EBQ:Radiograph vs MRI for Low Back Pain

From WikiMSK

Revision as of 12:24, 25 March 2022 by Jeremy (talk | contribs)
(diff) โ† Older revision | Latest revision (diff) | Newer revision โ†’ (diff)

This is ported content from Wiki Journal Club, used with permission
It is subject to the Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License .This is an exception to the WikiMSK license.It is not for modification and not for commercial use

Study: Jarvik et al.. Rapid magnetic resonance imaging vs radiographs for patients with low back pain: a randomized controlled trial. JAMA 2003. 289:2810-8. PMID: 12783911. DOI.

Clinical Question

In patients with non-traumatic low pain back, does rapid MRI compared with conventional radiography for diagnosis decrease back pain-related disability, patient perception of pain, or health care costs?

Bottom Line

There were no significant differences in back pain-related disability measured with modified Roland score, bothersome pain, pain frequency, or health care costs between radiography and MRI for non-traumatic low back pain. There was a nonsignificant trend towards more back surgeries in those who received MRIs.

Major Points

Low back pain is a common office complaint in both primary care physician clinics. In one study, approximately 8% of American adults reported at least 1 episode of severe acute low back pain within a one year period.[1] The prognosis is generally good, with a majority of patients improving with conservative management. Pain and disability often rapidly improves within the first month, with a return to work rate of 82% for those initially off work, and further improvements up to three months.[2] Clinicians are therefore faced with the decision to image or to empirically provide supportive care with analgesics and physical therapy. In those undergoing imaging, plain film radiographs have frequently been employed but the diagnostic yield is low in patients with acute non-traumatic low back pain. With advancements in MRI technology, rapid MRIs for lumbar spine have made it a potentially viable option as the initial imaging modality in patients with lower back pain. The excellent soft tissue resolution allows for better identifications of structural abnormalities which may or may not be symptomatic. While MRI may improve upon the diagnostic yield compared to plain film radiograph, further study would be required to demonstrate whether MRI would meaningfully improve patient outcomes, prompting this randomized trial.

Published in 2003, this study assessed 1250 patients for eligibility and ultimately randomized 380 patients with acute non-traumatic low back pain to either lumbar radiographs or to rapid MRI of the lumbar spine. The primary outcome was the modified Roland back pain disability score[3], which is a validated measure of disability related to back pain, assessed at 12 months from randomization. A variety of secondary outcomes were evaluated, including back pain frequently, amount of bothersome pain, and patient satisfaction with care, and the study also evaluated costs associated with the two imaging modalities. The investigators calculated they would need to recruit 372 patients to detect a meaningful between-group difference in the Roland score โ€” a 2-point difference is the minimum difference considered clinically significant. At 12 months, sufficient data was available from 89% of patients and demonstrated no difference in the primary outcome between groups: the modified Roland score was 8.75 in the radiograph group versus 9.34 in the MRI group (mean difference -0.59; 95% CI -1.69 to 0.87; P=0.53). Patient satisfaction was improved in the MRI group, and numerically more patients in the MRI group went on to receive spine surgery (10 versus 4; P=0.09). The authors conclude that MRI should not be routinely obtained in patients with acute non-traumatic back pain.

In the broader context, this randomized controlled trial suggests that despite physician and patient preference of MRI over conventional radiographs in the diagnosis of low back pain, MRI does not meaningful improve back pain-related disability or patient perception of pain, and raised the concern that perhaps MRI leads to unnecessary spine surgery. Modern guidelines now recommend against routine MR imaging for patients with nonspecific low back pain. The Choosing Wisely Campaign has highlighted several important issues surrounding the imaging of patients with back pain as well, including which red flag findings should prompt urgent imaging.[4]


ACP and APS low back pain management (2007, adapted)[5]

  • Clinicians should not routinely obtain imaging or other diagnostic tests in patients with nonspecific low back pain (strong recommendation, moderate-quality evidence).
  • Clinicians should perform diagnostic imaging and testing for patients with low back pain when severe or progressive neurologic deficits are present or when serious underlying conditions are suspected on the basis of history and physical examination (strong recommendation, moderate-quality evidence).
  • Clinicians should evaluate patients with persistent low back pain and signs or symptoms of radiculopathy or spinal stenosis with magnetic resonance imaging (preferred) or computed tomography only if they are potential candidates for surgery or epidural steroid injection (for suspected radiculopathy) (strong recommendation, moderate-quality evidence).


  • Multicenter, outcome collectors blinded, randomized, controlled trial
  • N=380 patients with acute non-traumatic low back pain
    • Radiograph (n=190)
    • MRI (n=190)
  • Setting: 4 centers in Washington
  • Enrollment: 1998-2000
  • Follow-up: 12 months
  • Analysis: Intention-to-treat
  • Primary outcome: Modified Roland back pain disability scale


Inclusion Criteria

  • Patients whose physicians ordered radiographs of the lumbar spine
  • Low back pain with or without radiating leg pain
  • Possess a telephone
  • At least age 18 years
  • Able to speak English
  • Able to provide written consent

Exclusion Criteria

  • Pregnant
  • Back pain after acute external trauma
  • Lumbar surgery within 1 year prior to enrollment
  • Metallic implants in the lumbar spine (eg, Harrington rods or pedicle screws)
  • Contraindications for MRI
  • If physician had initially ordered flexion or extension views or special views of the sacroiliac joints

Baseline Characteristics

Given for radiograph group, two groups similar

  • Mean age: 51.9 years
  • Mean BMI 28.5
  • Women: 55%
  • Race: White 81%, Black 12%, Other 10%
  • Married or living with another: 46/190
  • Less than high school education: 49/190
  • General Health Information:
    • Cigarette smoking: 28/190
    • Alcohol to relieve pain: 27/190
    • Any co-morbidities: 137/190
    • Unemployed, disabled or on leave: 29/190
    • Depression: 48/190
  • Recruited from general internal medicine or family practice group: 97/190
  • Previous episodes of back pain:
    • Continuous: 64/190
    • >5 episodes: 71/190
    • 1-5 episodes: 41/190
    • None previously: 13/190
    • Previous back operation: 12/190
    • Use of pain medications:162/190
    • Pain travels into leg:124/190
    • Pain travels below knee: 80/190


  • Randomized to rapid MRI within 1 week of study enrollment (sagittal and axial T2-weighted fast spin echo images, 1.5 T for 136/190 and 0.3 or 0.35 T for the rest) or radiographs as ordered by patient's physician โ€จ(161/190 of patient's had AP and lateral views only, 9/190 had additional such as obliques and for the rest, data is not available on number of views)
  • Images were interpreted by radiologists as part of their normal workflow.
  • Preliminary reports were faxed or physically delivered to the referring clinician.
  • Patients were contacted by phone or mail 1, 3, 6, and 9 months following randomization for follow-up and to request in-clinic physical examinations and functional scoring assessments
  • Modified Roland score was collected at months 3, 6, and 12 following randomization
  • Patient satisfaction was measured at months 1, 3, and 12


Comparisons are radiographs vs. MRI.

Primary Outcomes

Modified Roland score (at 12 months after randomization)
Larger numbers indicate more disability from low back pain.
8.75 vs 9.34 (95% CI โˆ’1.69 to 0.87; P = 0.53)
Modified Roland score (at 3 months after randomization)
8.6 vs. 10.4 (95% CI of the difference, โˆ’3.47 to โˆ’0.19; P = .03)

Secondary Outcomes

Back pain frequency and bothersomeness
4 separate symptoms on a 1-to-6 scale (1) leg pain; (2) numbness or tingling in the leg, foot, or groin; (3) weakness in leg or foot; and (4) back or leg pain while sitting, measured 12 months after randomization
Pain bothersomeness 9.75 vs. 9.68 (0.07, CI -0.88 to 1.22)
Pain frequency 10.21 to 10.09 (0.12, CI -0.69 to 1.37)
Medical Outcomes Study 36-Item Short Form Health Survey (SF-36), version 1 (measured at month 3 and 12 after randomization)
Bodily pain 52.59 vs. 51.34 (1.25, CI -4.46 to 4.96)
Physical function 63.77 vs. 61.04 (2.73, CI -4.09 to 6.22)
Role-physical 50.15 vs. 42.75 (7.40, CI -3.77 to 12.53)
Days of reduced or lost work at 12 months
Lost work 1.26 vs. 1.57 (-0.30, CI -0.92 to 1.15)
Limited activity 5.38 vs. 5.60 (-0.22, CI 1.69 to 1.84)
In bed 1.31 vs. 1.04 (0.28, CI-0.38 to 1.00)
Patient satisfaction with care
Measured with a modified version of the 11-item Deyo-Diehl patient satisfaction questionnaire, without question asking if patients thought they should have an imaging study and 2 added questions about reassurance that patients attributed to the imaging study, measured at 12 months after randomization.
7.34 vs. 7.04 (0.30, CI-0.42 to 0.99)
Patient elicited preference scores

Using a time trade off technique, anchored at 0 for death and 1 for perfect health, measured at baseline and 12 months after randomization

Baseline 0.78 vs. 0.75
12 months 0.83 vs. 0.86 (-0.03, 95% CI -1.12 to -0.02)
Societal cost of care
Office visits to conventional or alternative health care practitioners, drug use, including over-the-counter medications, hospitalizations, patient time spent obtaining health care for back pain and transportation, home care, and other expenses incurred while seeking health care
$2,059 vs. $2,380 (difference โˆ’$321; 95% CI, โˆ’$1100 to $458; P=0.42)
Number of patients who received lumbar spine surgery during follow up
4 vs. 10 (risk difference, 0.34; P = .09)

Adverse Events

No important adverse events in either group.


  • Authors did not clarify duration of low back pain in inclusion criteria. It is assumed the study is most applicable to chronic back pain as this is the population in which physicians would normally consider imaging.
  • Authors concluded against rapid MRI as initial test until effect on cost and unnecessary surgery can be clarified even though these differences did not reach significance in the trial.


Grants from the Agency for Healthcare Research and the National Institute for Arthritis and Musculoskeletal and Skin Diseases.

Further Reading

  1. โ†‘ Carey et al.. Acute severe low back pain. A population-based study of prevalence and care-seeking. Spine 1996. 21:339-44. PMID: 8742211. DOI.
  2. โ†‘ Pengel et al.. Acute low back pain: systematic review of its prognosis. BMJ (Clinical research ed.) 2003. 327:323. PMID: 12907487. DOI. Full Text.
  3. โ†‘ Roland & Morris. A study of the natural history of back pain. Part I: development of a reliable and sensitive measure of disability in low-back pain. Spine 1983. 8:141-4. PMID: 6222486. DOI.
  4. โ†‘ Choosing Wisely Campaign back pain highlights. Accessed November 1, 2018.
  5. โ†‘ Chou et al.. Diagnosis and treatment of low back pain: a joint clinical practice guideline from the American College of Physicians and the American Pain Society. Annals of internal medicine 2007. 147:478-91. PMID: 17909209. DOI.