Percutaneous cement augmentation of the osteoporotic compression fracture — balloon tamp height restoration and PMMA cement injection under fluoroscopic guidance
Vertebral compression fractures (VCFs) affect an estimated 25% of postmenopausal women over age 50 and are the most common osteoporotic fragility fracture, with roughly 1.5 million occurring annually in the United States alone. Correct patient selection — distinguishing an acute, non-healed, painful fracture from a chronic or incidentally-discovered one — is the single greatest determinant of procedural benefit and is the first checkpoint in every ISIS (Spine Intervention Society) and NASS coverage guideline.
The defining imaging finding that separates an acute, symptomatic fracture from an old, healed one is bone marrow edema on MRI:
• STIR (short-tau inversion recovery) sequence: hyperintense signal within the vertebral body indicates ongoing marrow edema/microfracture — the strongest predictor of pain relief from augmentation • T1-weighted: corresponding hypointensity of the edematous body • T2 fat-saturated: complements STIR; useful when STIR unavailable • Absence of edema (fatty/sclerotic marrow, no STIR signal): fracture is chronic — augmentation unlikely to relieve pain, and should generally be avoided • CT: defines fracture morphology, posterior wall integrity, and pedicle anatomy for procedural planning; a vacuum cleft sign (intravertebral gas) suggests non-union/avascular change and often predicts excellent cement fill • Radiographs (lateral): quantify vertebral height loss using the Genant semiquantitative grading (Grade 1: 20–25%, Grade 2: 25–40%, Grade 3: >40% height loss) and measure local kyphotic (Cobb) angle
When MRI is contraindicated (pacemaker, severe claustrophobia), bone scan (99mTc-MDP) with focal uptake at the fracture level is an accepted alternative to confirm metabolic activity.
A fracture without marrow edema on MRI is a contraindication to augmentation in most society guidelines (ISIS 2014, ASBMR) — cement placed into a biomechanically inert, already-healed vertebra offers no pain benefit and exposes the patient to procedural risk without corresponding gain.
Indications: painful VCF (VAS ≥ 4–5) refractory to 2–6 weeks of conservative management (analgesia, bracing, activity modification), progressive height loss/kyphosis on serial imaging, VCF causing impaired mobility or pulmonary compromise from severe kyphosis, and painful vertebral metastasis or myeloma deposit (oncologic vertebroplasty, a distinct but related indication).
Absolute contraindications: active local or systemic infection, uncorrected coagulopathy, allergy to PMMA components, asymptomatic/stable fracture, and retropulsed fragment causing significant canal compromise with neurologic deficit (relative — may favor open decompression/fusion instead).
The 2009 sham-controlled trials (Buchbinder et al. and Kallmes/INVEST) found no benefit of vertebroplasty over a sham procedure at 1 and 6 months, prompting substantial controversy. Subsequent, better-designed trials enrolling patients with more acute fractures (VERTOS II, VERTOS IV, and the 2016 Clark et al. VAPOUR trial) — which specifically required marrow edema and fracture duration <6 weeks — demonstrated clear, clinically meaningful pain reduction over sham/conservative care. The current consensus (NASS 2014 coverage policy, ISIS 2014) is that patient selection — acute fracture with edema, adequate pain, failed conservative therapy — is the decisive variable explaining the divergent trial results.
Safe, accurate needle placement is the technical foundation of the entire procedure: every subsequent step — balloon position, cement fill pattern, and leak risk — is dictated by where the trocar tip ends up within the vertebral body. The transpedicular approach exploits the pedicle as a protected bony corridor that keeps the needle away from the exiting nerve root and epidural venous plexus.
The patient is positioned prone on a radiolucent table with bolsters under the chest and pelvis to extend the spine slightly (this alone can partially reduce the fracture). Conscious sedation with local anesthetic infiltration to the periosteum of the pedicle is standard; general anesthesia is reserved for multilevel or anxious/uncooperative patients.
The C-arm is aligned to obtain a true AP "bullseye" view of the target pedicle — the pedicle oval appears symmetric and the spinous process is centered between the pedicles. The needle entry point is at the lateral aspect of the pedicle (10 o'clock on the left, 2 o'clock on the right for a transpedicular approach), angled roughly 10–15° medially so the trajectory travels through the pedicle's bony corridor rather than breaching its medial or lateral cortex, which would risk the neural foramen or paraspinal vessels respectively.
Transpedicular approach (preferred for lumbar levels with large pedicles, ≥8mm width): needle passes entirely through the cancellous bone of the pedicle, providing maximum protection from the exiting nerve root and segmental vessels; typically only a single unilateral needle is needed if the tip can be walked to reach the midline of the vertebral body (biconvexity of most bone cements allows adequate contralateral fill from a single trocar).
Extrapedicular (parapedicular/costovertebral) approach (preferred at thoracic levels, T1–T10, where pedicles are narrower, often <5mm): the needle passes lateral to the pedicle, along the costovertebral junction, entering the vertebral body just above the rib head — avoiding pedicle fracture risk in small thoracic pedicles.
Unipedicular vs. bipedicular: a single well-placed transpedicular needle reaching the midline achieves comparable biomechanical stiffness restoration to bipedicular (bilateral) access in most biomechanical studies, with less fluoroscopy time and cost; bipedicular access is favored when a single trocar cannot achieve adequate contralateral cement spread or in the kyphoplasty technique using two simultaneous balloons for symmetric height restoration.
Confirmation that the trocar tip has crossed the posterior vertebral body wall — and is therefore safely within the vertebral body rather than still in the epidural space — must always be verified on the lateral fluoroscopic view before any further advancement, balloon insertion, or cement injection.
Balloon kyphoplasty adds a mechanical height-restoration step absent from vertebroplasty: an inflatable bone tamp compacts cancellous bone to create a low-pressure cavity and, in appropriately mobile (non-united) fractures, partially reduces the endplate depression and local kyphosis before cement is ever introduced — theoretically lowering intraosseous pressure during injection and reducing leak risk relative to vertebroplasty.
Once the working cannula is confirmed in the anterior third of the vertebral body, the inflatable bone tamp (IBT) is advanced through it and positioned under lateral fluoroscopy. Inflation is performed with a calibrated syringe/manometer delivering dilute contrast, in small incremental steps (typically 30–50 psi at a time), continuously monitoring:
• Inflation pressure (psi) — plotted against injected volume; a sudden pressure drop suggests cortical breach • Balloon volume (mL) — tracked to avoid over-distension • Fluoroscopic balloon shape and position — inflation is halted immediately if the balloon approaches the lateral, anterior, or (critically) posterior cortex • Endplate elevation — visualized directly on lateral fluoro as the compressed endplate lifts
Maximum recommended inflation pressure is generally capped around 300–400 psi (per device IFU); balloons are engineered to preferentially compact bone rather than rupture, but posterior cortical contact is an absolute stop signal regardless of pressure reached.
Height restoration is most successful when the fracture is mobile (non-united), acute (typically <3 months, though later fractures can still respond), and has an intact posterior wall. Predictors of poor height restoration include:
• Chronic fracture with bony bridging/fibrous union — cavity creation is difficult and cement essentially fills a static defect rather than restoring geometry • Intravertebral vacuum cleft — often responds paradoxically well because it represents avascular, mobile segments • Severe (>70%) height loss — biomechanically harder to reduce fully • Posterior wall fracture — relative caution needed given the theoretically increased posterior leak risk during both balloon inflation and cement injection
After target height/angle correction (or maximum safe pressure) is reached, the balloon is deflated and withdrawn, leaving behind a well-defined, comparatively low-pressure cavity — the key mechanical rationale distinguishing kyphoplasty from vertebroplasty, in which cement is forced directly into intact, unyielding cancellous bone at higher injection pressures.
The FREE trial (Wardlaw et al., Lancet 2009) randomized 300 patients to balloon kyphoplasty vs. non-surgical care and found significantly greater improvement in SF-36 physical function and back pain at 1 month, sustained at 12 months, alongside a statistically significant height restoration advantage over conservative management.
Cement injection is the highest-risk step of the procedure: polymethylmethacrylate (PMMA) is delivered as a viscous, radiopaque paste that must fill the fracture cavity and interdigitate with surrounding trabecular bone while never being allowed to reach the posterior vertebral body wall, basivertebral venous plexus, or segmental veins — the pathways by which cement can enter the epidural space or the venous circulation.
PMMA bone cement is prepared by mixing a powder component (pre-polymerized PMMA beads, barium sulfate or zirconium dioxide radiopacifier, benzoyl peroxide initiator) with a liquid monomer (methyl methacrylate, hydroquinone stabilizer, N,N-dimethyl-p-toluidine accelerator). The exothermic polymerization reaction progresses through distinct phases:
• Mixing phase (0–2 min): low viscosity, "toothpaste"-like • Working phase (2–8 min, temperature and brand-dependent): viscosity rises steeply — cement is injected specifically during this high-viscosity phase to minimize the risk of low-viscosity cement tracking into veins • Setting/hardening phase (8–15 min): exothermic reaction peaks (60–90°C locally), cement fully hardens
Injection is performed under continuous, real-time lateral fluoroscopy (with intermittent AP checks) in small 0.5–1 mL aliquots, pausing after each aliquot to assess the cement front. A cement delivery system (bone cement gun or hydraulic remote injector) allows the operator's hands to remain out of the direct fluoroscopic beam.
Cement extravasation is classified by anatomic route:
• Type B (basivertebral vein): cement tracks posteriorly through the basivertebral venous channel toward the epidural space — the single most concerning type of leak due to proximity to the thecal sac and cord • Type S (segmental vein): cement enters paravertebral segmental veins, potentially embolizing to the pulmonary circulation • Type C (cortical/anterior): cement extrudes through an anterior or lateral cortical defect into paravertebral soft tissue — usually clinically silent • Type D (disc space): cement leaks through an endplate fracture into the adjacent intervertebral disc — associated with increased risk of adjacent-level fracture
Most leaks (60–90% in various series) are asymptomatic and detected only on post-procedure CT; symptomatic neurologic compromise from epidural leak occurs in well under 1% of cases. Injection is stopped immediately if cement is seen approaching the posterior third of the vertebral body, entering a venous channel, or tracking toward the neural foramen — the posterior 1/3 of the vertebral body on lateral fluoroscopy functions as a hard "no-go" boundary throughout injection.
Digital subtraction venography prior to cement injection (injecting a small test bolus of contrast through the trocar) can reveal a dominant venous outflow channel and prompt repositioning of the needle tip before committing to cement — a technique increasingly used in the highest-risk levels (T4–T8, or a fracture with a documented posterior wall defect).
The final phase evaluates what was achieved — pain relief, restored vertebral geometry, cement fill pattern — against what could go wrong, from clinically silent cement leaks to rare but life-threatening cement pulmonary embolism, while also confronting the well-documented association between cement augmentation and new fractures at adjacent vertebral levels.
Immediately post-procedure, a non-contrast CT of the treated level (or high-quality biplanar fluoro spot images) is obtained to document final cement distribution, confirm interdigitation with trabecular bone on both sides of the fracture line, and formally grade any extravasation. Clinically, most patients report substantial pain relief within hours to 1–2 days — a much faster trajectory than the 6–12 week natural history of conservative fracture healing — attributed to immediate mechanical stabilization of the micromotion at the fracture site (the same mechanism thought to generate fracture pain in the first place) rather than a purely analgesic effect of the procedure.
Height and kyphotic angle are re-measured on standing or supine lateral radiographs; kyphoplasty series typically report 30–50% restoration of lost height in mobile fractures, while vertebroplasty (no balloon) restores height only incidentally via prone positioning, if at all.
Cement pulmonary embolism (CPE) occurs when liquid cement enters the segmental/paravertebral venous system and travels via the vena cava to the pulmonary arterial tree, where it can harden in situ. Reported incidence varies enormously by detection method (0.06–4.6%; higher when every patient undergoes routine post-procedure chest CT versus only when clinically suspected). Most CPE is asymptomatic and found incidentally as linear radiopaque densities in a pulmonary artery branch; symptomatic CPE (dyspnea, pleuritic chest pain, hypoxia, and rarely hemodynamic collapse) is treated analogously to a small thromboembolism — anticoagulation is considered on a case-by-case basis given the mixed foreign-body/inflammatory nature of a cement embolus, and multidisciplinary discussion with pulmonology is recommended for any symptomatic case.
Other systemic risks include transient hypotension from the exothermic polymerization/monomer absorption (bone cement implantation syndrome — well described in arthroplasty, rarer in vertebral augmentation given smaller cement volumes), and, very rarely, fat embolism from marrow displacement.
One of the most persistently debated long-term risks is the development of new compression fractures at levels adjacent to the augmented vertebra, reported in roughly 10–20% of patients within the first year, with the highest risk in the first 60 days. Two (non-exclusive) hypotheses are debated in the biomechanical literature:
• Cement stiffness mismatch: PMMA is markedly stiffer than native osteoporotic trabecular bone, which may transfer additional axial load to the endplates of adjacent vertebrae, particularly if cement leaks into the adjacent disc space (Type D leak) • Natural disease trajectory: patients who sustain one osteoporotic VCF have, by definition, severely compromised bone mineral density and are simply at markedly elevated baseline risk of further fragility fractures regardless of any treatment received — several case-control and registry studies (and a 2018 Cochrane systematic review of vertebroplasty) found adjacent-level fracture rates statistically similar between augmented and conservatively-managed cohorts, supporting the natural-history explanation as dominant
Regardless of mechanism, all patients undergoing vertebral augmentation should be started or continued on osteoporosis pharmacotherapy (bisphosphonates, denosumab, or an anabolic agent such as teriparatide/romosozumab per endocrinology/ASBMR guidance) — cement augmentation treats the fracture, not the underlying disease.
A 2018 Cochrane review (Buchbinder et al.) concluded that vertebroplasty provides little to no clinically important benefit over sham for average VCF populations, while more selective, edema-positive, acutely painful cohorts in VERTOS IV and VAPOUR showed clear benefit — reinforcing that patient selection, not the procedure itself, is what the evidence actually turns on.