more, findings identified on the scan could have nothing to do with their pain and these patients can be referred for surgery.”

The total health costs for low back and neck pain increased 65 percent from $4,795 in 1997 to $6,096 in 2005, according to a 2008 paper in The Journal of the American Medical Association. Much of this cost, Dr. Chou said, is due to imaging and the prescriptions that follow from the readout.

Studies also suggest that many people have abnormalities on lumbar scans “but that does not mean that they are associated with their low back pain,” said Dr. Chou.

In one cross-sectional study, 36 percent of people over the age of 60 had evidence of a herniated disk diagnosed on a lumbar scan; 21 percent had spinal stenosis, and 90 percent had a degenerated or bulging disc. And none of them had symptoms.

“It is important to understand that the presence of imaging abnormalities need not mean that the abnormalities are responsible for symptoms,” the committee wrote.

Most patients with acute back pain get better within four weeks, Dr. Chou said. “Acute low back pain has a favorable natural history. The expected yield of routine imaging is low,” the committee said.

A review of 68,000 lumbar radiographs prescribed for patients between the ages 20 and 50, for example, identified a clinically unsuspected finding in one of every 2,500 patients.

Furthermore, the studies suggested that it rarely affects the treatment choices and could add to the anxiety levels of patients told that they have evidence of an abnormal scan.

These patients tend to report “more pain and worse overall health status than those who did not have a radiograph,” said Dr. Chou.

The working group was also concerned about the studies that found increased rates for spinal surgery on the heels of imaging. In a randomized clinical trial, people with low back pain who had a rapid MRI were sent into surgery twice as often as patients who had radiography. Another study found that patients who had an MRI within the first month of their complaints of low back pain had an eight-fold increase in surgery compared to those who had no imaging.

The group created a list of recommendations that includes holding off on imaging. If the low back pain remains severe after three months then imaging might be advised.

The committee wrote that imaging should only be ordered if there is a high degree of suspicion of cauda equina syndrome, vertebral on, or severe and progressive neurological deficits such as motor weakness.


“There is no question that we are over-imaging patients with low back pain,” said Jeffrey Jarvik, MD, MPH, professor of radiology and neurological surgery and director of the Comparable Effectiveness of Cost and Outcomes Research Center at the University of Washington in Seattle. “There is a lot of support for the recommendation that physicians should not order a scan right away for people with back pain. The natural history is favorable. More than 85 percent of patients are better within four weeks.”

Dr. Jarvik and his colleagues conducted a study on almost 150 asymptomatic people from the Veteran's Administration hospital in Seattle. No one in the study, published in 2005 in Spine, had a complaint of back pain. But when scanned, a large number of abnormal findings popped up on the images, including bulging disks, disk protrusions, nerve root compression, disk extrusions, and moderate to severe central stenosis.

“Without clinical symptoms, this information is meaningless,” said Dr. Jarvik. “The danger is that some of these findings are so common and physicians feel an obligation to do something about it.”

Pain specialist Anne Louise Oaklander, MD, PhD, associate professor of neurology at Harvard Medical School who is also a member of the Neurology Today editorial advisory board, said that “these kinds of guidelines are inevitable…we need to be governed by these guidelines but physicians should have some protections in place in the event that something is missed because they didn't scan.”

Dr. Oaklander said that the reasons that clinicians order scans for low back pain are complicated. “There is no right way or wrong way but it rests on different pressures in the decision-making process,” she said. “There is the pressure from patients; the pressure from insurance companies. There is societal pressure and finally pressure for physicians who fear that if they don't image they may be held liable if there is a structural lesion that they missed.”

“There needs to be a better way to determine who really needs scans,” she added.

She said that there are ways to cut down on the number of scans ordered. Physicians may order a new scan even if the patient had a scan already because they don't have access to it. Electronic versions of the images would allow physicians access to previous scans. Also, she said, the costs of scans could come down, noting that MRI scans are twice as expensive as CT scans.


Chou R, Qaseem A, Shekelle P, et al, for the Clinical Guidelines Committee of the American College of Physicians. Arch Intern Med 2011;154:181-189.
    Martin BI, Deyo RA, Hollingworth W, et al. Expenditures and health status among aduls with back and neck problems. JAMA 2008; 299:656-664.