CORR - 2026-07-20 - Journal Article
What Is the Diagnostic Accuracy of Clinical Examination and Dynamic Radiographs for Acute Vertebral Compression Fractures in Older Adults? A Prospective Study.
Yamamoto N, Shibata M, Konishi M, Tomita Y
Topics
Key Takeaway
A stepwise approach combining closed-fist percussion test (sensitivity 93%) and sitting lateral radiograph (vertebral-level specificity 99%, LR+ 51) increased posttest probability of acute VCF from 22% to 97%.
Summary Depth
Choose how much analysis to show on this article page.
Summary
This prospective study evaluated the diagnostic accuracy of clinical examination and dynamic radiographs against MRI (63%), CT (28%), or sequential radiographs (9%) as reference standards in 136 patients ≥50 years with acute thoracolumbar pain. The closed-fist percussion test yielded 93% sensitivity but only 20% specificity; sitting lateral radiographs achieved 77% sensitivity and 86% specificity at the patient level, and 99% specificity with LR+ of 51 at the vertebral level. Bayesian sequential integration of these tests raised posttest probability from 22% to 97%, supporting selective MRI use.
Key Limitation
The composite reference standard (MRI in only 63% of patients, with CT and sequential radiographs substituted in 37%) means the true sensitivity and specificity of the index tests are calculated against imperfect surrogates in over one-third of the cohort, potentially biasing all reported accuracy metrics.
Original Abstract
BACKGROUND
MRI is considered the reference standard for diagnosing acute vertebral compression fractures (VCFs), but its limited availability in many clinical settings necessitates alternative diagnostic strategies. The diagnostic accuracy of clinical examination and dynamic radiographs in a prospective setting remains insufficiently established.
QUESTIONS/PURPOSES
(1) What is the diagnostic accuracy of clinical examination findings for detecting acute VCFs? (2) What is the diagnostic accuracy of dynamic radiographs at the patient and vertebral levels? (3) Can a stepwise diagnostic approach integrating clinical examination and dynamic radiographs increase the posttest probability of acute VCF compared with the pretest probability?
METHODS
Between April 2025 and February 2026, a total of 136 patients age 50 years or older presented with acute low back or thoracolumbar back pain within 3 weeks of onset and were clinically suspected of having an acute VCF. All 136 patients were invited to participate in this prospective study evaluating the diagnostic accuracy of clinical examination and dynamic radiographs for the diagnosis of acute VCF. Of those, all 136 patients agreed to participate, and all 136 had full data sets that were analyzed. The mean ± SD age was 85 ± 8 years, and 71% (97 of 136) were female. We applied the following clinical examination elements in this order: clinical examination was performed first and included history of low-energy trauma, the closed-fist percussion test, and turning-over ability in bed. A positive closed-fist percussion test result was defined as localized pain anatomically corresponding to the suspected vertebral level. Dynamic radiographs were obtained after the initial clinical examination when clinical suspicion of VCF remained and were used as confirmatory tests. These included sitting lateral and supine flexion-extension lateral views. A positive dynamic radiographic test result was defined as a decrease in vertebral body height or a new morphologic change compared with the neutral supine image. MRI was used as the primary reference standard in 63% (86 of 136) of patients; 95% (82 of 86) of these patients underwent MRI within 7 days of the initial radiographic assessment. For patients unable to undergo MRI, CT was used in 28% (38 of 136) and sequential radiographs in 9% (12) as predefined alternative reference standards. A reference standard test was given to every study participant regardless of the result of the clinical examination or the results of dynamic radiographs. A positive MRI result was defined as signal change consistent with bone marrow edema within the vertebral body. A positive CT result was defined as vertebral body collapse or endplate or cortical disruption consistent with an acute compression fracture. A positive result on sequential radiographs was defined as progressive vertebral collapse or a new vertebral deformity during follow-up. Sensitivity, specificity, and likelihood ratios were calculated for each clinical examination maneuver at the patient level. For dynamic radiographs, these diagnostic accuracy measures were calculated at both the patient and vertebral levels. To answer the third research question, we used Bayesian updating to estimate whether a stepwise diagnostic approach integrating clinical examination and dynamic radiographs increased the posttest probability of acute VCF compared with the pretest probability.
RESULTS
The closed-fist percussion test demonstrated the highest sensitivity among clinical findings (93% [95% confidence interval (CI) 87% to 97%]) but low specificity (20% [95% CI 9% to 38%]). At the patient level, sitting radiographs showed the best overall diagnostic performance, with sensitivity of 77% (95% CI 65% to 85%) and specificity of 86% (95% CI 68% to 95%). At the vertebral level, sitting radiographs achieved very high specificity (99% [95% CI 98% to 99%]) and a strong positive likelihood ratio (positive likelihood ratio = 51), indicating strong rule-in capability. Flexion-extension radiographs demonstrated lower sensitivity (60% [95% CI 47% to 72%]) but high specificity (95% [95% CI 93% to 96%]). A sequential diagnostic approach increased the posttest probability of VCF from 22% to 97%.
CONCLUSION
The closed-fist percussion test may be useful for initial screening, whereas sitting radiographs provide strong confirmatory value. A stepwise diagnostic approach integrating clinical examination and dynamic radiographs may improve diagnostic efficiency and support more selective use of MRI in routine clinical practice; we have created a freely available online tool to help surgeons make these decisions at the point of care (https://norio-yamamoto.github.io/vcf-bayesian-simulator/).
LEVEL OF EVIDENCE
Level II, diagnostic study.