Orthopedics · Part Four · Seeing and Ruling Out
Lesson 32 / 44
The Missed Fracture: How Clinicians Screen for Vertebral Compression Fracture
An osteoporotic vertebral compression fracture is an uncommon but well characterized cause of back pain in older adults, and a careful clinician knows exactly how to screen for it.
A vertebral compression fracture is a collapse of the vertebral body itself, usually in bone thinned by osteoporosis, and it is the most common of the rare serious causes of back pain. Among 1,172 adults consulting primary care for acute low back pain, 11 had serious spinal pathology and 8 of those were fracture. The Unified Model of Tone treats the decision rules that catch them as assessment made explicit and reproducible.
Serious pathology in primary care back pain
11 of 1,172 patients, and 8 of the 11 were fracture
Incident radiographic vertebral fractures also diagnosed clinically
13.5 percent, in 4,398 older men
World Health Organization density bands
osteopenia at a T-score of minus 1 to minus 2.49, osteoporosis at minus 2.5 or below
New vertebral fracture in the year after one
19.2 percent
Vertebral compression fracture
The vertebral body is the squat block of bone at the front of each spinal segment, and it carries most of the compressive load passing down the spine. In a compression fracture that block loses height, most often at the front, which leaves a wedge shape on a lateral film. Nerve roots are usually untouched, so leg pain, numbness and weakness are typically absent.
Why thinned bone fails at ordinary loads
Inside the vertebral body sits a lattice of trabecular bone, a three-dimensional scaffold of struts that spreads compressive force across the whole block. Bone loss thins those struts and disconnects them from one another. Strength falls faster than mass does, because a lattice that has lost its cross-braces buckles at a load the intact scaffold would have carried.
01Bone as the pain generator
A vertebral compression fracture is a collapse of the vertebral body itself
A vertebral compression fracture is a fracture of the block of bone that carries the spine’s compressive load. It is uncommon. Among 1,172 consecutive adults consulting primary care for acute low back pain, 11 had serious spinal pathology of any kind, and 8 of those 11 were fracture (Henschke 2009).
Those numbers answer the question an older adult with new back pain quietly carries into the room. Most back pain comes from discs, facet joints, muscles, ligaments, nerve roots and fascia. The vertebral body still belongs on the list, because it is the most common of the rare serious causes.
In bone that has lost density a vertebra can fail under loads that would be harmless to a healthy spine. The pain that follows often looks and behaves like ordinary mechanical back pain, so the fracture can hide in plain sight. Recognizing that possibility is not alarmism. It is the difference between guessing and knowing.
What the base rates look like across settings
Where the person is sitting changes the number. A Cochrane review of eight diagnostic studies put the prevalence of vertebral fracture between 0.7 and 4.5 percent among people presenting to primary care with low back pain. In accident and emergency departments the same review found 6.5 to 11 percent (Williams 2013).
The complaint is identical and the prior probability is not. A good history in an older adult therefore asks about mechanism, prior fractures and bone health rather than stopping at where it hurts. The Red Flags Clinicians Screen For holds the wider list.
02Findings
What the research shows
From an inception cohort, a Cochrane review, two prospective bone cohorts, a multinational radiograph audit, a derivation study and two randomized trials.
03The quiet fracture
Most vertebral fractures are never diagnosed while they are happening
The majority of vertebral fractures pass without a clinical diagnosis. Baseline and follow-up spine films from 4,398 community-dwelling men aged 65 and over, taken an average of 4.6 years apart, held 237 new radiographic fractures in 197 men. Of those, 13.5 percent had also been diagnosed clinically at the time (Ensrud 2016).
Several things drive that gap. The symptoms resemble common mechanical back pain. Some fractures produce only mild discomfort. Many older adults already carry a long history of episodic back pain that can mask a new injury, and imaging is not always obtained at a first visit.
Pain location is a weak guide as well. The authors of the leading decision algorithm for this fracture note that its clinical picture typically involves asymptomatic findings and variable pain patterns (Roman 2010). A clinician who knows this never uses pain location alone to rule a fracture out.
Change carries more information than location. When a current episode differs from a familiar pattern in its onset, its severity or its mechanism, that difference is a prompt to reconsider rather than to reassure by habit. Instead of leaning on one finding, the presentation is read as a whole, which is precisely how the uncommon case gets caught.
Why the film itself can miss one
Radiographs miss these fractures at a measurable rate. In a multinational audit, thoracolumbar films from 2,451 postmenopausal women with osteoporosis were read at the local site and then centrally. Of the women, 789 had at least one vertebral fracture, and local reading against the central standard carried a false-negative rate of 34 percent (Delmas 2005).
The misses split two ways. Of 350 adjudicated discrepancies, 68 percent were fractures the local reader did not detect and 32 percent were equivocal wording in the report. False-negative rates reached 45.2 percent in North America and 29.5 percent across Europe, South Africa and Australia. The false-positive rate was 5 percent.
04Trabecular bone under load
Osteoporosis changes the internal lattice, and the vertebra then fails at loads a healthy spine ignores
Osteoporosis is a change in the internal architecture of bone before it is anything else. Healthy vertebral bone holds a lattice of trabecular struts that spreads compressive force across the whole block. As bone mineral density falls, that trabecular architecture deteriorates, internal support weakens, and the vertebra becomes less able to tolerate normal loading.
Forces that a strong spine shrugs off can then produce a structural collapse. Bending, lifting a light object or a bout of coughing all belong on that list.
Of 341 Rochester residents diagnosed with a first vertebral fracture over five years, 282 fractures followed moderate trauma or no trauma at all. Only 47 followed severe trauma, and 12 were pathologic (Cooper 1992). A minor mechanism sits inside that 83 percent, so the absence of a forceful one rules nothing out.
The same study set the overall age- and sex-adjusted incidence at 117 per 100,000 person-years. The rate in women, 145 per 100,000 person-years, ran to almost twice the rate in men at 73. Age and sex are therefore two of the strongest single inputs a clinician holds before anyone touches the patient.
What a bone density score settles, and what it does not
The T-score compares a person to a healthy young adult at peak bone mass. A score above minus 1 is normal. World Health Organization criteria set osteopenia at minus 1 to minus 2.49 and osteoporosis at minus 2.5 or below. Among 200,160 postmenopausal women screened in primary care, 39.6 percent fell in the first band and 7.2 percent in the second (Siris 2001). Osteoporosis carried a fracture rate about four times that of normal density, and osteopenia 1.8 times.
Those rate ratios describe risk per person, and the population arithmetic runs the other way. Following 616 postmenopausal women for a median of 5.6 years, 26.9 percent of all fractures arose in the osteoporotic group. The remaining 73.1 percent occurred in women without osteoporosis, 56.5 percent of them in the osteopenic band (Pasco 2006).
A reassuring density score is one input among several rather than a clearance. Bone integrity is not a footnote to a treatment plan. It is the first thing a good clinician establishes, because knowing the strength of the tissue at hand is what makes conservative care both effective and safe.
05The fracture decision rules
Combinations of findings, not single red flags, carry the diagnostic accuracy
Accuracy comes from combining features rather than from any one of them. A clinician weighs age, a known history of osteoporosis or osteopenia, the mechanism of injury, sudden onset, changes in posture and previous fractures together rather than in isolation. The signal in any single feature is modest. Combined, they sharpen accuracy considerably.
A pattern of sudden deep aching that worsens with standing and forward bending, and eases on sitting or lying down, arriving after a minor event, earns respect. None of that makes fracture likely in a given person. It means the clinician is thinking clearly about which bodies carry more risk and why.
The diagnostic literature states the same thing in numbers. A Cochrane review examined 29 groups of index tests across eight studies of vertebral fracture. In primary care only three were informative, and their estimates were wide. Significant trauma gave a positive likelihood ratio of 3.42 to 12.85, older age 3.69 to 9.39, and corticosteroid use 3.97 to 48.50 (Williams 2013).
Combined tests came out larger in magnitude and more precise than single flags. Many red flags also carry high false-positive rates that drive up imaging, and The Red Flags Clinicians Screen For carries that burden for the list as a whole.
The inception cohort found the same shape. Of 1,172 patients, 80.4 percent had at least one red flag, with a median of two. A rule built from four features, female sex, age above 70, significant trauma and prolonged corticosteroid use, produced an area under the curve of 0.834 for the presence of fracture (Henschke 2009).
The five-item cluster for vertebral compression fracture
One algorithm was derived specifically for this fracture, from routine clinical findings in more than 1,400 patients at an adult spine clinic (Roman 2010). Five items survived the analysis: age above 52 years, no leg pain, body mass index of 22 or below, no regular exercise, and female sex.
Two or fewer of the five gave a sensitivity of 0.95 and a negative likelihood ratio of 0.16, which lowers the probability usefully. Four of five gave a positive likelihood ratio of 9.6. The rule was derived rather than prospectively validated, and the authors leave open how it performs elsewhere.
What a validated rule buys
The clearest demonstration comes from the ankle. The Ottawa ankle rules were refined and then tested prospectively on 453 patients, where they identified all 50 malleolar zone fractures and all 19 midfoot zone fractures. The estimated reduction in radiography was 34 percent for the ankle series and 30 percent for the foot (Stiell 1993).
With the rule negative, the estimated probability of fracture was 0 percent, with an upper confidence bound of 0.8 percent. Fewer scans and no missed fractures arrived together, because the rule made the examination explicit and reproducible. The neck has its own pair, NEXUS and the Canadian C-spine rule, covered in Cervical Trauma and the Systematic Read.
06Referring out, reading back
Referring for imaging and interpreting the result in context is the portal of entry working
When the clinical picture warrants a scan, the correct move is to refer out for it and then read the result against the history and examination. The practice does not own an X-ray or MRI unit, and it does not need one. A clinician does not need to own the scanner to read the film well.
On a lateral film the reader looks for loss of vertebral body height, change at the endplates, and the wedge or biconcave shape that follows a collapse. On MRI, marrow signal separates an acute fracture from an old one. Those findings are then weighed against the history and examination that no radiology report can capture.
The audit data show why that second reading matters. In 350 adjudicated discrepancies, roughly a third came from equivocal terminology in the report rather than from a missed finding (Delmas 2005). Reading a report critically is clinical work.
Which fractures get caught
Severity drives detection. Among the new radiographic fractures in the older men studied, clinical diagnoses were made for 16.3 percent of those whose semiquantitative grade rose by two or more, against 13.5 percent overall. Running the comparison the other way, 86.5 percent of the fractures diagnosed clinically were confirmed radiographically, most of them severe (Ensrud 2016).
The mild ones slip past, and a fracture already present raises the risk of the next one whatever its grade. That is the argument for screening deliberately rather than waiting for a presentation dramatic enough to force the question.
Directing a recognized fracture to the pathway that manages it is never a limitation of conservative care. It is what a portal-of-entry profession exists to do, and When Conservative Care Stops sets the thresholds. It also protects the patient, because the high-velocity technique that suits healthy bone is matched to tissue capacity once osteoporosis has weakened the vertebrae.
Knowing bone quality lets a clinician modify technique, choose gentler approaches and work with older adults from understanding rather than from caution born of uncertainty.
07The year after a fracture
The twelve months after a vertebral fracture carry the highest risk of the next one
Once a fracture is recognized, management reaches well past the current episode. Placebo-arm data from four osteoporosis trials followed 2,725 postmenopausal women of mean age 74. New vertebral fractures appeared in 6.6 percent during the first year, and one or more fractures at baseline raised that risk about fivefold (Lindsay 2001).
The figure for those who fractured during the study is starker. Among the 381 women who sustained an incident vertebral fracture, 19.2 percent sustained another within the next twelve months, with a confidence interval running from 13.6 to 24.8 percent. Recognition of the first fracture is what makes that year actionable.
Fracture risk depends on more than density. A prevalent fracture roughly doubled risk independently of bone density, while each standard deviation of density lost raised it by 65 percent (Pasco 2006). The risk factors that shifted the odds in the primary care screening data were age, personal and family fracture history, smoking and cortisone use (Siris 2001).
What loading does to bone that has already thinned
Loading is the input bone responds to, and heavy loading has been tested in exactly the population people worry about. In the LIFTMOR trial, 101 postmenopausal women with a T-score below minus 1 were randomized to eight months of supervised high-intensity resistance and impact training or a home-based low-intensity program (Watson 2018).
The training arm gained 2.9 percent in lumbar spine bone mineral density while the control arm lost 1.2 percent. Femoral neck density, femoral neck cortical thickness and every measure of functional performance moved in the same direction. Height rose 0.2 cm in the training group and fell 0.2 cm in the controls.
One adverse event was reported across the whole trial, a minor low back spasm. This kind of training is not traditionally recommended for people with low bone mass, because of a perceived high risk of fracture. Bone and muscle rose together here, in a group whose density had already fallen. Progressive thoracic kyphosis is therefore never waved off as simple aging.
What the sham-controlled trial found
Cementing the fractured vertebra did not beat a sham procedure. In a double-blind trial, 78 participants with one or two painful osteoporotic vertebral fractures under twelve months old, unhealed on MRI, were randomized to vertebroplasty or a sham (Buchbinder 2009).
At three months, pain fell by 2.6 points of 10 after vertebroplasty and 1.9 after the sham. The adjusted between-group difference was 0.6, with a confidence interval from 0.7 below zero to 1.8 above it. Both groups improved at every assessment, and neither procedure separated from the other at any time point.
Much of the fear that surrounds back pain in later life comes from a nervous system settled into a protective state, and Why Recovery Differs takes up that variable. A plan that pairs safety with graded load, balance work and falls prevention is what quiets it.
08Claims corrected here
Three claims from the earlier version were corrected or removed
The earlier version stated that fewer than one third of these fractures are correctly diagnosed early, without a source. Two verified figures replace it. In older men, 13.5 percent of new radiographic vertebral fractures were diagnosed clinically at the time (Ensrud 2016). Local radiograph readings carried a 34 percent false-negative rate against a central standard (Delmas 2005).
The nutrition list of vitamin D, calcium and protein came off, because no trial cited here measured them. What the trials on this page did measure is loading, and the training numbers above report it.
The old description of pain felt across characteristic regions of the trunk is now stated the way its source states it, as variable pain patterns (Roman 2010).
09The model on missed fractures
What the Unified Model of Tone claims about screening for vertebral fracture
Everything above is established science, including the trial that came back null. What follows is this model’s reading, stated as ours rather than drawn from the papers cited.
Our model holds that assessment, and not delivery, is the true seat of accuracy. A validated decision rule is that principle in written form. It takes the reasoning a careful clinician performs privately and makes it explicit, reproducible and checkable by someone else.
The numbers on this page carry that claim. Single red flags gave likelihood ratios too imprecise to act on (Williams 2013). The five-item cluster reached 9.6 (Roman 2010) and the ankle rules reached full sensitivity while cutting radiographs by a third (Stiell 1993). Structure, not effort, produced the accuracy.
Escalation as correct dosing
Our model places every intervention on one continuous axis of magnitude, from the lightest sustained touch to the most invasive surgery. The rule that follows is to begin with the least invasive input that can carry the message. The model also says when to stop beginning there. Once a distortion has descended past what any surface input can reach, the larger magnitude is the correct one, and delay becomes its own kind of harm.
Referring an older adult with a suspected vertebral fracture for imaging is that rule executing. The referral is a dose increase chosen because the picture called for it, not a retreat from conservative care. Conservative First holds the full argument for the axis.
Reading the null trial through the model
The vertebroplasty result reads the same way. Cement was delivered to every participant whose fracture met the entry criteria, in a sample never stratified by the state each spine was in (Buchbinder 2009). Our model expects that design to average a well-matched input and a mismatched one, and to land near the middle.
It landed at 0.6 points of 10. Both arms improved substantially, which is what a body doing its own resolution looks like when an input adds little on top.
The prediction
From that follows a claim the fracture literature does not make. Our model predicts that older adults presenting with new back pain differ measurably in regulation before any fracture is present. Four measures recorded together in the same people will share one underlying factor rather than varying independently.
The four are pressure pain threshold over the suspect thoracolumbar segment, thoracolumbar active range of motion in degrees, and postural sway during a standardized reach. The fourth is time to return to resting heart rate variability after a graded loading task. Our model further predicts that adding those readouts to the five-item cluster adds information the cluster does not carry alone.
This is a claim about how vulnerability to fracture is organized rather than a claim about what treatment does. If pressure pain threshold, thoracolumbar range of motion, postural sway and time to return to baseline heart rate variability are shown to move together, the unification claim is confirmed.
10The tone reading
How vertebral fracture expresses tone
Every topic in this library expresses all of tone. In vertebral compression fracture three aspects carry the signature, because the bone that failed had been quietly losing the capacity to carry what it was asked to carry.
Load
Bone fails when demand exceeds what its lattice can carry. Of 341 first vertebral fracture diagnoses in one population, 83 percent followed moderate trauma or none at all.
Constraint
After a fracture the trunk splints. Available movement narrows, the thoracic curve deepens, and the narrowed range raises the load carried by the segments above and below.
Time course
Risk concentrates in time rather than spreading evenly. Of women who sustained one vertebral fracture, 19.2 percent sustained another inside the following year.
The remaining foundations run through this fracture as well. Set point: bone holds a remodeling balance, and the balance drifts toward resorption decades before a vertebra gives way. Input quality: a lattice that has lost cross-braces receives the same load as a poorer signal about how to rebuild. Gain: guarding after a fracture raises protective sensitivity well beyond what the healing bone requires. Prediction: expecting the spine to break next changes how a person stands, lifts and steps, and that expectation is measurable in gait. Coupling: bone density, trunk strength and balance move together, which is why a training program raised all three at once. Oscillation: the loading rhythm of daily activity is the signal bone remodels to, and losing that rhythm is itself an input. These are readings of one organization rather than separate systems, which is the core claim of the Unified Model of Tone.
11Across the library
How this page relates to the rest of the library
Screening for fracture sits inside a wider set of pages on ruling out the dangerous.
The diagnostic accuracy of the whole red-flag list, and the false-positive burden that comes with it.
How malignancy and infection are excluded on a scan, and the sensitivity each modality brings.
When a scan changes management, which is the question that decides whether to order one.
NEXUS and the Canadian C-spine rule, the two decision rules built for the neck after injury.
The age-stratified prevalence of spinal imaging findings in people reporting no symptoms at all.
The other bone-failure page, where repeated load rather than lost density exceeds what the pars can take.
Load as a measurable regulatory input, and what accumulating demand does to a tissue over time.
12Frequently asked
Questions about vertebral compression fracture and bone health
Can back pain be a hidden fracture?
It can, and it is uncommon. In an inception cohort of 1,172 adults consulting primary care for acute low back pain, 11 had serious spinal pathology and 8 of those were vertebral fracture. A Cochrane review put the prevalence between 0.7 and 4.5 percent in primary care and 6.5 to 11 percent in emergency departments. Clinicians screen for it in older adults by combining age, sex, prior fracture, mechanism, corticosteroid use and bone health rather than by reading pain location.
Why are vertebral fractures missed so often?
Because most of them are mild and mimic ordinary mechanical back pain. Across 4,398 men aged 65 and over, 237 new vertebral fractures appeared on follow-up films and only 13.5 percent had been diagnosed clinically at the time. Radiographs miss them as well. In a multinational audit of films from 2,451 women with osteoporosis, local reading carried a false-negative rate of 34 percent, split between undetected fractures and equivocal wording in the report. Severity drove detection, so the mild ones passed unrecorded.
What is a bone density T-score?
It is a measure of bone mineral density expressed against a young adult reference. Under World Health Organization criteria, a T-score of minus 1 to minus 2.49 is osteopenia and minus 2.5 or below is osteoporosis. A score above minus 1 is normal. In 200,160 postmenopausal women screened in primary care, 39.6 percent fell in the osteopenic band and 7.2 percent in the osteoporotic one. Osteoporosis carried a fracture rate about four times that of normal density, and osteopenia about 1.8 times.
Does a normal bone density result rule out fracture risk?
No, and that is one of the most useful findings in this literature. Following 616 postmenopausal women for a median of 5.6 years, 26.9 percent of fractures arose in the osteoporotic group and 73.1 percent in women without osteoporosis. More than half of all fractures, 56.5 percent, came from the osteopenic band. Density is one input among several, alongside age, prior fracture, balance, muscle strength and fall risk. Screening combines them rather than reading any one alone, and a prevalent fracture roughly doubled risk on its own.
Do clinicians use a checklist to decide about imaging?
They use validated combinations rather than any single sign. One derived cluster holds five items: age above 52, no leg pain, a body mass index of 22 or below, no regular exercise and female sex. Four of the five present gave a positive likelihood ratio of 9.6, and two or fewer gave a sensitivity of 0.95. A separate four-feature rule reached an area under the curve of 0.834. Single red flags performed far worse, with estimates too imprecise to act on alone.
Does the practice take its own X-rays?
No. This practice refers out for imaging and interprets the result alongside the history and examination. A clinician does not need to own the scanner to read the film well, and the audit data show why that second reading matters. Of 350 adjudicated discrepancies in one multinational study, roughly a third came from equivocal wording in the radiology report rather than from a missed finding. Referral, interpretation and coordination with radiology and medicine are all clinical skills in their own right.
What does the Unified Model of Tone say about screening for fracture?
That assessment, rather than delivery, is the true seat of accuracy, and that a validated decision rule is assessment written down where it can be checked. The model places every intervention on one continuous axis of magnitude and says to begin with the least invasive input that can carry the message. When the picture calls for more, escalation is correct dosing and delay becomes its own kind of harm. Referring an older adult for imaging is that rule executing rather than conservative care conceding.
13The sources
References
12 primary sources, each linked to its record. Figures quoted on this page were checked against the published abstract.
Related evidence