History

The classic symptoms from an acute vertebral compression fracture include deep pain with sudden onset, midline location, and exacerbation by motion and standing. These fractures may occur with little or no trauma. The pain often diminishes little in the first week and then fades gradually over the next several weeks to 3 months.¹ Lateral radiation may be present, but persistent radiation of the pain in a radiculopathic pattern is rare. When there is substantial kyphosis, patients may also suffer from difficulty breathing, anterior chest wall pain, and gastrointestinal discomfort.

In addition to evaluating the nature of the pain, the interview should include assessment of the effect of the pain on activities of daily living and sense of well-being. The practitioner may choose to quantify these factors using such questionnaires as the Oswestry Disability Questionnaire, Roland-Morris Disability Questionnaire, and Medical Outcomes Study 36-Item Short-Form Health Survey (SF-36). Such indices can be used in conjunction with a visual analogue pain scale to follow the success of treatment.

The physician should evaluate ongoing or potential treatment of the underlying disease process. For osteoporotic patients, the interview is an opportunity to make sure that appropriate treatment of the osteoporosis is under way. The severity of osteoporosis and the use of antiosteoporosis medications substantially alter the risk of additional fractures. For patients with neoplasm, timing of the treatment may be affected by ongoing chemotherapy and radiation.

The physician should evaluate the patient’s general health and coexisting medical problems to perform the procedure safely. This evaluation allows the practitioner to determine the feasibility and the safest and most effective means of performing the procedure. Consultation with an anesthesiologist may be warranted. The physician should search for possible contraindications to the procedure (e.g., use of warfarin) and for allergies to planned pharmaceutical use (e.g., prophylactic antibiotic).

Physical Examination

The practitioner should perform a directed physical examination. This examination includes inspection of the back, palpation for focal areas of tenderness, and correlation of the site of pain with anatomic landmarks. In difficult cases, examination of the sites of pain and tenderness can be performed with fluoroscopic assistance to localize the pain to specific anatomic structures. Although point tenderness at the spinous process of the fractured vertebra is the typical finding in unhealed compression fractures, the absence of typical focal tenderness does not preclude the presence of such a fracture. Imaging is a more reliable test than this portion of the physical examination.⁶ Assessment of lower extremity neurologic function is especially important in patients with symptoms suggestive of myelopathy, radiculopathy, or spinal stenosis. Evaluation of the heart and lungs is necessary for the safe use of sedation for the procedure.

Imaging

Imaging of the spine is undertaken in all cases to confirm the clinical diagnosis and to plan the procedure. Radiographs are the imaging study of choice for the initial imaging evaluation (Fig. 173-1). Anteroposterior (AP) and lateral views should be obtained. This study allows the practitioner to confirm the presence of fracture, determine the location of fractures, assess the degree of height loss and kyphosis, and identify anatomic variants. Whenever possible, comparison to prior studies is valuable. A single radiograph almost never allows the practitioner to distinguish a new or unhealed fracture from a chronic, healed fracture. With prior radiographs (“old gold”) for comparison, new compression fractures can be identified. If serial radiographs taken several weeks apart show a new compression fracture in a patient with classic history and symptoms, no further imaging may be necessary. However, since back pain is a difficult diagnostic problem, most patients benefit from advanced imaging.

FIGURE 173-1 Lateral radiograph of the lumbar spine. The loss of height and wedge deformity signal compression fractures at L1 to L4. It is not possible to determine the age of these fractures from a single radiograph.

Magnetic resonance imaging (MRI) is the test of choice for this additional evaluation. The main goals of MRI are to (1) distinguish new or unhealed fractures from healed fractures, (2) identify other causes of pain, and (3) evaluate the risk of symptomatic cement leakage during the procedure. For these purposes, the most useful sequence is a T2-weighted sequence with fat suppression. We prefer using inversion recovery (i.e., a short tau inversion time recovery, or STIR, sequence) to achieve the fat suppression, since it is less prone to artifact than frequency-selective fat saturation techniques. On these MRI sequences, recent or unhealed fractures show a hyperintense signal within the bone marrow (Fig. 173-2). This signal may be present in poorly marginated areas filling much of the vertebral body or as discrete curvilinear fracture clefts. In contrast, healed fractures have a marrow signal similar to that of unfractured vertebra. These sequences also provide an adequate evaluation of the spinal canal and show most alternative causes of pain that might mimic fracture. Additional imaging with a sagittal T1-weighted sequence and axial T1- and T2-weighted images may be helpful. In most patients with a typical clinical history and physical examination for osteoporotic compression fracture, the STIR sequence alone is adequate, and the practitioner can reserve the full set of MRI sequences for patients whose pain is atypical or who have STIR abnormalities that require clarification. We have found value in obtaining an additional MRI shortly before the procedure when either (1) there has been worsening of the patient’s pain since the initial evaluation or (2) the imaging study used for the initial evaluation is more than 3 months old.⁷

FIGURE 173-2 MRI with a sagittal STIR sequence. The hyperintense signal in the T7 vertebral body is consistent with an unhealed compression fracture. While there is also wedge deformity at T11 and T12, the absence of marrow signal abnormality identifies these compression fractures as old and healed.

In patients who cannot tolerate MRI (e.g., those with a pacemaker), bone scan is the test of choice. Bone scan does not provide any evaluation of the soft tissues or spinal canal, but it does allow the differentiation of healed and recent fractures. The recent fractures take up the injected technetium 99m–medronate tracer in much higher concentrations (Fig. 173-3). The relative sensitivity and specificity of bone scan and MRI for this purpose have not been established.

FIGURE 173-3 Bone scan. Areas of increased uptake in the L1 and L2 vertebrae represent bone turnover in recent or unhealed fractures.

CT is not necessary for most osteoporotic compression fractures but is useful for preprocedure evaluation of burst fractures and metastases. In these circumstances, there may be substantial loss of integrity of the posterior vertebral body cortex (Fig. 173-4). This loss increases the risk of the procedure, specifically the risk of posterior leakage of cement or posterior displacement of bone or tumor.⁸ CT provides the best means for assessing this cortex. CT is also the test of choice for postprocedure evaluation of unexpected symptoms (see the later section on risks).

FIGURE 173-4 Loss of posterior vertebral body integrity. A, MRI shows metastasis with a pathologic compression fracture at T1. B, CT scan demonstrates more clearly the extensive destruction of the posterior cortex.

Sedation

Analgesia is necessary to perform vertebroplasty and kyphoplasty safely and with minimal discomfort. This process begins with the informed consent. Fully informing the patient of what to expect in the operating room or procedure suite facilitates patient comfort, cooperation, and satisfaction. Satisfactory analgesia usually can then be achieved with local anesthetics and moderate sedation, but in some cases general endotracheal anesthesia is needed. Vertebroplasty has generally been performed with moderate sedation, while kyphoplasty has more commonly been performed with general anesthesia. At our institution, both are usually performed using moderate sedation with intravenous midazolam and fentanyl. Continuous patient monitoring is performed with electrocardiography, blood pressure measurement, and pulse oximetry. The drug delivery and monitoring are performed by certified nursing personnel. In patients with substantial preexisting respiratory or cardiac disease, an anesthesiologist is asked to evaluate the patient and determine whether monitored anesthesia care is warranted.

Patient Positioning

The ideal patient position for thoracic and lumbar procedures is prone. This position allows the practitioner to place the needle from either side and simplifies positioning of the biplane fluoroscopy unit or C-arm. Furthermore, this position and proper cushion support maximize extension of the fractured segments, promoting kyphosis reduction.⁹ Although the patient can be rotated slightly, substantial patient angulation can make this a more difficult procedure. A special table and cushions may be used to support the head and body, but good results can be achieved with a flat fluoroscopy table and careful placement of cushions. The patient’s arms should be placed sufficiently toward the head to keep them out of the path of the fluoroscope. Patients can usually be placed in the prone position despite their painful compression fracture. If necessary, moderate sedation can be initiated before the patient is moved. Care must be taken when transferring aged or osteoporotic patients, since transfer alone may cause new fractures.

Antibiotic Prophylaxis and Skin Preparation

The risk of infection is minimized with the meticulous use of operating room procedures for sterile preparation of the skin, draping, operator scrub, and sterile gowns, masks, and gloves. Antibiotic prophylaxis is usually provided. The most common form is the intravenous administration of cefazolin (or clindamycin for those with cephalosporin allergy) just before or during the procedure. Alternatively, tobramycin may be added directly to the PMMA.

Image Guidance

Imaging guidance is necessary to ensure the success and safety of vertebroplasty and kyphoplasty. Generally, needle placement and cement placement are performed with fluoroscopy, but in rare cases, CT may be necessary. Fluoroscopy has inherently lower contrast resolution and makes greater demands on imaging technique, on the operator’s knowledge of radiographic anatomy, and on the operator’s ability to integrate planar data into a three-dimensional picture. However, the compelling reasons for using fluoroscopy whenever possible are the lower radiation dose and the ability to monitor the cement infusion continuously as the injection proceeds.

There are no established minimum standards for the fluoroscopic equipment used in vertebroplasty. However, one key to avoiding complications in this procedure is the ability to see the needles, bone landmarks, and opacified cement at all times. At least one C-arm unit is necessary. A dedicated biplane fluoroscopy suite is optimal. The quality of these fluoroscopic systems varies widely, and many older C-arm systems used for general orthopedic work put the operator at a disadvantage. The newer image intensifiers and digital fluoroscopy systems available from numerous vendors provide the best image quality and lowest radiation exposure. Shielding to minimize operator exposure to scatter radiation is necessary if the practitioner performs this procedure routinely.

Needle Placement

Once the patient is properly positioned in the fluoroscopy unit, the needle trajectory can be planned. There are two principal approaches for needle placement: transpedicular and parapedicular. The transpedicular approach takes the needle directly through the full length of the pedicle. It has the advantage of a long intraosseous path protecting the nerve roots and blood vessels from direct injury and protecting the soft tissues from cement leakage along the tract. However, in pedicles that have a long axis parallel to the sagittal plane, this approach limits the practitioner’s ability to achieve a final needle tip position near the midline. A parapedicular approach may then be the technique of choice. Here, the needle is directed along the lateral cortex of the pedicle, and then additional angulation is applied as the needle pierces the cortex at the junction of the pedicle and vertebral body.

For either approach, there are two image guidance strategies. The “end-on” approach uses fluoroscope angulation to place the barrel of the pedicle cortex over the target site in the vertebra body. This view simplifies trajectory planning and is the simplest way of preventing the needle from violating the medial cortex of the pedicle. Alternatively, the AP approach may be used. This view may be easier to achieve with some fluoroscopy systems and in some cases provides superior delineation of the pedicle margins. Once an approach has been selected, the skin is anesthetized with a small amount of subcutaneous lidocaine or bupivacaine. Following the planned needle trajectory, the periosteum is anesthetized in a similar fashion. The planned trajectory can be assessed and adjusted from fluoroscopic views of this smaller needle at the bone surface. A small nick is made in the skin, and an 11- or 13-gauge needle is placed.

End-on Technique

For the end-on technique, the operator rotates the fluoroscopy unit into an ipsilateral oblique view that places the fluoroscopy beam and needle tract perfectly parallel to each other (Fig. 173-5). The image intensifier is first rotated to a true AP position, aligning the spinous process midway between the pedicles. The craniocaudal angulation is changed to bring the pedicles to the midportion of the vertebral body. The image intensifier is then rotated to bring the ipsilateral pedicle so that its medial cortex is at the middle third of the vertebral body. This rotation can only be continued as long as the medial cortex of the pedicle remains clearly visible. The needle is placed so that it is “end on” to the image intensifier and appears as a small circle or ellipse. Once the needle has been advanced to the bone surface, small corrections in the craniocaudal angulation can be made using a true lateral view (care is needed to angle this view so that the pedicles project over each other). The needle is then advanced through the pedicle, maintaining the end-on appearance of the needle. Once the needle has traversed the pedicle, it is advanced using the lateral view to the junction of the anterior and middle thirds of the vertebral body.

FIGURE 173-5 Needle placement with end-on approach. A, Initial AP view. B, Craniocaudal angulation to place the pedicle at the vertical middle of the vertebral body. C, Ipsilateral oblique view to place the pedicle over the medial aspect of the vertebral body. D, Ipsilateral oblique view with the needle (black circle) in proper alignment. This view is maintained as the needle is advanced through the pedicle. Care is taken to avoid the medial cortex of the pedicle. E and F, Lateral and AP views with final needle location.

AP Technique

For the AP technique, the craniocaudal angulation is adjusted to bring the vertebral body end plates perpendicular to the image (Fig. 173-6). The skin entry site is placed about 1 cm superior and lateral to the center of the pedicle but must be adjusted for the thickness of the posterior paraspinal soft tissues. The needle is advanced anteriorly, medially, and caudally. By the time it reaches the bone surface, the tip should project over the upper outer cortex of the pedicle. Again, small corrections in the craniocaudal angulation can be made using a true lateral view. The needle is advanced so that its tip projects over the center of the pedicle on both AP and lateral views. The tip should project over the medial pedicle cortex as the needle traverses the posterior third of the vertebral body. The final position is again adjusted on the lateral view.

FIGURE 173-6 Needle placement with the AP approach. A, AP view with the site of skin entry (black dot). B, Needle appearance as it enters the posterior pedicle surface. C, Needle appearance when it has been advanced to the midpoint of pedicle on the lateral view. If the needle tip is more medial at this point, there is risk for traversing the medial pedicle cortex.

The image guidance for a parapedicular approach is quite similar. The needle is placed lateral and superior to the lateral cortex of the pedicle and enters the vertebral body at its junction with the pedicle. The position at which bone is encountered (i.e., at the junction of the pedicle and vertebral body) is more anterior on the lateral view.

The advance of the large needle is achieved with either using a drilling motion or tapping with a small hammer. The needle position should be checked frequently in the oblique (or AP) view for two reasons. First, it is important to ensure that the needle does not violate the medial cortex of the pedicle. Second, needle trajectory cannot be markedly altered once the needle has passed beyond a small distance into the pedicle. Accurate placement is also aided by positioning the needle centrally in the field of view to avoid inaccuracies from parallax.

The planning for the lateral position of the tip should be adjusted for the procedure to be performed and the shape of the vertebral body. Vertebroplasty for osteoporotic vertebral compression fractures was initially performed using two needles and a bipedicular approach. This technique allows reliable filling of both halves of the vertebral body, independent of the precise lateral position of the needle tip within its half of the vertebral body. However, some practitioners have adopted a unipedicular approach. This technique requires a single needle and decreases the risk and time associated with needle placement. For the unipedicular approach, a near-midline position of the needle tip provides more reliable filling of the vertebral body. In my experience, both techniques work well, and my colleagues and I base our approach on the number of levels to be treated and the configuration of the fractured vertebra. For example, in a severely collapsed vertebra, there is often substantially more marrow space at the lateral aspects of the vertebral body. In this circumstance, a bipedicular approach with lateral positioning of the needle tip provides reliable bilateral filling with little risk of symptomatic cement leakage.¹⁰ With more lateral tip positions, it is important to remember that the vertebral body is round and that the anterior cortex of the vertebra along the needle trajectory is posterior to the radiopaque line of anterior cortex on the fluoroscopic image.