The treatment of discogenic low back pain is one of the most challenging clinical problems to the spinal surgeon. Undoubtedly, nonoperative treatment, including bed rest, exercise, traction, drug therapy, transcutaneous electrical nerve stimulation, manipulation, etc. may be effective for some patients. However, a substantial number of patients experience no improvement with these therapeutic attempts [25]. For these patients, surgical management may be necessary because of their poor response and functional limitations. A novel, valid, and minimally invasive method of management for discogenic low back pain is therefore desirable.

The pathogenesis of discogenic low back pain is extremely complicated and poorly understood. Our previous study discovered that the distinct pathological characteristic of the painful discs was the formation of a zone of vascularized granulation tissue with extensive innervation extending from the outer layer of the annulus fibrosus into the nucleus pulposus along a torn fissure as shown on CT discography [21]. Our findings suggested that the zone of granulation tissue may be the culprit responsible for causing the discogenic low back pain. With this assumption in mind, it is presumed that if the nerve fibers and nerve endings growing into the disc along the tear could be devitalized, discogenic pain would be alleviated or abated.

The treatment efficacy of patients with a single level discogenic pain was compared with that of two levels. According to a minimum two-point change between pre- and post-treatment VAS scores, of the 24 patients treated with intradiscal MB injection, a successful result was achieved in 17 (94%) of 18 patients treated at a single level and four (67%) of six patients treated at two levels. The VAS scores in patients injected at a single level were improved by a mean reduction of 5.65 (t=12.511, P=0.000), and in patients injected at two levels by a mean reduction of 4.40 (t=4.439, P=0.007). The ODI changes in patients injected at single level was improved by a mean reduction of 35.72 (t=12.265, P=0.000), and in patients injected at two levels it was improved by a mean reduction of 24.50 (t=3.246, P=0.029) (Table 2).

In recent decades, surgical fusion of the lumbar spine has been increasingly applied for chronic low back pain. However, the reported results vary considerably in different studies, and the complication rate after fusion surgery in the lumbar spine is not negligible [7–9, 14, 18, 22]. As an alternative treatment, intradiscal electrothermal annuloplasty (IDET) has recently been advocated [13, 19, 26–28]. This procedure involves the introduction of a flexible electrode into the disc to encircle the nucleus pulposus around the lateral and posterior annulus. Electric current is passed through the electrode to produce heat to coagulate collagen and nociceptive nerve fibers. However, the mechanism of relieving pain has been denied by a recent experimental study [6], though a successful outcome of IDET has been reported in 40–80% of patients [13, 19, 26–28].

The treatment results achieved in the current study were similar or superior to those obtained by fusion surgery or IDET [7–10, 14, 22], as shown by the difference between pre- and post-treatment VAS scores [7–10, 13, 14, 19, 22, 26–28]. In addition, based on the ODI, marked improvement was observed in 87% of the patients. Although the follow-up time in the current study was short, there was no difference in the VAS and the ODI scores between 3 months and those of 12 months (or more) after the treatment for all the patients. This observation suggests that the therapeutic efficacy of intradiscal MB injection is stable and credible.

The pathology responsible for chronic discogenic low back pain is within the disc property itself. The only feature of these painful discs that is distinctive from other asymptomatic degenerated discs is the positive pain response during discography [2, 10, 12, 16, 23, 29]. Mooney [17] maintained that the disc should be the primary structure responsible in producing low back pain. Our previous study found [21] that the pathologic features of discs obtained from the patients with discogenic low back pain were the formation of a zone of vascularized granulation tissue with extensive innervation in fissures extending from the outer part of the annulus into the nucleus pulposus. The study clearly demonstrated that in the zone of vascularized granulation tissue there were abundant SP-immunoreactive nerve fibers which had been thought to be nociceptive. Most likely the nociceptors within the granulation tissue could be excited by local inflammation to produce a painful response. Thus, the sharp painful sensation elicited during injection of contrast medium should be attributed to the innervation into the fissures within the disc structure. The sharp increase of intradiscal pressure following the injection of the contrast medium doubtlessly produces an abrupt irritation of the nociceptive nerve elements, hence excruciating pain, as evidenced by immediate subsidence of pain when pressure declines as the contrast medium leaks out of the disc from the tears [21].

According to the results of our previous study [21], theoretically, if a drug or chemical compound could destroy the nerve endings or nociceptors ingrown into the painful disc along the tear, the rational treatment of discogenic low back pain should be aimed at disrupting the pathway of nerve conduction. Intradermal MB injection has been shown by electron microscopy to be able to destroy dermal nerve endings [3, 5]. It led us to think that it is reasonable to use intradiscal MB injection for alleviating discogenic low back pain. A successful outcome has finally been achieved in relieving pain and improving disability. This result may indicate that MB indeed has devitalized the nerve endings or nociceptors in the painful discs. Theoretically, devitalized nerve fibers or nerve endings are able to regenerate under certain conditions, and consequently leading to the recurrence of symptoms. Practically, our observation revealed that the pain relief of intradiscal MB injection is long lasting. The underlying mechanisms of a prolonged effect may include that MB destroys nerve endings or nociceptors within discs, sustains action ascribed to the absence of blood vessels in the disc, and alleviates inflammatory response by inhibition of vasodilatation [20, 30]. Further evidences are required to clarify the accurate mechanisms of efficacy of intradiscal MB injection.

Although MB might leak into the spinal canal through annular tears, no side effects or complications were discovered in the group of patients treated with intradiscal MB injection. We presumed that MB may act only on nerve endings or nociceptors, because no evidence of nerve root injury was found. The exact mechanisms of the effect of intradiscal MB injection are not yet clearly understood. Because of high incidence and high disability rate of chronic discogenic low back pain, further clinical and experimental studies appear extremely necessary.