One hundred thirty-five patients out of 156 who were diagnosed with HLD underwent ELELD between December of 2007 and September of 2014 in Bundang Medical Center. All patients were male, with average age of 22 (20-25 years old). Average follow up period were 16 months. Magnetic resonance imaging (MRI) and computed tomography (CT) were done to all the patients. Surgery was done on patients who had conservative therapy for at least 3 months without improvement on lower back pain, radiating pain and radiological findings matched with clinical symptoms. Additionally, plain radiography was taken to check if iliac crest is higher than the space between L5 and S1 and enlarged transverse process prevents the working channel to get inserted to spinal epidural space. We excluded the patient who underwent laminectomy, discectomy, or endoscopic transforaminal lumbar discectomy.

The herniation was categorized as central, postero-lateral and foraminal based on the radiological findings, whether the herniated disc was migrated or not; and if migrated, severity was measured based on the posterior disc space; and if not migrated, severity was measured based on the 50% of area of lumbar spinal canal.

The surgery was done on prone position with slight flexion and local anesthesia at the site of the herniation. C-arm was placed to observe anterior-posterior plane perpendicular to lateral plane of the upper and lower vertebrates of the target disc. Basal line was set with posterior facet line (Fig. 1) and incision was made at the baseline or 1.5 cm above the baseline as incision below the baseline gives greater chance of abdominal perforation, which as 13 to 18 cm away from the central line based on the patient’s circumstance of lumbar region and height of iliac crest. Surgery on the herniated disc between L5 and S1, extreme lateral approach could be limited due to the iliac crest and the transverse process; thus 9 to 11cm from the central line was the optimal position.

Skin incision (0.5-1 cm) was made to the injection site after local anesthesia was applied. 18 gauge needle was injected with c-arm image to inject the needle to the posterior facet by extreme lateral approach. The angle of approach was 5 to 20 degrees to the ground (Fig. 2). During the initial injection, the angle of the needle was toward the abdomen to minimize peritoneum perforation even if the needle was touching peritoneum, and changed the angle upwards to ease the access to the vertebral space at the facet joint. The position of the needle should be at the inner pedicular line from the anterior posterior view and at the intervertebral foramen from the lateral view (Fig. 3). Omnipaque 2mL was injected to the epidural space to identify relative position of the needle from the exiting nerve root and the traversing nerve root from the anterior posterior view and from the lateral view. The needle is replaced by a guide wire followed by an obturator and a final working cannula in that sequence (Fig. 4). Using the slanted working channel, the wider surgical view was provided and helped preventing nerve root damage. When the working channel passed through the exiting nerve root, slanted side was turned toward the nerve root. After that, slanted side was turned 180 degrees making the slanted side to face the traversing nerve root to minimize the damage. When endoscope with two tubing system with 6mm inner diameter and 20 degrees slope (Yeung Endoscopic Spine Surgery System [YESS]) was inserted to epidural space, blue colored lesion could be identified. During the surgery, saline mixed with antibiotics was used continuously to prevent the infection and the damage from burn. When transforaminal ligament was obstructing, the ligament was removed, and identified the colored protruding nucleus populous and removed with pituitary forcep until the nerve was decompressed and performed hemostasis using radiofrequency trigger-flex bipolar. When the HLD was removed and the decompression of the traversing nerve root was observed via endoscope, the surgery was finished. Patients were allowed to communicate and when the patient complained about symptoms related to irritating nerve root or any other neurological symptoms, the surgery were stopped immediately and continued after the problem was fixed. When epidural hemorrhage occurred during the surgery, radiofrequency trigger-flex bipolar was used to stop the bleeding and if not corrected, pressure was applied with epinephrine soaked gel foam to the hemorrhagic site and increased the tubing pressure to wash with saline. There were no serious complications including damaged nerve root, internal organ or massive hemorrhage.

The postoperative follow up study was carried out by means of outpatient clinic and telephone interview, the clinical result was evaluated through MacNab’s criteria and visual analogue scale where the ‘excellent’ and ‘good’ score was considered as successful result. The statistical testing was carried out by analysis of variance (ANOVA) using SPSS for windows (Release 7.5.1; SPSS Inc., Chicago, IL, USA).

Sixty seven percent of patients complained about lower back pain and radicular pain, and 33% only complained about the radicular pain. Sixty five point seven percent of patients had sensory dysfunction and motor weakness was observed in 40.3%. Eighty seven percent were positive in the straight leg raising test. Patents with osteophyte due to the apophyseal ring fracture were excluded from this study. Average period of lower back pain until surgery was 2.3 years and 4.5 months for radicular pain. Seventy four point one percent of patients had surgery between L4 and L5, 19.3% had surgery on the disc between L5 and the S1, and 6.7% had surgery on the disc between L3 and L4 (Table 1). The direction of herniation was categorized into central, posterolateral and foraminal herniation and presented at 24.4%, 63% and 12.6% respectively. There was no patient with extreme lateral lumbar herniation (Table 2). In the group of patients without migration, 17.6% showed spinal canal compromise more than 50%. Twenty point six point showed mild migrated disc under the height of the posterior lumbar disc, and 6.6% showed severe migrated disc (Table 3).

The surgery was evaluated using MacNab method and excellent or good scores were considered as successful surgery. Overall, the average visual analogue scale (VAS) of all patients before surgery was 8.92±1.20 and VAS at the final follow up was 2.38±1.55 (p<0.05) (Fig. 5). Group without migrated disc with 50% or more protrudes disc and group of mid-severe vertical migrated disc showed significant decrease in the success rate of the surgery (p<0.05) (Table 4). Post-operative complications included dysesthesia, discitis, hemorrhage, cerebrospinal fluid leakage, decrease of lumbar disc, and nerve root damage (Table 5). Twenty percent of the patients showed dysesthesia after surgery, but most patients except for 4 recovered after 3 to 4 weeks, and 2 patients had spinal fluid leak, but these were corrected with conservatory therapy such as bed rest. Discitis and decrease in the height of the lumbar disc were observed with one patient, but discitis was treated with antibiotics and only conservative therapy. Therefore, posterior lumbar interbody fusion was not performed because of lack of severe lower back pain. Two patients had severe lower back pain due to the decrease in lumbar disc height after surgery, but corrected with posterior lumbar interbody fusion. Out of 8 patients who had ‘fair’ result, 2 had multiple herniations, 3 had lasting dysesthesia after surgery, and 3 had incomplete removal of nucleus pulposus. Patients with ‘fair’ result had chiropractic therapy after surgery and reoperation was not performed. Out of 6 patients with bad result, one patient presented with decrease of disc space due to discitis who had mild herniation without disc migration, 2 patients received posterior interbody fusion as they presented with severe lower back pain due to severely decreased disc space who had severe central herniation with 50% or more protrusion of intrathecal. One patient who had mild migration of the herniated disc presented with motor and sensory dysfunction due to nerve root damage during the surgery and two patients received conventional lumbar discectomy because of continued pain due to incomplete removal of nucleus pulposus who had severely migrated herniated disc. During the follow up period, there was no relapse or serious complications such as blood vessel damage and peritoneum perforation.