Lumbar spinal fusion is the most common treatment method of lumbar spine disorders. From the early posterolateral fusion, various spinal fusion procedures have gradually appeared. On the one hand, the continuous improvements of surgical procedures suggest the deeper understanding of the anatomical structure and the instrument improvement; on the other hand, it also shows that there is no ideal fusion technique at present. Posterior lumbar interbody fusion (PLIF) surgery could achieve interbody spinal fusion while decompressing the spinal canal directly and thoroughly. However, PLIF can bring great damage to the paravertebral soft tissue and bone structure and can greatly interfere with dural sac and nerve roots. TLIF can reduce the iatrogenic injuries and avoid excessive traction on nerve tissue. However, for the TLIF, most of the ipsilateral facet joints and ligamentum flavum need to be removed before decompression and implantation of the fusion cage. ALIF, OLIF, and LLIF entering intervertebral space from the lateral or anterior side can avoid the destruction of posterior muscles and osseous structures, without encroaching on the spinal canal. However, the spinal surgeons are not familiar with the anatomical structure, and thus, the injuries of the important organs, such as the abdominal cavity and large blood vessels, might occur during the operation . The decompression of these procedures is indirect, with limited indications, and sometimes it might be necessary to change the body position to obtain auxiliary pedicle screw fixation. In 2002, Foley et al.  proposed MISTLIF to reduce the damage of muscle and bone structure. This technique uses a dilation system to establish a surgical access through the intramuscular space, and then, the decompression and interbody fusion are performed from the intervertebral foramen. However, the working channel is long and narrow, the anatomical structure is difficult to identify, and the operation might be inconvenient, so the early learning curve is steep, and the incidence of complications is high. Nerve injuries, surgical site infection, hematoma, and the leakage of cerebrospina are complications commonly reported in the literature . In addition, muscle atrophy and ischemic necrosis might occur due to the continuous squeezing of the working channel . In recent years, the spinal endoscopy has been widely used in the lumbar spine and achieved satisfactory clinical outcomes. Besides, some surgeons have begun to perform lumbar fusion under the endoscope [3, 4, 12] and have achieved satisfactory clinical outcomes.
In this study, the two groups both achieved satisfactory clinical results, and there were no serious complications in the interbody fusion, indicating that these two types of procedures were safe and effective in the treatment of lumbar degenerative diseases.
The average length of hospital stay in the PETLIF group (2.4 days) was shorter than that in the MISTLIF group (4.5 days), with a significant difference, which indicated that the PETLIF surgery could promote the recovery and shorten the hospital stay due to less surgical trauma.
The VAS-L score in the two groups significantly decreased after operation, indicating that PETLIF and MISTLIF both can achieve complete spinal canal decompression. The VAS-B score in the two groups also significantly decreased after operation, but the postoperative VAS-B score in the PETLIF group was significantly lower than that in the MISTLIF group, for which the reason might be that the PETLIF procedure caused less surgical trauma to the soft tissue and bone structure than the MISTLIF procedure.
As we all know, bone grafts play an important role in interbody fusion. Studies [13, 14] suggested that bone grafts and intervertebral cages should cover at least 30% of the surface area of the endplate to provide sufficient stability while allowing early movement. Autologous bone has always been the gold standard for bone grafts. Traditional fusion surgery can obtain enough autologous bone for grafting, but the amount of autologous bone collected by PETLIF surgery is insufficient, which cannot meet the needs of intervertebral bone grafting. Various sources of bone grafts have already been reported in the literature . We implanted bone harvest from facet joint mixed with allogeneic bone into the cage and obtained satisfactory bone fusion during the follow-up period.
Nerve root injury is a serious complication of endoscopic surgery, especially the injury of the exiting nerve roots, which deserves special attention . The intervertebral foraminal stenosis and prolonged operation time are the main risk factors for nerve root injury . Obvious intervertebral stenosis and level II of spondylolisthesis should be regarded as relative contraindications, and the reduction of intervertebral height and changes of local anatomical structure result in Kambin triangle changes, which makes it difficult to establish working channels and puts the nerve roots in danger . Anatomical studies have shown that the exiting nerve roots are adjacent to the upper facet joints at the disc level, so the removal of part facet joint to expand Kambin triangle is a key step to reduce nerve root injuries . In addition, the long operation time is another high-risk factor that leads to nerve injuries. Partial resection of the upper facet joint can achieve lateral recess decompression while expanding the surgical channel, which can reduce the operation time and the risk of nerve damage. Moreover, skilled surgical techniques can significantly reduce the operation time and the impact of instruments on the existing nerve roots.
Postoperative headache is a common complication of spinal endoscopic procedure. Endoscopic surgery requires continuous infusion of normal saline into the spinal canal, which leads to an increase in epidural pressure and changes in cerebrospinal fluid. Higuchi et al.  used MRI to study the degree and duration of compression of the dural sac after the injection of normal saline into epidural space, and the results showed that changes in the cerebrospinal fluid (CSF) volume were related to the volume of normal saline. The compression degree lasted at least 30 min after saline injection. Joh et al.  measured cervical epidural pressure in 28 patients undergoing spinal endoscopic surgery and found that with the continuous epidural perfusion, the epidural pressure continued to rise, and high pressure was closely related to intraoperative cervical pain. Choi et al.  reported that during spinal endoscopic surgery, elevated epidural pressure can cause cervical pain and even induce epilepsy. In this study, two patients had headaches, of which one had mild symptoms, and the symptoms were relieved after symptomatic treatment, which might be caused by the increased epidural pressure by the continuous perfusion, and the other one had severe headache, which could not be alleviated by the symptomatic treatment. Brain CT scan showed high-density shadows in the brain sulcus and cistern, and MRA excluded the lesions such as intracranial aneurysm, vascular malformation, and venous thrombosis leading to spontaneous intracranial hemorrhage. Therefore, subarachnoid hemorrhage was considered. The reasons might be as follows: firstly, after epidural perfusion of saline, the increased intracranial pressure might lead to passive expansion and congestion of cerebral blood vessels, which might increase the permeability of blood vessels and blood cell exudation and eventually induce intracranial hemorrhage; secondly, after the dural sac is compressed, it might cause the occlusion of the epidural vein, resulting in increased intracranial venous pressure, causing the intracranial vein to dilate and congest and eventually rupture and hemorrhage.
This study has several limitations. First of all, the sample size is small and the follow-up time is short. In the future, clinical observations with multiple centers, large sample size, and long-term follow-up are needed to evaluate the clinical efficacy of PETLIF in the treatment of lumbar degenerative disease. Secondly, there are many minimally invasive surgical procedures for lumbar fusion, including OLIF, ALIF, and other anterior lumbar fusion, but only MISTLIF was used as the control group in this study. Therefore, it is necessary to compare with other procedures in the future. Last but not least, this study is a retrospective case–control study, and a higher-level evidence is required in the future.