Dorsal Root Ganglion Stimulation

Evidence status, current to August 2026. Randomized comparative evidence remains limited to ACCURATE and to lower-extremity CRPS I/II. Other pain territories and diagnoses are off-label in the United States and rely mainly on observational evidence, and the labeling separately restricts the vertebral levels at which leads may be placed. Two things change faster than this page: device labeling with its MRI conditions, and device availability. DRG stimulation is supplied in the United States by a single manufacturer, so confirm current availability and the current instructions for use before building a plan around it.

1.The Rationale of the Target

The dorsal root ganglion holds the pseudounipolar cell bodies of the primary sensory neurons. Each soma gives off a single stem axon that bifurcates into a peripheral and a central branch, and that branch point is the T-junction. Experimental work supports several candidate mechanisms: injury-related ectopic firing arising in the ganglion itself, altered ion-channel expression, satellite-glial signaling around the soma, and frequency-dependent conduction failure at the T-junction, where a high-frequency train arriving from the periphery may fail to propagate centrally. These models make the DRG an attractive target, but no single mechanism has been established as the clinical explanation for analgesia.

That makes the DRG a rational, and unusually favorable, stimulation target. Each ganglion contributes to a dermatomal territory with overlap between adjacent levels, so one lead can cover a focal territory that broad dorsal-column stimulation cannot selectively reach. And because the ganglion sits in the bony neural foramen in only a thin layer of CSF, the electrode-to-target distance is small and mechanically stable, so stimulation works at much lower energy than dorsal-column SCS and shows far less postural variation (the steep distance-dependent swings that plague a lead facing the moving cord are largely absent at the foramen). Ex vivo and animal work shows that ganglionic field stimulation can alter excitability and conduction, while human benefit may also involve spinal and supraspinal network effects. Treat these as converging hypotheses rather than proof that stimulation simply switches off ganglion output, and note that much of the preclinical DRG literature, including the ex vivo field-stimulation work cited below, was produced with manufacturer involvement. That does not invalidate it. It does mean the mechanistic story and the commercial story grew up together, and it should be weighted accordingly.

2.The ACCURATE Trial

The pivotal evidence is the ACCURATE randomized comparative trial (Deer et al., 2017): 152 patients with CRPS or peripheral causalgia of the lower extremity, randomized to DRG stimulation versus traditional dorsal-column SCS. DRG met non-inferiority and then superiority for treatment success (≥ 50% pain relief without a stimulation-related neurological deficit): 81.2% versus 55.7% at 3 months, and 74.2% versus 53.0% at 12 months. DRG also produced less postural variation in paresthesia and more anatomically targeted coverage. ACCURATE is the basis for the FDA approval of the DRG system (Axium, originally Spinal Modulation, then St. Jude Medical, now Abbott's Proclaim DRG) in 2016 for moderate-to-severe chronic intractable pain of the lower limbs in adults with CRPS types I and II; CRPS II is the condition historically called causalgia.

ACCURATE was open-label, manufacturer-sponsored, and compared DRG with tonic dorsal-column SCS, not burst or 10 kHz stimulation. It therefore supports a diagnosis- and comparator-specific result through 12 months, without estimating benefit against sham. The denominator also matters: patients who did not proceed to permanent implantation were counted as treatment failures. The published evaluable analyses reported 56/69 versus 39/70 successes at 3 months and 49/66 versus 35/66 at 12 months; these are neither implanted-only response rates nor complete 76-per-arm randomized denominators. Consider attrition and the highly selected study population when counseling. Read the FDA summary of safety and effectiveness data for P150004 alongside the publication.

ACCURATE sits on top of earlier prospective work: a small feasibility study (Deer, 2013) and the Liem multicenter prospective trials, which reported sustained relief across CRPS, failed-back, and other neuropathic conditions at 6 and 12 months. The NACC appropriate-use consensus (2019) consolidated this into best-practice guidance on selection, lead placement, strain relief, and complication mitigation, endorsing DRG stimulation chiefly for focal/regional neuropathic pain and most strongly for lower-limb CRPS, consistent with the label.

3.Candidate Selection: Focal Neuropathic Pain

The unifying theme of the good DRG candidate is focal, mappable neuropathic pain confined to one or a few dermatomes or a peripheral-nerve territory, the situations conventional SCS captures least reliably:

• CRPS types I and II of the lower limb: the only on-label, randomized-evidence population. Upper-limb CRPS is off-label and rests on observational data.
• Focal or dermatomal neuropathic pain that broad dorsal-column coverage cannot selectively reach.
• Groin and pelvic pain, including post-herniorrhaphy (inguinal) neuralgia.
• Foot pain and other discrete distal territories.
• Post-amputation stump and phantom limb pain.

An honest caveat on scope: in the United States the on-label, RCT-supported indication is lower-limb CRPS types I and II. Groin, pelvic, post-amputation, and foot-pain uses without qualifying lower-limb CRPS are off-label, as are the knee, post-thoracotomy, and post-mastectomy applications reported in the literature; all rest on case series and retrospective reviews rather than randomized data. They are reasonable to consider in the right focal-neuropathic patient, but they should be presented as lower-evidence extensions, not as established indications. Note that the thoracic applications also require lead placement outside the labeled vertebral levels (Section 5), which is a separate and more serious departure than an off-label diagnosis.

4.How the Lead Is Placed

The epidural space is entered percutaneously with loss-of-resistance technique, but the lead is not simply steered laterally as an SCS lead would be. A curved delivery sheath and a steerable guidewire are advanced through the needle and used to direct the lead toward and out through the target foramen under fluoroscopy; in the lumbar spine the needle is usually introduced from the contralateral side, one to two levels below the target, so the sheath can be directed across the epidural space toward the target foramen, and a lateral image confirms the lead exits at the dorsal aspect of the foramen. The added skill over a conventional percutaneous SCS lead is this sheath-directed foraminal egress and the loop that follows it.

The defining technical step is the strain-relief loop formed in the epidural space just medial to the foramen. Once foraminal position is confirmed, the sheath is backed out and rotated to point rostrally and the lead advanced under AP fluoroscopy until a cranially oriented loop extrudes without displacing the foraminal segment; the guidewire is then partially withdrawn to soften the intraspinal portion and the lead advanced further to create the shepherd's crook double curve, which is thought to stabilize the electrode by pinching it against the pedicle. The loop decouples spinal and body motion from the delicate foraminal segment and is the principal defense against migration and lead fracture. It is also described in the literature as an S-curve; the terminology varies, the maneuver does not.

The loop is not the only defense. The lead is also anchored at the fascia, and the two protect different segments: the anchor fixes the extraspinal course, the loop absorbs motion at the foraminal segment. Both are deliberate steps, and over-tightening an anchor on a lead of this caliber risks the very fracture the fine lead is prone to; take the specific anchoring technique from the current instructions for use and from the NACC and ASPN best-practice documents. The delivery hardware has also been revised over time, so confirm the components and steps for the system stocked in your hospital. Entry side, entry level, and approach vary with the target level, particularly at the sacral levels.

S1 is the recognized hard level, and the reason is anatomic: the S1 ganglion frequently sits medially or intraspinally rather than within the foramen, so there is less foraminal purchase for the lead and less epidural room in which to extrude a loop, and the technical difficulty of the loop and the higher early migration rates at S1 follow from that. Ganglion position relative to the foramen varies by level and by patient, so review the preoperative imaging for the specific target rather than assuming a textbook intraforaminal ganglion. Current US Abbott DRG safety information requires the patient to remain awake and conversant during lead placement to provide feedback that helps reduce nerve injury risk. Do not equate the absence of a paresthesia-mapping requirement with permission for general anesthesia; follow the exact device instructions and a specialist anesthesia plan.

One or more leads are selected to match the pain map within the system's allowed configuration. Because of the thin CSF layer, programming uses low amplitudes; sub-perception (paresthesia-free) stimulation is often achievable, because many patients report analgesia at amplitudes below their perception threshold, and the captured territory is stable across positions: coverage assessed lying down tends to hold when the patient stands. Candidacy is confirmed by a trial, and the same periprocedural discipline applies as for SCS: DRG placement is a high bleeding-risk neuraxial procedure for anticoagulation purposes, and the infection-prevention principles are identical.

5.Complications, Limitations, and Device Scope

The complication profile centers on the lead. The foraminal segment is short and mobile and the lead is finer in caliber than an SCS lead, and migration and fracture dominated the early DRG experience. There is a genuine tension in the evidence worth stating plainly: ACCURATE found no significant difference in serious or device-related adverse events, but procedure-related adverse events were more frequent with DRG, while observational and registry series reported migration and fracture rates that concerned early adopters enough to drive the technique changes above. A 2025 systematic review and meta-analysis of non-infectious complications (13 studies, 634 patients) pooled all complications at 37%, device-related complications at 27%, lead fracture at 6%, lead migration at 6% and explantation at 12%, and concluded that the safety profile is broadly comparable to that of conventional SCS. Quote ranges rather than single figures when consenting, because rates vary widely by series, level, and era, and cross-check against the NACC and ASPN documents. Standardized strain-relief-loop technique is believed to reduce migration and fracture, which is why loop formation is treated as the crux of the operation rather than an afterthought, but the supporting evidence is technical notes and before-and-after series rather than controlled comparison. Dural puncture with post-dural-puncture headache is an access-related risk amplified by foraminal manipulation, infection is the usual implanted-device risk, and there is a real learning curve; the sensible prerequisite is to be fully comfortable with conventional percutaneous SCS lead placement first.

Two scope limitations should be stated plainly, and they are separate limitations. The first is diagnostic: the FDA indication is moderate-to-severe chronic intractable pain of the lower limbs in adults with CRPS types I and II, framed in the label as an aid in management. Upper-limb CRPS is not on label, and all other diagnoses are off-label. The second is a restriction on where the leads may be placed: the labeling states that the safety and efficacy of leads implanted above the T10 vertebral level have not been evaluated, so cervical and upper-thoracic placement sits outside the evaluated anatomy rather than merely outside the indication. That is the more serious of the two departures, and it is the one to check against the current instructions for use for the exact system before offering a thoracic or cervical target. And like any implanted neuromodulation system, the device carries MRI-conditional labeling that has evolved by generation; confirm the current conditions for the exact system, leads, and field strength rather than assuming, since a fractured or retained lead can change MR eligibility. DRG is a poor fit for diffuse pain, which remains conventional SCS territory. Axial low back pain deserves a more careful answer than a flat no: in the Liem prospective cohort, back-pain relief started respectably near 58% at six months but fell to roughly 42% at one year, while foot-pain relief held near 80%, the pattern you would predict from a focal target. There is an off-label literature on L2 DRG stimulation for axial and discogenic low back pain, on the rationale that L2 carries sinuvertebral afferents from the lower lumbar discs. Treat it as investigational rather than as either an established option or a settled contraindication.