The Role of Vertebroplasty in Steroid-Induced Vertebral Osteoporotic Fractures

Background Data: Glucocorticoid-induced osteoporosis is a well-known significant health problem worldwide that causes morbidity and mortality. Glucocorticoid-induced vertebral compression fracture is one of the most common types of osteoporotic fractures associated with significant morbidities such as severe agonizing pain, limited mobility, and spinal deformity. Percutaneous vertebroplasty (PV) can be performed in the treatment of refractory back pain in these cases of osteoporotic vertebral fractures (OVFs) without significant complication when conservative treatment fails. Purpose: To evaluate the clinical and radiographic outcomes of thoracolumbar OVFs treated with PV in adult osteoporotic patients with long-term corticosteroid therapy. Study Design: Retrospective clinical case series. Patients and Methods: Twenty-eight patients with painful steroid-induced OVFs underwent vertebroplasty in 61 vertebral levels. Inclusion and exclusion criteria were applied. Preoperatively, all patients were subjected to intensive diagnostic workups, including history taking and clinical and radiological examinations, such as CT scan and MRI. The procedure was guided by C-arm and considered complete when the unfilled area was less than 25% of the vertebral body height in the lateral radiograph. Visual Analogue Scale (VAS) and Oswestry Disability Index (ODI) were used to assess preand postoperative back pain and functional status of our patients. Results: Seventeen patients (60.71%) were males and eleven patients (39.28%) were females. The mean age was 57 ± 5.04 (range, 49–68) years. The mean follow-up of the patients was 38.4 2± 11.16 (range, 24–60) months. Overall, 61 levels were reported including 10 patients (39%) with a single level and 18 patients (61%) with two levels or more as follows: two levels in eight patients, three levels in six patients, four levels in three patients, and one patient with five levels. The most common affected region was the thoracolumbar junction (T11, T12, and L1) in 38.69%. Back pain VAS decreased from 7.29 ± 1.04 before vertebroplasty to 3.25 ± 0.75 one week after vertebroplasty, 1.68 + −0.66 at 12 months postoperatively, and 3.11 ± 1.13 at final follow-up 24 months postoperatively (p < 0.001). ODI improved from 40.82 ± 12.32 (range, 14–66) preoperatively to 16.68 ± 3.19 (range, 10–24) at 12 months postoperatively and 20.92 ± 4.66 (range, 10–30) at final follow-up 24 months postoperatively (p < 0.00)


INTRODUCTION
Glucocorticoid-induced osteoporosis is a wellknown significant health problem worldwide leading to morbidity and mortality. Glucocorticoidinduced vertebral compression fracture (G-OVF) is a common type of osteoporotic fractures associated with significant morbidities such as severe pain, limited mobility, and spinal deformity. 1 The pathogenesis of steroid-induced osteoporosis is complex, including reduced bone formation and increased bone resorption. The fracture risk in cases of OVFs is mainly related to time and dosage of glucocorticoids (i.e., the higher the dosage and the longer the treatment, the greater the risk). 23 Primary osteoporotic vertebral compression fracture affects mainly the elderly and postmenopausal women, 33 whereas G-OVF affects men more frequently and the affected age group is relatively younger. 39 The main presenting symptom in patients with G-OVFs is back pain which is associated in most cases with limited daily activity and kyphotic deformity; however, vertebral height loss is the main radiological sign. 35 Patients with glucocorticoid osteoporosis, such as chronic obstructive pulmonary disease, rheumatoid arthritis, and leprosy, cannot tolerate bracing. 35 Regardless of the use of corticosteroids, the inflammation itself has a harmful effect on bone remodeling by increasing bone resorption and decreasing the formation. Rheumatoid arthritis alone doubles the vertebral fractures. 24 The risk of fractures is increased up to 2.6 and the global prevalence of vertebral fractures in patients receiving long-term glucocorticoids has been reported to be 37%. 35,36 Percutaneous vertebroplasty (PV) was introduced in 1987 by Gailbert and Diamond for the treatment of large spinal hemangioma. It is an image-guided minimally invasive procedure through which the polymethylmethacrylate (PMMA) cement is injected into the fractured vertebra. 7 The surgical metallic fixation of the osteoporotic spine with long-term use of glucocorticoids has usually failed because of poor bone quality and comorbidities that preclude prolonged anesthesia. 23,39 VP should be considered in patients who suffer from severe persistent pain despite adequate conservative treatment for at least three weeks. 23 This study aims to evaluate clinically and radiographically thoracolumbar OVFs treated with VP in adult osteoporotic patients with longterm use of corticosteroid therapy.

PATIENTS AND METHODS
We retrospectively reviewed the medical records of all subacute cases of OVFs due to longterm use of steroid therapies that underwent operations in the Orthopedic Department, Kafr El-Sheikh University Hospital, and Neurosurgery Department, Zagazig University Hospital, between January 2012 and December 2017. We traced 28 patients with painful OVFs who underwent VP at 61 levels with complete clinical and radiological data. The mean follow-up of the patients was 38.42 ± 11.16 (range, 24-60 months).
Inclusion criteria for the current study were as follows: (1) patients recently diagnosed with OVFs due to long-term therapy of corticosteroids and failed medical treatment for 3-6 weeks with persistence of pain; (2) confirmed cases of osteoporosis by Dual Energy X-ray Absorptiometry (DEXA) with T-score of −2.5 or less; (3) patients with preoperative pain intensity according to Visual Analogue Scale (VAS) ≥ 5; (4) levels from T5 down to L5, and patients with incomplete data. On the other hand, exclusion criteria included the following: (1) patients with G-OVFs for more than three months; (2) patients with infection, coagulopathies, tumor, or trauma; (3) those who had a neurological deficit. Relative contraindications included improved cases on conservative treatment with VAS less than 5, vertebral body collapse more than 75%, posterior vertebral wall destruction involvement (for fear of cement leakage), and any level above T5 because image guidance is difficult due to shoulder shadow, making the procedure unsafe with the risk of cement leakage and neurological deficit, the smaller targets are more difficult to cannulate accurately, and the risk of neurologic damage from needle placement or PMMA migration increases. Preoperatively, routinely, an intensive diagnostic workup was performed in all patients. It included medical history, clinical assessment including VAS for the back pain and ODI for functional status, and radiological examination such plain radiographs, CT scan, and MRI. MRI thoracolumbar spine was essential for diagnosing recent fractures (vertebral edema), excluding malignant lesions, and detecting canal compromise. Surgical Technique Patients were placed prone on a radiolucent table, and the procedure was performed under local anesthesia and sedation. Under C-arm guidance in anteroposterior (AP) and lateral views, a large Jamshidi needle, typically 10 or 11 G for lumbar and 8 G for thoracic vertebra with a beveled tip, was used to puncture the fractured vertebra through the pedicles. The needle moves parallel to the superior and inferior edges of the pedicle or in a slightly descending course through the pedicle. The needle was advanced toward the anterior third of the vertebral body under fluoroscopic guidance; it needed 5 to 7 minutes to mix the radiopaque cement to be ready and inject into the vertebral body using small 1 cc insulin syringes or a cement gun under continuous fluoroscopic monitoring (Figures 1 and 2). In most cases, between 3 to 4 mm cement has been used (3 to 4 insulin syringes) per level. The procedure was terminated immediately if cement reached the posterior quarter of the vertebral body or if significant leakage into the disc space occurred. If the unfilled area was more than 25% of the vertebral body height in the lateral radiograph, another needle was inserted from the contralateral side with the needle aiming at the unfilled area. The procedure was considered complete when the unfilled area was less than 25% of the vertebral body height in the lateral radiograph. Skin closure was done using single sutures ( Figure 3). Postoperatively, following PVP, the patient was kept in the supine position in bed for 4 hours and after that, could ambulate with assistance and a postoperative X-ray was performed. The patient was observed and discharged on the same day. Postoperative Outcome Evaluation Clinical Assessment. All patients were evaluated at our outpatient clinic at 1 week, 6 weeks, 6 months, 1 year, and 2 years postoperatively or as recommended in complicated patients. Overall back pain and functional status (quality of life) were assessed with VAS and ODI, respectively, and any neurological or systemic complications were reported. Radiological Evaluation. Postoperative plain radiographs and CT-scans were used to measure the anterior vertebral height of cemented vertebrae, the quantity of bone cement used, cement extravasation, and new level vertebral fractures.

Statistical Analysis
Collected data were statistically compared using tests from the Statistical Package for the Social Sciences (SPSS) program version 17 (Chicago, Illinois, USA). Analysis of Variance (ANOVA) test was used and p value ≤ 0.05 is considered significant.

RESULTS
Twenty-eight patients, with 61 levels, were retrospectively included in this study. Seventeen patients (60.71%) were males and 11 patients (39.28%) were females. The mean age was 57 ± 5.04 (range, 49-68) years. Ten patients (39%) had a single level and 18 patients (61%) had two levels or more (2 levels in 8 patients, 3 levels in 6 patients, 4 levels in 3 patients, and 5 levels in 1 patient) ( Table 1). Treated levels were distributed nonuniformly from T5 to L5, and 24 (39.33%) of the 61 fractures occurred at the thoracolumbar junction (T11 to L1) ( Table 2). The patient with the affected five levels was contagious and managed in consecutive sessions. The mean cement volume was 3.58 ± 0.557 mm (range, 3-5.8 mm) cement per vertebra. Sixteen patients (57%) had previously sustained at least one fragility fracture, and 43% were actively receiving osteoporosis treatment (primarily oral bisphosphonates) at the time of entry into the study. All patients had previously received systemic steroid medications for at least 6 months, and 61% were currently taking steroids. The preoperative mean anterior vertebral body height was 17.22 ± 2.10 and increased to 17.57 ± 2.80 one month postoperatively; the difference was statistically insignificant (p = 0.408). Pain assessed by the VAS significantly decreased from 7.29 ± 1.04 before vertebroplasty to 3.25 ± 0.75 one week after surgery (p < 0.001) and improved to 3.11 ± 1.13 at the final follow-up (Table 3) (p < 0.001). ODI improved significantly from 40.82 ± 12.32 (range, 14-66) preoperatively to 25.64 ± 5.22 (range, 16-36) postoperatively (p < 0.0001) and 20.92 ± 4.66 (range, 10-30) at the final follow-up (p < 0.0001) (Tables 4 and 5). Asymptomatic cement extravasation was seen in 7 levels (11, 47%). Four of the 7 leaks were in adjacent disc spaces, and each leak had been anticipated preoperatively based on the recognition of endplate disruption as seen on preoperative CT. Two small leaks in the anterolateral segmental vein were detected intraoperatively but did not embolize or preclude the completion of the procedure. Another small ventrolateral soft-tissue leak was detected on scans made a week postoperatively. No cement leak was symptomatic. There were no significant relationships between cement extravasation and the quantity of cement used, the number of levels augmented, and other locations (thoracic or lumbar) (p > 0.05 for all comparisons). No serious complications were reported in this study, except for one case of wound infection and another case of postoperative radiculopathy, both relieved by medical treatment.  20.92 ± 4.66 (10-30) *Pre-versus all postoperative p < 0.000 **Pre-versus all postoperative p < 0.000

DISCUSSION
Glucocorticoid-induced OVFs are one of the most common types of osteoporotic fractures in clinical practice but have received little attention. It has been estimated that up to 17% of OVFs in the population can be attributed to the chronic oral corticosteroids use, even at small doses as low as 2.5 mg per day. 6 One of the most possible side effects of the usage of glucocorticoids is the trabecular bone mass loss. 11,28 Back pain at different intensities is the main presenting symptom in all patients of G-OVFs. Mild-to-moderate degree of pain in cases of acute vertebral osteoporotic fracture can be relieved by bed rest, bracing, analgesics, anti-osteoporotic drugs (bisphosphonates or denosumab), and anabolic drugs as tripartite (Forteo); 23,38 however, advanced cases with compression fractures mostly need intervention. 23 PV should be considered in patients who suffer from severe persistent pain even after at least three weeks of conservative treatment. 1,23 Twenty-eight patients (17 males and 11 females) with 61 vertebral levels were treated with PV. The mean age was 62.50 ± 7.09 years, which is consistent with the literature reporting that patients of osteoporotic spine due to long-term use of corticosteroid therapy treated with PV were significantly younger and more likely to be males compared to those with primary osteoporosis. 23,30 Severe persistent back pain was the main presenting symptom in all patients in the current study; the pain was assessed using VAS which was significantly (p < 0.001) decreased from a mean of 7.29 ± 1.04 preoperatively to a mean of 3.17 ± 0.77 6 weeks postoperatively, which is in agreement with most studies addressing the same issue. 23,30 Although in this study, a significant decrease of VAS was observed at the final follow-up at 24 months compared to preop VAS value, we reported a significant increase of VAS value at the final follow-up compared to that at 12 months; this final increase of VAS was mainly due to the nature of secondary osteoporosis, maintenance of corticosteroids in high dose, and the newly developed vertebral compression fractures during the two-year follow-up. The efficacy of vertebroplasty in OVFs in relieving pain and improving the functional status has been discussed in many studies with reported effectiveness up to 90%. 23 The mechanisms of pain relief following PV include the following: stopping vertebral micromotion which could be achieved by cement injection and stabilization of microfractures 20,40 and cement polymerization process, which can lead to a high temperature sufficient to cause protein denaturation, cell necrosis, and nerve ablation. 5,13 In this study, ODI improved significantly from a mean of 40.82 ± 12.32 (14-66) preoperatively to a mean of 16.68 ± 3.19 (10-24) at 12 months postoperatively and 20.92 ± 4.66 (10-30) two years postoperatively (p < 0.000), consistent with other similar studies. 23 Although PV has a successful overall outcome in the literature, it is associated with some local and systemic complications with an overall rate of < 1% in those osteoporotic fractures. 3 The most common complications are cement leakage beyond the vertebral body margins, into vascular and surrounding structures, new levels of vertebral compression fractures, pulmonary distress, and infection. 17 Patients with steroid-induced OVFs experienced the same degree of significant pain relief and complications after PVP as those not receiving corticosteroids. 23

Cement Leak
In clinical practice, leakage is more common and frequent in pathological osteolytic vertebral fractures, and mostly the cement leak noticed in steroid-induced osteoporotic OVFs is local and asymptomatic 16,30 ; however, it could be serious in cases of extravasation into the vertebral canal or the neural foramen in large volume causing cord or nerve root compression. 18,38 Many factors are controlling the percentage of cement leak during VP, such as the volume and viscosity of the cement, vertebral bone status, the The EGYPTIAN SPINE Journal pressure used, and even the operator. Jima et al. 14 reported cement leak in 8% of PVP procedures in cases of VCFs, while Yang et al. 40 noticed an overall higher percentage of cement leakage up to 36.44%. In this study, asymptomatic cement extravasation was seen in 7 of the 61 levels treated (11.49%), which is comparable to other series 23 with a reported cement leakage rate of 13.51% into the disc space and the paravertebral tissue. Four of the 7 leaks were in adjacent disc spaces and two small leaks into an anterolateral segmental vein, detected intraoperatively but did not embolize the operation, and the procedure was finished smoothly. Another small anterolateral soft-tissue leak was detected on the postoperative scan. No cement leak was symptomatic or serious. In general, the presence of cortical disruption, the amount of PMMA bone cement and its viscosity are important independent risk factors for the occurrence of cement leak 13,18 but severe vertebral fracture with marked height loss and the presence of vertebral clefts are specific risk factors for intradiscal cement leak. 31 The mean volume of the injected cement was 3.58 ± 0.557 ml (range, 3-5.8 ml), and this correlated with most studies reporting that a volume of about 3 ml of cement was enough to alleviate the associated pain. 19,25 Hitwatashi et al. 10 found that the volume of the injected cement may also be important when discussing the leakage; the more the volume, the more the risk of leakage, especially with low-viscosity cement. Although low-viscosity cement is good for spreading into the trabecular bone, the risk of extravasation is high. In contrast, higher-viscosity cement forms a more solid mass with more trabecular disruption but reduces leakage incidence; therefore, cement with doughy consistency and medium viscosity is recommended. Still, in cases of severe types of fractures with cortical disruption, the high viscosity one is the solution to minimize cement leak. 40 Intradiscal cement leakage should be avoided because there is an association between adjacent level compression and intradiscal cement leak. 12,29 New Vertebral fracture The occurrence of new OVFs in patients of steroidinduced vertebral fractures was discussed widely; some clinical studies 11,24,36 reported an increased incidence of new OVFs following vertebroplasty and others 15,23 stated that the rate is the same without a difference. Koch et al. 23 and others 9,15 reported that patients with G-OVFs experience levels of pain relief and development of new fractures after PV similar to those with primary osteoporotic vertebral fractures. There was another study 28 stating that adverse effects on bone are mainly due to underlying autoimmune or inflammatory disorders. The previous results support the use of PV to treat OVFs unresponsive to conservative management in patients receiving long-term corticosteroid therapy. Numerous studies in the literature suggest a higher rate of new vertebral fractures after vertebroplasty in G-OVFs compared to those not using steroids. 11.24,30,36 Syed et al. 36 reported that the rate was twice that of patients not taking steroids and others studies stated that the incidence of newly developed vertebral fractures increased by approximately 30%-50% in patients using corticosteroids for more than 3 months, which is consistent with the rate in our study indicating that new vertebral fractures developed in 39% of the patients. Steinbuch et al. 35 demonstrated that the increased rates of symptomatic new vertebral fractures were closely associated with higher dosages, longer durations, and continuous usage of glucocorticoids rather than the procedure of vertebroplasty; Kaji et al. 16 observed that the risk of fractures for the same bone mineral density (BMD) was higher in G-OVFs than in postmenopausal or senile osteoporosis. Due to the nature of secondary osteoporosis and the maintenance usage of large doses of corticosteroids in our patients, it is expected to find new vertebral compression fractures in a high percentage during the follow-up and we also The EGYPTIAN SPINE Journal reported a mild increase of the VAS value and ODI at final 24-month follow-up after they were being decreased at 12 months postoperatively (VAS was 1.68±0.6 and 3.11±1.13 at 12 months and 24 months; ODI was 16.89 ± 3.15 and 21.20 ± 0.914 at 12 and 24 months postoperatively, resp.). Moreover, the effects of PV to the adjacent nonaugmented levels in patients with G-OVFs were also discussed in detail. Some clinical studies 8,26 reported an increased rate of new OVFs after vertebroplasty due to the pressure created by the stiffness of the cemented vertebrae on the adjacent ones and others 22,31 demonstrated that there was no relation and excluded PV from the risk factors for new OVFs and they developed with conservative treatment. The mechanical effects of vertebroplasty on injured and adjacent vertebrae may also be influenced by the characteristics of the treated spine, such as BMD, the type, fracture severity, and disc degeneration. 22 Therefore, PV is not a risk factor for new osteoporotic vertebral compression fracture, which occurs even with conservative treatment. 22 In summary, vertebroplasty is generally a safe and effective procedure to control pain associated with fractures of the spinal column. Many complications can occur and are likely to pass underreported or without serious effect. Proper technique can minimize the risk of PMMA migration. The learning curve should be moderately quick and safe if physicians familiarize themselves with these axioms. Due to the progressive nature of the disease, the effect of PV decline with time in patients with G-OVFs. Large prospective, randomized, controlled studies with long-term follow-up are important and warranted for proper evaluation of the procedure for G-OVFs and to report any possible complications from cement-bone junction reaction, new and recurrent vertebral fractures. Minimizing PV complications depends on the following factors: good selection of the patients, good learning curve, high radiologic resolution, and proper cement polymerization state.

CONCLUSION
This study suggests that fast and substantial pain relief and quality of life improvement could be achieved after percutaneous vertebroplasty in most patients of glucocorticoid induced osteoporotic vertebral fractures. These improvements could be maintained up to one year, however this effect decline with time due to the progressive nature of the underlying disease.