In 1968, Wiltse first described the paraspinal sacrospinalis-splitting approach , and later he replaced the originally designed bilateral incisions with a single median incision and expanded the indications of this approach .
Via the anatomical avascular space existing between the multifidus and the longissimus muscles, this procedure provides an access to the intervertebral foramen, the facet joints, and the transverse processes. Although it still belongs to a type of open approach, Wiltse paraspinal approach not only avoids extensive detaching of the paraspinal muscle, electrical-burn damage and continuous retractor compression for neurovascular supply, but also preserve the posterior osseous structure including the zygapophysial joint, and the posterior ligamentous complex to the maximum extent.
By imaging, histology and electrophysiological assessment, Liu et al  found that the Wiltse approach for thoracolumbar fractures was accompanied with a lower incidence of multifidus atrophy and denervation, and less fatty infiltration compared with conventional posterior open approach. Pedicle screw fixation via Wiltse paraspinal approach has been considered as an effective and minimally invasive approach in treatment of thoracolumbar fracture with neurologic intact for the advantages of simple operation and less trauma [12, 13].
In 1995, Mathews et al.  apply percutaneous pedicle targeting originally designed for temporary external fixation to perform an entirely percutaneous lumbar pedicle fixation, and later Foley et al.  suggested improved percutaneous techniques.
Following each incision which was 1.5-2cm long and down to the deep fascia, the insertion of puncture needle and the placement of screws were carried out under the guidance of the fluoroscopy with small damage to paravertebral muscle, which avoids iatrogenic injuries to the posterior ligamentous complex, minimizes surgical exposure, reduces intraoperative blood loss, and has a better postoperative pain score than open surgery, thereby improving the clinical outcome. After measuring cross-sectional areas of the multifidus muscle by the preoperative and postoperative MRI, Kim et al  detected significant decrease in open pedicle fixation group but no statistical difference in percutaneous fixation group. An increasing number of studies suggested that percutaneous pedicle screw fixation has the advantages of limited injury, less bleeding, shorter hospital stay, reliable immediate fixation and better pain improvement [17–19]. A recent meta-analysis also confirmed those study results . The present study compared the clinical and radiological effects of the two procedures for mono-segmental without neurological injury. Meanwhile, perioperative parameters, operative and post-operative costs and complications related to the two procedures also were compared.
Spinal fusion has always been taken as the role of the stabilising procedure for a long time , but some scholars suggested that in traditional open surgery for thoracolumbar fracture no significant differences existed in clinical or radiologic outcome between the fusion group and no fusion group, and fusion procedure was associated with increased operative duration and larger intraoperative bleeding . A recent systematic review and meta-analysis also drew similar conclusions . Furthermore, compared with open short-segment pedicle screw fixation with posterolateral fusion or no fusion, percutaneous pedicle screw placement without fusion could provide not only similar radiologic outcome but also earlier pain relief and functional improvement [18, 24]. Liu et al.  also reported that no statistically difference existed between the two groups with no fusion in radiologic outcome and the accuracy rate of screw placement, however, Wiltse paraspinal approach had apparent advantages over the conventional open method in operative time, intraoperative blood loss, postoperative drainage, and postoperative pain improvement. In our study, no fusion was completed in the two groups. We found that in both groups the Cobb angles and the anterior vertebral height were statistically improved 1 week and 1 year after surgery compared with pre-operative data, and there was no statistical difference between the two groups before and after surgery. Pedicle screw fixation after hyperextended position reduction under general anesthesia could correct kyphosis and restore the anterior vertebral height, but loss of correction and kyphotic angle increase is an inevitable problem observed in the follow-up of the surgical treatment for thoracolumbar fractures. The follow-up results indicated that the percutaneous and Wiltse paraspinal approach groups did not show significant differences in short and long-term radiological effect, and that both approaches can bring satisfactory reduction effects.
This is different from the results of Fan et al. . In their study, uniaxial screws were used in both surgical procedures, and the results showed that the correction of the deformity in the paraspinal group was significantly better than that in the percutaneous group. They think that the skin and muscles in the percutaneous procedure may hinder indirect reduction, while the paraspinal approach provides the nail placement and reduction in nearly direct vision. Fitschen-oestern et al.  found that percutaneous surgery and open surgery showed similar early results in reduction and reduction loss, and there was no significant difference in reduction loss between the two surgeries over time. There was no significant difference between uniaxial and multiaxial screws. In this study, uniaxial screws were also used in both groups. Therefore, we believe that the different conclusions may be caused by different surgical experience and habits. In this study, all patients underwent surgery by the same team of spine surgeons, all of whom were senior associate chief physicians.
In this study, no significant difference was detected in functional and symptom outcome between the two groups at final follow-up. Compared with the pre-operative values, the post-operative VAS and ODI values were significantly improved in both groups. However, the percutaneous group showed significantly lower pain score than the paraspinal group at 3 days after surgery, which indicated that the percutaneous approach resulted in better short-term pain improvement. The percutaneous approach was superior to the paraspinal approach in terms of the lengths of incisions, intraoperative blood loss and duration of postoperative hospitalization. Early studies have shown that the distance of superficial points of the intermuscular plane and the midline decrease from S1 to L1 . Actually the intermuscular space appears curvilinear in the axial plane, whose concavity facing the spinal elements and convexity facing the lateral spine. From S1 to L1 the intermuscular space is to rotate towards the spinous process plane around the outer edge of the facet joint, and then at the level of L1 the superficial point is close to the spinous process . After blunt separation of the longissimus and multifidus muscles via Wiltse paraspinal approach, deep retractors are needed to pull the two muscles apart to clearly expose the operative field. Lateral traction is relatively easy, while medial traction can cause great soft tissue tension due to spinous process obstruction, which actually do much damage to neurovascular structure of soft tissue lateral and medial to the facet joint line respectively. At the same time, the authors also found that some muscle Spaces could not be identified on MRI, which meant that the muscle spaces of some patients could not be accurately identified during surgery . Actually sometimes it is difficult to find the intermuscular plane accurately during operation and the blood vessels passing through the muscle are damaged, resulting in increased bleeding. In the anatomy report of thoracolumbar junction, Wang et al. pointed out that after the lumbar and dorsal fascia is cut open, the intermuscular space can't be found immediately . Instead, we first see is the aponeurosis of erector spinae covering it, which is made up of the tendon of the thoracic longissimus muscle. In order to expose the intermuscular space, part of the longissimus tendon is often severed. Moreover, the Wiltse paraspinal approach procedure may cause some damage to the intertransverse muscle, the cephalic part of the facet joints, the posterior medial ramus of the spinal nerve that innervates the paravertebral soft tissue and the related ligaments. A human cadaver research by Regev et al.  indicated that the percutaneous approach was superior to the Wiltse paraspinal approach in terms of preserving the integrity of the multifidus innervation at the adjacent cranial level, when inserting pedicle screws. The percutaneous approach can clearly expose the entry point after the insertion of the working sleeve, freeing the assistant's hands. When the minimally invasive percutaneous sleeves are inserted through the puncture hole and the diameters increase gradually, it can compress the bleeding point in the muscle and stop the bleeding. The blood loss is less than the former approach. Therefore, the percutaneous approach conforms more to the minimally invasive concept that has been popular, which brings about faster recovery as a consequence of less trauma.
The current study also showed that the operation time of the percutaneous group was longer than that of the paraspinal group, but there was no no statistical difference between the two groups (P>0.05). The lack of anatomic markers results in more operative procedure and more intraoperative C-arm monitoring, which requires surgical physicians with enough anatomical imaging knowledge basis and experience in minimally invasive operation, especially in early cases. We also found that the fluoroscopy numbers were 18.52±3.54 in percutaneous group and 6.31±2.18 in paraspinal group (P<0.05). While there is no consensus on the risks of long-term low-dose X-ray exposure, the link between radiation and skin cancer and cataracts has long been a concern for surgeons. In addition, percutaneous approach is associated with more instruments and more complicated procedures, which leads to prolonged operation time to some extent. In the present study, we also compared the operative and post-operative costs of the percutaneous group (32.34±0.92 thousand yuan) and the paraspinal group (28.63±0.64 thousand yuan), with a significant difference (p<0.05). Since different post-operative hospitalization time and similar treatment prescriptions between the two groups, the main reason for the cost difference was due to expensive hollow screw used in the percutaneous group. With the accumulation of surgical experience and the development of percutaneous procedure equipment, especially the gradual application of navigation, virtual reality and other technologies in surgery, the time of percutaneous pedicle screw surgery will be further shortened, the number of fluoroscopy will be significantly reduced, and intraoperative blood loss will be reduced accordingly, implant will be cheaper with better biomechanical performance and its advantages will be further highlighted.
In terms of safety, Chapman et al.  reported that the mal-placement rates of percutaneous group were similar to that of the open group after measurement for postoperative CT imaging for a series of 1609 screws by Gertzbein-Rao grading system. Similarly, some scholars argued that no statistically significant differences about the accuracy existed between the traditional open insertion and percutaneous placement according to Zdichavsky’s scoring system . Even Raley et al.  suggested that fluoroscopic guided percutaneous pedicle screw insertion in the thoracolumbar spine is associated with a low misplacement rate and an extremely low rate of complications compared with the high rates published in the literature for open method. Liu et al.  also discovered that no statistically difference existed between the Wiltse paraspinal and conventional open groups with no fusion in the accuracy rate of screw placement. We found no significant difference in accuracy of pedicle screw placement between the two groups, which is consistent with that published in literature .
In our study, two cases of fat liquefaction exudation suffered delayed incision healing in percutaneous group. One case of superficial surgical site infection who had a long history of type 2 diabetes, and two cases of skin edge necrosis were found in paraspinal group. Operation can raise the blood glucose levels of patients, which exerts influence on wound healing of diabetic cases. Guzman et al.  analyzed 2,568,994 degenerative lumbar spine procedures and stressed that diabetic patients had a higher risk of complications, including surgical site infection, than nondiabetic ones after lumbar spine surgery. The paraspinal approach requires a median incision similar to the traditional open method, and subcutaneous tissues separated, resulting in more trauma and bleeding, which increased the incidence of microorganisms entering the both superficial and deep wound site.
For the one case of delayed incision healing in the percutaneous group, we thought that the cause of fat liquefaction exudation might have been that the skin incision was not made lateral enough to the pedicle projection so that no appropriate angulation of the puncture needle can be made when inserting into the pedicle. More adjustment was required, pressing surrounding fat tissue. Raley et al.  clearly pointed out that a certain distance should be kept lateral to the outer edge of the pedicle projection during incision selection, so as to ensure that the puncture needle enters from the outer edge of the pedicle projection, and the adjustment of puncture direction can be reduced when passing through the pedicle. Meanwhile, in the paraspinal group, the surgeon sometimes deliberately reduces the length of the incision to pursue minimal invasive outcome, but in order to expose the surgical field sufficiently, he has to pull the skin forcefully during the operation, causing skin edge necrosis.
The guide wire breakage was thought as a result of the screw deviating from the trajectory of guide wire, which was transected when the screw was inserted. What's more, repeated use of the guide wire could also be one of the reasons. The anterior wall of the vertebra was broken because the guide wire was advanced by the tap unexpectedly. The great vessels were not injured by guide wire, which should be attributed to the fall as a result of the effect of gravity. This wire acts as a guide for all subsequent instruments, so it is imperative to keep the guide wire in place without inadvertently withdrawing or advancing it. To avoid this complication, prono-supination is required to withdrawn the needle. Keep the trajectory of guide wire by two needle holders such that it is not transected or bent by the tap or screws. The guide wire should be removed as soon as the screw tip exceeds the posterior vertebral wall.