Reference: Chou R, Fu R, Carrino JA, Deyo RA. Imaging strategies for low-back pain: systematic review and meta-analysis. Lancet. 2009;373(9662):463-472.
Clinical questions: In patients with low back pain (LBP) who do not have indications of a serious underlying condition, does routine, immediate lumbar imaging result in improved patient outcomes when compared with clinical care without immediate imaging?
Data sources: Studies were identified by searching MEDLINE (1966 through first week of August 2008) and the Cochrane Central Register of Controlled Trials (third quarter of 2008). The reference lists of identified studies were manually reviewed for additional citations. The search terms spine, low-back pain, diagnostic imaging, and randomized controlled trials were used in both databases. The complete search strategy was made available as an online supplement.
Study selection: The search criteria were applied to the articles obtained from the electronic searches and the subsequent manual searches with no language restrictions. This systematic review and meta-analysis included randomized, controlled trials that compared immediate, routine lumbar imaging (or routine provision of imaging findings) with usual clinical care without immediate lumbar imaging (or not routinely providing results of imaging) for LBP without indications of serious underlying conditions.
Data extraction: Data extraction and assessment of study quality were well described. The trials assessed one or more of the following outcomes: pain, function, mental health, quality of life, patient satisfaction, and overall patient-reported improvement. Two reviewers independently appraised citations considered potentially relevant, with disagreements between reviewers resolved by consensus. Two independent reviewers abstracted data from the trials and assessed quality with modified Cochrane Back Review Group criteria. The criterion for blinding of patients and providers was excluded because of lack of applicability to imaging studies. In addition, the criterion of co-intervention similarity was excluded because a potential effect of different imaging strategies is to alter subsequent treatment decisions. As a result of excluding these criteria, quality ratings were based on the remaining 8 criteria. The authors resolved disagreements about quality ratings through discussion and consensus. Trials that met 4 or more of the 8 criteria were classified as higher quality, whereas those that met 3 or fewer of the 8 criteria were classified as lower quality. In addition, the authors categorized duration of symptoms as acute (<4 weeks), subacute (4-12 weeks), or chronic (>12 weeks). The investigators also contacted the study authors for additional data if included outcomes were not published or if median (rather than mean) outcomes were reported. Statistical analysis was conducted on the primary outcomes of improvement in pain or function. Secondary outcomes of improvement in mental health, quality of life, patient satisfaction, and overall improvement were also analyzed. Outcomes were categorized as short term (≤ 3 months), long term (>6 months to ≤ 1 year), or extended (>1 year). For continuous outcomes, standardized mean differences (SMDs) of interventions for change between baseline and follow-up measurements were calculated. In studies reporting the same pain (visual analog scale [VAS] or Short Form-36 bodily pain score) or function (Roland-Morris Disability Questionnaire [RDQ]) outcomes, weighted mean differences (WMDs) were calculated. In all analyses, lower pain and function scores indicated better outcomes. For quality-of-life and mental health outcomes, higher scores indicated improved outcomes. All statistical analyses were performed with Stata 10.0. For outcomes in which SMDs were calculated, values of 0.2 to 0.5 were considered small, 0.5 to 0.8 were considered moderate, and values greater than 0.8 were considered large. For WMDs, mean improvements of 5 to 10 points on a 100-point scale (or equivalent) were considered small, 10-point to 20-point changes were considered moderate, and changes greater than 20 points were considered large. For the RDQ, mean improvements of 1 to 2 points were termed small, and improvements of 2 to 5 points were termed moderate.
Main results: The total number of citations identified using the search criteria was 479 articles and abstracts. Of these, 466 were excluded because either they were not randomized trials or they did not use imaging strategies for LBP. At this step, 13 articles were retrieved for further analysis. This analysis resulted in 3 additional articles being excluded (1 was not a randomized trial and the other 2 compared 2 imaging techniques rather than immediate imaging versus no imaging). The final step resulted in the inclusion of 6 trials reported in 10 publications for the meta-analysis. In the studies meeting the inclusion criteria, 4 assessed lumbar radiography and 2 assessed magnetic resonance imaging (MRI) or computed tomography (CT) scans. In these 6 trials, 1804 patients were randomly assigned to the intervention group. The duration of patient follow-up ranged from 3 weeks to 2 years. In addition, 1 trial excluded patients with sciatica or other radiculopathy symptoms, whereas another did not report the proportion of patients with these symptoms. In the other 4 studies, the proportion of patients with sciatica or radiculopathy ranged from 24% to 44%. Of the included trials, 3 compared immediate lumbar radiography with usual clinical care without immediate radiography, and a fourth study compared immediate lumbar radiography and a brief educational intervention with lumbar radiography if no improvement was seen by 3 weeks. The final 2 studies assessed advanced imaging modalities. Specifically, one group compared immediate MRI or CT with usual clinical care without advanced imaging in patients with primarily chronic LBP (82% with LBP for >3 months) who were referred to a surgeon. In the other advanced imaging study, all patients with LBP for <3 weeks underwent MRI and were then randomized to routine notification of results or to notification of results only if clinically indicated. With respect to study quality, 5 trials met at least 4 of the 8 predetermined quality criteria, leading to a classification of higher quality. In addition, 5 trials were included in the primary meta-analysis on pain or function improvement at 1 or more follow-up periods. With regard to short-term and long-term improvements in pain, no differences were noted between routine, immediate lumbar imaging and usual clinical care without immediate imaging ( Table 1 ). In studies using the VAS pain score, the WMD (0.62, 95% confidence interval [CI] = 0.03, 1.21) at short-term follow-up slightly favored no immediate imaging. No differences in outcome were seen in studies using the Short Form-36 bodily pain score. No improvements in function at short-term or long-term follow-up were noted between imaging strategies. Specifically, short-term function measured with the RDQ in 3 studies showed a WMD of 0.48 points (95% CI = -1.39, 2.35) between imaging strategies, whereas long-term function in 3 studies, also measured with the RDQ, showed a WMD of 0.33 points (95% CI = -0.65, 1.32). One included trial reported pain outcomes at extended (2-year) follow-up and found no differences between imaging strategies for pain (Short Form-36 bodily pain or Aberdeen pain score), with SMDs of -2.7 (95% CI = -6.17, 0.79) and -1.6 (-4.04, 0.84), respectively. The outcomes between immediate imaging and usual clinical care without immediate imaging did not differ for short-term follow-up in those studies reporting quality of life (SMD = -0.10, 95% CI = -0.53, 0.34), mental health (SMD = 0.12, 95% CI = -0.37, 0.62), or overall improvement (mean risk ratio = 0.83, 95% CI = 0.65, 1.06). In those studies reporting long-term follow-up periods, similar results can be seen for quality of life (SMD = -0.15, 95% CI = -0.33, 0.04) and mental health (SMD = 0.01, 95% CI = -0.32, 0.34). In the study reporting extended follow-up, immediate imaging was not better in terms of improving quality of life (SMD = 0.02, 95% CI = -0.02, 0.07) or mental health (SMD = -1.50, 95% CI = -4.09, 1.09) when compared with usual clinical care without immediate imaging. In the included studies, no cases of cancer, infection, cauda equina syndrome, or other serious diagnoses were reported in patients randomly assigned to either imaging strategy.
Conclusions: Available evidence indicates that immediate, routine lumbar spine imaging in patients with LBP and without features indicating a serious underlying condition did not improve outcomes compared with usual clinical care without immediate imaging. Clinical care without immediate imaging seems to result in no increased odds of failure in identifying serious underlying conditions in patients without risk factors for these conditions. In addition to lacking clinical benefit, routine lumbar imaging is associated with radiation exposure (radiography and CT) and increased direct expenses for patients and may lead to unnecessary procedures. This evidence confirms that clinicians should refrain from routine, immediate lumbar imaging in primary care patients with nonspecific, acute or subacute LBP and no indications of underlying serious conditions. Specific consideration of patient expectations about the value of imaging was not addressed here; however, this aspect must be considered to avoid unnecessary imaging while also meeting patient expectations and increasing patient satisfaction.