Risk factors of neurological deficit and pulmonary cement embolism after percutaneous vertebroplasty

With the aging of the population, the incidence of osteoporotic vertebral compression fractures (OVCFs) is increasing and is becoming a major healthcare concern [1].

Percutaneous vertebroplasty (PVP) has gained widespread acceptance and is implemented broadly, mainly as a treatment method for painful OVCFs [2‐6]. Its benefit over conservative treatment has been reported in a large and high-quality randomized clinical trial [7]. However, there are also reports that put into doubt the efficacy of vertebroplasty as a treatment for VCFs. Some studies have found that PVP had no efficacy compared with sham-operated patients [8‐10]. Later reports by the same authors showed that vertebroplasty had no or limited value in long-term follow-up studies [10‐12].

The complication rate of PVP is low, and has been reported to be 1.6% to 3.8% in meta-analyses [13, 14]. Severe complications of PVP are rare; they are restricted to case reports and mainly comprise sequelae of excessive cement leakage, such as paraplegia [15], neurologic deficits [16, 17], cardiac perforation [18, 19], and even death [20].

The rate of occurrence of cement leakage itself appears to vary, with reported incidences ranging from less than 5% to more than 80% [21‐24]. When assessed using computed tomography (CT) scanning, which is known to be substantially superior to intraoperative fluoroscopy or postoperative radiography for detection of cement leakage [24, 25], the incidence of leakage was found to be 63% to 87% [21‐25].

PVP can also increase the risk of fractures of adjacent vertebrae [26‐30]. The incidence of postoperative infections in patients undergoing PVP is low [31].

In this study, we analyzed clinically significant complications after cement leakage in 3175 patients treated with PVP for symptomatic OVCFs.

This study included 2127 (67%) women and 1048 men (33%), 60 to 98 years old (mean age, 76 years), all with at least 6 months of follow-up. A total of 3812 vertebroplasty procedures were performed in 3175 patients, among which 2644 had a single fracture, 436 had two vertebral fractures, 84 had three fractures, and 11 had four fractures (Table 1). Most operations were performed in subacute stage (56.7%), and chronic stage (40.7%); only 2.6% in acute stage (Table 1). Patient number of cement leakage corresponded to each characteristic were detailed listed (Table 1). There was no significantly statistical difference of types of cement leakage between age, gender, number of fracture treated, and time from fracture to vertebroplasty.

Parameters related to imaging and technical characteristics, including the method of approach, amount of bone cement injected, and local or general anesthesia used were recorded (Table 2). Most operation were performed using unipedicular approach (98.4%) and bipedicular procedures were performed if poor filling of vertebra after unipedicular approach. The volume of cement injected were less than 3.5 cc in 320 patients (8.4%), 3.5 cc to 7 cc in 2821 patients (74%), and more than 7 cc in patients (17.6%). General anesthesia were used only in 153 patients (4.8%) with poor surgical compliance. The levels treated were from T5 to L5; four in T5, ten in T6, eight in T7, 19 in T8, 21 in T9, 36 in T10, 72 in T11, 1170 in T12, 1187 in L1, 601 in L2, 304 in L3, 272 in L4, and 108 in L5; with an emphasis on T12 and L1 (62%).

Osteoporosis is a systemic disease that results from progressive bone mineral loss and changes in bony component, leaving the spinal vertebrae vulnerable to VCFs. Painful osteoporotic VCFs can be a devastating burden for patients, as they impair physical function and quality of life. Moreover, VCFs can lead to progressive sagittal spine deformities and changes in spinal biomechanics, which are believed to contribute to a five-fold increased risk of further fracture [32]. Conservative treatment for the pain caused by VCFs includes analgesic medication, bed rest, and back braces; however, these therapies do not address spinal deformities [1]. Furthermore, pain and disability may be prolonged while the fractured vertebral body heals [33]. PVP in osteoporotic VCFs, involving the percutaneous injection of cement directly into the fractured vertebra, is effective in ameliorating VCF-associated pain [34‐36]. The limitation of this procedure is the substantial risk of extravertebral cement leakage after high-pressure cement injection, with severe clinical complications including spinal cord compression [15, 37‐40], radicular pain [41], and systemic embolism [42‐46]. Local leakage of PMMA is frequent, and varies up to more than 70% [1, 47‐51]; however, in most cases, the leakage does not produce any symptoms [47‐49].

In our review, cement leaks were noted in 69.1% of all performed procedures, but only 0.72% of patients had clinical complications. These included neurological deficit in four patients (0.26%) who achieved partial or full recovery after decompression surgery, and pulmonary embolism in nine patients (0.46%) who needed postoperative oxygen support.

There were statistical differences in neurological deficits between the thoracic and lumbar spine (Table 4), leading us to believe that cement leakage is not tolerated in the narrowing thoracic spinal canal. It was of interest that only four of 26 patients with a type-C posterior leak had clinical symptoms. We thought that the kyphotic spinal curvature might have made the canal wider. The four patients with a neurological deficit achieved partial or full recovery after decompression surgery.

In our series, the nine patients with symptomatic pulmonary embolism were all in the type-S leak group (Table 5), and this may be related to the cement embolism pathophysiology. The PMMA flows out through the basivertebral vein and then through the anterior external vertebral venous plexus, which leads to the pulmonary veins via the segmental spinal veins, the vena radicularis magna, the azygos vein, and the accessory hemiazygos vein [52]. A symptomatic pulmonary embolism following vertebroplasty can occur either by migration of acrylic or migration of fat and bone marrow cells [53]. On the one hand, the majority of radiologically detected PMMA migration into lung vessels is asymptomatic and, on the other, fat embolism is more frequent than PMMA embolism [54]. According to the observation, in cases of symptomatic pulmonary embolism with detectable PMMA in lung vessels after vertebroplasty, fat embolism occurring simultaneously may also contribute to the clinical manifestation of pulmonary embolism. Early symptoms caused by the passing of emboli through the pulmonary artery occur 6 s after injection, which should be distinguished from the very rapid decrease in heart rate and the fall in arterial pressure caused by a nerve reflex response occurs about 2 s after injection.

One study [55] recommended no treatment besides clinical follow-up for asymptomatic patients with peripheral pulmonary cement embolism, and initial heparinization and a following 6-month coumarin therapy for symptomatic pulmonary cement embolism.

Pulmonary embolism occurs in 4–26% of patients [56‐59], but it occurred in only nine patients in our study. While symptoms were resolved after oxygen support and anticoagulation therapy, as recommended [60], no patient required open cardiac surgery to remove the cement from the lungs or from the right perforated ventricle.

Two limitations were in our study. First, this is a retrospective single center study. Second, only X-ray image was not enough to detect cement leakage and result in under estimate the leakage rate.

PVP with PMMA is a common and effective treatment modality for painful osteoporotic vertebral compression fractures, but is not as simple and risk-free as suggested in many radiology and spinal surgery studies. Vertebroplasty is a minimally invasive surgical procedure with a high rate of cement leakage (69.1%), but a low incidence of clinical complications (0.72%). When injecting a high volume of cement, physicians should be aware of type S leaks if there are signs of pulmonary embolism, including decreased blood pressure, tachycardia, and dyspnea. Awareness of possible neurological symptoms should be heightened when there is cement leakage into the thoracic spinal canal.