Anatomy

The DREZ is the region of the spinal cord that contains the dorsolateral fasciculus of Lissauer and Rexed laminae I to V (Fig. 130-1). The anatomy of the DREZ in both the cervical spine and the lumbosacral outflow region has been studied in detail by Sindou and others.^3,7–11 The dorsal nerve roots enter the dorsolateral spinal cord as a linear arrangement of small rootlets.¹² The division into rootlets occurs approximately 1 cm before entering the dorsolateral spinal cord, and the rootlets may travel in a subpial plane for 1 mm before entering the cord.³ Microanatomic studies have indicated that the cervical dorsal nerves divide into an average of 7.7 rootlets prior to entering the lateral sulcus, with fewer rootlets in the upper cervical spine.^8,9 The length of the rootlets generally increases in the lower cervical spine.¹³ Within the rootlets, small fibers, including the nociceptive afferents, are laterally located, while the larger fibers from somatic receptors enter a medial location. A portion of the lateral small fibers proceed to enter Lissauer’s tract prior to terminating in laminae I and II of the dorsal horn. Lissauer’s tract modulates synaptic transmission and is composed of a medial part with a large percentage of unmyelinated fibers from the lateral rootlets and lateral portion that contains propriospinal fibers.^14–16

FIGURE 130-1 Cross-sectional diagram of the spinal cord demonstrating the locations of the Rexed laminae within the dorsal gray matter of the cord: posteromarginal nucleus (I), substantia gelatinosa (2), nucleus proprius (III and IV), and the Rexed laminae (V and VI). The arrowhead on the left marks the depth of the DREZ lesion as discussed by Sindou.³ This is typically 3 mm in depth in the region of the brachial plexus.

Figure 130-2 presents a diagram of the DREZ in the cervical region that illustrates the three characteristic geometric measures of the exiting rootlets, including (for a given cervical segment) the angle of exit of the superiormost and inferiormost rootlets with respect to the cord and the vertical length of the DREZ along the cord. Alleyene et al. measured these quantities from the C3 to T1 segments in a set of 10 adult cadavers (128 cervical roots).⁷ The average angle with respect to the cord for the inferior rootlets ranged from 116 degrees (C5) to 156 degrees (T1), while the superior angle ranged from 132 degrees (C6) to 166 degrees (T1). Similarly, Xiang et al. explored the cervical DREZ region in a series of 20 adult cadavers (200 dorsal cervical roots, C5-T1).⁹ The average inferior angle the inferior rootlets made with the axis of the spinal cord (180 degrees, the angle of Alleyne et al.⁷) varied gradually from 65.6 ± 3.6 degrees at C5 to 19.8 ± 2.9 degrees at T1.⁹ In addition, these authors measured the distances from the posterior median sulcus to the posterior lateral sulcus (the anatomic sulcus at which the root fibers enter the cord). This quantity varied from 3.54 ± 0.12 mm at C5 to 2.23 ± 0.28 mm at the T1 level. The average distance of the DREZ from the median sulcus is 2.95 mm, with the distance decreasing in the lower cervical spine.^8,9

FIGURE 130-2 Box 1: Definitions of the superior (S) and inferior (I) rootlet angles, and the length of the dorsal root entry zone (D) as defined by Alleyne⁷. Box 2: Definitions of the superior (S’) and inferior (I’) angles as defined by Xiang⁹.

In general, thoracic region DREZ procedures are rarely performed, for example, only when there is an SCI resulting in segmental pain. However, DREZ procedures in the conus region are more frequently performed, specifically for avulsion of lumbar and sacral roots or trauma to the region.^5,17

The caudal termination of the spinal cord is called the conus medullaris. This conical structure occurs caudal to the sites of lumbar and sacral nerve root emergence and is located near the L1/L2 vertebrae in adults.¹² The angle of the lumbar nerve roots with respect to the thecal sac decreases from L1 to S1. In an analysis of magnetic resonance images of healthy volunteers, the average nerve root angle decreased from 40.9 degrees at L1 to 17.1 degrees at S1.¹¹ The size of the ventral and dorsal sacral roots has been shown to decrease in size from S1 to S5, with an average S1 nerve root diameter of 1.70 mm (ventral) and 2.39 mm (dorsal) decreasing to 0.17 mm (ventral) and 0.40 mm (dorsal) by S5.¹⁰

Indications

DREZ ablation has been applied as a treatment for pain from several etiologies, including neoplasm, trauma, and infection. The application of the DREZ procedure for treatment of pain following BPA has been the most widely described application and is associated with the best results.^18–23 Avulsion or deafferentation pain, which occurs following BPA injuries, is commonly described by patients as a constant crushing-type sensation, with can be episodic and involve pain in the hand or bursts of pain traveling down the arm.²⁴ The severity of this pain can be disabling and resistant to most medical therapies. Sensation is also abnormal in the affected limb, and these patients report phantom-like sensations, possibly related to somatosensory reorganization after deafferentation.²⁴ Deafferentation-related changes in higher central nervous system structures such as the cuneate nucleus or the ventral caudal nucleus of the thalamus can take decades to evolve. These changes represent the slow deafferentation process that occurs with second- and third-order relay neurons after an injury and imply slow plasticity changes in higher brain structures.²⁵

Htut et al. found that the intensity of the patient’s ongoing chronic pain rating after BPA was correlated with the number of roots avulsed when measured at a mean of 4 years after injury.²⁴ In a cohort of 76 patients, 56% of patients with BPA injuries also described referred sensations mapping to an unaffected dermatome at some point after their injury. The intensity of these sensations declined over time in the majority. It has been hypothesized that the pain after BPA occurs secondary to abnormal network activity in the deafferented dorsal horn; however, there is some evidence that neighboring injured roots might also play a role as pain mediators.^21,26,27 Bertelli and Ghizoni studied 10 patients with chronic neuropathic pain after brachial plexus injuries and selectively blocked nonavulsed roots under computed tomography guidance.²⁷ The patients all exhibited decreases in pain ratings in this small, uncontrolled study, but inflammatory-driven central sensitization from neighboring roots might have future implications for the extent of DREZ lesions.

Guenot et al. performed single-unit analysis of the spinal dorsal horn in patients with various types of neuropathic pain in the context of the DREZ procedure.²⁸ They studied three categories of patients, including patients with peripheral nerve injury, true deafferentation injury from BPA, and disabling spasticity. They found that the coefficient of variation in firing rates was highest for the patients with BPA for cells recorded in the dorsal horn. In addition, BPA patients and spasticity patients had similar patterns of firing behavior, whereas the peripheral nerve injury group demonstrated the most “nonrandom” forms of firing discharges for these cells, manifested in a serial dependency of the interspike intervals.²⁸

Animal models for BPA have also been developed to probe the possible molecular causes of the subsequent pain syndrome. Quintão et al. studied BPA in mice and examined the effects of blocking tumor necrosis factor A with a selective antibody.²⁹ This study demonstrated abolishment of the hypernociception that occurs after injury through intravenous injections of the blocking antibody, thereby pointing toward a possible pathway that could be blocked as a therapy.²⁹ However, in general, neuropathic pain syndromes developing after BPA have been difficult to treat medically, and even newer, novel treatments fail to produce drastic reductions in pain.³⁰

Similar to avulsion-type injuries, the pain following spinal cord or cauda equina injury is a common phenomenon, with a large percentage of patients reporting severe pain.^31,32 An estimated 65% of patients with SCI experience pain and 50% of patients with SCI have a level of pain severe enough to interfere with rehabilitation.^33,34 Two types of pain following SCI have been described: nociceptive and neuropathic. Nociceptive pain is thought to be the result of activation of peripheral nociceptors due to ongoing tissue damage and is generally considered to be responsive to nonsteroidal anti-inflammatory drugs and other treatments such as physical therapy.³⁵ In contrast, neuropathic pain is the result of an abnormally functioning nervous system.³⁵ Neuropathic pain can develop below, as well as above, the level of the lesion and is thought to be related to spinothalamic tract dysfunction.^36,37

The mechanism of neuropathic pain following SCI has been proposed to involve neuronal hyperactivity and hyperexcitability.³⁵ In addition, there appear to be differences in the deafferentation process that SCI patients demonstrate with or without the development of neuropathic pain. In patients with neuropathic pain below the spinal cord lesion, the electroencephalogram peak frequency is significantly lower than that in patients without neuropathic pain.³⁷ Interestingly, in patients suffering from neuropathic pain after SCI, there also appears to be a significant reorganization of primary somatosensory cortex (S1) compared to the arrangement in SCI patients without an associated below-the-injury pain syndrome.³⁸ Wrigley et al. compared the receptive fields for little finger, thumb, and lip brushing derived during functional magnetic resonance imaging for 20 subjects after SCI (10 with and 10 without neuropathic pain), along with 21 controls without SCI. The SCI patients had suffered complete thoracic level injuries that had occurred 2 to 37 years prior to the study. The receptive field for little finger sensation was statistically different in the neuropathic pain patients, having migrated toward the leg representation.³⁸ This effect was correlated with the intensity of pain the subjects experienced.

Hulsebosch et al. hypothesized that the types of neuropathic pain arising after SCIs have distinct mechanisms of production. Neuropathic pain “below the level” may have its origin in an inflammatory cascade after the injury involving the release of glutamate, chemokines, and the activation of microglial cells in the dorsal horn, eventually leading to neighboring neuronal hyperexcitability.³⁹ It has been further proposed that the chronic neuropathic pain developing “at the level” of the injury is likely due to persistent activation of several intracellular kinase pathways, including mitogen-activated protein kinase, leading to persistent activation of the cyclic adenosine monophosphate response element–binding transcription factor. This transcription factor leads to products that phosphorylate N-methyl-d-aspartic acid–activated channels, tipping the neurons toward a hyperexcitable state.³⁹ Finally, “above the level” pain conditions may involve a peripheral sensitization mechanism. The below-the-level type of pain does respond to the DREZ procedure, as outlined later; however, future therapies for this condition and its associated central sensitization might include viral-mediated gene transfer or stem cell transplantation as their basis.⁴⁰

In an effort to find alternatives to pharmacologic treatments that are accompanied by systemic side effects, DREZ ablation has been evaluated in the setting of pain resulting from SCI.^41–45 It has been reported by Sindou and colleagues that DREZ ablation is only effective in segmental pain corresponding to the level of the spinal cord lesion or adjacent pathologic processes such as gliosis or cavitation.^3,43 Diffuse, bilateral, and primarily sacral pain levels are less likely to respond.⁴⁶

There are other less commonly reported indications for the DREZ operation. DREZ ablation has been reported for the treatment of postherpetic neuralgia, occipital neuralgia, phantom limb pain, and radiation-induced plexopathy.^{23,42,47–50} Amputation of a limb or BPA can also result in a chronic form of phantom limb pain.^51–54