Question: In children presenting with low back pain, what is the rate of documented diagnosis and how well does a diagnostic algorithm perform in determining causes?
Design: 2-year cohort study.
Setting: A university medical center in California.
Patients: 73 patients <18 years of age (mean age, 14 y; 68% girls) who had back pain lasting >3 months and no previous spinal surgery. Patients were followed for up to 2 years (mean follow-up period, 22 mo).
Diagnostic strategy: Patients were evaluated according to a diagnostic algorithm: The initial diagnostic workup included history taking, physical examination, and plain radiographs. Patients with negative results on plain radiographs and no neurological findings received bone scans, single-photon-emission computed tomography scans, and erythrocyte sedimentation rate and complete blood-cell count testing. Patients who had positive results on bone scans and negative results on plain radiographs underwent computed tomography scans to determine the presence of osseous tumors or spondylolysis. Patients who had abnormal neurological findings or worsening symptoms at the time of follow-up underwent magnetic resonance imaging. Patients who had abnormal results on tests for erythrocyte sedimentation rate received antinuclear antibody and HLA-B27 testing to determine the presence of spondyloarthropathy.
Main outcome measures: Final diagnosis or no diagnosis after 2 years of follow-up.
Main results: 13 patients (18%) had definitive diagnoses. In 10 of these patients, the diagnosis was made with use of plain radiographs: 8 patients were diagnosed with spondylolysis with or without spondylolisthesis, and 2 patients had Scheuermann disease. In 2 patients with negative results on plain radiographs and positive results on bone scans, computed tomography scans showed a nondisplaced L5 pars defect with no associated spondylolisthesis in 1 patient and an osteoid osteoma in the other patient. In a patient with negative results on plain radiographs and positive neurological findings, the magnetic resonance imaging scan showed lumbar disc herniation with nerve root compression. 3 patients with initial positive laboratory tests ended the study with no definitive diagnosis. The diagnosis for 3 patients who received surgery had been made on the basis of the initial clinical examination or the results of plain radiographs. 57 patients ended follow-up with no diagnosis and no positive radiographic or laboratory findings.
Conclusion: In children presenting with low back pain, the rate of diagnosis was low. Most diagnoses were made on the basis of the clinical examination at the time of the initial workup or on the basis of the results of plain radiographs or bone scans.
Contrary to the common wisdom, this study by Bhatia confirms that most children with back pain have no identifiable cause. The type of evaluation required to rule out a serious condition is important but controversial. This study details a simple algorithm for the clinician to utilize when seeing a child with back pain of greater than three months' duration. This study excluded children with acute back pain, and therefore this workup cannot be extrapolated to that patient population.
While once it was thought that all children with back pain require an exhaustive and costly workup, this and other recent studies1 over the past decade have confirmed that a more limited assessment is adequate. The extent of the testing depends on symptoms and radiographic findings. Physical examination and plain radiographs are a mainstay in the initial evaluation of the child with back pain.
The authors utilized single-photon-emission computed tomography as a means of further evaluation in the child with normal radiographic results and a normal neurologic examination. Why single-photon-emission computed tomography was needed in the evaluation of children with normal radiographic results was not discussed. The study did not determine the most effective algorithm for children with back pain, as there was no control or comparison group. Specifically, magnetic resonance imaging, if cost-effective and available, might replace single-photon-emission computed tomography as an initial diagnostic test in these children.
This prospective study of children with back pain provides a working algorithm on which clinicians can base their assessment. What this study does not provide is the answer to the question of which children require tests beyond physical and radiographic examination. Further study is also required to determine the most clinically and cost-effective diagnostic test when further imaging is deemed appropriate.
Feldman DS, Hedden DM, Wright JG. The use of bone scan to investigate back pain in children and adolescents. J Pediatr Orthop.2000;20:790-5.20790
2000
[PubMed][CrossRef]