Predictive factors of volumetric reduction in lumbar disc herniation treated by O2-O3 chemiodiscolysis.

PURPOSE: Aim of this study is to assess the effectiveness of O2-O3 percutaneous chemiodiscolysis by evaluating volumetric changes in lumbar disc herniation on magnetic resonance imaging, in order to identify possible pre-treatment factors affecting such changes
Methods: Between January 2014 and December 2017, a total of 87 patients with low back pain and 103 lumbar disc herniations with MRI confirmation were considered for O2-O3 chemiodiscolysis. The volume of each herniated disc was determined before and after the treatment.
RESULTS: Multiple linear regression analysis showed a strong correlation between post-treatment LDH volume percent change and both pre-treatment LDH volume and pre-treatment EQ-VAS (p<0.05), while age showed only a weak positive correlation with post-treatment LDH volume percent change (p<0.1). No association was found for other factors, such as sex and herniation disc level.
CONCLUSIONS: In conclusion, age, baseline LDH volume and self-assessed disease severity score could represent three easy accessible outcome predictive parameters to consider when intradiscal O2-O3 chemiodiscolysis is envisaged. Better results after intradiscal O2-O3 chemiodiscolysis were obtained in older patients with higher pre-treatment LDH volume and low-moderate pre-treatment EQ-VAS.

= Lumbar Disc Herniation

= Magnetic Resonance Imaging

= EuroQol Visual Analogue Scale

= Cross Sectional Area

= Diffusion-Weighted Imaging

= Apparent Diffusion Coefficient

Introduction

Degenerative disease of the intervertebral disc presenting as symptomatic lumbar disc herniation represents one of the most frequent causes of low back pain and functional limitation in adult population, with variable onset (ranging from acute to chronic) and symptoms intensity. Magnetic resonance imaging is one of the most largely used imaging tool in many diagnostic and interventional settings thanks to its intrinsic excellent soft tissue contrast and the absence of ionizing radiation compared to CT (1-9).

Despite disc herniation can resorb spontaneously and associated symptoms can be relieved by conservative therapy, indications for surgical approach should be considered in case of drug-resistant persistent pain. At present the most appropriate surgical choice remains unclear, with recent evidences suggesting that minimally invasive techniques can be adopted prior to invasive surgery. Among these techniques, percutaneous intradiscal ozone injection has successfully and safely been used, with encouraging results in different clinical trials on large and representative populations (10, 11). Indeed ozone, a gas normally present in the atmosphere, can be used in combination with oxygen (O2-O3) at low concentrations for chemiodiscolysis; its main indication is low back pain (independently from radicular pain) refractory to 4-6 weeks of conservative therapy, in absence of motor deficits. Percutaneous CT-guided intradiscal injection is performed with thin needle (18-22gauge), and it is usually complemented by peri-ganglionic injection of corticosteroids and anaesthetics (12). This combination ensures immediate pain relief and allows for ozone to display its beneficial effects, including reversion of local venous stasis, anti-inflammatory effect and anti-nociceptor analgesic activity. O2-O3 chemiodiscolysis is a cost-effective procedure, with low complication rate and the possibility of being repeated in the same patient without precluding recourse to different surgical techniques in case of failure. However, despite its many advantages, O2-O3 chemiodiscolysis still raises some concerns in patients’ selection and outcome optimization. Aim of this study is to assess the effectiveness of O2-O3 percutaneous chemiodiscolysis by evaluating volumetric changes in lumbar disc herniation (LDH) on magnetic resonance imaging (MRI), in order to identify possible pre-treatment factors affecting such changes.

Materials and Methods

Patients’ selection

Between January 2014 and December 2017 we prospectively recruited 146 patients (86 males, 58,9%; 60 females, 41,1%; mean age 54,6y, DS±13,9y, range 24–83y) with low back pain following a dermatomal distribution and MRI confirmation of single- or multiple-level LDH, who were considered for O2-O3 chemiodiscolysis; 14 patients presented with disc herniation at two levels, while 3 patients at three levels. Patients referred to our department for well-discriminated persistent dermatomal pain with positive Lasègue test. The inclusion criteria were: single- or multiple-level LDH at MRI examination; 3-months long conservative treatment failure; age >18y and <85y. Exclusion criteria were: significant structural deformity of the spine, severe vertebral osteoarthrosis, fractures or calcified hernias (13); neurological motor deficit or cauda equina syndrome; infectious, inflammatory or neoplastic bone lesions (14-16); former contraindications to undergo MRI examination; absolute contraindications to CT-guided procedure, including allergy to proposed drugs and pregnancy (17). The presence of bleeding disorders, glucose-6-phosphate dehydrogenase deficiency, hyperthyroidism, severe anaemia, myasthenia, recent myocardial infarction and history of mental disorders were considered relative contraindications, therefore evaluated on a case-by-case basis. Finally, 87 patients (48 males, 55,2%; 39 females, 44,8%; mean age 53,7y, DS±13,1y, range 24–83y) were eligible for chemiodiscolysis, and a total number of 103 LDH were considered for statistical purposes (L5 n=25; L4 n=56; L3 n=15; L2 n=7). Before minimally invasive treatment, each patient was asked to rate disease severity by using the self-reported EuroQol Visual Analogue Scale (EQ-VAS) for pain intensity measurement. After therapy, all patients undergone chemiodiscolysis performed a follow-up lumbar spine MRI examination within 6 months from treatment (mean interval 19,7w; DS±7,2w; range 10-29w). The study was formerly approved by local Institutional Board, and written informed consent was preliminarily obtained from each patient.

Intradiscal O2-O3 chemiodiscolysis

Intradiscal O2-O3 chemiodiscolysis is carried out in a one-day surgery regimen; preliminary CT examination is performed to identify the infiltration point and measure the distance from the conjugate foramen. The patient is placed on the imaging bed in prone position, with a cushion to reduce the physiological lumbar lordosis. CT scans are focused on the level to be treated and the entry point is marked with demographic pencil; the needle is introduced with an angle of 45° from the skin plane under CT guidance. Treatment consisted of 15ml O2-O3 injection (concentration: 30μg/ml), partly injected directly into the herniated disc (3-4ml) and partly in the periradicular/paravertebral soft tissues (11-12ml); using the same needle, the procedure is concluded with intraforaminal injection of corticosteroid (1mL, 40mg methylprednisolone) and local anaesthetic (1mL, 0.5% marcaine). A final CT scan to check gas distribution was also performed in order to exclude possible complications.

MRI examination and volume changes assessment

Pre- and post-operative lumbar spine MRI examinations carried on a 1.5T scanner included a sagittal TSE-T2w sequence (TE=100ms; TR=3000ms; NA=3; FA=90; slice thickness=3mm; spacing between slices=0), along with other axial and sagittal conventional T1w and T2w sequences. Two experienced neuroradiologists in consensus, different from the ones of team that performed the chemiodiscolysis, provided LDH measurements on raw MRI images. Lateral insertion positions of LDH were defined as the points showing the greatest contrast between the structures, while intervertebral disc height was defined as the distance between the middle portions of each endplate at the level of the affected disc. The contained affected disc was defined as the material located in the intervertebral space behind the intervertebral height line; conversely, the outer disc herniation was included until its borders became visually indistinguishable. Disc signal intensity and type of disc herniation (protruded, contained, extruded, sequestered, free fragment and migrated disc herniation) were also considered. For both time-points (pre- and post-operative MRI examination), LDH volume was assessed by hand-drawing height and cross sectional area (CSA) of the herniated disc located between the lateral margins of the vertebral pedicles (18), and computed by using the OsiriX “Compute Volume Tool”. An example of LDH measurements before and after intradiscal O2-O3 chemiodiscolysis is shown in Figure 1.

Figure 1.

Sagittal and axial TSE-T2w MRI images on the most representative level, showing an example of LDH measurements before (a-b) and after (c-d) intradiscal O2-O3 chemiodiscolysis.

Statistical analysis

Multiple linear regression analysis was used to describe the relationship between LDH volume variation after intradiscal O2-O3 chemiodiscolysis and independent variables including age, sex, disc level, LDH volume at MRI diagnosis and pre-treatment self-assessed score of disease severity (EQ-VAS). To explain variation in treatment response that can be attributed to each explanatory variable, linear regression analysis was then applied; goodness of fit of regression was analysed using R2 along with p-value. All statistical analysis was performed using XLSTAT software v.2019.1 (Addinsoft).

Results

LDH mean volume before treatment at baseline MRI examination was 0,58cm3(DS±0,29cm3), versus 0,43cm3(DS±0,23cm3) after treatment at MRI follow-up (mean variation -0,15±٠,19; -20,4%, DS±44,6%); in 87,2% cases a reduction in LDH volumes was computed, in 10,9% cases there was a slight increase in LDH size, while in the remaining 1,9% cases no significant variation was observed. Concerning self clinical assessment, mean EQ-VAS at baseline was 23,4 (DS±18,1), versus 59,2 (DS±25,1) after treatment (mean difference +35,7±30,9); 81,4% patients experiences an improvement in perception of disease severity, 9,7% patients complained a worsening in clinical symptoms, while the remaining 8,9% patients did not referred significant variation.

Multiple linear regression analysis showed a strong correlation between post-treatment LDH volume percent change and both pre-treatment LDH volume and pre-treatment EQ-VAS (p<0.05), while age showed only a weak positive correlation with post-treatment LDH volume percent change (p<0.1). Adjusted R2 was 0.114 for the regression equation “Y= -0.112810 -0.392345X1 +0.00580527X2” (with Y=volume percent change; X1= pre-treatment LDH volume; X2= pre-treatment EQ-VAS). No correlation was found for sex and herniation disc level. Multiple linear regression analysis results are shown in Table 1. Linear correlation for bivariate data showed a significant negative correlation between post-treatment LDH volume percent change and pre-treatment LDH volume (p=0.013), as well as a significant positive correlation between post-treatment LDH volume percent change and pre-treatment EQ-VAS (p=0.021). Bivariate linear correlation analysis results are resumed in Table 2, and plotted in Figure 2.

Table 1.

Multiple linear regression analysis results of patients’ variables (pre-treatment LDH volume, age and pre-treatment EQ-VAS, respectively) with post-treatment LDH volume percent change (p<0.05)

	Post-treatment LDH volume percent change
Patients’ variables	b	p-valueLower limit	95% CI
			Upper limit
Pre-treatment LDH volume	-0.394368	0.008 (p<0.05)	-0.682803	-0.105933
Age	-0.00551925	0.088 (p>0.05)	-0.0118802	0.000841720
Pre-treatment EQ-VAS	0.00526179	0.026 (p<0.05)	0.000651827	0.00987176

Legend: LDH=lumbar disc herniation; b=regression coefficient; CI=confidence interval; EQ-VAS=EuroQol Visual Analogue Scale.

Table 2.

Bivariate linear correlation analysis results of patients’ pre-treatment variables with post-treatment LDH volume percent change (p<0.05).

	Post-treatment LDH volume percent change
Patients’ variables	r	R2	p-value
Pre-treatment LDH volume	-0.243	0.059	0.013
Age	-0.187	0.035	0.057
Pre-treatment EQ-VAS	0.226	0.051	0.021

Legend: LDH=lumbar disc herniation; r=correlation coefficient; R2=R-squared; EQ-VAS=EuroQol Visual Analogue Scale.

Figure 2.

Bivariate linear correlation showing: (a) significant inverse correlation between post-treatment LDH volume percent change and pre-treatment LDH volume; (b) lack of significant association between post-treatment LDH volume percent change and patients’ age; (c) significant positive correlation between post-treatment LDH volume percent change and pre-treatment EQ-VAS.

Discussion

Back pain is one of the most common health problems affecting western population (19), and can be attributed to a wide range of disorders recognizing different aetiologies (8, 20-23). Among the possible aetiologies, LDH represents the most common cause (up to half cases of low back pain associated to sciatic symptoms are generated by disc herniation) (13, 16). The mechanism underlying low back radicular pain is multifactorial. Beyond the clinical evaluation, imaging plays a fundamental role in the initial assessment and cost-effective patient management; CT and MR imaging are considered the modality of choice, as in the study of other pathologies (24-34).

Indeed, LDH determine both variable direct compression of nerve root/dorsal root ganglion and indirect compression on perineural vessels, causing local venous stasis and ischemia. Moreover the injured herniated disc determines focal inflammation with subsequent involvement of facet capsule, epidural tissue surrounding nerve root and nerve root itself; the subsequent inflammatory cascade sensitizes neural nociceptors and reduces pressure pain threshold (14, 16, 35).

LDH natural history is variable, and it is well known that disc herniation can resorb spontaneously and associated symptoms can be relieved by conservative therapy only (18). Therefore indications for surgical approach should be considered in case of drug-resistant persistent pain, when symptoms of LDH are refractory to initial conservative management (36, 37). However, before resorting to invasive surgery, minimally invasive approaches are often envisaged as a viable alternative to the more conventional open surgical techniques as they offer patients several benefits (including minor tissue trauma, reduced pain, and prompt recovery). Among these techniques, intradiscal O2-O3 chemiodiscolysis is one of the most used due to the beneficial antiseptic, immunomodulating, analgesic and anti-inflammatory ozone properties (38). However at present literature evidences from O2-O3 chemiodiscolysis compared to conventional surgical techniques are still controversial, with surgical treatment providing faster and greater relief from back pain in patients with LDH (despite no significant difference in results stability over time was observed at mid-term and long-term follow-up) (37). This variability in O2-O3 therapy outcome can be at least in part attributed to a challenging patient selection (39).

In this study we analysed potentially predictive factors of LDH volumetric reduction in order to improve patient selection and intradiscal O2-O3 chemiodiscolysis results, highlighting the possible role of age, baseline LDH volume and EQ-VAS score in predicting treatment outcome.

On the one hand, our results showed a strong correlation between post-treatment LDH volume percent change and both pre-treatment LDH volume and pre-treatment EQ-VAS (p<0.05). Inverse correlation between post-treatment LDH volume percent change and pre-treatment LDH volume suggest that O2-O3 chemiodiscolysis is more effective on larger disc herniation, probably when trans-ligamentous extension and posterior longitudinal ligament disruption makes unlikely an eventual spontaneous reduction (40). At the same time, positive correlation between post-treatment LDH volume percent change and pre-treatment EQ-VAS suggest that O2-O3 chemiodiscolysis should be preferred in case of mild clinical symptoms and moderate functional limitations, reserving more invasive approaches to patients with higher disease severity who experiment only limited benefits from minimally-invasive strategies. However it should always be considered that self-rating scales underlie some possible limitations regarding the interpretation of the results, which are mainly related to difficulties in determining how large a change in a certain score should be before it can be considered relevant for the patient.

On the other hand, age showed only a weak positive correlation with post-treatment LDH volume percent change (p<0.1). In recent times age was proposed as one of the factors influencing LDH resorption (41); indeed herniated disc in older patients is supposed to be more fibrotic and dehydrated compared to the one of younger patients, negatively influencing spontaneous resorption. It is already known that, when considering acute disc herniation typical of younger patients, disc material is hyperintense on T2w sequences in more than 70% cases, with higher tendency to resolution (16). Conversely, when acute symptoms appear in a context of chronic disc herniation, inflammatory response is more involved as chronically damaged disc tissue elicits intense inflammatory changes when exposed to epidural fat/vasculature (42). According to this finding, in our sample the large majority of LDHs in older patients showed a low T2w-hyperintensity (presumably related to poor water content) (14, 15). Probably due to this difference, the LDH volume percent change after intradiscal O2-O3 chemiodiscolysis was more satisfying in older patients, where analgesic and anti-inflammatory ozone properties played a prominent role over direct disc dehydration.

Our experience further confirms the pivotal role of MRI in providing crucial data regarding pre-treatment patients’ stratification based on LDH morphological features, thus not limiting this technique to simple diagnostic purposes. To this purpose recent evidences also suggested a possible role for diffusion-weighted imaging (DWI) and relative apparent diffusion coefficient (ADC) maps, already commonly used in different clinical settings to predict disease evolution and clinical outcome of various central nervous system disorders (43-45). In particular, LDH with lower ADC values were associated to poor improvement in clinical symptoms after O2-O3 chemiodiscolysis (30, 46). However DWI sequences were not available for our patients, therefore further studies are required to assess the possible role of pre-treatment ADC values for patients’ selection. Some more limitations to the study should also be considered. First the results are referring to a 6-months long follow-up, and LDH volume reduction stability over time should be monitored on a mid/long-term follow-up; moreover, some additional risk factors that could somehow influence symptoms relapse (including for example weight fluctuations or work/professional activity) should also be considered in future analysis. Finally, it would be useful to assess the reproducibility of the presented results with different minimally invasive therapies other than intradiscal O2-O3 chemiodiscolysis.

In conclusion the flourishing of non-surgical therapeutic strategies for LDH should benefit of a more precise patients’ selection based on clinical data and imaging features, in order to optimize treatment results. Age, baseline LDH volume and self-assessed disease severity score could represent three easy and rapidly accessible outcome predictive parameters to consider when intradiscal O2-O3 chemiodiscolysis is envisaged, with better outcome after intradiscal O2-O3 chemiodiscolysis in older patients with higher pre-treatment LDH volume and low-moderate pre-treatment EQ-VAS.