Spinal cord stimulation, from diagnosis-oriented to mechanism-based treatment

There is a considerable increase of randomized placebo-controlled trials (RCTs) for the treatment of chronic pain in recent years (Attal et al., 2006). Most of these trials assess drug efficacy. Spinal cord stimulation (SCS) has been shown to be helpful in patients with failed back surgery and radicular leg pain (Kumar et al., 2007), arterial occlusive disease – related pain and Complex Regional Pain Syndrome type I (CRPS) (Kemler et al., 2004). This evidence usually comes from comparative studies, where the best conventional treatment is the comparator. Placebo-controlled studies cannot be performed with SCS, because a therapeutic effect is achieved only when subjective paresthesias cover the painful area, and subthreshold stimulation is not beneficial. The pain relieving mechanism of SCS is not fully understood. Besides segmental mechanisms with A-ß fiber and dorsal column stimulation and local modulation in the superficial layers of the dorsal horn, neurohumoral effects with the release of inhibitory transmitters like GABA or glycine and supraspinal effects are postulated (Rasche et al., 2005).

Despite its doubtless beneficial effects, SCS became never a routine treatment for chronic CRPS. The reasons are the invasive nature, the high costs and the high rate of treatment failures. Unfortunately predictors for a positive treatment response in order to improve the cost/efficacy ratio of SCS have been unknown. In their investigation published in this issue of the European Journal of Pain Van Eijs and colleagues (2010) build on their earlier study (Kemler et al., 2001), in which they demonstrated the positive effects of SCS for treatment of chronic CRPS pain on a group level. In the actual study they followed-up their 36 patients treated with SCS for CRPS I for up to 1 year. While analysis failed to demonstrate an effect of patient´s age, pain intensity and sensory loss, the presence and the amount of brush-evoked allodynia proved to be a significant negative prognostic factor for long-term SCS treatment response. Thus, the paper demonstrates and proves for the first time, what many clinicians already have observed in their patients: while the constant pain is usually very well reduced by SCS, allodynia is not.

Allodynia to brushing the skin occurs when impulses from low-threshold mechano-afferent fibers (A-ß-fibers) get access to the nociceptive system – either in the spinal cord by excitation of spinal nociceptive and wide dynamic range neurons (Koltzenburg et al., 1994) or by reorganization of the brain (Maihofner et al., 2004). As described above, SCS now excites exactly these low-threshold primary afferent A-ß fibers and the dorsal columns in the spinal cord – in order to enhance inhibition as a prerequisite for the suppression of ongoing pain. If allodynia is established, exactly this has fatal consequences since this low-threshold mechanical afferent input has pathological access to the nociceptive system. Thereby SCS analgesia will be levelled by the competition of analgesic and allodynia action. These findings of Van Eijs and colleagues are therefore not only important for clinicians for patients selection, they also improve our understanding of chronic pain by practically supporting theoretical considerations from many experimental pain studies.

In contrast to allodynia, the presence of non-painful sensory loss was unrelated to SCS efficacy although in general sensory loss might indicate Aß or dorsal column dysfunction and therefore a loss of the substrate for SCS to induce pain relief. Maybe this fact supports recent findings that sensory loss in subjects without nerve lesion (CRPS I) must not indicate loss of afferent fibers or function but rather indicate changes of central touch processing during ongoing pain (Geber et al., 2008).

The study has, however, also limitations. The first one is the limited number of subjects for such an association study. Obviously the indication for SCS implantation in CRPS is so rare that higher numbers are hard to achieve in single centre studies. Multicentre studies are highly recommended. The second one is the fact that the same patients group as published in 2001 (Kemler et al., 2001) was investigated. In fact, this renders the study retrospective since the original study was not designed to investigate treatment failure but treatment success.

Nevertheless, the results presented from Van Eijs and colleagues have implications for chronic CRPS management – since they help to select patients based on their symptoms. This will help to improve cost – efficacy relation of SCS and to avoid treatment failures. And it might be a first step into an individualized and mechanisms-oriented treatment of chronic pain (Baron et al., 2009), in the case of CRPS – without brush-evoked allodynia. As long as we cannot adhere to such a restriction of SCS use since confirming data are missing, we must at least honestly inform the individual patient about their expectations and chances of symptom relief in cases of combined spontaneous pain and allodynia. In these cases further treatment for allodynia in addition to SCS might be necessary for sufficient pain relief.