The study was approved by the Korea University Institutional Review Board in 2015. We retrospectively collected and analyzed data from April 2009 to March 2013. Patients who underwent PDD by L'DISQ (U&I Co., Uijeongbu, Korea) were analyzed. The inclusion criteria of the study were as follows: radicular symptoms and signs with for ≥1 month despite conservative management (e.g., physiotherapy, selective nerve root blocks, and medication), and single-level lumbar HNP with neural contact as seen on lumbosacral spine magnetic resonance imaging (MRI) or MRI evidence of disc herniation with EMG evidence of lumbar radiculopathy. Patients with a progressive neurologic deficit including a cauda equine syndrome, previous lumbar disc operation, history of psychological disorder and radiological evidence of other pathologies that explain the clinical presentation (e.g., spinal stenosis and spinal cord tumor) were excluded.

We investigated the patients' medical records and MRI. We recorded the clinical characteristics and physical examination of the patients such as sex, age, smoking, duration of pain, and the result of the straight leg raising (SLR) test. The duration of pain that was longer than 6 months was defined as chronic pain. The straight leg raising test was considered positive when pain occurred before 60° passive elevation from horizontal [9].

One independent physician analyzed the MRI without previous knowledge of the patients' symptoms. The MRI images were classified by HNP level, HNP type, HNP zone, nerve root compromise grade, disc degeneration grade, and the high intensity zone (HIZ). Classification of HNP type, HNP zone followed the standards used by the American Society of Spine Radiology, the American Society of Neuroradiology, and the North American Spine Society [10]. HNP type was divided into 'protrusion', 'extrusion', and 'sequestration'. HNP zone consisted of 'central', 'subarticular', 'foraminal', and 'extraforaminal'. Nerve root compromise grade was classified as grade 0 (normal), grade 1 (contact), grade 2 (deviation), and grade 3 (compression) [11]. Disc degenerative grade was assessed by 5-scale degenerative grade by Pfirrmann et al. [12]. Grade 4 and 5 degeneration were defined as severe degeneration. The HIZ was defined as the local area with a high intensity signal in the posterior annulus fibrosus in the sagittal T2-weighted image [13].

PDD was performed on an outpatient basis. Preoperatively, prophylactic intravenous antibiotic was administered 30 minutes before the procedure. We monitored patients with electrocardiogram, pulse oximetry, and automated blood pressures. Sedative (20 mg of profopol) was administered as necessary during anesthetization of the skin and subcutaneous fascia.

The patients were placed in a prone position with the back mildly flexed on a surgical table. Fluoroscopic examination of the lumbar spine was performed to confirm intervertebral disc level and determine the appropriate level of needle [6]. We used a standard posterior lateral approach to the intervertebral disc, as previously described [14]; however, we modified the technique to approach the intervertebral disc more laterally so that the introducer needle would contact the disc margin rather than the midline.

A 25-gauge needle was first inserted into the ipsilateral side of target disc nucleus and 1 to 2 mL of contrast was injected to outline the disc herniation. Then, C-arm was rotated to position the lateral margin of the contralateral superior articular process as visualized in the oblique position. This typically required rotating the C-arm 15° from a zero-degree lateral projection. The skin and subcutaneous tissue were infiltrated with 1% lidocaine. We manually curved the 16-gauge introducer needle approximately 15° in the distal, 1 cm from the distal tip. The introducer needle was advanced toward the lateral edge of the superior articular process. Once the lateral edge was touched, we rotated the needle to point toward the midline. As the needle tip was directed toward the midline, the antero-posterior (AP) projection was checked intermittently to assure that the needle tip was always lateral of the medial border of the pedicle. The patient was asked to report any buttock or leg pain. The advancement of the needle was controlled by rotating the direction of the bent needle tip. Entering the herniation was identified by a sudden loss of resistance. After confirming the introducer needle position with the lateral and AP view, the wand was advanced to the center of the herniated disc under fluoroscopic monitoring (Figs. 1, 2). Before ablation, motor nerve stimulation was performed to confirm that the needle was not close to the nerve root. During the ablation, the tip of the wand was continuously rotated and moved back and forth to increase the ablated volume of the disc [6]. The entire procedure was monitored and evaluated by the C-arm fluoroscopy.

The effectiveness of PDD was investigated using the visual analogue scale (VAS) and the Oswestry Disability Index (ODI). Baseline data was obtained before the procedure. After the procedure, patients visited our outpatient department and answered the VAS and the ODI after 1-, 3-, 6-, and 12-month. We defined the successful outcome as >50% reduction of the VAS.

The improvement of pain and disability after percutaneous disc decompression were analyzed by the Wilcoxon signed-rank test. The Mann-Whitney U test, the Kruskal-Wallis test were used for analysis of difference between the clinical outcome and the presence of the patients' clinical and radiological data such as sex, smoking, duration of pain, SLR test, HNP level, HNP type, HNP zone, nerve root compromise grade, disc degeneration grade and HIZ. The Spearman correlation analysis was used to measure the correlation between age and the clinical outcomes. Univariate logistic regression analysis was performed on each variables to determine the crude odds ratios of successful outcome. After that, multivariate analysis was performed by logistic regression using the above-mentioned independent variables, and the successful outcome of PDD as the dependent variable. Statistical analysis were performed using SPSS Statistics ver. 20 software (IBM, Armonk, NY, USA). Significance was determined when the p-value was <0.05.

PDDs were performed in 109 patients from April 2009 to March 2013. Follow-up data were obtained in 83 patients. Seven patients were excluded due to prior lumbar surgery and 7 patients were excluded because of multi-level lumbar HNP. Sixty-nine patients, with mean age of 42.71 years (standard deviation of 12.76 years) were included in this study where 43 (62.3%) were male. Mean duration of symptoms was 19.90 months, ranging from 1 to 124 months. The basic clinical characteristics are shown in Table 1. There were no significant differences of baseline VAS and ODI according to subgroup analysis (Table 1).

Clinical outcome was investigated through the medical records. The VAS score decreased significantly from 6.59±2.00 to 2.46±2.00 at 1 year evaluation (p<0.01) (Fig. 3). The ODI score significantly decreased from 46.60±16.80 to 25.02±16.68 at 1 year follow-up (p<0.01) (Fig. 4). Eight patients who had prolonged pain underwent conventional microdiscectomy after PDD. There were no major complications including nerve damage, epidural hematoma or spondylodiscitis.

The change of VAS according to the factors were presented in Table 2. The improvement of pain (ΔVAS) between the baseline and 1 year follow-up was calculated. There were no significant differences in baseline VAS among subgroups. One year after the PDD, ΔVAS of patients with their pain <6 months was significantly greater than that with chronic pain patients (p=0.03). ΔVAS was significantly different between the 3 groups by HNP zone (p=0.04). Post hoc comparison showed that ΔVAS of the subarticular HNP was considerably better than that of the central HNP (p=0.015). We were unable to compare the effectiveness in the extraforaminal zone since there were no patients enrolled. Difference of ΔVAS by SLR test showed borderline significance between the 2 groups (p=0.06). The following factors showed no significant differences with the ΔVAS: sex, smoking, HNP level, HNP type, nerve root compromise grade, disc degeneration grade, and HIZ. Age was not significantly correlated with ΔVAS by the Spearman correlation analysis (r=0.012, p=0.93).

The improvement of disability (ΔODI) between baseline and 1 year follow-up was calculated. The statistical differences of ΔODI among the various factors were analyzed. There were no significant differences in baseline ODI among subgroups. ODI of the patients with HIZ was significantly better than that of the patients without HIZ (p=0.04). Difference of ΔODI by duration of pain showed borderline significance between the 2 groups (p=0.05). ΔODI was not significantly different among sex, smoking, SLR test, HNP level, HNP type, HNP zone, nerve root compromise grade, and disc degeneration grade (Table 2). Age was not significantly correlated with ΔODI by the Spearman correlation analysis (r=-0.156, p=0.23).

The associations between the successful outcome of PDD and patients' clinical and radiological characteristics were summarized in Table 3. Univariate analysis indicated significant differences for pain duration, SLR test, and HNP zone. Nerve root compromise grade was not calculated for the analysis because of small sample size in the deviation grade. When a dependent variable was set as a binary outcome by the presence of successful outcome after PDD, multivariate binary logistic regression analysis identified 5 independent predictors: pain duration of <6 months (odds ratio [OR], 14.036; 95% confidence interval [CI], 2.134-92.309), positive SLR test (OR, 8.425; 95% CI, 1.548-45.852), the extruded HNP (OR, 0.106; 95% CI, 0.013-0.899), the sequestrated HNP (OR, 0.037; 95% CI, 0.002-0.674), and the subarticular HNP (OR, 10.876; 95% CI, 1.705-69.361). HIZ (OR, 14.588; 95% CI, 0.769-276.674) showed borderline significance (p=0.074). No significant associations were observed for sex, young age (<40 years), smoking, HNP level, and severe disc degeneration.