With the aging of the global population, the number of patients with spinal fractures, especially osteoporotic spinal fractures, has increased year by year [20, 21], and patients with Kummell disease have also increased clinically. If one patient without manifests asymptomatic for a few months or even a few years after experiencing mild trauma, and then has pain in the same site, with worsened symptoms, which gradually progresses to kyphosis, and even appears neurological symptoms, it is referred to post-traumatic vertebral delayed osteonecrosis or Kummell disease. The disease often occurs in the elderly, more common in the thoracolumbar spine, the most common spinal segment is the T12 vertebra, male to female ratio of about 1:10[22, 23].
Intravertebral vacuum cleft (IVC) is a characteristic imaging manifestation of Kummell disease. Libicher et al [24] found that IVC showed a sensitivity of up to 85% and a specificity of 99% in the diagnosis of Kummell disease. However, IVC is not the basis for the diagnosis of Kummell disease. It can be not only found in Kummell disease, but also found in acute fracture, infection, primary neoplasm or metastasis etc [2, 25]. IVC is formed by gas or liquid accumulation in the vertebral body. It is usually located in the middle of the vertebral bodies or adjacent to the compression endplate in the coronal view, and is most commonly located in the anterior edge of the vertebral body in the sagittal view [26]. Hence, for Kummell disease, bone deficiency in the anterior edge of vertebral body is very common clinically.
Currently, the pathogenesis of Kummell disease is still unclear. In 1951, Steel et al. [27] first proposed a possible mechanism of formation of cancellous bone fracture and small hematoma caused by tiny spinal trauma, which further led to osteonecrosis and eventually collapse of the vertebral body. With the deepening of research in recent years, hypotheses of pseudo-arthrosis and ischemic necrosis in the vertebral body have become the mainstream. Kim et al. [28] believed that IVC indicated the existence of dynamic activities of non-union pseudo-articular joints in the micro-fracture end, which further resulted in osteonecrosis and vertebral collapse, rather than ischemic necrosis. Furthermore, biopsies of the IVC demonstrated fibrocartilage within a fibrous stroma, findings that represent a pseudo-synovium of a pseudarthrosis. This false joint is indicative of a permanent failure of bone healing after fracture [29, 30]. However, more scholars favor that Kummell disease is caused by ischemic necrosis of the vertebral body. Yu et al. [26] and Matzaroglou et al. [31] performed biopsy of Kummell diseased vertebral bodies, and found trabecular ischemic necrosis. All of these results supported the hypothesis of vertebral ischemic necrosis.
With the rapid development of minimally invasive spine surgery, vertebroplasty and kyphoplasty have become the main methods for the treatment of Kummell disease without neurological damage due to advantages such as small trauma, good tolerance, rapid pain relief and effective deformity correction, etc [3, 11–13, 32, 33]. In theory, kyphoplasty is superior to vertebroplasty, with a better vertebral height reduction and a lower probability of bone cement leakage. However, for Kummell disease, especially for phases II and III patients, the most important aspect of vertebral height restoration is to use the overextending position to fully open IVC instead of reduction using a balloon. At the same time, when the bone cement is targeted and injected into the IVC, bone cement leakage is unlikely to occur since the IVC walls are composed of hardened bone, except that it is prone to occur anterior bone cement leakage in patients with anterior vertebral body deficiency. Krauss et al. [32] implemented vertebroplasty in 44 cases with osteoporotic fractures plus IVC and 148 cases with osteoporotic fractures without IVC, and found that the bone cement leakage rate was much lower in the patients with IVC. In addition, for Kummell disease, the use of a balloon may rupture the thin endplate, and result in intervertebral leakage of bone cement, and thus increase the probability of adjacent vertebral fractures. Therefore, in clinical practice, vertebroplasty is more commonly used instead of kyphoplasty after the overextending position, which is also more economical.
Although the use of bone cement has achieved satisfactory clinical results [3, 33, 34], the catastrophic complication of bone cement displacement during or after surgery is inevitably worrying. Even if the postoperative results are satisfactory, there is still concern of bone cement displacement during follow-up. Once bone cement displacement occurs, patients often need to undergo open posterior, anterior, or even anterior and posterior revision surgery to remove the displaced bone cement, reconstruct spinal stability, restore spinal sequence and fusion [16, 17, 35]. Thus, some scholars intended to apply short or long segmental fixation combined with vertebroplasty to avoid bone cement displacement. However, compared with using the bone cement alone, it has the disadvantages of large surgical trauma, loss of spinal mobility and high cost [3, 36].
The reason for postoperative bone cement displacement for Kummell disease is still unclear. We summarized and analyzed the Kummell disease patients with bone cement displacement after VP or KP who were treated in our center or reported in nearly all literature in recent years[14–18, 35, 37], and found that all these patients suffered from bone deficiency at vertebral anterior border, and the bone cement shifted from the front of the deficient vertebra to ventral part. In the vertebroplasty, the bone cement tends to distribute on the vertebral anterior border under the action of the push pressure, but is easy to leak into the front of the vertebral body under pressure in case of bone deficiency at vertebral anterior border. In the meantime, even with perfect postoperative kyphosis correction, the patients may still be in mild kyphosis compared with the status before the disease onset. Hence, in postoperative activity, the barycenter of the diseased vertebra is moved forward, and the stress can squeeze the bone cement to produce a force for anterior displacement, thereby results in anterior displacement of bone cement since there is no bone barrier due to bone deficiency at vertebral anterior border. Wang et al. [15] and Tsai et al. [16] also believed that the integrity of the anterior cortex is an important reason for anterior displacement of bone cement. In addition, the permeation of bone cement into trabecular bone plays an important role in preventing leakage and displacement, both before and after the procedure [38]. It has been confirmed by pathology that IVC is a dead space composed of necrotic bone tissue. After filling with bone cement, it is difficult for the bone cement to infiltrate into the normal trabecular structure, meaning that the bone cement and the surrounding bone tissue are non-union, which is another important reason for cement dislodgment [17, 18]. Tsai et al. [16] believed PMMA cement in vertebroplasty merely a space occupying material without mechanical interlock and biocompatibility and, thus, make the potential for dislodgment.
In order to interlock the bone cement and the surrounding bone tissue to avoid bone cement displacement during or after surgery, we have developed a novel bone cement screw system. The key technology is that after the bone cement screw is inserted in the IVC, the bone cement will slowly release through the lateral hole at the front end of the screw and completely fill the IVC during injection, which tends to integrate the bone cement and the screw, while maintaining the screw in the vertebral body. Hence, the cement screw acts as a “bridge” to interlock the bone cement with the surrounding bone tissue, and even makes it connect to the pedicle that is the most rigid part, which strengthens the linkage, and the screws are unlikely to loosen and shift as well. Furthermore, the tail-free design of the screws allows the bone cement screw to completely locate in the bone structure, which is unlikely to wear the muscles and other soft tissue during the activity while acting as a bridge. At the end of the 3-year follow-up, none of the 27patients had bone cement displacement and screw loosening, with a VAS improvement rate of 73.12% and ODI improvement rate of 71.38%. These preliminary results reveal that the screw system is safe and effective, and plays a good short and medium-term therapeutic effect on patients with Kummell disease and bone deficiency at vertebral anterior border, while its long-term efficacy needs further observation. In addition, the development of a screw system that enables the growth of bone tissue in it may provide an even better stability.