Scanning an aching back does not make it better, and the trials show why

For ordinary low back pain with no warning signs of a serious underlying condition, having an X-ray, CT or MRI early does not lead to less pain or better function than not having one. That is the pooled result of six randomised trials in 1,804 patients, and it has been the conclusion of the field for more than fifteen years [s1].

The trials compared immediate lumbar imaging against usual clinical care without immediate imaging. At short-term follow-up, up to three months, the difference in pain was a standardised mean difference of 0.19 (95% CI −0.01 to 0.39) and the difference in function 0.11 (−0.29 to 0.50) — neither statistically significant, and in the direction that does not favour imaging [s1]. At six to twelve months the estimates were closer to nothing still: −0.04 (−0.15 to 0.07) for pain and 0.01 (−0.17 to 0.19) for function [s1]. Quality of life, mental health, overall patient-reported improvement and satisfaction with care did not differ significantly either [s1]. The review's authors concluded that clinicians should refrain from routine, immediate lumbar imaging in patients with acute or subacute low back pain and no features suggesting a serious underlying condition [s1].

Why a picture of the spine tells you so little

The intuition behind scanning is that pain must have a visible cause and that finding it must help. The problem is that the things scans find are extremely common in people whose backs do not hurt.

A systematic review of 33 studies covering 3,110 asymptomatic individuals — people with no back pain who were imaged anyway — modelled how often each degenerative finding appears by age [s2]. Disc degeneration was present in 37% of 20-year-olds and 96% of 80-year-olds [s2]. Disc bulges rose from 30% of those aged 20 to 84% of those aged 80 [s2]. Disc protrusion went from 29% to 43% across the same span, and annular fissures from 19% to 29% [s2].

The authors' reading is the one that matters clinically: many imaging-based degenerative features are likely part of normal ageing and unassociated with pain, and must be interpreted in the context of the patient's clinical condition rather than treated as an explanation on their own [s2].

That is the mechanism behind the null trial result. If a scan of a painless 50-year-old back looks much like a scan of a painful one, then a finding on the painful one is weak evidence about the source of the pain — and a treatment plan built on it is being aimed at something that may be incidental.

What "no indication of a serious condition" means

None of this applies to the situations imaging exists for. The trials and the guidance both carve out patients with historical or clinical features suggesting serious underlying problems [s1]. The Australian primary-care literature describes these as "red flag" features, and the standing position is that imaging should not take place unless they are present, because outside those situations it does not improve patient outcomes [s4]. Deciding which features count is a clinical judgment, and it is the judgment the whole question turns on — not a decision a reader can make from an article.

The gap between the evidence and what happens in clinics

Imaging for uncomplicated back pain has been discouraged for a long time and still happens. A cross-sectional analysis of Australian GP registrars' consultations offers an unusually clean measurement of the rate. Across 2,333 early-career GPs — a 96.0% response rate — covering 325,058 consultations and 508,316 patient problems, new-onset non-specific low back pain accounted for 3,066 problems, or 0.6% of the total [s4]. Imaging was ordered for 450 of them: 15% [s4].

Ordering rose sharply with patient age. Compared with patients aged 15 to 34, the adjusted odds of imaging were 1.55 for those aged 35 to 64 (P = 0.002) and 2.32 for those over 65 (P < 0.001) [s4]. It was also higher when the registrar sought in-consultation help from a supervisor (OR 1.74, P = 0.009) and when follow-up was scheduled, and lower when the patient was referred onwards (OR 0.48, P < 0.001) or prescribed medication (OR 0.63, P < 0.001) [s4]. The authors describe the approach as considered but the overall prevalence as likely above optimal levels [s4].

Fifteen per cent is not a scandalous figure, and some of those 450 will have had red flags. It is still a useful marker of how much of the ordering is driven by things other than the imaging question itself — age, uncertainty, the shape of the consultation.

What guidelines recommend instead

The American College of Physicians guideline on non-invasive treatment of low back pain opens from the same starting point: most patients with acute or subacute low back pain improve over time regardless of treatment [s3]. Its first recommendation is for non-pharmacological treatment — superficial heat, rated moderate-quality evidence, and massage, acupuncture or spinal manipulation, all rated low-quality evidence — with non-steroidal anti-inflammatory drugs or skeletal muscle relaxants if a drug is wanted [s3]. For chronic low back pain, the guideline puts exercise, multidisciplinary rehabilitation, acupuncture and mindfulness-based stress reduction first, at moderate-quality evidence, ahead of drug treatment [s3].

Note how modest the evidence grades are. "Strong recommendation, low-quality evidence" is a recurring combination in this guideline [s3], and it reflects a field where the alternatives to doing nothing are not well separated from each other. That is a real limitation of the advice, not a reason to prefer the scan — the scan was tested against usual care and did not win [s1].

What to watch

The imaging question is largely settled at the level of randomised evidence and largely unsettled at the level of practice. The interesting measurements now are the ones like the Australian registrar study: not whether imaging helps, but what actually drives the order — patient age, clinician uncertainty, the presence of a supervisor, the expectation in the room. Those are the variables an intervention could plausibly move.

This article is informational and is not medical advice. Whether imaging is appropriate for a particular episode of back pain is a clinical decision that depends on features an article cannot assess.

Sources

• Imaging strategies for low-back pain: systematic review and meta-analysis — The Lancet, 2009-02-07
• Systematic literature review of imaging features of spinal degeneration in asymptomatic populations — American Journal of Neuroradiology, 2014-11-27
• Noninvasive Treatments for Acute, Subacute, and Chronic Low Back Pain: A Clinical Practice Guideline From the American College of Physicians — Annals of Internal Medicine, 2017-02-13
• Imaging in low back pain: a cross-sectional analysis of Australian early-career general practitioners' ordering of imaging for non-specific low back pain — Journal of Primary Health Care, 2025-07-31