Central or peripheral?

It should be remembered that the patient’s report of pain, the nature and severity, does not always distinguish between central and peripheral neuropathic pain; to the patient it feels the same irrespective of the location. Some tools can be used to distinguish neuropathic pain from non-neuropathic pain, which have usually been developed and tested on patients with peripheral neuropathic pain (Bennett et al., 2007) and have not been established in patients with central neuropathic pain with the exception of multiple sclerosis (MS) (Rog et al., 2007). Careful examination, quantitative testing and response to prior treatments can help elucidate the possible cause.

It is perhaps useful to refresh some of the definitions which are important in neuropathic pain (Merskey & Bogduk, 1994):

International Association for the Study of Pain definitions

Occasionally the sensations the patient reports might seem bizarre and may not fit the common definition of ‘pain’; this is immaterial to the patient. Some sensations can be extremely distressing even if we might not describe them as pain per se. We should also remember that patients with central neuropathic pain who usually have functional impairment, grossly altered biomechanics and ongoing secondary changes associated with neurological damage, are more likely to have co-existing pathologies and musculoskeletal pain problems. Each patient should be considered individually and the pain problem can be very complex. It should also be remembered that neuropathic pain may be persistent, paroxysmal (come and go without explanation), evoked (dependent on a stimulus) or any combination or all of these. Some peripheral neuropathic pains may start with a relatively trivial injury and the patient may have had trouble convincing health-care professionals of the veracity of their condition and consequently the condition goes undiagnosed for some time. A careful examination of the area is essential.

Central neuropathic pain

Central neuropathic pain can arise from primary injury or dysfunction within the central nervous system (CNS) (Merskey & Bogduk, 1994) and can arise at any level or even from more than one level. The recently suggested prerequisites for a diagnosis of the condition are:

It is important to distinguish the changes which may occur as a result of neuroplastic changes after damage to the peripheral nervous system (PNS) or which may occur in chronic pain states which may begin with an injury in another part of the body other than the CNS. In these conditions the dysfunction of the CNS is considered to be secondary to ongoing nociception rather than from a primary source in the CNS.

Central neuropathic pain states can be usefully classified into three main groups: (1) pain associated with progressive neurological conditions, e.g. MS; (2) pain following stroke; and (3) pain following spinal cord injury. Another group which does not necessarily fall into these is neuropathic pain associated with HIV infection. This may be due to damage caused by the HIV virus itself or by neuropathy as a result of the antiretroviral treatment (Cox & Rice, 2008). This last group will not be considered here.

The prevalence of pain for all causes in MS has been estimated at between 43% and 70% (Moulin et al., 1988; Solaro et al., 2004). One of the best estimates for the presence of central neuropathic pain is that over 27% of patients report the condition (Osterberg et al., 2005). It is commonly widespread with an increased prevalence in the lower limbs and a variable clinical picture but a low report of paroxysmal pain. A small proportion of patients report pain at the onset of their MS but in general the incidence for central neuropathic pain syndrome (CNPS) is reported to be higher in the early years. However, this may be an artefact of diagnosis, as additional musculoskeletal pain and pain associated with spasticity may predominate in the later stages of the disease masking the true prevalence of neuropathic pain. The actual cause of CNPS in MS is difficult to determine owing to the disseminated nature of the disease. Using MRI, Osterberg et al. (2005) demonstrated hyperintensity of activity in the lateral and medial thalamic regions in one-third of MS patients with CNPS and concluded that, although there is an indication that the cause of the pain may share some similarities with central stroke pain, they could not conclude that lesions in the thalamus were the cause of the pain and postulated that lesions in the spinal cord, particularly the spino- and quintothalamic pathways, are likely to be the cause. A previous hypothesis also suggested the importance of lesions in the neospinothalamic pathway, which may become hyperexcitable following lesioning.

Pain is also seen in Parkinson’s disease but the exact reason for this is yet to be determined. It is suspected that the basal ganglia and dopaminergic systems are involved in the processing of nociception to higher centres (see Ch. 6).

Post-stroke pain is the commonest CNPS seen in the population because of the common occurrence of stroke. Andersen et al. (1995) followed 207 new stroke patients who survived for 6 months and were able to participate in a quantitative sensory testing protocol. In this study abnormal sensory signs were common (47%) and 8% of patients were diagnosed with post-stroke pain. There does not appear to be a higher prevalence in either ischaemic or haemorrhagic strokes, but because more people sustain ischaemic strokes CNPS is more commonly seen in this group of patients in clinical practice (Andersen et al., 1995).

The original description of a possible cause of post-stroke pain was made over one hundred years ago (Dejerine & Roussy, 1906) and since that time the role of lesions of the thalamus in ‘thalamic pain’ has been well documented. The literature on the importance of lesions in and around the thalamus has grown. Lateral medullary infarctions are more likely to result in damage to the spinothalamic and trigeminothalamic pathways and have the highest incidence in the development of CNPS (MacGowan et al., 1997). The incidence of pain following damage to the thalamus is also high. In an MRI study of people with post-stroke pain a high proportion of them had thalamic lesions (>60%), but multiple lesions were seen in almost all of the patients and no thalamic involvement was demonstrated in others, so the specificity of the location is difficult to demonstrate (Bowsher et al., 1998).

Pain following spinal cord injury

About two-thirds or more of people who sustain a spinal cord injury report persistent pain, but this can be due to a number of reasons, not all of which are due to CNPS (Siddall et al., 2002). A proposed classification for spinal cord injury pain describes two broad groups: nociceptive and neuropathic. Nociceptive is further broken down into musculoskeletal and visceral (Siddall et al., 2002). Neuropathic is subdivided into:

In a 5-year follow-up of 73 patients with spinal cord injury Siddall et al. found 81% of patients reported pain; of these 41% had neuropathic pain at the level of the lesion and 34% had neuropathic pain below the level of lesion. Most patients reported more than one type of pain with musculoskeletal pain being the commonest (59%), although the least severe (Siddall et al., 2003).

At level neuropathic pain corresponds to the segmental level of the injury often involving two dermatomes above or below. At level neuropathic pain can be a result of damage to the nerve roots or the spinal cord and is characterized by allodynia and hyperalgesia in the affected dermatomes. The physiology is the same as by which peripheral neuropathic pain develops following damage or constriction of peripheral nerve roots in the affected area.

Below level pain is purported to be caused by changes in the spinal cord in or near the area of injury as described above and also as a result of what is sometimes termed a ‘supraspinal generator’. A loss of neurological input to higher centres, especially the thalamus, may result in abnormal activity, in particular spontaneous activity. However, the nature of the pain, which is often persistent without the spontaneous outburst of pain, suggests that there is more than one area of the brain implicated as a supraspinal generator.

Peripheral physiological changes