The study protocol was approved by the medical ethical committee of our institution.

139 consecutive patients (75 M, 64 F; aged 18-71 years) entered the study.

Inclusion criteria : monoradicular pain, lumbar disc herniation shown by CT or MRI, herniation site congruous with the neurological level, and Oswestry disability index over 30%. All patients complained of pain since at least 8 weeks and they had been submitted to conservative therapy for 2-4 weeks, with no or poor clinical improvement.

Patients with major neurological deficits and other spine disorders (degenerative, infective, previous spinal surgery) were excluded.

Patients were randomly assigned to two groups. 66 patients (group A) underwent intradiscal and intraforaminal injection of Triamcinolone Acetonide 2 ml (40mg/ml) and Ropivacaine 2% 2-4ml. 73 patients (group B) underwent the same treatment with add of 5-7 ml of O2-O3 mixture injection.

Procedures
All procedures were performed on an outpatient basis.

A 9 cm or 15 cm 22G spinal needle was positionned by means of a paravertebral approach (Fig. 1). The side of the injection was chosen on the basis of the main location of symptoms. In some cases an interlaminar approach was preferred (Fig. 2a, Fig. 2b, Fig 2c).
After local anaesthesia, the needle was advanced to the intraforaminal space. After confirmation of the position, the drugs were injected at this level (Fig. 3a). The needle was subsequently advanced towards the centre of the disc, in order to inject the drugs inside the nucleus pulposus. When the needle entered the disc, a soft resistance was felt. Before injecting CT confirmed needle position. Postprocedural CT scan was acquired to check for O2-O3 distribution (Fig. 3b, Fig. 3c). If epidural diffusion was absent or poor, the needle was pulled out of the disc and repositioned in the epidural space, and O2-O3 was injected again.

Outcome evaluation
To determine the efficacy of the procedures, a 6 months follow-up was performed.
We administered to all patients the Oswestry pain questionnaire the same day of the procedure, after two weeks, after three months and after six months. The Oswestry disability index was calculated to determine patients outcome.

Patients and clinicians who tested them were blinded whether the treatment received.

The treatment was considered successful when the Oswestry disability index at follow-up was 0-20%. If the index was more than 20% the treatment was considerate unsuccessful, even if the patients had moderate improvement.
The Chi-squared test was used for comparison.