Background

For almost 40 years stimulation of peripheral nerves has been used for the control of neuropathic pain. Like spinal cord stimulation (SCS), the mechanism of peripheral nerve stimulation (PNS) is believed to have its basis in the gate control theory of pain.¹ Although PNS and SCS have been accepted techniques for the treatment of neuropathic pain, SCS has become more widely used. A number of factors have prevented the evolution of PNS, not least of which is a perception that PNS is an orphan modality. This has resulted in PNS never receiving the medical attention that it deserves from either manufacturers of implantable devices or implanting surgeons. Unfortunately, appropriate investigation of its scope and application remains latent.

Open surgical implantation of PNS electrodes, which was described between the early 1970s and 1990s, presaged the ongoing interest at selected centers in the United States and elsewhere to manage peripheral neuropathic pain by PNS.^2–13

Another handicap for the development of PNS is the lack of any coordinated effort by the implanting physicians who have a vested interest in furthering the scope of PNS to engage in a dialog with the U.S. Food and Drug Administration (FDA). For example, a dialog to extend the current approval for the radiofrequency (RF) interface to include approval for an implantable power source implantable pulse generator (IPG) is wanting. Only one manufacturer, Medtronic (Minneapolis, Minn), has FDA approval to provide an electrode in conjunction with an RF external generator for PNS. All IPGs, whether Medtronic, Advanced Neuromodulation Systems (ANS, Plano, Tex), or Boston Scientific (Valencia, Calif), that are used in conjunction with PNS are considered “off-label” indications. The PNS (On-Point) electrode itself was originally developed as a quadripolar electrode for use in SCS. Recently PNS has received a boost through the efforts to develop occipital nerve stimulation.^14,15

Definition

“Neuropathic pain is pain initiated or caused by a primary lesion, dysfunction, or transitory perturbation in the central or peripheral nervous system.”¹⁶

Indications

PNS is indicated for the treatment of neuropathic pain in the distribution of a peripheral nerve or nerve trunk that is chronic and unresponsive to conventional medical management (CMM). Loss of function, an inability to participate in exercise therapy, and the nonresponse to local anesthetic or sympathetic blocks are considerations for PNS. Cases of neuropathic pain arising from a plexus injury or mononeuropathies from various causes may have in addition a partial or complete sensory loss that is within a particular nerve distribution. Common indications for open PNS are shown in Table 17-1.

Table 17-1 Indications for Open (Surgical) PNS Implant Using the FDA-Approved On-Point (Paddle) Electrode

FDA, Food and Drug Administration; PNS, Peripheral nerve stimulation.

A number of conditions amenable to PNS are as follows: occipital neuralgia;^17,18 postherpetic neuralgia;^19,20 postherniorrhaphy pain;²¹ complex regional pain syndrome (CRPS);²² cluster headache;^23–26 chronic daily headache;^18,27 coccygodynia;²⁸ fibromyalgia;²⁹ cervicogenic pain;^30,31 and migraine.³² Neurogenic pain following surgery for tarsal or carpel tunnel and postherpetic pain in a peripheral nerve distribution on the face, trunk, or limb are obvious indications for PNS. As a consequence of the foregoing indications, the contemporary unavailability of dedicated electrode designs should stimulate the engineering of nerve-specific electrode interfaces. Other potential sites for PNS are the sphenopalatine ganglion (SPG)²⁶ and other autonomic nervous system targets.

As is customary in every prospective case of SCS, it is essential to obtain a psychological evaluation for all potential PNS patients. This has been summarized by Doleys.³³

Although a trial of neurostimulation always precedes implantation of an SCS, in the case of open PNS the success and stability of this technique in most cases does not warrant the risk of infection from having an externalized connection to a pulse generator for 48 to 72 hours. In addition, the high success rate of the modality precludes this initial step. This approach does not apply to percutaneous applications, in which case a trial is always mandated.

Contraindications

PNS is not associated with many contraindications other than the aspects that apply to all surgeries. Patients with bleeding diatheses or those in whom the discontinuation of anticoagulants is contraindicated are obviously excluded from PNS. Active infection, particularly in cases in which the possibility of bacteremia is high, and patients whose medical condition or malignancy may require serial magnetic resonance imaging (MRI) studies would preclude PNS at the present time.

Neuroanatomy

The axon is the functional unit of a peripheral nerve. Both afferent and efferent axons with their Schwann cells are enclosed in a delicate layer of endoneurial tissue (endoneurium). This is connective tissue that allows the free diffusion of fluids to and from neural structures. Each bundle of axons is enclosed by the perineurium. Cell bodies in the dorsal root ganglion are the source of an axon with a long branch that extends to its peripheral functional source and a shorter branch that passes from its cell body to the spinal cord. Sensory axons are unipolar and transmit sensory information from receptors in the periphery to second-order neurons in the spinal cord. On the other hand, motor neurons arise from the cell bodies in the ventral horn of the spinal cord and in contrast are multipolar with many dendrites. In addition, an axon carries impulses peripherally to activate their specific effector organs. Both dendrites and cell bodies of these neurons are highly specialized to integrate postsynaptic currents that modulate effector organs.

Myelinated nerve fibers have many concentric laminae that form from a single Schwann cell. The nodes of Ranvier are interruptions in the myelin sheath where the inward currents during depolarization are regenerated. An axon of a sensory neuron varies in diameter from as little as 2 µm to 11.75 µm.^34,35 To facilitate regional distribution and therefore sensory coverage, nerve fibers divide into many branches, thereby allowing the innervation of a significant tissue mass by a single neuron. Clinically this results in referred pain that may originate in a single neuron being transmitted by branches to other tissues in the same region. The axon reflex is another mechanism that allows pain to be felt in undisturbed tissue. In this case antidromic transmission passes to other adjacent tissue, causing an expansion of the painful area. Table 17-2 lists the diameters of nerve fibers and their conduction velocities and function. The fascicular anatomy within nerve trunks is shown in Fig. 17-1. Figs. 17-1 and 17-2 show nerve fibers grouped within a thin laminated sheet (epineurium) that covers the axons.

Fig. 17-1 Plexiform pattern of fascicular anatomy in a peripheral nerve. The dots demonstrate the expanding distribution of nerve fibers with distal projection.

Fig. 17-2 A, Transverse section of median nerve at the elbow. B, Groups of fasciculi are shown as circles. Letters AMPL reflect orientation. The letters within each fascicle are abbreviations of the distal motor and sensory nerves in the arm and forearm.

A collection of nerve fibers (axon bundles) are known as fascicles. Each fascicle containing many axons is encased by a connective tissue layer and perineurium. The entire nerve is contained within a loose outer covering, the epineurium. Although fascicles vary in size from 0.04 to 4 mm, the majority are found between 0.04 and 2 mm in diameter. As nerves proceed distally, their fascicles begin to divide into smaller and smaller units and become more numerous. In addition, this organization takes on a topographically discrete nature, particularly in mixed (motor and sensory) nerves, and is responsible for providing an intimate view of the fascicular architecture.^36–39 For example, in the ulnar nerve behind the medial epicondyle many nerve fibers are grouped into a single fascicle. A similar arrangement is found in the radial nerve in the spinal groove, the axillary nerve behind the humerus, and the common peroneal nerve in the lower thigh.

The histology of nerve fibers has considerable bearing on the ability to selectively stimulate the sensory or motor nerve fibers. The cross sectional area of a nerve trunk is comprised of 25% to 75% epineurial tissue, the highest amount being in the sciatic nerve in the gluteal region and the lowest in the ulnar nerve at the medial epicondyle. This characteristic influences the effect of neurostimulation. The greater the thickness (higher impedance), the greater is the attenuation of the electric field. In a similar manner, this effects the diffusion of local anesthetics and therefore the amount necessary to achieve their mechanism of action at the axon.

Blood Supply

The vasa nervorum provide nutrition to peripheral nerves derived from collateral vessels, which are branches of adjacent veins and arteries (Fig. 17-3). Because of the dynamic nature of tissues and the translational movement of nerves, the vasa nervorum are quite tortuous. The magnitude of this movement increases in the vicinity of joints. There is considerable variation of the collateral blood supply throughout the length of each nerve. This has the effect of creating various watershed zones in each nerve between collateral sources. These zones of relatively poor nutrition may jeopardize the integrity of the nerve and cause increased stress such that extraneous compression or handling may compromise nerve function. In spite of the foregoing hazards, Ogata and Naito 1996⁴⁰ and Smith⁴¹ report that a reduced interneural blood flow during and/or after, for example, surgical resection of a nerve, is generally reestablished within 3 days.

Fig. 17-3 Dissection of the axilla showing collateral blood supply from the axillary artery supplying several adjacent nerves.

Many of the complications that resulted from nerve stimulation during its evolution can be largely attributed to the morbidity induced at the time of electrode placement, implanter inexperience, and the overall slow technical development of PNS systems. Some of the blame can be laid at the feet of electrode-neural interface, an aspect that has been largely resolved.

Certainly compression injury or contusion to a nerve is a greatly diminished factor underlying contemporary PNS morbidity. Preexisting trauma caused by congenital, occupational, or incidental trauma is now largely responsible for any associated PNS morbidity.

Physics and Physiology Underlying Neurostimulation

The mechanism by which PNS achieves it effect is the activation of peripheral low threshold Aβ fibers, which in turn inhibit activity of small-diameter nociceptor Aδ and C fibers. Modulation of this nociceptor input is either direct through the selective activation of Aβ fibers, giving rise to inhibitory postsynaptic potentials or indirect via inhibitory spinal interneurons.⁴² Peripheral Aβ fiber activity most likely engages the medial lemniscal pathways, which in turn provide input to the ventral posterior medial nucleus of the thalamus, thereby overriding afferent input from spinothalamic tracts.^43–45 These proposed mechanisms would have the effect of modulating activity from central sensitization at the dorsal horn. As a consequence, sensitization of supratentorial structures such as those involved in both cognitive and perceptual dimensions in the limbic system could render neurostimulation less effective or absent. Thus early application of this modality would be crucial to achieving its maximum effective potential. In contrast to SCS, PNS electrodes are placed on nerves affected by whatever is the ongoing neuropathic process; the primary action of peripheral neurostimulation tends to be direct inhibition of the first-order nociceptor and not second-order or higher neurons. It is clear that both allodynia and hyperalgesia, if present, are frequently reduced or eliminated by PNS. This suggests that sensitization can be subverted by suppressing peripheral nociceptor activity in the spinal cord and overriding nociceptor input at the thalamus. Given the constraints of slowing any temporal attempt on the progression of nociception, PNS may be more effective than SCS in providing inhibition at an interneuronal level.

Equipment

Only the On-Point and Quad Plus electrodes manufactured by Medtronic (Minneapolis Minn) are approved by the FDA for open PNS (Fig. 17-4). Other electrodes such as the Resume, Resume II, or Resume TL from Medtronic; the Artisan from Boston Scientific (Valencia, Calif); and the Quattrode and Axxess 6 from St. Jude Systems (Plano, Tex) are also used off label. For percutaneous PNS applications and trial, cylindrical wire type leads such as the Quad, Quad Plus, or Quad Compact from Medtronic and the Quattrode, Octrode, or Axxess from St. Jude, and Artisan from Boston Scientific are also used in an off-label basis.

Fig. 17-4 Two electrodes, Quad and On-Point, approved by the Food and Drug Administration for use in peripheral nerve stimulation.

Although the RF-coupled transmitter receiver made by Medtronic is approved by the FDA as a power supply for PNS, IPGs made by St. Jude Systems and Boston Scientific and Medtronic are used as power sources for PNS as off-label indications.

Surgical Technique

This description is confined to the open surgical approach to a peripheral nerve. If a trial of PNS is desired, this may be undertaken using a percutaneous lead that is passed through a needle to the selected nerve. If ultrasound is used for guidance, a percutaneous lead can be introduced precisely to the target nerve. A 2- to 3-day trial would be sufficient to determine whether function is improved or that allodynia/hyperalgesia, if present, is reduced and to what degree pain relief is evident. At least 50% or more symptomatic improvement and, more important, functional restoration should be demonstrated. An alternative approach to a percutaneous trial is surgical exposure (first surgical stage) of the affected nerve; fixation of the plate (paddle) electrode; and passage of the extension, which is connected to an external pulse generator. This will allow the patient to try the device for a period of 2 to 3 days. Demonstration of pain relief is one thing; but more important, the restoration of function in the affected limb is the measure that should be used to determine a successful outcome. At second-stage surgery the wound is reopened; the extension is removed from the skin site; and, after creating a tunnel, the lead extension is passed to a site where a pocket for the IPG has been made. For PNS of the radial, median, and ulnar nerves, a pocket for the IPG is usually made beneath the clavicle. The device can be secured to the fascia overlying the pectoralis major. In thin individuals the IPG can be placed beneath the fascia, but it must not exceed the depth that allows interrogation of a rechargeable generator. Nonrechargeable IPGs can be sutured to the fascia overlying the ribs beneath the pectoralis major muscle. Because of the current limitations of electrode design, the only nerves in the lower extremity amenable to PNS are the femoral, sciatic, common peroneal, and tibial.

Access to the sciatic nerve is by an incision in the lateral thigh posterior to the iliotibial band. Dissection is carried down to the hamstring compartment, where the sciatic nerve lies between the adductor magnus and long head of biceps. Once identified, the sciatic nerve is carefully dissected from its bed while retaining its vasa nervorum and protecting any related muscular branches for a length sufficient to accommodate two On-Point electrodes (Figs. 17-5 and 17-6). Each electrode is placed adjacent to the tibial and peroneal components. To stabilize the two On-Point electrodes, their adjacent Gore-Tex skirts are sutured together using 4.0 nylon sutures. The two electrodes are then wrapped around the nerve as a “sandwich,” making sure that each bank of contacts is adjacent to the peroneal and tibial divisions, respectively. The free Gore-Tex edges are then tacked to the epineurium using 4.0 to 5.0 nylon sutures, taking care not to constrict the nerve. A pocket can be made either deep to the iliotibial band or under the deep fascia, whichever seems most appropriate at the time. The IPG can be retained in situ using 2.0 braided nylon or silk sutures to the fascia. This position has the advantage of avoiding the need to pass a long extension from the midthigh to a pocket in the buttock.

Fig. 17-5 Shown are two On-Point electrodes placed so the contacts are in opposition to the tibial and common peroneal components of the sciatic nerve.

Fig. 17-6 A prototype electrode is shown attached to the sciatic nerve.

Because of the small current requirements of PNS, it is common to implant a nonrechargeable IPG. For most purposes a life of 9 to 12 years can be expected. Depending on the surgeon’s preference, pockets for the IPGs can be made in the gluteal area³⁰ and abdominal wall.⁴⁶

Irrespective of where the IPG is implanted, the pocket should be deep enough to prevent erosion through the skin. It should not be too deep to interfere with programming (if it is a programmable IPG), and its location should be in an area where there is minimal mobility. Likewise, to minimize mechanical damage the extension should pass, at the most, around only one joint.

For PNS in the upper limb, both the ulnar and median nerves are found superficial and medial to the biceps brachii and medial head of the triceps (Fig. 17-7). The best surgical site for the radial nerve is in the spiral groove between the lateral and medial heads at the triceps and is found by dissecting the intermuscular septum. In each case, while protecting the collateral blood supply and preserving the motor branches to the triceps, a length of nerve sufficient to accommodate the On-Point electrode is dissected free. This is retained using 4.0 nylon sutures to the epineurium.

Fig. 17-7 Common implant sites on the torso and thigh for sourcing power of the implantable pulse generator for peripheral nerve stimulation (PNS) of nerves in the upper and lower extremities. A, Median/ulnar nerve PNS generator placement. B, Femoral nerve PNS generator placement. C, Sciatic nerve PNS generator placement.

Patient Management and Evaluation

To allow for healing and stabilization of the neural electrode interface, exercise therapy is withheld for a period of 6 weeks. Rehabilitation of the affected extremity together with general physical therapy should then be undertaken. Therapy should be continued until optimal function is established. It may be necessary to reprogram the neurostimulator immediately and subsequently following implant, particularly in patients who may have required significant medication, including opioids, for the management of their neuropathic pain. With the anticipated improvement of function during exercise therapy, it should be possible to gradually decrease this dependence. However, in some cases weaning from opioids may require a significant investment of time.

Because most nerves that are currently amenable to neurostimulation are mixed nerves and because fascicular selective electrodes are not available, the plexuslike fascicular arrangement significantly influences sensory thresholds. Therefore only a small window between sensory and motor stimulation is available. Under ideal circumstances provision of a bipole to a number of sensory fascicles would not only reduce collateral stimulation of motor fascicles but could also allow for the smallest possible current to achieve a sensory threshold for paresthesia in the nerves involved in neuropathic pain. Stimulation currents used for PNS are significantly less than those required for SCS. Under most circumstances the amplitude required to achieve therapeutic stimulation is 0.5 to 7 mA, a pulse width of 120 to 180 msec, and a frequency that varies between 40 and 90 Hz.

Outcomes Evidence

No prospective studies of PNS are currently available. A few case series and clinical reports provide an insight into the value of PNS. The most important from a functional point of view are described here. Hassenbusch and associates¹¹ described 30 patients with CRPS II in whom symptoms had been present, in some cases for many years. Although 35% noted a decrement in analgesia over the first 2 years after implant, residual analgesia persisted at 15 years. Also noteworthy is the fact that 23% of these patients went back to work full or part time. Of the eight most recently published studies,^47,48 a study of 17 patients by Novak and Mackinnon in 2000⁴⁹ described 60% (10 patients) having good to excellent pain relief, 4 with fair pain relief, and 3 with no adequate change in their symptoms at 21 months. However, a third of their patients returned to work. Eisenberg, Waisbrod, and Gerbershagen⁵⁰ looked at data from three institutions: the Red Cross Center, Mainz; the Institute for Back Care, Bad Kreuznach; and the Linn Medical Center, Haifa. Earlier data (1985 to 1986) were compared with recent (1993 to 1995) data at the Red Cross, Mainz. Seventy-eight percent of the patients were regarded as good, and 22% as poor out of 46 total. It is also important to realize that these patients had failed CMM.

Mobbs, Nair, and Blum⁵¹ published data on 38 patients (M = F). Assessment was based on pain relief, narcotic use, function, and activities of daily living. Sixty percent had significant improvement in pain and return to work. No distinction between the response from workers’ compensation patients or noncompensable patients was found (p >0.05). Verbal pain scores were 5.1 with SD = 2.73, a figure that was acknowledged by an independent evaluation at 31 months.

From the foregoing, it is clear that PNS can be associated with good outcomes. Although the current indications for PNS include neuropathic pain resulting from trauma, surgical injury, and CRPS, new indications such as migraine and cluster headaches, trigeminal neuralgia, and fibromyalgia will have to await the development-specific electrodes. Certainly the relatively low invasive nature of open PNS will increase the attractiveness of this modality. The technique is reversible and testable; and, in comparison with ablative surgical options, there is no debate over its duration of effect. It has become an important clinical alternative. Complications associated with PNS are low and include local infection, hardware erosion, technical failure due to breakage or disconnection, fracture of current electrodes, and displacement.^52–54

The almost complete absence of site-specific PNS electrode development has seriously handicapped progress of this modality. New electrodes with special arrays are in stark contrast to the level of sophistication now reached by functional electrical stimulation. Alternative spacing and lower profile need research if PNS is to be optimized. One completely new concept, an electrode termed BION,^55–57 has the capability of enlarging the scope and indications for PNS (Fig. 17-8).

Fig. 17-8 BION peripheral nerve stimulation. This BION microstimulator image is the property of Boston Scientific and was provided by Boston Scientific in response to an unsolicited request. The BION received CE Mark; however the company no longer markets the device in Europe. It was never approved by the Food and Drug Administration for general medical use but was granted an exemption for clinical trials, which have since been discontinued.

Continuation of these developments, in particular obtaining approval from the FDA with specific research proposals, will ultimately enable a pallet of neural electrode interfaces that will place PNS at the forefront of neuromodulation of neuropathic pain.

Cost-effectiveness and Future of Peripheral Nerve Stimulation

It should be clear from the foregoing that the results of managing neuropathic pain by PNS can be very effective. Unfortunately, the Psychological Pain Assessment Scale (PPAS) as a subjective measure is a poor determinant of any functional improvement that could be expected from the use of this modality. The use of quality-of-life measures is a better and more objective determination of device effectiveness. At least the debate regarding the role of pain measures is increasing.^58,59

The need for prospective randomized controlled studies in relation to PNS is urgently needed. From such studies cost-effectiveness ratios can be derived, and more stringent characteristics can be applied to patient selection.

Following a review of 6000 citations by Health Technology Assessment,⁶⁰ it was determined that there is significant value in terms of function, symptomatic relief, and cost-effectiveness for the use of SCS for CRPS, neuropathic, ischemic, and low back pain.

As already discussed, the scope of PNS remains largely unchartered and will rely on a considerable effort in the development and evolution of neural electrode interfaces. As already mentioned, the introduction of a self-contained electrode power source, the BION (Advanced Bionics, Valencia, Calif) is remarkable.^55–57 Being without a lead or a separate implantable neurostimulator, it can be introduced through a small incision and easily directed to its target nerve. The device is small (3.3 mL, 27-mL length) with cathode and anode at each end. This device can be recharged externally and reprogrammed by telemetry. It is capable of a wide range of stimulation parameters—pulse width 1000 msec, a rate up to 1000 Hz, and an amplitude 12 mA. This is but a foretaste of the type of technical innovation that could propel PNS into the future. Although the scope of PNS has recently broadened by the use of percutaneous and subcutaneous leads, the need for dedicated electrodes on specific cranial nerves and other mixed peripheral nerves requires selective stimulation of afferent fibers with minimal effect on the motor and vasomotor components. Because of the highly selective distribution of PNS, in cases of SCS that lack a more regional topographical effect, PNS can be added to SCS to significantly improve the overall analgesia in the affected region.

Summary

In summary, the stability and low current requirements of PNS systems, particularly in the light of future technological developments, will place PNS at the forefront of neuromodulation technology that will effectively modulate neuropathic pain not otherwise achievable by CMM.